2019-05-01 00:33:44 Hello, all 2019-05-01 03:48:26 Hello friends, when I add a regular user's files under home directory stored under /home to the lbu backup targets using 'lbu add /home/john/.ssh' etc.. lbu restores them on boot with regular user as owner, however, that users home directory is owned by root. This is undesirable as many programs expect write access to $HOME. If I lbu add '/home/john/', lbu will add everything stored, including undesired objects like .cache/. 2019-05-01 03:49:21 Right now, I've got a derpy start job where I chown the users directory back for themselves. I wonder if I should submit a bug/ look into writing a patch, or is this somewhat intended? I can see why this sort of thing happened 2019-05-01 04:25:15 c705: https://bugs.alpinelinux.org/ 2019-05-01 06:37:43 c705: its not intended. I am not sure how to fix it though 2019-05-01 12:16:32 c705: use lbu exclude on the undesired object :/ 2019-05-01 14:08:54 i think we need store the permissions in the intermediate dirs in the apkovl 2019-05-01 14:10:00 iow, we need make tar behave more like cpio 2019-05-01 14:10:18 i think there was an option to tar to exclude contents of given directories 2019-05-01 14:25:57 How do I pull in all the -lang packages for my installed pkgs? 2019-05-01 14:27:55 Cogitri: apk add lang 2019-05-01 14:28:30 or apk add lang-no for eg. norwegian 2019-05-01 14:28:48 Ah, that was easy, thanks! 2019-05-01 14:30:01 yw, you can do the same with "docs" 2019-05-01 16:15:44 <_sfiguser> hello all guys, what is the most minimal gnu/linux firewall? i want to run an alpine container with a firewall, but don't know if iptables is the most minimal choice 2019-05-01 16:32:15 anything else i can do for this PR or just let it simmer? https://github.com/alpinelinux/aports/pull/7320 2019-05-01 16:37:48 _sfiguser: iptables / nftables 2019-05-01 16:38:47 nftables aims to replace iptables, it does have cleaner syntax but I don't find it to be well-documented 2019-05-01 16:43:52 Anyway, minimal in what sense? I don't think it matters much if you're just going to be leveraging netfilter 2019-05-01 16:44:08 also, a container might not be the best way to run a firewall 2019-05-01 17:46:53 how can i add man of lbu? thx 2019-05-01 17:48:04 <_ikke_> appll: there is no man page yet 2019-05-01 17:52:53 got it. thanks. _ikke_ 2019-05-01 17:53:34 found https://wiki.alpinelinux.org/wiki/Alpine_local_backup 2019-05-01 18:13:13 HI. I'm trying to install prosody ( apk add prosody ) but I get ERROR: unsatisfiable constraints: prosody (missing): required by: world[prosody] 2019-05-01 18:13:29 what does that mean and how can I resolve this problem 2019-05-01 18:15:58 yann-kaelig: what 'apk search prosody' says 2019-05-01 18:16:45 mps: hi, no output 2019-05-01 18:16:58 yann-kaelig: is community enabled in your /etc/apk/repositories file? Which alpine version do you use? 2019-05-01 18:18:15 bratkartoffel: hi, no community is not enable, 3.9.3 2019-05-01 18:18:23 yann-kaelig: what bratkartoffel told 2019-05-01 18:18:54 Ok, can you please add the repository to that file? Just take the existing "main" and replace the last part by "community" 2019-05-01 18:19:03 then run an "apk update" and the "apk add" should run fine 2019-05-01 18:19:51 bratkartoffel: thx, now that work. good to know that community is not enable by default 2019-05-01 18:35:45 yann-kaelig: I thought community was enabled and testing was not by default 2019-05-01 18:47:58 <_ikke_> nope, only main is enabled by default 2019-05-01 19:23:51 Docker got community enabled by default 2019-05-01 19:34:47 <_sfiguser> guys can alpine linux work as a router distribution? if yes, where can i learn how to configure a router? or what software do you use typically? 2019-05-01 19:42:02 _sfiguser: yes, AL has its roots in LEAF. There are some guides on the wiki - https://wiki.alpinelinux.org/wiki/Fault_Tolerant_Routing_with_Alpine_Linux 2019-05-01 19:42:03 Depends on the services you want to offer - routing demon (frr), dhcp, dns, firewall, vpn (wireguard)... 2019-05-01 19:45:56 s/demon/daemon/ 2019-05-01 19:46:52 _sfiguser: https://wiki.alpinelinux.org/wiki/Linux_Router_with_VPN_on_a_Raspberry_Pi 2019-05-01 19:47:08 <_sfiguser> TBK[m], what is LEAF? 2019-05-01 19:47:48 _sfiguser https://en.wikipedia.org/wiki/LEAF_Project 2019-05-01 20:56:54 hi 2019-05-01 20:57:09 i am looking for cmdline parameters in alpoine linux ... 2019-05-01 20:57:21 need set ip=dhcp on specific interface ... is it possible ? 2019-05-01 20:58:15 like boot parameter 2019-05-01 20:58:27 marnej: it is not specific to alpine, it is general for kernel 2019-05-01 20:58:52 ok thx :-) 2019-05-01 20:59:39 so there is no script ... in alpine :-) 2019-05-01 20:59:53 script? 2019-05-01 21:00:41 marnej: it is not specific to alpine linux but your bootloader 2019-05-01 21:01:29 cool,. i got it now :-) 2019-05-01 21:01:39 ip=:::::interface:dhcp 2019-05-01 21:02:46 i thought, that it is alpine specific - some script, read /proc/cmdline and act .... 2019-05-01 21:03:57 btw. thx for fast response : 2019-05-01 21:06:54 marnej: alpine uses initramfs and there you can change or add handling of boot however you need 2019-05-01 21:13:34 mps: thx! so initramfs is taking care of :-) 2019-05-01 21:14:37 yes, you have to unpack it, make changes, then pack it again 2019-05-01 21:15:24 but if you only want to add kernel parameters then you don't need to 'mess' with initramfs 2019-05-01 21:16:52 it is enough to add them to /etc/update-extlinux.conf and run update-extlinux 2019-05-01 21:19:52 tcely: thanks for approving that PR 2019-05-01 21:21:14 m4t: you're welcome 2019-05-01 21:26:30 what is update-extlinux ? :-) 2019-05-01 21:27:27 i am adding parameters to my netboot config 2019-05-01 21:27:59 command line utility to make new boot config 2019-05-01 21:38:26 marnej: extlinux is Alpine's default bootloader on BIOS 2019-05-01 21:38:50 doggone: ah thx :-) 2019-05-01 21:38:55 update-extlinux is a script that generates a config for it, like update grub 2019-05-01 21:40:04 ahh, extlinux = syslinux ? 2019-05-01 21:40:53 ok, i see now 2019-05-01 21:40:59 { EXTLINUX - SYSLINUX for ext2/ext3 filesystems } 2019-05-01 22:05:30 how can I run abuild without cleaning up the source and package dirs at the end? 2019-05-01 22:06:00 abuild -r ? 2019-05-01 22:06:05 use the -K switch 2019-05-01 22:06:19 see -k and -K switches on --help 2019-05-01 22:06:50 ah yes, exactly, the -K switch is what I need 2019-05-01 22:07:05 it is not yet documented here: https://wiki.alpinelinux.org/wiki/Abuild_and_Helpers 2019-05-01 22:07:24 That is very outdated 2019-05-01 22:07:50 can someone tpaste the current help? 2019-05-01 22:08:06 http://ix.io/1HN0 2019-05-01 22:08:39 thanks! 2019-05-01 22:35:07 midasi: The abuild wiki section now reflects the current version of abuild 2019-05-01 22:38:33 and what about ssh_key cmdline parameter ? 2019-05-01 22:38:48 it is handled by alpine ? 2019-05-02 10:38:36 marnej: in short, yes 2019-05-02 12:12:31 anyone have ready-made setup for multiple public net interfaces for Alpine 2019-05-02 12:14:57 mps: yes 2019-05-02 12:15:43 :) where is it? 2019-05-02 12:16:19 I found https://wiki.alpinelinux.org/wiki/Fault_Tolerant_Routing_with_Alpine_Linux but this is not what I need 2019-05-02 12:16:31 i have a couple of hosts 2019-05-02 12:16:48 so i have "setup" 2019-05-02 12:17:10 what do you need? are the multiple public ip's in same subnet? 2019-05-02 12:17:48 no, two eth to separate net 2019-05-02 12:18:18 one of them needs to be the default gateway 2019-05-02 12:18:33 but you need to set up some policy roiting 2019-05-02 12:18:35 routing* 2019-05-02 12:18:37 not important, but could 2019-05-02 12:19:13 so that correct gateway is used depending on the source address 2019-05-02 12:19:42 no, just to answer to packet at interface to which is incoming packet arrived 2019-05-02 12:20:24 I can write such script but asked if something is integrated to alpine already to not diverge too much 2019-05-02 12:20:46 i think pingu can do that for you 2019-05-02 12:21:31 mind you, I wrote a guide long time ago about linux ip and qos http://arvanta.net/mps/linux-tc.html 2019-05-02 12:22:05 cool :) 2019-05-02 12:23:05 I could write config scripts but want to integrate it somehow to alpine networking 2019-05-02 12:24:29 i'm not sure we have anything particular for that 2019-05-02 12:24:56 i wrote pingu, to do failover, so it solves the problem for you 2019-05-02 12:25:05 you just dont need configure any ping host 2019-05-02 12:25:27 np, I will look then if I can do something to make it useful for alpine 2019-05-02 12:25:41 that would be good 2019-05-02 12:26:02 and wil look for pingu, didn't know it exists 2019-05-02 12:26:37 thanks for hints 2019-05-02 12:43:07 https://git.alpinelinux.org/hosted/pingu/about/ 2019-05-02 12:43:21 i think you can set it up with empty interface definitions 2019-05-02 12:44:27 very nice, and in C. 2019-05-02 12:45:02 I'll steal some code from it for something different I need :) 2019-05-02 12:45:37 its a daemon, so it works with dynamic added interfaces (ppp0, eth over USB etc) 2019-05-02 12:45:44 sure, its opensource 2019-05-02 12:47:23 will look deeply but not now, I have other tasks for $dayjob and for alpine in todo list 2019-05-02 16:30:34 hey everybody. I was wondering where should I submit issues (bugs) with alpine packages. 2019-05-02 16:31:02 DannyB: Please see https://bugs.alpinelinux.org/projects/alpine/issues 2019-05-02 16:31:43 DannyB: bugs.alpinelinux.org 2019-05-02 16:32:01 Oh - I saw this, but I thought this is only for alpine linux itself. Is it ok to post there for specific packages? 2019-05-02 16:33:02 DannyB: Use the 'aports' category 2019-05-02 16:33:19 thank you 2019-05-02 16:35:42 Welcome 2019-05-02 16:48:56 the email confirmation for bugs.alpinelinux.org does not seem to function. is it typical for it to take several minutes to send the confirmation email? 2019-05-02 16:53:13 never mind, got it. 2019-05-02 17:49:53 hi everybody 2019-05-02 17:50:38 I try to edit /etc/network/interfaces but when put :i on vi, say "i is not implemented" 2019-05-02 17:50:49 Hello there, Classic_ 2019-05-02 17:51:06 hi 2019-05-02 17:51:11 Don't do `:i`, just press i to insert 2019-05-02 17:51:36 great! 2019-05-02 17:52:01 another question, why my alpine don`t get ip over dhcp? 2019-05-02 17:54:23 Hey all -- is there an equivilent to `telinit u` for alpine ( openrc )? Please feel free to just point me at docs. 2019-05-02 17:54:43 Classsic_▸ if you don't have it configured to get an ip over dhcp, then it won't. what does your interfaces file look like? 2019-05-02 17:56:38 caryatid: kill -HUP 1 2019-05-02 17:57:06 oh no, you asked for -u 2019-05-02 17:57:55 yah. I wanna replace init essentially -- basically doing an even crazier version of: https://github.com/marcan/takeover.sh 2019-05-02 18:01:11 telinit is sysvinit specific 2019-05-02 18:09:27 mps: totally. Thanks. 2019-05-02 19:16:59 hi 2019-05-02 19:17:06 I follow this: https://wiki.alpinelinux.org/wiki/VirtualBox_shared_folders 2019-05-02 19:17:27 but can´t install virtualbox-guest-modules-virthardened 2019-05-02 19:18:09 is not found, and enable all repos.----> http://dl-3.alpinelinux.org/alpine/v3.9/main http://dl-3.alpinelinux.org/alpine/v3.9/community http://dl-3.alpinelinux.org/alpine/edge/main http://dl-3.alpinelinux.org/alpine/edge/community http://dl-3.alpinelinux.org/alpine/edge/testing 2019-05-02 19:20:44 Classsic_▸ that wiki looks like it could use some TLC. There's a message at the top that says "as of 3.8..." 2019-05-02 19:20:48 apk add virtualbox-guest-additions virtualbox-guest-modules-virt 2019-05-02 19:20:53 community repo 2019-05-02 19:21:11 then reboot, install module, and you can then do the mount. 2019-05-02 19:21:24 after that is a horizontal rule, which seems to delineate the old and new instructions. 2019-05-02 19:23:57 thanks, i`ll try 2019-05-02 19:23:58 just updated the wiki. 2019-05-02 19:24:02 should be more clear 2019-05-02 19:24:20 https://pkgs.alpinelinux.org/packages?name=virtualbox-guest-modules*&branch=v3.7 vs https://pkgs.alpinelinux.org/packages?name=virtualbox-guest-modules*&branch=v3.8 helps confirm this. 2019-05-02 19:24:32 basically, they dropped the hardened kernel a couple versions ago. 2019-05-02 19:59:09 Hello once again friends. I have a Yubikey that I'm trying to get working in Alpine - usb device works under root but userland cannot access the device. Yubikey docs say add a rule under udev/rules.d/ but Alpine has mdev. 2019-05-02 19:59:49 c705: install eudev 2019-05-02 19:59:52 How do I do the equivelant of "adding a udev rule" in mdev? If you could point me at some documentation, I'll buy you a virtual beer. Thank you in advanced 2019-05-02 20:00:08 mps: I'd prefer to keep using busybox if I can 2019-05-02 20:00:42 also I, but not sure if it will work for hot plug devices 2019-05-02 20:01:25 mps, Poop. I'll give it an hour or so and see if anyone else has suggestions, but I'll look into udev. Thank man 2019-05-02 20:01:32 mps: Or lady. Whatever ;) 2019-05-02 20:01:38 maybe to add '/sbin/mdev' to /proc/sys/kernel/hotplug 2019-05-02 20:02:28 well, man ;) 2019-05-02 20:02:30 mps: /proc/sys/kernel/hotplug has /sbin/mdev already 2019-05-02 20:02:47 I mean, it works as root, just user has no access 2019-05-02 20:03:30 then try to edit /etc/mdev.conf and add appropriate rules there 2019-05-02 20:03:35 I had it working before I lost my old OS. I forget if I used udev or if I figured something else out 2019-05-02 20:04:10 mps: Yes, but how does the syntax of mdev.conf compare with rules in rules.d/ ? 2019-05-02 20:05:48 yes, it is different but with try and error maybe you could find working solution 2019-05-02 20:08:45 mps: I suppose I could, but I'm hoping someone here has some knowledge so that I don't have to resort to that 2019-05-02 20:10:38 I didn't tried so don't know, my use cases are simple. If you make any progress please inform us or make guide/wiki 2019-05-02 20:15:58 mps: 10-4 sir. I suppose you're just using udev>? 2019-05-02 20:16:19 right 2019-05-02 20:17:25 X window and some of it's siblings depends on udev 2019-05-02 20:19:03 mps: Lovely. I supose I'll just do udev then, since I'll be installing X some of the time anyways 2019-05-02 20:19:10 Than ks for your help 2019-05-02 20:19:25 you are welcome 2019-05-02 21:22:42 to run a cmd after reboot, can i add it to /etc/rc.local? 2019-05-02 21:23:10 didn't find the info from https://wiki.alpinelinux.org/wiki/Alpine_Linux_Init_System 2019-05-02 21:23:17 appll: /etc/local.d/ 2019-05-02 21:23:31 read README file there 2019-05-02 21:23:34 (jeez, move to runit-init or s6 already) 2019-05-02 21:23:46 appll: If you enable /etc/init.d/local, you can dump a script ending in .stop in /etc/local.d/ and have it run on reboot 2019-05-02 21:24:28 got it. Thank you mps, c705. It is nice too there is README 2019-05-02 21:24:36 bougyman: runit-init or s6 don't have the facilities to do rc.local 2019-05-02 21:24:47 seriously, who's choice is openrc. 2019-05-02 21:24:52 north1: that's completely false. 2019-05-02 21:25:00 in runit, you put it in /etc/runit/2 2019-05-02 21:25:06 in s6 there's a facility as well 2019-05-02 21:26:16 bougyman: so you write your own mechanism 2019-05-02 21:26:20 so openrc also has it 2019-05-02 21:26:21 north1: I've packaged runit-init for multiple flavors, and I follow s6 and the supervision mailing list closely. rc.local is a solved problem. 2019-05-02 21:26:36 bougyman: yes, openrc solved it for openrc using local.d 2019-05-02 21:26:42 runit solved in their own way 2019-05-02 21:27:07 In any case, openrc is as antiquated in design as SysV. 2019-05-02 21:27:23 That's why I question its use in something generally modern in all other ways (alpine) 2019-05-02 21:27:58 that said, I run all my alpine with runit-init, not openrc. So it's not like it's a hard forklift. 2019-05-02 21:27:59 nothing is too old, as it serves its funcs 2019-05-02 21:28:24 just a pita when the package manager isn't in sync with the init/supervision system. 2019-05-02 21:28:41 I thought at one point alpine was considering s6? 2019-05-02 21:29:28 Maybe 66 would be a better choice somewhere down the line 2019-05-02 21:30:11 s6-rc is nice, but a bit clunky to use IMHO 2019-05-02 21:35:50 What's the problem with openrc? You say "antique design" I say "well-tested" 2019-05-02 21:37:05 c705: sysVinit-ish scripts tend to be kinda dirty, tracking services via PIDfiles is kind of terrible compared to using proper supervision like runit/s6-rc do 2019-05-02 21:38:11 Although https://lwn.net/Articles/773459/ might better that 2019-05-02 23:02:43 bougyman: add supervisor=s6 to your conf.d file and openrc will use s6 for you afaik 2019-05-02 23:07:37 https://wiki.alpinelinux.org/wiki/Writing_Init_Scripts#Services_supervised_by_s6 2019-05-02 23:08:15 ^^ why is the "${s6_service_path}" variable not being used when checking / creating the link ? 2019-05-02 23:13:05 https://gist.github.com/tcely/f35ecb97e760e4e6ce379113041247ed/revisions 2019-05-03 02:50:45 hey, so i'm trying to mount a nfs share in my alpine box, but i'm getting connection refused. it mounts in my other linux vms. any ideas? 2019-05-03 06:44:29 Any way to list packages that were manually installed ? 2019-05-03 07:25:19 north1: `grep '=' /etc/apk/world` and ignore garbage after '>' 2019-05-03 07:25:43 that where I reached :) 2019-05-03 07:26:10 thanks for the tip but i doubt i have only 5 packages in world... 2019-05-03 07:27:16 ah, cat is enough 2019-05-03 07:27:17 thanks :D 2019-05-03 07:27:18 oh, so I didn't understood your question probably 2019-05-03 07:27:47 you mean from local signed repo? 2019-05-03 07:28:24 I thought installed with '--allow-untrusted' 2019-05-03 07:28:40 no, all packages manually installed 2019-05-03 07:29:23 ah, then have no idea 2019-05-03 10:52:54 hey 2019-05-03 10:53:04 where can I find dmi-decode ? 2019-05-03 10:53:59 ryszard: apk add dmidecode 2019-05-03 10:55:21 hmm 2019-05-03 10:55:33 yeah, I made mistake 2019-05-03 10:55:41 was searching for dmi-decode 2019-05-03 10:55:43 :"p 2019-05-03 10:56:17 and made second mistake when used pkgs.alpinelinux.org 2019-05-03 10:56:24 when not used asterisk 2019-05-03 10:56:35 mps: thanks 2019-05-03 10:56:57 np, :) 2019-05-03 14:03:51 is you ntp working ? 2019-05-03 14:04:10 when I'm trying to start ntp it returns: tls connect failed: 172.217.16.4 (www.google.com): ssl verify memory setup failure 2019-05-03 14:32:57 Hi. Is there a possiblility to list files of a not installed package via apk? 2019-05-03 14:39:35 pkg.a.o ? 2019-05-03 14:39:51 not via apk 2019-05-03 14:40:08 but there is a tool that queries pkgs.a.o via shell 2019-05-03 14:44:08 <_ikke_> apk-files 2019-05-03 14:45:19 i've got a little script that runs on an embedded workstation: http://ix.io/1HUr 2019-05-03 14:46:08 one issue i'm having is if /dev/ttyUSB0 goes away those xdotool lines will constantly exec 2019-05-03 14:47:03 i can't seem to find a way to pass exit codes through pipes, which would be ideal here 2019-05-03 14:52:21 geosmin: how do you mean? set -o pipefail or something else? 2019-05-03 14:54:14 tcely: i was trying to keep it POSIX 2019-05-03 14:55:52 i could add a line above the three main ones 'if [-f /dev/ttyUSB0]: ' and then after '[while ! [ -f /dev/ttyUSB0 ]:; sleep 2;' but that seems a bit ugly 2019-05-03 15:45:44 when I have removed 's' from 'https' to 'http' of 'constraints from "http://www.google.com"' in /etc/ntpd.conf it started to work 2019-05-03 15:49:29 why then in defualt installation https is used when it doesn't work ? 2019-05-03 16:03:11 ryszard: ntpd to google.com ? 2019-05-03 16:06:12 ryszard: works fine for me on 3.8 2019-05-03 16:06:39 prob. messed up your ssl installation 2019-05-03 16:06:47 ssl_client 2019-05-03 16:19:52 I have 3.9 2019-05-03 16:20:25 how can I properly stop and start ntpd on alpine ? 2019-05-03 16:22:29 service ntpd stop || service ntpd start || service ntpd restart... 2019-05-03 16:27:24 TBK[m]: thanks 2019-05-03 16:27:57 and as I guess 'service' uses OpenRC files from /etc/init.d ? 2019-05-03 16:34:41 are you sure that your openntpd works with 'https'? I have found bug reported on Alpine https://bugs.alpinelinux.org/issues/9635 2019-05-03 16:51:47 ryszard: service or rc-service, yes (/etc/init.d/openntpd). Can't answer the second part. I do not use openntpd. 2019-05-03 17:14:17 interesting, I have installed libressl - openntpd started to work 2019-05-03 17:14:31 after remove of libressl it still working after restrat 2019-05-03 17:15:12 I'm little confused 2019-05-03 17:15:31 will reboot the machine and check again 2019-05-03 17:18:40 its a weird magic 2019-05-03 17:18:43 still works 2019-05-03 17:20:43 maybe libressl installation overwrite sth ? 2019-05-03 17:21:03 ok, another question 2019-05-03 17:21:11 what is base ssl implementation in Alpine ? 2019-05-03 17:22:51 ^ ncopa might answer this question: libressl with openntpd on 3.9 ? 2019-05-03 17:23:42 ryszard : that what I guess earlier. afaik, openssl should be used instead of libresssl 2019-05-03 17:25:04 ah it is libtls-standalone 2019-05-03 17:25:11 ryszard: i think it happens because openssl has a 'replaces=libressl' 2019-05-03 17:25:16 the openssl libs 2019-05-03 17:25:24 so openssl will overwrite some things 2019-05-03 17:26:10 https://github.com/alpinelinux/aports/pull/5406#issuecomment-489084440 2019-05-03 17:26:29 we need to fix libressl/openssl so they can be installed in parallel 2019-05-03 17:26:42 apk list --installed *ssl* returned: libssl1.1-1.1.1b-r1 x86_64 {openssl} (OpenSSL) and ssl_client-1.29.3-r10 x86_64 {busybox} (GPL-2.0) 2019-05-03 17:28:04 then can I just replace openssl with libressl or other packages have dependencies on libssl1.1-1.1.1b-r1 anyway ? 2019-05-03 17:28:51 no, you cannot just replace openssl with libressl 2019-05-03 17:28:56 they are not ABI compatible 2019-05-03 17:29:05 uhm 2019-05-03 17:29:34 this one is overwritten: etc/ssl/openssl.cnf 2019-05-03 17:29:49 not sure if they are compatible 2019-05-03 17:30:31 ca-certificates will overwrite etc/ssl/cert.pem 2019-05-03 17:30:40 ok then in my scenario looks like installation of libressl overwritten etc/ssl/openssl.cnf file and it looks like it was broken/corrupted/wrong because now openntpd works 2019-05-03 17:31:08 if you installed the packages that overwrote the mentioned files 2019-05-03 17:31:11 and then removed them 2019-05-03 17:31:16 then those files woudl be missing 2019-05-03 17:31:38 (re)-installing libressl would make it work 2019-05-03 17:31:39 how could those be missig when I have vanilla installation 2019-05-03 17:31:39 but... 2019-05-03 17:31:46 im just guessing here 2019-05-03 17:31:56 it may be that there is some broken dependency in openntpd 2019-05-03 17:32:38 I just flagged openntpd a new version was released 18 days ago + it needs an openrc subpkg 2019-05-03 17:32:38 yeah for me it is also incomprehensible, but instalaltion and deinstallation of libressl fixed the problem, 2019-05-03 17:51:08 apk upgrade does not like upgrading packages installed as `cmd:` - what to do about it/ 2019-05-03 17:51:29 <_ikke_> not like as in? 2019-05-03 17:51:44 example: https://tpaste.us/lZm5 2019-05-03 17:51:50 I'm asking for a friend, fwiw 2019-05-03 17:51:56 :) 2019-05-03 17:53:27 apk --force-broken-world fix -r? 2019-05-03 17:53:31 or something along those lines? 2019-05-03 17:59:17 we fixed it, nevermind :) 2019-05-03 18:12:10 <_ikke_> what was the fix? 2019-05-03 22:02:43 _ikke_: apk --force-broken-world del cmd: cmd: 2019-05-03 22:02:55 and then reinstalled them by the package name without using so: / cmd: 2019-05-04 01:14:33 is there a usermod command in alpine? 2019-05-04 01:15:01 I need to give a particular user a specific uid for file access 2019-05-04 01:15:38 vizo: Could adduser -u 2019-05-04 01:16:17 hmm thanks 2019-05-04 01:26:38 I notice ssh-agent isn't started when user logs in. What is the cononical way to enable this? One-liner in .profile or is there a better way? 2019-05-04 01:27:32 hmm how do I add users to a group or dos it have to be done when creating a user? 2019-05-04 01:29:10 vizo: `addgroup` 2019-05-04 01:31:25 Cogitri: thanks 2019-05-04 01:40:27 err where do I put polkit rules? localauthority.confd ? 2019-05-04 01:40:49 trying to do ths https://www.poftut.com/use-virt-manager-libvirt-normal-user-without-root-privileges-without-asking-password/ 2019-05-04 01:41:51 I don't think that works with the current version of polkit we have in our repos 2019-05-04 01:42:26 hmm suggestions? 2019-05-04 01:43:18 ah found it 2019-05-04 01:43:32 at the end of the wiki 2019-05-04 01:43:43 Wait for https://github.com/alpinelinux/aports/pull/7374 to be merged, I guess 2019-05-04 01:43:55 why so differnet than everything else? 2019-05-04 01:45:35 well other distros I mean 2019-05-04 01:45:50 curious coming over from gentoo 2019-05-04 01:51:30 You mean with the stuff like addgroup? Because Alpine uses busybox instead of coreutils 2019-05-04 01:52:33 ah that and the polkit thing 2019-05-04 01:52:54 I plan on using this OS for my KVM host 2019-05-04 01:54:21 Well, polkit is due to us not having upgraded it yet, so that should change soon-ish hopefully 2019-05-04 01:54:34 oh cool 2019-05-04 01:55:02 You'll either have to cope with busybox or install tools which replace part of its functionality with something you're more used to (coreutils, findutils, grep etc.) 2019-05-04 01:55:19 ok good to know 2019-05-04 01:55:45 since I want to keep it barebones as possible I can wait 2019-05-04 02:05:16 hmm the ssh part of the https://wiki.alpinelinux.org/wiki/KVM does not seem to work 2019-05-04 02:11:40 nevermind user error 2019-05-04 03:08:36 is there away to change to a differnt version of ps in alpine? 2019-05-04 03:08:51 Install procps 2019-05-04 03:09:33 north1: thanks 2019-05-04 03:12:18 hot damn I just deployed alpine as a opennebula kvm node 2019-05-04 03:12:34 thanks you guys!! 2019-05-04 03:16:08 but sisnce I forgot to tak notes I have to do it all over again 2019-05-04 08:20:03 How is Alpine so fast? 2019-05-04 08:20:53 I usually use a CuoDuo machine for a point of reference, and only on Alpine, no other distro has even just the framebuffer alone been so fast it even looks different when text flows across it 2019-05-04 08:21:24 we only ship the bare-bones with our images, and apk is tiny + written in C 2019-05-04 08:21:35 there are a couple other factors that play into it as well, for that sake 2019-05-04 08:21:56 but I've run other minimal distros as well 2019-05-04 08:22:01 such as Debian netinst 2019-05-04 08:22:04 and gentoo 2019-05-04 08:22:36 have not seen the same levels of speed (Gentoo coming very close) 2019-05-04 08:23:15 could it have something to do with the musl libc being used? 2019-05-04 08:27:51 hi. whats matter with php pkg? its only 5 and in 3.3 branch 2019-05-04 08:28:11 https://wiki.alpinelinux.org/wiki/Nginx_with_PHP#PHP7_Installation is it dropped? 2019-05-04 08:34:21 oh. its in community. sorry 2019-05-04 08:35:09 But pacman is also written in C right? How come apk is faster? 2019-05-04 08:35:31 agris: I believe musl libc is one of the cause. 2019-05-04 08:36:05 pickfire, How did the Alpine team get musl libc working for the entire OS? 2019-05-04 08:36:12 Gentoo is still having trouble with that 2019-05-04 08:36:32 Lots of patching, mostly 2019-05-04 08:36:33 No, I have trouble using IME on alpine with musl libc. 2019-05-04 08:36:37 I can compile a cross-compiler or a program or two, but the whole system 2019-05-04 08:36:46 agris: Some patching really 2019-05-04 08:36:50 if I try the whole system it's littered with bugs 2019-05-04 08:36:52 gentoo is fine for more than a year 2019-05-04 08:37:04 agris: I don't think Alpine does cross, Void Linux does lots of crossing though 2019-05-04 08:37:40 And no, musl most likely isn't the cause for APK being faster. Musl can be a bit faster sometimes, but overall hasn't received as much optimization as glibc has. 2019-05-04 08:37:57 (And even so, it doesn't make _that_ much of an impact) 2019-05-04 08:40:36 not talking about apk specificly 2019-05-04 08:40:54 more talking about overall system responsiveness and the framebuffer 2019-05-04 08:42:42 Hello everyone 2019-05-04 08:43:09 what CFLAGs is alpine built against 2019-05-04 08:43:17 I'm in trouble compiling a compression/archiving program under Alpine Linux 2019-05-04 08:43:30 agris: -Os -fomit-framepointer 2019-05-04 08:43:35 it's the (famous?) zpaq by Matt Mahoney 2019-05-04 08:43:43 it compiles correctly 2019-05-04 08:43:53 but coredumps during use :( 2019-05-04 08:44:06 hmm 2019-05-04 08:44:09 I've used it on a lot of other systems 2019-05-04 08:44:12 optimize small 2019-05-04 08:44:21 I've never used that 2019-05-04 08:44:27 compiled under all linuxes 2019-05-04 08:44:31 but this one fails 2019-05-04 08:44:51 my guess it's something about the musl library? 2019-05-04 08:44:56 Hm, It's kinda hard to debug that via IRC, Wyk72 2019-05-04 08:45:00 I know 2019-05-04 08:45:09 maybe someone has a trick :) 2019-05-04 08:45:24 it hangs opening a new pthread 2019-05-04 08:45:55 someone said it's because the stack size is wrong (musl allocates 80Kb stack, while glibc allocates 8Mb) 2019-05-04 08:46:06 but I'm no expert 2019-05-04 08:46:41 Thread 2 "zpaq" received signal SIGSEGV, Segmentation fault. [Switching to LWP 2317] 0x00007ffff7fac5c0 in libzpaq::Predictor::Predictor(libzpaq::ZPAQL&) () (gdb) 2019-05-04 08:47:21 also guys. '# CONFIG_FEATURE_SH_NOFORK is not set' why not? 2019-05-04 08:47:30 in busybox 2019-05-04 08:50:51 Wyk72: What alpine version do you use? 2019-05-04 08:52:20 On newer musls you can compile with `-Wl,-z,stack-size=2097152` (or any other number) for bigger stacks 2019-05-04 08:52:41 tried that too 2019-05-04 08:52:48 but still coredumps :( 2019-05-04 08:52:56 will try that again 2019-05-04 08:53:19 I think I that needs musl >=1.21 though 2019-05-04 09:19:54 it's alpine 3.9.3 2019-05-04 09:20:00 64bit 2019-05-04 09:26:10 iirc, stack size is increased in musl 1.1.22 2019-05-04 09:26:54 Yup, that changed the default a bit, but it's still pretty small 2019-05-04 09:27:45 if the program have bug stack size will not help however big 2019-05-04 09:28:55 re earlier talk about speed, ime musl is a noticable faster than glibc 2019-05-04 09:30:06 openwrt switched to musl some time ago and it is still fast and small 2019-05-04 09:32:39 my interest to Alpine started because of musl, I tested some programs on debian building them with glibc and musl and have seen a difference 2019-05-04 09:47:13 when I issue apk upgrade it returns: 3 errors; 1973 MiB in 519 packages 2019-05-04 09:47:19 how can I check what those errors are ? 2019-05-04 09:47:35 can't find any related option in its help 2019-05-04 09:48:47 try `apk fix` 2019-05-04 09:49:25 but how can I check what is wrong ? 2019-05-04 09:51:40 try with '-v' 2019-05-04 09:52:39 it usually means some pkg tried to overwrite a file in package which would be removed later. 2019-05-04 09:53:05 it still coredumps when opening a new pthread and I have no idea why 2019-05-04 09:53:09 no matter the stack size 2019-05-04 09:53:21 --verbose doesn't help 2019-05-04 09:53:32 it's the only program I need for backups 2019-05-04 10:56:49 I'm trying to boot Alpine Linux on NanoPi NEO2 (Allwinner H5) board but I'm always getting kernel panic on bootup. 2019-05-04 10:57:15 I'm using the Generic ARM rootfs for aarch64, I compile my own u-boot and dtb 2019-05-04 10:58:02 here's the log, I appended `earlycon console=ttyS0,115200n8` to the kernel cmd. https://hastebin.com/vipupaqeke.log 2019-05-04 11:00:22 granra: sunxi-mmc 1c0f000.mmc: ignoring dependency for device, assuming no driver 2019-05-04 11:00:31 no driver for mmc 2019-05-04 11:00:50 did you looked is it built for this kernel 2019-05-04 11:05:21 you should look is the sunxi-mmc in initramfs, and if not add it there, I presume you know how to do that 2019-05-04 11:09:06 mps: Thanks, I'll take a look 2019-05-04 11:15:37 mps: initramfs contains /lib/modules/4.19.34-0-vanilla/kernel/drivers/mmc/host/sunxi-mmc.ko 2019-05-04 11:17:01 hmm, then it should be loaded by kernel. maybe there is bug somewhere 2019-05-04 11:17:38 I don't have such board so can't test 2019-05-04 11:22:04 I understand, I've been having a bit trouble with this board in fedora and opensuse where it can't read solv files in their package managers (dnf and zypper) but at least I manage to boot it. 2019-05-04 11:22:12 how would I go about debugging this on my own? 2019-05-04 11:25:05 I usually build u-boot and kernel with needed storage drivers in kernel, not as module 2019-05-04 11:26:58 this way it is a lot easier, don't have to play with initramfs for every change 2019-05-04 11:27:07 I'll give that a try, thanks 2019-05-04 11:30:31 have you heard about it? https://www.bleepingcomputer.com/news/software/firefox-addons-being-disabled-due-to-an-expired-certificate/ 2019-05-04 11:30:40 <_ikke_> yes 2019-05-04 11:50:19 nice fail :| 2019-05-04 12:17:15 Seems like they've rolled out a fix now though 2019-05-04 12:36:02 <_ikke_> You need to have some kind of remote management plugin running for that hough 2019-05-04 12:36:03 <_ikke_> though 2019-05-04 12:36:18 <_ikke_> Normandy 2019-05-04 12:36:24 <_ikke_> https://wiki.mozilla.org/Firefox/Normandy/PreferenceRollout 2019-05-04 13:06:35 Wyk72: sounds like bug in application. Do you have a backtrace? 2019-05-04 13:07:44 Or link to the sources where I segfaults? I’m afk but can maybe look at the source if there is something evident 2019-05-04 13:08:17 You can also report it upstream 2019-05-04 16:42:01 looks like libelf is abandoned upstream and source should be set to some net archive 2019-05-04 16:42:39 wrong channel, sorry 2019-05-04 18:05:15 is there already a release date for the 3.9.4 release? 2019-05-04 18:16:27 <_ikke_> I think 3.10 has priority now 2019-05-04 18:16:50 still better...what's the ETA? 2019-05-04 18:17:15 <_ikke_> Soon (TM) 2019-05-04 18:17:34 <_ikke_> The builders have just been set up, so they are now building all the packages 2019-05-04 18:18:46 Thanks, that's good to know! 2019-05-04 19:21:16 hi. help. openrc during boot displays "caching service dependencies" and thats it. no action for long time. how to debug ? 2019-05-04 19:24:08 sr20det: https://bugs.alpinelinux.org/issues/9960 2019-05-04 19:25:47 but it taking hours in 3ghz vps. anyway thanks 2019-05-04 19:26:31 you can reboot it in single mode 2019-05-04 19:26:56 chroot into disk and add haveged to runlevel 2019-05-04 19:27:37 why is such a thing is default? 2019-05-04 19:29:41 sr20det: kernel developers decided to move this issue to distribution packagers 2019-05-04 19:30:01 and distro packagers move the issue to users 2019-05-04 19:30:18 and users move it back to us :) 2019-05-04 19:30:26 :D 2019-05-04 19:30:53 but its a good explanation from mps. 2019-05-04 19:32:00 we can start good ping-pong with users ;) 2019-05-04 19:34:36 back to seriousness, IMO decision of trusting HW RNG should be given to users and not distributors or kernel dev's 2019-05-04 19:36:08 mps: but like was suggested in that bug report, I think an RNG source should be suggested in the installer 2019-05-04 19:37:54 i think there is no reason to lock up login shell . just dont start sensitive services 2019-05-04 19:37:58 doggone: right, I proposed there to offer choice to user during installation or upgrade 2019-05-04 19:38:47 or to inform user about issue during install or upgrade, at least 2019-05-04 19:39:00 and the fck write to log what is happening 2019-05-04 19:39:24 I'm bad in shell scripting so don't dare to add this option to installer 2019-05-04 19:41:03 almost everyone will use ssh and that requires entropy 2019-05-04 19:41:15 so almost every installation will be 'sensitive' to it 2019-05-04 19:41:58 and I guess it's the kernel that should notify that its "entropy-starved" 2019-05-04 19:43:03 there could be a "setup-entropy" script analogous to "setup-ntp" or "setup-sshd" 2019-05-04 19:43:58 printf '#!/bin/sh\necho please wait few hours for rng shenanigans\nexec /bin/real.init' > /bin/init 2019-05-04 19:45:06 echo "do you trust your HWRNG manufacturer [y/n]" 2019-05-04 19:51:24 random.trust_cpu do nothing 2019-05-04 20:31:43 We're using Alpine Java containers such as openjdk:8u181-jre-alpine. I assume the OpenJDK is and will be packaged into these containers and nothing that requires the Subscription License which Oracle suddenly imposes. Can anyone comment, please. 2019-05-04 20:32:38 For Java 8, can we go beyond version 202 with Alpine for a commercial project? 2019-05-04 20:33:41 IANAL, but hope such issues will not appear 2019-05-04 20:34:06 mps: it's oracle, you never know 2019-05-04 20:35:16 (I have a hope for java, to disappear ;) ) 2019-05-04 20:36:35 doesn't change the fact people use it 2019-05-04 20:36:40 a lot of people 2019-05-04 20:36:59 mps, did you wish Java or Oracle to disappear? 2019-05-04 20:37:26 yes, I know, even I use it but dislike it 2019-05-04 20:37:31 for me, both, but primarily oracle 2019-05-04 20:37:46 anyway - I may have misunderstood, but the latest openjdk8-jre package we have is 8.212.04 (in 3.9 and edge) 2019-05-04 20:37:52 Lol, good then. 2019-05-04 20:38:38 ok, I think openjdk will stay open 2019-05-04 20:39:04 Yes, Oracle allows use of <=202 without subscription license, but this is OracleSDK. 2019-05-04 20:39:54 They may decide to do the same to other JDK vendors if they should win the case against Google. 2019-05-04 20:39:54 yeah, this isn't Oracle JDK, and iirc these aren't Oracle's OpenJDK builds either 2019-05-04 20:40:08 oh yeah, i don't doubt it 2019-05-04 20:40:16 knowing oracle, anyway 2019-05-04 20:42:03 So "let's copyright APIs and kill open source in favor of the big enterprises". I really hope the Supreme Court is smart enough to not let it happen. 2019-05-04 20:42:09 Thank you guys. 2019-05-04 20:42:47 I mean, it's not like Oracle hasn't pulled *** moves in the past, but yeah 2019-05-04 20:42:57 we're not lawyers, and we're not psychic - we don't know what oracle will do 2019-05-04 20:43:24 danieli: nice told 2019-05-04 20:43:36 Yeah, I understand, sure. 2019-05-04 20:43:47 Google also thought they were on the safe side. 2019-05-04 20:43:50 I don't see open-source Java disappearing any time soon 2019-05-04 20:44:40 I hope you are right. I wouldn't use it for new projects anymore, though. 2019-05-04 20:44:58 The risk is high with Oracle at the helm. 2019-05-04 20:45:07 the issue is with Android's java specifically, if that's what you're aiming at 2019-05-04 20:45:33 Yes, I know. These are currently two separate issues. 2019-05-04 20:45:38 mm. 2019-05-04 20:45:46 it's a whole mess 2019-05-04 20:46:25 but as to the future of java, i think you'd be a lot wiser if you asked in some java community 2019-05-04 20:46:33 But if they win one case, they will probably go after the others with similar reasons. Look at Apache Harmony, for example. They had to give up. 2019-05-04 20:47:07 Yeah, will do. Thanks! 2019-05-04 20:47:35 bye 2019-05-04 20:48:45 there was some arm chips that could accelerate java? 2019-05-04 21:30:22 is it possible to install alpine-linux without usb/cd? 2019-05-04 21:30:48 sr20det: ya jazelle then thumbee but i don't think it was used much outside of old feature phones etc. https://en.wikipedia.org/wiki/Jazelle 2019-05-04 21:31:16 boot into some sort of RAM kernel that installs the OS? 2019-05-04 21:31:37 markweston: https://boot.alpinelinux.org/ ? 2019-05-04 21:34:39 i mean, yeah, you could netboot 2019-05-04 21:35:11 or /part ;) 2019-05-04 22:02:15 uhm, even faketime doesn't help with mozilla addons fiasco 2019-05-04 22:05:18 mps: the extension thing? 2019-05-04 22:05:59 I think that should be fixed already, shouldn't it be? 2019-05-04 22:06:52 Ah, https://discourse.mozilla.org/t/certificate-issue-causing-add-ons-to-be-disabled-or-fail-to-install/39047 2019-05-04 22:07:10 You have to enable Firefox Preferences -> Privacy & Security -> Allow Firefox to install and run studies. 2019-05-04 22:07:58 If changing the date before expiration doesn't work none of the other work around options were anywhere near as easy. 2019-05-04 22:08:09 Cogitri: oh nice 2019-05-04 22:11:43 I read this about studies, but didn't tried because someone wrote that the account is needed on mozilla for this 2019-05-04 22:12:41 No, it's not required 2019-05-04 22:13:03 ah, thanks, will try now 2019-05-04 22:13:05 Just remember to disable it again afterwards, that switch allows Mozilla to remotely install plugins for testing purposes 2019-05-04 23:42:44 I hopy this is a catalyst to give life to a better browser 2019-05-04 23:43:04 Uh? 2019-05-04 23:43:17 I hope* 2019-05-04 23:43:22 the firefox fiasco right now 2019-05-04 23:43:45 choosing a browser is like choosing between the plague and cholera 2019-05-04 23:44:29 either you pick something like Falkon or qutebrowser, that can't really keep up with web standards and don't support extensions like umatrix and https everywhere 2019-05-04 23:44:51 or you choose Firefox or Chromium and the likes, which are riddled with telemetry, DRM and other junk 2019-05-04 23:45:38 My hope is that somebody one day implements - Atlantis https://www.microsoft.com/en-us/research/publication/atlantis-robust-extensible-execution-environments-for-web-applications/ 2019-05-04 23:47:07 doggone: Well, you can disable EME and Telemetry in FF though. 2019-05-04 23:47:46 QtWebengine/Webkit2gtk based browsers do slowly catch up though, Epiphany is kind of nice to use already (but extension support keeps me with other browsers) 2019-05-04 23:57:49 Cogitri: well, you can disable many things (take a look at pyllyukko's user.js and despair) but the average user won't do that and get screwed over 2019-05-04 23:58:43 and god forbid you take an OpenBSD-ish look at the code (which is impossible since it consists of millions lines of code) and scrutinize all the ways they mess up 2019-05-04 23:58:45 True that 2019-05-05 00:05:18 I tend to stick with Firefox, Opera or Vivaldi 2019-05-05 00:07:20 I tend to not use proprietary browsers 2019-05-05 00:07:57 I use Chromium right now though 2019-05-05 08:52:41 hey, anybody recently tried to install virtualbox guest additions in alpine? 2019-05-05 08:52:54 I followed https://wiki.alpinelinux.org/wiki/VirtualBox_shared_folders 2019-05-05 08:52:56 doesn work 2019-05-05 11:45:09 what is difference between 'apk list *pkg_subname*' and 'apk list --available *pkg_subname*' 2019-05-05 11:46:17 <_ikke_> from the help output, I would assume that with --availabe, it would not show installed packages 2019-05-05 12:01:21 ok I made experiments 2019-05-05 12:01:47 there is package named openssl-dbg, I don't have it in my system 2019-05-05 12:02:12 'apk list --available openssl-dbg' doesn't show it when 'apk list openssl-dbg' shows 2019-05-05 12:02:48 and for opposite situation with already installed package 2019-05-05 12:02:57 named openssl-doc 2019-05-05 12:03:24 'apk list --available openssl-doc' doesn't show it when 'apk list openssl-doc' shows (but with [installed] suffix) 2019-05-05 12:03:37 then result is pretty same 2019-05-05 12:06:28 where can I find Alpine Linux apk definition of 'available' ? 2019-05-05 12:13:01 ryszard: apk-tools is, unfortunately, poorly documented currently you could take a look at the source code though 2019-05-05 12:13:06 https://git.alpinelinux.org/apk-tools/tree/src/list.c#n147 2019-05-05 12:16:15 yeah the help is sometimes funny '-a, --available List available packages only' it could be anything like '-r, --rounded List rounded packages only' 2019-05-05 12:16:18 :p 2019-05-05 12:17:48 I see 'APK_REPOSITORY_CACHED' then as I guess -a will list only packages that have been donwloaded to some local storage/cache for offline use ? 2019-05-05 15:33:27 anyone have success in running alpine armhf or armv7 under qemu, with disk image 2019-05-05 15:36:01 Not me at least :c 2019-05-05 15:36:12 qemu-system-arm -bios u-boot.bin .... starts and loads kernel and initramfs from disk image but after boot it cannot load disk driver 2019-05-05 15:36:37 this works for aarch64 2019-05-05 15:36:51 maybe I should ask u-boot people 2019-05-05 15:38:24 Cogitri: just this part missing to have script which install arm{hf,v7} in qemu 2019-05-05 15:41:02 seeing your intense work on alpine I wanted to prepare these two for you 2019-05-05 15:43:03 Oh, thank you! :) 2019-05-05 15:43:12 That's very nice of you :) 2019-05-05 15:45:14 you are welcome, want to help where I can, but this disk drivers doesn't work or I don't understand difference between armhf and aarch64 2019-05-05 15:45:48 maybe I should try with external kernel/initramfs/dtbs 2019-05-05 15:50:51 yes, it works with external kernel etc 2019-05-05 16:04:07 eh, I think I found problem, arm{hf,v7} vanilla kernel don't have PCI drivers enabled 2019-05-05 19:38:07 When does `apk upgrade` also fetch the APKINDEX? Just after a certain time or is there more logic behind it? 2019-05-05 19:38:20 <_ikke_> I think only time 2019-05-05 19:38:26 time 2019-05-05 19:38:29 <_ikke_> which is a fairly recent addition 2019-05-05 19:39:01 Ah, alright. Thanks :) 2019-05-05 19:41:21 Cogitri: https://git.alpinelinux.org/apk-tools/commit/?id=8a28c6d0d4f81406c67e76b2bcdbe491aa734717 2019-05-05 19:46:52 Cool 2019-05-05 19:48:00 Neat, FF 66.0.4 is out now which should fix the addon fiasco for everyone 2019-05-05 20:30:35 seems one of my buntus got fixed automatically, another one didn't 2019-05-05 20:32:58 cogitri, in alpine? 2019-05-05 20:33:50 i see 62 in main, 65 in testing 2019-05-05 20:33:52 No, in upstream, I don't think it's in the repos yet 2019-05-05 20:34:13 oh no maybe 62 is just installed 2019-05-05 20:34:19 i don't know how to read apk policy output :-p 2019-05-05 20:34:24 <_ikke_> TBB: buntus? 2019-05-05 20:36:13 ubuntu boxes i would guess 2019-05-05 20:50:52 those exactly 2019-05-05 22:48:11 has anyone has ethernet issues when using alpine with raspberry pi before? 2019-05-05 22:48:35 no ip on boot, udhcpc gives "Network is down" error 2019-05-05 22:48:54 using a raspberry pi 3+ 2019-05-05 22:48:58 any known missing drivers? 2019-05-05 22:52:20 i don't recall whether i ever tried it with the wired ethernet but i have a pi3 working with wifi 2019-05-05 22:53:03 hmm, so wireless drivers should be fine 2019-05-05 22:53:20 actually i think i did during install 2019-05-05 22:53:29 just because getting wifi up during install is a pain 2019-05-05 22:53:45 but i'm not 100% sure 2019-05-05 22:54:07 must have gotten wpa_supplicant from some connection, haha 2019-05-05 22:54:15 unless you added the package to the image 2019-05-05 22:57:46 i didn't, so yeah i must have used ethernet 2019-05-05 22:58:19 sadly i'm not using the pi at all still 5 months later :/ 2019-05-05 22:59:23 got it for retro gaming but alpine arm was missing lots of the dependency libs for emulators so i went thru the pain of building them myself, then most of the ones i wanted to run were were to slow on it :( 2019-05-05 22:59:44 and the more mainstream (raspian-based) approaches just take forever to boot :-p 2019-05-05 23:02:02 oh, wpa_supplicant is actually packaged with the image 2019-05-05 23:02:10 i'll give wireless a shot 2019-05-05 23:02:29 ah then i'm not sure again :-p 2019-05-05 23:09:06 success! 2019-05-05 23:12:14 :) 2019-05-06 07:16:19 seems like some recent update related to polkit broke my ability to mount external drives from pcmanfm :-( 2019-05-06 07:17:08 pkaction just says 2019-05-06 07:17:09 Error getting authority: Error initializing authority: Error calling StartServiceByName for org.freedesktop.PolicyKit1: GDBus.Error:org.freedesktop.DBus.Error.Spawn.ChildExited: Launch helper exited with unknown return code 1 2019-05-06 07:21:42 You're on edge, yes? Mind running `/usr/lib/polkit-1/polkitd`? 2019-05-06 07:21:54 yep 2019-05-06 07:22:15 kay, gotta reboot 2019-05-06 07:23:49 think i figured it out myself 2019-05-06 07:23:59 apparently, it now requires a user "polkitd" 2019-05-06 07:24:12 so i read the APKBUILD and recreated the same users, groups and permissions 2019-05-06 07:24:41 now pkaction works okay 2019-05-06 07:26:21 and mounting works, too 2019-05-06 07:49:36 Oh? Doesn't polkit create that user now? :o 2019-05-06 07:51:52 apparently, the user and group are added in pre-install, and permissions are changed in the APKBUILD 2019-05-06 07:52:02 but neither of them happened when i updated the package 2019-05-06 07:52:22 dbus was complaining that there's no such user 2019-05-06 07:53:07 Ah, I guess dbus might have to be restarted for stuff to go into effect 2019-05-06 08:56:41 mornin guys, I just rebuilt an image I haven't touch in like a month or two and now it throws a usermod not found error. it did exist/work the last time I used it. alpine 3.9 that it, did it recently change? 2019-05-06 08:57:24 <_ikke_> steven: is the shadow package installed in that image? 2019-05-06 08:59:30 if you tell me how to check it I'll tell you right away 2019-05-06 08:59:47 apk info | grep shadow 2019-05-06 08:59:50 apk info gives me a installed size 2019-05-06 09:00:09 that returns nothing TBB 2019-05-06 09:00:12 so I assume it is not 2019-05-06 09:00:36 yup, it's not 2019-05-06 09:00:46 <_ikke_> apk info -e shadow && echo yes || echo no 2019-05-06 09:01:17 <_ikke_> apk list -I shadow 2019-05-06 09:01:34 was that removed recently TBB _ikke_ ? like in a .* upgrade or something? 2019-05-06 09:01:49 <_ikke_> It's not part of base install 2019-05-06 09:02:07 nope, it's been in the shadow package for ages, and as ikke said, base install has busybox user utilities instead 2019-05-06 09:02:21 well was it at some point in time? (I am not using alpine directly but rather an image that uses alpine so it might as well been an issue with that base image) 2019-05-06 09:02:40 cool cool, so I blame the base package then, install it manually and say thanks for the help :D 2019-05-06 09:02:40 I don't think it was 2019-05-06 09:02:41 <_ikke_> Yes, might be that it got changed in the image 2019-05-06 09:02:50 s/package/image/ 2019-05-06 09:03:06 oh yeah, what image are we talking about? 2019-05-06 09:03:36 <_ikke_> I assume a docker image 2019-05-06 09:04:24 obviously I am talking about docker, yes. sorry. I just mentally have that connection and always tend to ignore the fact that ppl use it outside of docker, too 2019-05-06 09:05:07 yeah, for me, for example, the first association is a squashfs format filesystem image :) 2019-05-06 09:05:21 (worked on those for 6 years) 2019-05-06 09:05:54 <_ikke_> I associate it with an iso image 2019-05-06 09:11:40 alrighty, working agian. thanks guys!¬ 2019-05-06 11:14:19 anyone here? ;D 2019-05-06 11:14:39 <_ikke_> nope, nothing to see here 2019-05-06 11:14:44 hehe 2019-05-06 11:14:45 <_ikke_> ;-) 2019-05-06 11:15:27 Do you happen to know what the problem is when the start-stop-daemon(init scripts) returns : ¨failed to exec %s : No such file or directory 2019-05-06 11:16:01 <_ikke_> Sounds like an error in the init script 2019-05-06 11:16:09 yeah but its a simple script 2019-05-06 11:16:13 and the file is there 2019-05-06 11:16:32 <_ikke_> the fact that is %s means that something is going wrong 2019-05-06 11:16:43 <_ikke_> Can you share the init script? 2019-05-06 11:16:47 yes 2019-05-06 11:16:49 sec 2019-05-06 11:16:52 #!/sbin/openrc-run 2019-05-06 11:16:52 name=$RC_SVCNAME 2019-05-06 11:16:52 command_args="-r" 2019-05-06 11:16:52 command="/opt/test" 2019-05-06 11:16:52 command_background=true 2019-05-06 11:16:52 pidfile="/run/$RC_SVCNAME.pid" 2019-05-06 11:17:13 the %s is what I wrote here, its not part of the error message 2019-05-06 11:17:19 <_ikke_> Please use a paste service next time :) 2019-05-06 11:17:24 sorry 2019-05-06 11:17:46 <_ikke_> what happens if you manually execute /opt/test? 2019-05-06 11:17:53 sec 2019-05-06 11:18:05 Is RC_SVCNAME set from name? 2019-05-06 11:18:53 the name of the script in /etc/init.d is ¨test¨ 2019-05-06 11:19:15 from filename yes 2019-05-06 11:19:20 so you can symlink a script 2019-05-06 11:19:34 @_ikke_ when I manually execute nothing happens 2019-05-06 11:20:01 <_ikke_> Peter5532: whow are you executing it? 2019-05-06 11:20:33 <_ikke_> s/whow/how/ 2019-05-06 11:20:55 well I cd into /opt/ and then I type ¨test -r¨ 2019-05-06 11:21:14 the argument is required by the app 2019-05-06 11:21:16 <_ikke_> Peter5532: You are executing the built-in shell test command 2019-05-06 11:21:22 <_ikke_> try ./test -r 2019-05-06 11:21:50 when I type that it says ¨-ash: ./test: not found¨ 2019-05-06 11:22:11 <_ikke_> It cannot execute it 2019-05-06 11:22:16 <_ikke_> what kind of file is it? 2019-05-06 11:22:18 what is it? 2019-05-06 11:22:25 some binary ? 2019-05-06 11:22:43 it´s a daemon program written in Freepascal 2019-05-06 11:23:00 that will probably not work on musl 2019-05-06 11:23:01 <_ikke_> Probably not compiled against musl 2019-05-06 11:23:34 there is no way I can run this? 2019-05-06 11:23:47 compile it against musl 2019-05-06 11:24:42 you can try to install gcompat 2019-05-06 11:25:19 check permissions first. 2019-05-06 11:25:46 it could be something simple ;-) 2019-05-06 11:26:01 freepascal application in /opt? i doubht it :) 2019-05-06 11:26:42 permissions are ok, only root user with execute 2019-05-06 11:26:48 regarding opt, I put it there 2019-05-06 11:27:10 what does file say? 2019-05-06 11:27:22 Peter5532: apk add gcompat 2019-05-06 11:27:30 see what happens when you run it again 2019-05-06 11:28:18 which repo is this in 2019-05-06 11:28:28 community 2019-05-06 11:28:34 ok sec 2019-05-06 11:29:39 yes it works now 2019-05-06 11:29:41 ;D 2019-05-06 11:29:50 thank you 2019-05-06 11:30:40 we accept euros and dollars 2019-05-06 11:31:09 give me link 2019-05-06 11:31:17 :) 2019-05-06 11:31:29 thank you is more than enough :) 2019-05-06 11:32:01 lol 2019-05-06 11:32:01 yes thank you very much, Im total noob in linux 2019-05-06 11:32:19 <_ikke_> Note that it's usually better to compile against musl, but there are cases where that's not feasible, that's when you have to rely on gcompat 2019-05-06 11:32:37 is there a performance difference when compiling against gcompat? 2019-05-06 11:33:01 you cannot compile againt gcompat 2019-05-06 11:33:17 its a layer on top of musl to support glibc features 2019-05-06 11:33:36 I see, does it affect the application tho? 2019-05-06 11:33:45 it could be 2019-05-06 11:34:00 its not advised to use it. 2019-05-06 11:34:05 its a hack 2019-05-06 11:34:41 if you have the source, you should definitely build it with alpine sdk 2019-05-06 11:35:20 I will check into that, thank you 2019-05-06 11:35:39 <_ikke_> But that requires freepascal to be available as well 2019-05-06 11:35:42 <_ikke_> on alpine 2019-05-06 11:38:00 doesn't someone mentioned intention to package it few weeks ago 2019-05-06 11:39:03 <_ikke_> no mention here or in #alpine-devel 2019-05-06 11:39:19 or I read about related to something else 2019-05-06 11:39:59 hmmm, I looked in void linux packages, I remember now 2019-05-06 11:44:47 any news about firefox upgrade for alpine, I think fiasco is fixed in 66.x 2019-05-06 11:48:46 mps: I was working on upgrading firefox to 66.x but gave up and just switched to chromium 2019-05-06 11:49:02 mps, mozilla had a 66.0.4 update released yesterday 2019-05-06 11:50:59 Invader_Bork: I read that, I'm asking for Alpine 2019-05-06 11:51:31 ok, will look at firefox-esr upgrade 2019-05-06 11:52:24 interestingly, my personal build of esr for aarch64 don't have problem 2019-05-06 11:56:52 esr with fix is already in community 2019-05-06 13:15:54 hello, everyone. ncopa: hey, have you had a chance to take a look at https://bugs.alpinelinux.org/issues/10306 ? 2019-05-06 13:48:41 FreeBDSM: maybe in a bit 2019-05-06 15:06:06 hmm did the cert change for docs.alpinelinux.org? 2019-05-06 15:06:56 I am getting a warning and it is missing docs.alpinelinux.org in the "Certificate Subject Alt Name" field. 2019-05-06 15:08:09 TBK[m]: I don't think we're doing docs.a.o anymore, until beta.docs.a.o is out of beta anyway 2019-05-06 15:08:32 so yeah, if you're using httpseverywhere, try https://beta.docs.alpinelinux.org/ 2019-05-06 15:09:23 _ikke_: should probably add docs.a.o to the LE TLS cert for beta.docs.a.o 2019-05-06 15:10:25 danieli: would be better with a redirect or simple index.html with a link + beta statement 2019-05-06 15:11:53 <_ikke_> danieli: docs.a.o should probably use the wildcard cert 2019-05-06 15:12:56 the redirect from docs.a.o to beta.docs.a.o is obvious enough in itself, no? 2019-05-06 15:13:01 _ikke_: mm, might as well 2019-05-06 15:45:01 ncopa: pingo 2019-05-06 15:58:03 FreeBDSM: im not sure that we can upgrade to version 2 in a stable branch 2019-05-06 16:02:35 FreeBDSM: someone need to investigate. Im not sure libjpeg-turbo is affected 2019-05-06 16:02:51 we will probably need to backport patches 2019-05-06 19:43:18 hi i try to build my own kernel but it's aborted with this error https://pastebin.com/aFgkNqqj 2019-05-06 19:53:23 nice, built new firefox-esr from aports locally and all extensions are back 2019-05-06 19:54:15 Tuxist: looks likke you need to adjust it for newer kernel 2019-05-06 19:54:38 ok 2019-05-06 19:59:40 unforunately i dont have time or energy to help you now, sorry :-/ 2019-05-06 19:59:43 thanks https://patchwork.kernel.org/patch/10250477/ 2019-05-06 20:00:07 yeah, thats it 2019-05-06 20:02:55 ncopa: it works now i hope that will fix lxc :-D 2019-05-06 20:10:52 sorry for stupid question, but how to manage config updates? apart from vim -d /etc/prog.conf /etc/prog.conf.apk-new 2019-05-06 20:17:37 sr20det: how would you like it better 2019-05-06 20:20:47 magically? 2019-05-06 20:21:42 afaik no OS handles it magically. a solution might be to do it with a "/etc/prog.d/file.conf" kind of setup 2019-05-06 20:22:43 but that's not always desireable (I dislike the convoluted way dovecot does that, for example) 2019-05-06 20:22:58 and anyway it's up to the upstream dev 2019-05-06 20:23:39 but nobody is going to risk overwriting and messing up your config, possibly ruining your security 2019-05-06 22:34:26 Will the arm64 (aarch64) build of Alpine run on a Raspberry Pi 3? 2019-05-06 22:34:44 positive 2019-05-06 22:34:58 The Debian image shows "armv7" so I was confused. Thanks! 2019-05-06 23:14:10 I enabled edge by replacing "v3.9" with "edge" in my /etc/apk/repositories, but when I do `apk add wireguard-rpi` it says it's unsatisfiable. 2019-05-06 23:14:17 https://pkgs.alpinelinux.org/package/edge/testing/aarch64/wireguard-rpi 2019-05-06 23:14:29 The package is right here so I'm not sure why it isn't finding it 2019-05-06 23:15:01 Did you enable edge for main, community, and testing? 2019-05-06 23:15:11 fyber: I think wireguard is in testing 2019-05-06 23:16:11 mps: that pkgs.a.o link would say you are correct 2019-05-06 23:17:17 yes, I see now when looked at url. I know it is in testing but wasn't sure for rpi version 2019-05-06 23:17:21 tcely: That was it, I didn't notice testing was commented out. 2019-05-06 23:18:11 and more errors, this time it says lib/modules doesn't exist :( 2019-05-06 23:20:46 You probably need to upgrade to the kernel from edge first 2019-05-06 23:21:34 Hmm. I thought I did that with an `apk upgrade` 2019-05-06 23:21:39 is there something else I need to do? 2019-05-06 23:21:55 yeah my kernel isn't the right version 2019-05-06 23:22:00 Usually a reboot is needed to run the new kernel 2019-05-06 23:24:22 tcely: No? Just rebooted and I'm still on .34, wireguard was looking for .40 2019-05-06 23:25:27 How is your system booting? If .40 is installed, I would expect it to be selected as newer than .34 2019-05-06 23:25:43 .40 probably isn't installed 2019-05-06 23:26:48 the wiki says "apk upgrade" "sync" "reboot" which didn't install it 2019-05-06 23:28:40 Which kernel(s) are installed according to apk info -e ? 2019-05-06 23:30:13 No kernels? 2019-05-06 23:30:17 That command is blank 2019-05-06 23:30:50 though when I boot up it says 4.19.34-0 2019-05-06 23:31:00 I'm doing this from memory, I might have messed it up. 2019-05-06 23:31:51 Try -a 2019-05-06 23:32:48 there's no kernel listed there 2019-05-06 23:33:03 if I do apk info linux-rpi then it shows that .40 is (available?) 2019-05-06 23:34:02 yeah, available but not installed if not shown with -e -a 2019-05-06 23:34:21 If I install it I get a bunch of errors 2019-05-06 23:34:41 which mirror are you using? 2019-05-06 23:34:45 a bunch of stuff failing to rename, it can't find /lib/modules, etc 2019-05-06 23:34:51 TBK[m]: dl-3 2019-05-06 23:35:49 it is up to date accoring to https://mirrors.alpinelinux.org/#mirror4 2019-05-06 23:36:13 just to confirm you did run?: apk update 2019-05-06 23:36:22 yup 2019-05-06 23:36:24 it just says "OK" 2019-05-06 23:36:55 oh, update, yeah that too 2019-05-06 23:37:31 TBK[m]: It finds .40, but just fails to install if I try to install it 2019-05-06 23:37:35 probably nothing I can do about it 2019-05-06 23:37:42 just a broken package :( 2019-05-06 23:47:23 I will roll a RPi3 with AL edge and test 2019-05-06 23:50:39 I'm running in diskless mode with no persistence if that makes a difference. (lbu commit -d before reboot instead) 2019-05-07 00:29:14 hmm something is wonky with linux-firmware-brcm and linux-rpi 2019-05-07 01:02:03 TBK[m]: any progress? 2019-05-07 01:26:32 fyber: nope, do not know what is going on with linux-firmware-brcm and linux-rpi. clandmeter & ncopa will have to look at it. 2019-05-07 04:12:57 arp -a 2019-05-07 04:13:00 er 2019-05-07 04:17:50 lol oops 2019-05-07 06:13:48 fyber: you are running tmpfs? 2019-05-07 06:14:25 Then you cannot install kernel or modules 2019-05-07 08:21:04 hello 2019-05-07 08:21:17 anybody encounter problems with snmpd and alpine? 2019-05-07 08:21:23 i can't start the daemon 2019-05-07 08:21:57 https://pastebin.com/2jbkR0aM 2019-05-07 08:22:18 snmpd version is 5.8 2019-05-07 08:22:56 <_ikke_> What version of Alpine? 2019-05-07 08:23:24 3.9.3 2019-05-07 09:05:55 any ideea? 2019-05-07 11:30:16 SoulRaven : thanks for reporting. it seems there is a bug going on. I'm checking 2019-05-07 13:58:57 SoulRaven, tmhoang: I had a problem with snmpd, which looks similar: https://bugs.alpinelinux.org/issues/9723 2019-05-07 14:01:36 You could verify that by stracing the process, e.g. with "strace -e bind ..." 2019-05-07 16:49:36 Hello. `apk add package` adds package to world and installs. What is the canonical/easiest way to add a package WITHOUT adding to world, ie I want a package to only survive for this boot 2019-05-07 16:53:58 I don't know if there's a way to add it for just this boot, but you can use the --virtual option to remove packages as a bundle, along with --no-cache to prevent them from being cached in /var/cache/apk 2019-05-07 16:54:09 Is there a particular use case that you have in mind? 2019-05-07 16:54:27 --virtual will still at that virtual to world, ofc :) 2019-05-07 16:54:37 i'm not sure if there's a straight-forward way -- what's your use case for it? 2019-05-07 16:55:06 s/at/add 2019-05-07 16:57:45 My goal right now is to have a minimal base apkovol and I run from RAM. If my system needs X, I add it one time just for that session with script 2019-05-07 16:58:08 Right now, I'm just not lbu committing, but I'm looking for a more elegant solution' 2019-05-07 17:01:34 That would seem to be about as elegant as you can get without snapshots in a virtual machine system. Depending on how you have the scripts set up, maybe you can start the script with "apk add --virtual script-pkgs ..." 2019-05-07 17:02:11 Then at the end of the script have a line "apk del script-pkgs" to clean up 2019-05-07 17:03:12 That's kind of messy. I'd prefer option where it just doesn't write the world file. 2019-05-07 17:03:45 --no-cache is a thing, I would imagine something like that would exist in apk, but perhaps i'm wrong 2019-05-07 17:04:11 Then you really need a VM or a Python virtualenv-like system 2019-05-07 17:04:31 The point is this will be my KVM hypervisor 2019-05-07 17:04:38 Sometimes I need X, sometimes I don't 2019-05-07 17:05:22 What task are you trying to accomplish on a hypervisor that needs X? 2019-05-07 17:06:14 Front-end GUI development 2019-05-07 17:06:38 Windows pen-testing? Idk, I can make a list if you'd like 2019-05-07 17:06:39 Do that in a VM / container ? 2019-05-07 17:07:15 Agreed -- that sort of thing should be done in a VM or container, leaving the hypervisor level minimal 2019-05-07 17:07:27 I'm trying to turn Alpine into a KVM hypervisor that will enable me to work with VMs / containers in an emperal sort of way 2019-05-07 17:07:37 I want the KVM host to be emperal as well 2019-05-07 17:08:07 Alpine will be the host, so that I can run other guests and do my development work 2019-05-07 17:08:08 You're trying too hard, it sounds like to me. 2019-05-07 17:08:19 c705: you might like rancheros with an Alpine console instead 2019-05-07 17:08:46 The host doesn't need to have anything ephemeral 2019-05-07 17:09:10 Concentrate the emphemera in the various VMs 2019-05-07 17:09:47 For security reasons, I want the host ephemeral 2019-05-07 17:10:10 You might want to make the hypervisor layer deliberately difficult to do much in 2019-05-07 17:10:22 To remove the temptation to do anything in there 2019-05-07 17:12:07 I'd prefer not to go into the justifications of why I want to do the things I do, I'm just looking for a solution to my issue 2019-05-07 17:12:21 Which is to install packages one-time and not add them to /etc/apk/world 2019-05-07 17:13:15 When nothing outside /opt & /home survives reboots you don't write important things all over the system. :-) 2019-05-07 17:13:18 I can awk/sed them out later, but writing then removing from a file is dumb. If nothing is built into apk maybe I'll patch it and submit as a feature 2019-05-07 17:13:28 Then, no, there isn't anything short of a VM. 2019-05-07 17:13:45 tcely: /home does not survive and /etc does. Not sure what you are talking about 2019-05-07 17:14:44 c705: my setup for docker host. Everything runs in containers with any persisted data in volumes 2019-05-07 17:16:17 tcely: That's nice. If I can somehow write to your .profile in your home from inside a container using an exploit, and you lbu commit that, I have control of your sessions. 2019-05-07 17:16:43 The point is to be empheral. /home is built from a static skeleton dir every boot 2019-05-07 17:19:17 Does app die horribly if world is immutable? 2019-05-07 17:19:27 apk 2019-05-07 17:19:50 <_ikke_> I can imagine it would not like that 2019-05-07 17:20:06 <_ikke_> But it might be ok as long as you dont add / delete packages 2019-05-07 17:22:46 My only other crazy idea is adding inside a chroot that is not preserved 2019-05-07 17:23:39 tcely: Making world immutable seems like a radical solution to what I think is a very simple problem 2019-05-07 17:23:58 mepholic: nothing new about tcsh release, last msg in your issue is 9 day old 2019-05-07 18:01:37 I'm trying to install numactl-dev, which exists according to https://pkgs.alpinelinux.org/package/edge/main/x86/numactl-dev - however, I can't find the package under apk ("apk search numa") 2019-05-07 18:01:54 This is my first day playing with Alpine Linux, so I apologize if the answer is something really simple 2019-05-07 18:02:05 Thanks in advance 2019-05-07 18:02:53 prizzle: try apk --update add numactl-dev 2019-05-07 18:04:00 tcely: same issue - unsatisfiable constraints 2019-05-07 18:04:21 It says numactl-dev (missing) is required by world[numactl-dev] which I imagine is because it can't find the package 2019-05-07 18:05:13 prizzle: what is in /etc/apk/repositories, v3.9 or edge 2019-05-07 18:06:23 mps: Most of the lines are commented out - here's a paste https://hastebin.com/fezuboceje.cpp 2019-05-07 18:06:41 Should I try uncommenting testing? 2019-05-07 18:07:14 you only have the v3.9 main repo 2019-05-07 18:07:26 you found the package you want in the edge main repo 2019-05-07 18:07:28 no, it is in edge and you have v3.9 2019-05-07 18:08:09 you have to upgrade your packages to edge (development) alpine 2019-05-07 18:08:17 Can I set packages on "hold" when installed via meta image? 2019-05-07 18:08:20 If I add http://dl-cdn.alpinelinux.org/alpine/edge/main and then update, will that work? 2019-05-07 18:08:49 ofc, if you are ready to use development version of distro 2019-05-07 18:09:29 Yeah I am open to playing with it, I'm not attached to anything in this VM 2019-05-07 18:09:30 mixing stable and development is not wise 2019-05-07 18:10:18 better option is to replace 'v3.9' with 'edge' and 'apk update' 'apk upgrade' 2019-05-07 18:10:29 mpls: Thanks, I did that just now 2019-05-07 18:10:38 Things look to be building now, fun 2019-05-07 18:10:59 I appreciate the help guys! Thanks 2019-05-07 18:18:39 Anyone with an idea? 2019-05-07 18:28:24 velix: you mean metapackage? 2019-05-07 18:28:30 exactly. 2019-05-07 18:30:04 same way you'd pin a regular one? 2019-05-07 18:30:27 oh wait 2019-05-07 18:30:29 I see what you mean 2019-05-07 18:31:11 in some cases, it will work, like apk add xfce4=4.13.0 2019-05-07 18:32:04 Oh okay... f.e. I'm installing "apk add abc-dev" and it installs "abc", too. When I remove abc-dev, abc also is gone, but I'd love to keep abc. 2019-05-07 18:32:17 So I just need to apk add abc ? 2019-05-07 18:32:25 oh, that's what you mean 2019-05-07 18:32:26 <_ikke_> yes 2019-05-07 18:33:30 Okay. I think, apt/dpkg has a similar functionality: manually installed packages won't get autoremoved. 2019-05-07 18:33:38 Thanks, I'll try this. 2019-05-07 18:34:02 I thought you meant version pinning, not installing as explicit 2019-05-07 18:34:42 so as _ikke_ says, you will have to "apk add abc", yes 2019-05-07 18:35:02 doggone: Nah, you're right. I like both answers ;) 2019-05-07 18:35:09 Both are useful. +1+1 = +2 2019-05-07 18:35:20 actually 2.00001 (float) 2019-05-07 18:35:30 Sorry, too much coffee. 2019-05-07 18:36:33 <_ikke_> velix: to give more context. apk add will add to /etc/apk/world, which is the list of explicitly installed packages 2019-05-07 18:36:58 <_ikke_> any package that's not in that list will automatically get removed as soon as it's no longer needed as a dep 2019-05-07 18:38:28 I've just tried it and it works (warning: cotainer stuff): https://bpaste.net/show/5664f284793c 2019-05-07 18:38:43 The first lines are a baaad hack for a python bug. 2019-05-07 18:39:14 oops, that was the old one. 2019-05-07 18:41:31 mps yeah 2019-05-07 18:41:51 I realized after I started that discussion that I know basically nothing about tcsh 2019-05-07 18:42:03 and have nothing of value to contribute other than "release plz" 2019-05-07 18:45:20 mepholic: to be fair, also I don't have idea to upgrade it or not. I didn't experienced any of bugs mentioned in commits so not sure what to do 2019-05-07 18:46:26 I already added /etc files csh.login and csh.cshrc and pushed it 2019-05-07 18:47:56 thought to upgrade it for 3.10 just in case someone have noticed these issues 2019-05-07 22:44:04 Hey, as anybody had any problems with the package linux-firmware-brcm, as somehow, every single time I want to install that package the system can't rename the installed files 2019-05-07 22:44:43 quadrat: RPi? 2019-05-07 22:47:34 TBK[m], yes 2019-05-07 22:47:53 (a 3b) 2019-05-07 22:48:38 Have you setup persistent storage? https://wiki.alpinelinux.org/wiki/Raspberry_Pi#Persistent_storage 2019-05-07 22:48:57 I included the dir in the lbu backup program 2019-05-07 22:51:35 TBK[m], shoudn't that be enough? 2019-05-07 22:51:46 tmpfs/diskless mode does not support firmware/kernel upgrades 2019-05-07 22:52:20 ahh thanks :), and one last question then, same thing with kernel modules? wanna try nftables out 2019-05-07 22:52:35 but the kernel module is not there 2019-05-07 22:53:16 yes 2019-05-07 22:53:37 thanks :) so have to setup the persistent storage then. 2019-05-08 00:03:40 clandmeter: Yup, that explains it. I'll try to do a regular install. Thanks! 2019-05-08 00:43:27 ... and in the regular install, ethernet is broken :( 2019-05-08 00:46:25 ah, my modules aren't loading 2019-05-08 01:14:28 The wiki article about sys-based boot doesn't work (completely wrecks the kernel). The random forum post linked at the bottom under "resources" works flawlessly. 2019-05-08 02:25:12 tcely: it's odd that you have dockerhub-powerdns which builds pdns from scratch, but you also maintain the pdns apks... is there a reason you do that? (just curious) 2019-05-08 03:46:20 is there a qsort_r()? 2019-05-08 03:46:50 There is an independent implementation of it for musl 2019-05-08 04:30:15 no filesystem could mount root 2019-05-08 04:31:05 I am attempting to install alpine on hardware in a partition (for multiboot) 2019-05-08 04:31:42 I had to follow a instruction list that I put together myself. The multiboot page on the wiki tells me it is out of date. 2019-05-08 04:32:10 This is the xen version 3.9.3 2019-05-08 04:32:22 one partition for / and /boot 2019-05-08 04:32:32 I'm using grub not syslinux 2019-05-08 04:32:45 so it is a bit off the path from the usual install 2019-05-08 04:33:56 I even tried generating the grub.cfg on devuan, thinking maybe the alpine version of grub is different. Devuan identifies the partition as alpine linux and generates two entries (normal and fallback?) 2019-05-08 04:34:20 (and then, of course, re-installing grub so that the devuan grub.cfg will be picked up) 2019-05-08 04:35:10 that message sounds backwards, inside-out, sideways. I think they mean something more like could not find a file system to mount on root. 2019-05-08 04:35:40 at any rate, if anyone here has a suggestion for getting this to work, I'd appreciate it. Or even a clue or two. 2019-05-08 07:24:04 nvm. Turns out devuan did not generate the initrd entry in grub.cfg. 2019-05-08 07:25:04 it may be that alpine is not using some standard approach to naming the files in /boot, but idk that for sure. 2019-05-08 10:03:35 iggy: it's simple really, one came before the other and once the work was done automated builds aren't much trouble to keep going. 2019-05-08 11:46:09 problem now is that the system hangs early in the boot. It enumerates the usb devices, then just hangs. This may have something to do with my KVM switch, although 2 other linux systems on the same machine work OK with it. 2019-05-08 11:47:19 If I remove and replace the usb cable for the KVM while the system is "hanging," there will be a re-enumeration of the usb devices. So it is not truly hung per se, but the boot process just sits there, like it is waiting for something. 2019-05-08 11:48:22 The kernels on the other 2 systems are similar, 4.9.x kernels, whereas alpine seems to be at 4.19.28. It is more advanced, yet unable to boot whereas the other 2 boot ok. 2019-05-08 11:49:07 I'm looking around for hints on how to work through this problem, and it looks like building a debugging kernel is about the only solution. Is there a better, faster one? 2019-05-08 11:50:26 systemdlete: waiting for entropy problem? 2019-05-08 11:52:34 possibly. I note that after some time (sometimes), there will be messages about random, etc. 2019-05-08 11:52:52 This is real hw, so there shouldn't be a lack of real entropy 2019-05-08 11:54:33 are you aware of this bug report https://bugs.alpinelinux.org/issues/9960 2019-05-08 11:56:48 ah, thanks. That's just the kind of thing I would not know off the top of my head. But you did! 2019-05-08 11:57:36 I'll try the suggested solution. It is a test machine anyway, so I don't care about security issues. 2019-05-08 11:58:27 I'm figuring this issue will be resolved by the time I roll out to my main machine, else I guess I could fall back to an earlier alpine release? 2019-05-08 12:02:59 there is explanation how to enable rdrand if you trust manufacturer of CPU 2019-05-08 12:04:40 yes, I just tried random.trust_cpu=on on the line in grub with "linux" but it did not make a difference 2019-05-08 12:06:49 btw, I don't follow how haveged would help me, given that I don't get to the init phase or the runtime phase. It is clearly stuck in the kernel start-up 2019-05-08 12:08:11 "early boot" problem 2019-05-08 12:08:41 are you a udev user by any chance? 2019-05-08 12:09:26 I am using whatever the default is. I did not go to any lengths to change the udev/eudev 2019-05-08 12:10:09 I did not intentionally try to change it, no. 2019-05-08 12:11:22 okay, Alpine uses mdev by default, I'm not very familiar with it, but I remember USB devices being a bit of a problem when you use udev (as you would whenever you're using a desktop), there are some additional services that need to be started that, I imagine, trigger some sort of a re-evaluation of devices 2019-05-08 12:11:51 It's a desktop PC! 2019-05-08 12:12:12 But I never get to the init stage. 2019-05-08 12:12:39 I'm talking about messages coming from the kernel, not messages emitted by init scripts (rc) 2019-05-08 12:13:12 ah, -that- early 2019-05-08 12:13:14 TBB, are you saying I am loading udev because it's a desktop? 2019-05-08 12:13:15 systemdlete: ah, so. maybe it waits for root FS 2019-05-08 12:13:17 yes! 2019-05-08 12:13:40 TBB: udev-trigger on sysinit and udev-postmount on local for udev users 2019-05-08 12:14:41 mps? 2019-05-08 12:15:19 maybe 'root=/dev/device" is wrong or missing driver for device 2019-05-08 12:15:41 north1: yes, those two exactly 2019-05-08 12:16:09 root= is a bit picky about what it accepts isn't it 2019-05-08 12:16:29 whenever I have wrong 'root=' or missing driver I have something similar like you described 2019-05-08 12:17:28 I just looked, and now I see the message "random: crng init done" 2019-05-08 12:17:38 but nothing else has happened 2019-05-08 12:18:24 Do I need to specify a special driver for ext3? Currently, it is loading ext2, as per grub 2019-05-08 12:19:44 root= is specified 3 ways in one entry. There is the root='hd0,gpt2' (or the like) command, then there is the UUID spec, and again on the "linux" command line (kernel) where it is /dev/sda2 2019-05-08 12:20:31 And that is generated by grub-mkconfig on devuan (which missed the initrd line, which I later added, to get past THAT problem) 2019-05-08 12:20:59 I double checked the UUID against the blkid output, and it matches 2019-05-08 12:21:43 I can post my grub.cfg, or at least the alpine section 2019-05-08 12:26:12 I'm not good at grub debugging but maybe someone else can help 2019-05-08 12:26:46 TBB, north1, mps: Here's st interesting: While the boot is sitting there, if I hit the space bar, the boot spits out more messages saying it is disconnecting the usb and then goes through the enumeration again. I can do this repeatedly, and the boot proceeds no farther. Just loops. 2019-05-08 12:27:44 I DO have a KVM switch because I have 2 PCs. I thought maybe this was important. But keep in mind that two other, recent distros, are able to boot through the same KVM on that same PC (it's multiboot) 2019-05-08 12:28:51 The KVM could be slightly wonky, given it was a $10 Fry's Electronics special. I researched the switch and came up bupkas. 2019-05-08 12:29:21 KVM shouldn't be problem, but who knows at the end 2019-05-08 12:29:36 It's a Bytecc 4x1 HDMI mouse keyboard (USB) KVM-4UHNM. 2019-05-08 12:29:53 mps: I agree, but I thought I should be upfront about this. 2019-05-08 12:30:28 who generated initramfs for alpine, devuan or alpine 2019-05-08 12:30:31 The install image has no such issue with the KVM. It seems it is only the kernel that is installed to the disk. 2019-05-08 12:31:00 me, but I followed the instructions on the alpine wiki 2019-05-08 12:31:43 I mean, in alpine or debian 2019-05-08 12:31:59 As far as Devuan and Hyperbola, the other 2 distros, I have not touched those directly. They have all the automationh for that sort of thing. 2019-05-08 12:32:40 Oh god, mps. Alpine of course. I wouldn't dare try to generate the image with another distro's tools. Especially if I might need assistance later. 2019-05-08 12:32:46 do you have 'APPEND' line in grub.cfg 2019-05-08 12:32:51 Now, the grub.cfg is a different deal... 2019-05-08 12:33:11 Not sure. I don't think so. 2019-05-08 12:34:47 something like this 'APPEND root=/dev/sdb2 modules=sd-mod,usb-storage,ext4 nomodeset rootfstype=ext4' 2019-05-08 12:39:01 Nope. But here is my grub.cfg, generated by Devuan (because I could not get alpine to install grub): pastebin.com/mqjdcfec 2019-05-08 12:45:38 mps: I think "APPEND" is only a grub legacy command, not in grub2. 2019-05-08 12:45:50 but the same info could be tacked on to the "linux" line 2019-05-08 12:46:21 this line 'linux /boot/vmlinuz-vanilla root=/dev/sda2', maybe you missing 'modules=sd-mod,usb-storage,ext4' 2019-05-08 12:47:21 do you see above the "linux" command where it says "insmod ext2" -- I think that takes care of it 2019-05-08 12:47:28 i.e. 'linux /boot/vmlinuz-vanilla root=/dev/sda2 modules=sd-mod,usb-storage,ext4' 2019-05-08 12:47:46 I can certainly try that. But I am only using ext3, not ext4 2019-05-08 12:48:02 insmod is for grub to access ext2/3/4 fs 2019-05-08 12:48:30 ok, give me a few minutes... 2019-05-08 12:48:33 ext4 is used for all ext FS, 2,3 and 4 2019-05-08 12:52:14 systemdlete: btw, it is better to use text only paste service like tpaste.us 2019-05-08 12:52:43 systemdlete : nice name. was missing 'e' intentional ? 2019-05-08 12:54:25 tmhoang: Very intentional. I wanted there to be no mistake as to what I meant... 2019-05-08 12:54:43 mps: pastebin is text only as far as I know. 2019-05-08 12:55:00 and even with the update, it is still hanging 2019-05-08 12:57:42 although, I wonder if needed modules are in the modules file. 2019-05-08 12:59:26 syslinux and grub are not analogous in terms of all the different moving parts, so coming up with an "analogous" solution in grub based merely off the facts of syslinux configuration may not map 100%. So it is possible that grub won't work well for alpine. 2019-05-08 13:00:03 Yet... I think if one tried hard enough, it is still possible. 2019-05-08 13:01:26 one could say grub is an overengineered, overcomplicated solution to a relatively simple problem... 2019-05-08 13:01:47 grub 1 or grub 2? The latter... I'd agree 100%. 2019-05-08 13:02:09 grub 1 was built to launch an OS. grub 2 was built to launch an OS so it could launch an OS. 2019-05-08 13:02:36 I long for the days of LILO and grub legacy... 2019-05-08 13:02:57 grub2 naturally, grub1 is obsoleted 2019-05-08 13:03:30 syslinux is not so bad, either. But I need to be able to settle on one boot manager since all my systems live on one box. I could try chainloading syslinux I suppose... 2019-05-08 13:06:17 Yes. I think that is what I shall do. I will try chainloading from grub2 to syslinux for alpine. I know this is possible; it is documented somewhere IIRC. But right now, I must run and buy a new air bed. My old one blew its last hole this morning, and I would like to get some sleep sometime. So I am hurrying out to Walmart to buy a replacement. 2019-05-08 13:06:32 When I return, I may sleep a bit, then work on the chainloading idea. 2019-05-08 13:07:37 This will allow for utilizing alpine's own tools for creating its environment, and just a tweak to bring the alpine partitions into unison with the other distros on the disk. 2019-05-08 13:07:54 And less Q&A and noise on the IRC channel. 2019-05-08 13:09:17 TBB, mps, north1: Thanks for all your feedback. 2019-05-08 13:09:42 I'll come back and report on whatever success I might have. 2019-05-08 16:50:37 has it ever been considered to rename the apk package manager to something else to avoid ambiguity with android's app packaging? 2019-05-08 16:58:02 why shouldn't android have to rename it? alpine did it first afterall :) 2019-05-08 16:58:54 either way, i'd hope all people dealing with alpine knows that '.apk' files are the '.deb' or '.rpm' files of alpine 2019-05-08 17:18:45 Alpine was first? 2019-05-08 17:20:11 leonardus: Alpine Linux was released in 2005, Android in 2008 according to wikipedia 2019-05-08 17:20:35 oh, wow 2019-05-08 17:22:11 I feel like they can coexist either way, Alpine and Android don't really have the same target audience 2019-05-08 17:22:58 danieli: pmOS? 2019-05-08 17:23:04 that's not alpine though 2019-05-08 17:23:12 it's *based on* it, sure 2019-05-08 17:23:30 but they're distinct, different projects (Alpine and pmOS) 2019-05-08 17:25:19 yes, I know, but anyway ;) 2019-05-08 23:04:14 hey 2019-05-08 23:04:15 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2019-5021 2019-05-08 23:05:27 when did this get a fix? 2019-05-08 23:06:14 FreeBDSM: this is discussed today on #alpine-devel 2019-05-08 23:06:44 mps: do you have a gist of what was discussed? 2019-05-08 23:07:21 uhm, you can look at irc log, but can't remember right url right now 2019-05-08 23:08:05 here is I think https://dev.alpinelinux.org/irclogs/%23alpine-devel-2019-05.log 2019-05-08 23:08:36 search for cve keyword 2019-05-08 23:13:00 yeah, I've grepped the log and read relevant parts, no info there 2019-05-08 23:17:36 so no fix yet 2019-05-08 23:17:57 right, but is discussed at least and dev's now know for it 2019-05-08 23:19:50 actually, I think that for distributed images root password is not relevant 2019-05-08 23:32:01 it only applies to services that uses PAM (or otherwise) the shadow file as an auth backend 2019-05-08 23:32:33 even if people for some reason did both install openssh sshd *and* expose port 22, the default openssh-server in alpine doesn't use pam 2019-05-09 01:29:07 Yeah, I don't understand how that rates a 9.8 CVSS 2019-05-09 01:30:02 AV:N is completely wrong there, IMO. If you were dumb enough to install SSH on a docker container, that's a you problem :) 2019-05-09 01:31:48 i wouldn't say it's dumb, but it seems like it's blown way out of proportion and misconstrued by media online for clicks 2019-05-09 01:32:00 it's not the first nor the last time that'll happen, not by a long shot 2019-05-09 01:32:06 i have SSH in an SSH image 2019-05-09 01:32:13 still doesn't use PAM 2019-05-09 01:32:28 https://pkgs.alpinelinux.org/package/edge/main/x86_64/openssh-server-pam 2019-05-09 01:33:32 doesn't the default openssh config also disable root login? 2019-05-09 01:33:50 let me check, i'm pretty sure that's a yes 2019-05-09 01:34:17 hmhm, looks like it's using the default opensshd one, let me dig 2019-05-09 01:36:34 the openssh default is prohibit-password by the looks of it 2019-05-09 01:36:38 so that's it 2019-05-09 06:02:52 FreeBDSM: it was fixed with https://github.com/docker-library/official-images/pull/5516 2019-05-09 08:20:10 any php users here? GLOB_BRACE for glob() isn't defined :( 2019-05-09 08:32:56 that is a glibc extension and not portable 2019-05-09 08:33:45 https://github.com/wp-cli/wp-cli/issues/4353 2019-05-09 11:35:21 Is there an official announcement for CVE-2019-5021 that I can refer to? 2019-05-09 11:36:00 The alpine-security ML is marked as inactive, and there are no announcements on alpine-announce or twitter 2019-05-09 11:36:09 <_ikke_> captn3m0: We are working on one 2019-05-09 11:36:56 any chance the alpine-security list is being revived? 2019-05-09 11:38:16 captn3m0: this is the work-in-progress announcement: https://wwwtest.alpinelinux.org/posts/Docker-Image-root-User-Hard-Coded-Credential-Vulnerability.html 2019-05-09 11:38:31 i am considering add a section at the botton "how could this happen?" 2019-05-09 11:38:59 but not sure if i should drop that and post the announcement as is 2019-05-09 11:44:08 I like the announcement as-is. 2019-05-09 11:44:26 Very much helpful 2019-05-09 12:07:45 this is what I am considering to add at the bottom: http://tpaste.us/Q1vO 2019-05-09 12:09:15 Looks like a good addition. 2019-05-09 12:09:19 typo: s/This makes is possible/This makes it possible 2019-05-09 12:18:59 captn3m0: thanks! 2019-05-09 12:20:21 this is the updated proposal: https://wwwtest.alpinelinux.org/posts/Docker-image-vulnerability-CVE-2019-5021.html 2019-05-09 12:21:26 can somebody re-read that and verify that I didnt have other typos? someone native english preferable 2019-05-09 12:23:34 Looks ok to me 2019-05-09 12:44:09 it is live https://alpinelinux.org/posts/Docker-image-vulnerability-CVE-2019-5021.html 2019-05-09 12:44:29 <_ikke_> Hackernews submission: https://news.ycombinator.com/edit?id=19867737 2019-05-09 13:02:27 amazing, guys 2019-05-09 14:05:27 thanks everyone for helping the the response out 2019-05-09 14:06:41 :D we did it alpine 2019-05-09 14:07:18 👍 2019-05-09 14:58:05 mepholic: tcsh 6.21.00 is released 2019-05-09 14:58:29 just pushed upgrade to aports 2019-05-09 16:38:23 the audacity to take cheap shots, wow https://twitter.com/pid_eins/status/1126458346515705857 2019-05-09 16:40:13 uhg, i can't get anything to update on edge because of some conflict hell with readline8 2019-05-09 16:44:08 danieli: but password-based root login and passwordless login are disabled in SSH so you'd need "physical" access >_> 2019-05-09 16:44:12 danieli: His comment doesn't even make sense 2019-05-09 16:44:47 our opensshd doesn't even use pam, even if you for some reason installed (*and* ran it in a container) 2019-05-09 16:45:05 and yes, it's prohibit-password by default 2019-05-09 16:45:31 um, looks like world file was corrupt 2019-05-09 16:45:42 north1: it's a cheap shot, nothing more, nothing less 2019-05-09 16:46:25 probably best to ignore idiots like Poettering 2019-05-09 16:49:09 it could pose a problem if you're using eg. postfix with pam auth 2019-05-09 16:50:56 hey.. personal attacks aren't cool :( don't stoop there 2019-05-09 16:51:07 that's an utterly broken configuration 2019-05-09 16:51:09 and hm, true 2019-05-09 16:51:27 people are imagining nonsensical (esp in a docker container) configurations that are insecure for tons of other reasons 2019-05-09 16:51:37 as a stretch to claim this is a docker/alpine vuln 2019-05-09 16:51:39 it's not 2019-05-09 16:51:54 well yeah, i don't think docker is a security freak's choice of set-up 2019-05-09 16:52:02 but I don't think that's so broken 2019-05-09 16:52:08 it's a configuration that could make doing certain other completely idiotic things mildly more dangerous than they already were 2019-05-09 17:20:55 ncopa, current firefox is badly crashing for me :( 2019-05-09 17:28:28 it looks very bad. it's a js-induced crash on many sites that takes the form of foo=ptr1[(long)ptr2] 2019-05-09 17:37:27 trying a complete upgrade in case there are any dependencies that weren't updated that are responsible for the crash 2019-05-09 17:37:45 anyone interested in looking at this with me? 2019-05-09 17:40:25 dalias: which firefox? firefox-esr or firefox 2019-05-09 17:45:50 firefox 2019-05-09 17:48:53 iirc, firefox 66.0.4 from testing have an issue with where it crashes on some sites and it is not yet fixed afaik 2019-05-09 17:49:16 s/with /with js sites/ 2019-05-09 17:49:17 yes that's what i'm experiencing :( 2019-05-09 17:49:36 at least blocking js makes it go away 2019-05-09 17:49:58 not sure if that means it's js tho; it might be some manipulation the js is doing to the dom/presentation 2019-05-09 17:50:11 that puts things into a crashing state 2019-05-09 17:50:22 it's almost surely memory corruption or UAF tho 2019-05-09 17:50:35 because it's using a pointer value as an index into an array 2019-05-09 17:50:39 so it looks very very bad 2019-05-09 17:50:48 I think ncopa worked on this but can't remember if this is fixed or not 2019-05-09 17:51:03 it's not 2019-05-09 17:51:40 right, from you writing it is obvious it is not fixed :) 2019-05-09 17:51:50 :) 2019-05-09 17:51:55 was 65 affected? 2019-05-09 17:52:16 it would be nice if there were an archive of older package versions... 2019-05-09 17:52:38 isn't 65 disappeared from repo? 2019-05-09 17:52:44 exactly 2019-05-09 17:52:46 :/ 2019-05-09 17:53:15 since they're signed tho we could use any copy somebody still has 2019-05-09 17:53:45 this is a pretty much "desktop alpine is unusable" level of severity :( 2019-05-09 17:54:29 If you will build it yourself than you can 'downgrade' (git checkout) 65 and try 2019-05-09 17:55:18 well, firefox-esr works for me, or I'm lucky to not visit sites where it can crash 2019-05-09 17:57:45 esr should work fine but reverting back to an ancient version is going to be problematic for lots of reasons and probably can't use profile/settings 2019-05-09 17:58:12 installing it in addition (if that's possible) might be a temp workaround for accessing affected sites 2019-05-09 18:01:27 last time I tried coinstall didn't worked, there could be only one ;) 2019-05-09 18:02:30 who reported the crash before? 2019-05-09 18:03:58 sorry, can't remember 2019-05-09 18:11:26 opened on tracker https://bugs.alpinelinux.org/issues/10441 2019-05-09 18:18:48 that's interesting, and kind of deterring 2019-05-09 18:19:28 ? 2019-05-09 18:20:35 watch this be nothing to do with alpine or musl and just an idiotic UAF bug that's been in firefox for 10 years and possibly exploited 2019-05-09 18:20:53 :-P 2019-05-09 18:25:43 wouldn't surprise me, and i meant that it being old and deep inside an obscure library is deterring 2019-05-09 18:27:23 hi all 2019-05-09 18:27:46 hi 2019-05-09 18:28:01 hello 2019-05-09 18:28:06 If you're using IPv6, you should also add the following special IPv6 addresses to your /etc/hosts file: 2019-05-09 18:28:14 ::1 localhost ipv6-localhost ipv6-loopback 2019-05-09 18:28:16 fe00::0 ipv6-localnet 2019-05-09 18:28:18 ff00::0 ipv6-mcastprefix 2019-05-09 18:28:20 ff02::1 ipv6-allnodes 2019-05-09 18:28:27 ff02::2 ipv6-allrouters 2019-05-09 18:28:31 ff02::3 ipv6-allhosts 2019-05-09 18:28:33 why? 2019-05-09 18:28:35 north1: danieli 2019-05-09 18:28:42 all are as new as me? 2019-05-09 18:28:44 i am using alpine perhpas first time 2019-05-09 18:28:54 dalias: do we know if it is apline specific? or musl specific? fwiw ff-66 works fine on the site you listed for me... 2019-05-09 18:29:02 please use a paste service to post 3 or more lines 2019-05-09 18:30:07 rfs613, alpine package of it? 2019-05-09 18:30:32 dalias: no, different distro entirely 2019-05-09 18:31:05 well obviously whatever the bug is isn't showing up as a crash with other dists/builds 2019-05-09 18:31:19 so if it's a UAF, the allocator pattern is such that something different gets clobbered, etc 2019-05-09 18:31:56 pie3, why? 2019-05-09 18:32:56 dalias: gotcha. 2019-05-09 18:33:26 ip maddr works in alpine? 2019-05-09 18:35:20 mps, did you hear that panfrost is going in 5.2? 2019-05-09 18:36:48 Invader_Bork: I had a hope for 5.1 but nothing yet is announced even for 5.2 afaik 2019-05-09 18:37:02 pie3: you'll likely need to install iproute2 package. Otherwise you get busybox version of "ip" which is limited. 2019-05-09 18:37:32 rsal: how to install that 2019-05-09 18:37:41 mps, https://www.phoronix.com/scan.php?page=news_item&px=Linux-5.2-Early-Look 2019-05-09 18:38:04 apk add iproute2 2019-05-09 18:38:09 pie3: 'apk add iproute2' 2019-05-09 18:38:10 mps, https://www.phoronix.com/scan.php?page=news_item&px=Panfrost-DRM-For-Linux-5.2 2019-05-09 18:38:16 these were my sources 2019-05-09 18:38:42 Invader_Bork: yes, just found it 2019-05-09 18:38:51 it gives an error rfs613 2019-05-09 18:39:19 unsatisfiable constraints 2019-05-09 18:39:56 actually, I was (and still I'm) busy with my $dayjob and making upgrades and fixes for alpine and didn't looked at it 2019-05-09 18:40:10 pie3: have you setup your apk repos? 2019-05-09 18:40:25 i dont know how to do it rfs613 2019-05-09 18:40:56 https://wiki.alpinelinux.org/wiki/Tutorials_and_Howtos 2019-05-09 18:40:58 is it listed here? 2019-05-09 18:42:57 pie3: the command for this is setup-apkrepos. But start with "Post Install" on https://wiki.alpinelinux.org/wiki/Installation 2019-05-09 18:43:43 here specifically https://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management 2019-05-09 18:44:56 it seems mirrors are not working 2019-05-09 18:45:01 hence choose f 2019-05-09 18:45:11 mostly "bad address" 2019-05-09 18:45:15 ipv6 only? 2019-05-09 18:45:24 what it means? 2019-05-09 18:45:26 you'll need networking working to access mirrors 2019-05-09 18:45:35 Invader_Bork: so, interesting time coming for us, i.e. nights with building and hair tearing :) 2019-05-09 18:46:02 er.. what sort of network did you configure for the alpine instance in question? 2019-05-09 18:47:35 how to get it working rfs613 ? 2019-05-09 18:47:45 i havent configured any network so far 2019-05-09 18:48:19 oh, is it because i have installed/using alpine on qemu? 2019-05-09 18:48:42 you have to configure networking for apk to be able to access mirrors on the internet 2019-05-09 18:49:11 how to configure? 2019-05-09 18:49:20 uhm, let me find the wiki page, one second 2019-05-09 18:49:26 ok 2019-05-09 18:49:48 https://wiki.alpinelinux.org/wiki/Configure_Networking 2019-05-09 18:49:51 yup 2019-05-09 18:49:54 "If you're using IPv6, you should also add the following special IPv6 addresses to your /etc/hosts file: " 2019-05-09 18:49:57 how to do this? 2019-05-09 18:50:04 i mean how to add 2019-05-09 18:50:12 i am not sure whether i am with ipv4 or v6 2019-05-09 18:50:26 if you don't know, it's probably ipv4 with dhcp 2019-05-09 18:50:37 how to verify? 2019-05-09 18:50:59 or why you are sure that it si ipv4/dhcp 2019-05-09 18:51:08 i'm not sure, i'm saying it's *probably* that 2019-05-09 18:56:40 alpine has gui? 2019-05-09 18:56:47 you can install one, yes 2019-05-09 18:56:55 where can i find one? 2019-05-09 18:57:10 is it also lightweight like alpine itself? 2019-05-09 18:57:14 er, there are lots, you could look for instructions to install one on the wiki 2019-05-09 19:18:42 dalias: thanks for your report. also fabled did some investigation today. I have been busy with this CVE drama unfortunately 2019-05-09 19:21:54 danieli: i take it as a compliment that @pid_eins admit that alpine exists :) 2019-05-09 19:22:03 ncopa: ha, good point 2019-05-09 19:22:42 ncopa, yeah understandable 2019-05-09 19:22:50 it's such bullshit :/ 2019-05-09 19:22:59 whereas this firefox issue might be a *real* CVE 2019-05-09 19:23:17 not a "it's insecure if you do something utterly stupid and nondefault on top of the default config" CVE 2019-05-09 19:25:13 I think it's at least partially attributable to CVSS 9.8 and clickbait headlines 2019-05-09 19:26:38 dalias: i know, people dont care about if CVE ir real or not 2019-05-09 19:26:49 and yes, the firefox issue is way more serious 2019-05-09 19:27:16 dalias: it is not big news that firefox (or other browsers) have a serious issue, but news is that the alpine have a minor issue 2019-05-09 19:27:43 security theater 2019-05-09 19:28:34 :-P 2019-05-09 19:29:14 dalias: fabled tried to get a backtrace but was unable due to rust debuginfo bug 2019-05-09 19:29:26 :/ 2019-05-09 19:29:48 has it been determined what function the crash is in? 2019-05-09 19:29:54 he said that this is the suspect https://github.com/rust-lang/rust/issues/41775 2019-05-09 19:30:53 ah 2019-05-09 19:31:02 suspect for the bad debug info 2019-05-09 19:31:10 i thought you meant for regression in ff 2019-05-09 19:31:25 no, the bad debug info 2019-05-09 19:32:39 at least we have it recorded in the bugtracker now, where we can post updates on progress 2019-05-09 19:32:54 unfortunately i dont have time to look at it :-( 2019-05-09 19:42:26 mps, haha yes, but there is a reward at the end 2019-05-09 19:43:38 Invader_Bork: yes, I hope we will have good and fast drivers for arm finally 2019-05-09 19:44:10 ncopa, i wonder if it's practical to review the commits to the relevant components of firefox in that range 2019-05-09 19:44:21 and see if something stands out as utterly wrong 2019-05-09 19:44:55 I will start testing when the 5.2 rc-{1,2} will be released 2019-05-09 19:47:35 hmmm, do we have epub viewer in alpine or I'm blind to find it 2019-05-09 19:47:40 dalias: do you know which is the last known good version? 2019-05-09 19:47:58 mps, awesome. i see that in mesa they mentioned support only for opengl. when i'll find some free time i'll dig a bit to see what's going on with vulkan 2019-05-09 19:49:22 Invader_Bork: I built mesa successfully for test but didn't made package 2019-05-09 19:50:22 I tried their kernel with panfrost from gitlab and it worked although I didn't noticed speed boost 2019-05-09 19:50:37 speaking of epub readers, anyone know of a tui one? 2019-05-09 19:50:46 dalias: i noticed that there is a new release of firefox out. I will try upgrade to that and see if that fixes it 2019-05-09 19:51:13 mps, most importantly, do gears run haha? 2019-05-09 19:52:01 will try after alpine 3.10 released 2019-05-09 19:52:39 when's 3.10 set to release? 2019-05-09 19:53:10 ncopa, ok thanks 2019-05-09 19:53:11 this month, i think 2019-05-09 19:53:40 <_ikke_> ncopa is going on holiday for a couple of weeks, I suspect it will be after that 2019-05-09 19:54:02 are there any big upcoming changes in 3.10? 2019-05-09 19:54:59 <_ikke_> I'm not aware of any big spectacular changes, mostly upgrades 2019-05-09 19:55:21 <_ikke_> python 3.7 2019-05-09 19:55:45 "python 3.7", oh nice i've been waiting on that 2019-05-09 21:23:46 some feedback for the aarch64 installation on the overdrive 2019-05-09 21:24:02 It seems that the system successfully installed, but that there was no efi entry made 2019-05-09 21:28:33 ... which actually makes me wonder - if there is no efibootmgr, how does one create a boot entry? 2019-05-09 22:16:16 hey 2019-05-09 22:16:18 I have such trivial code: https://pastebin.com/raw/NQbEqhSY 2019-05-09 22:16:23 ehn I run it scanf is executed as a very first one and after press Enter key after providing data both printf executes 2019-05-09 22:16:33 do you have an idea why ? 2019-05-10 02:02:21 I'm getting a little bit further now. I had forgotten(?) to mkinitfs using my custom setup (for multi-booting). So now I get to the same place as before, where the EARLY boot process of the kernel is enumerating USB devices and hangs, but I see a message scroll by informing me that I am being dropped into the initramfs emergency recovery shell, but I don't actually get a prompt. I also see a 2019-05-10 02:02:21 message saying it can't access tty. Do I need a tty driver? 2019-05-10 02:03:06 (Actually, pardon me, it says it is launching the shell, not that I am being dropped into it) 2019-05-10 02:14:35 I suppose I should also mention that I cannot see the error that causes the shell to be launched. Since I cannot get into the shell, I cannot interrogate the log files or other materials to sense what went wrong. 2019-05-10 02:16:07 The error has scrolled beyond the 24 lines of the sort-of "virtual terminal" (my monitor actualy has 1920 lines, or more, but the video is set for something like 24 lines) 2019-05-10 03:13:49 Is anyone else seeing problems with dl-cdn right now? 2019-05-10 03:14:15 My apk update is just sitting there 2019-05-10 04:16:30 Ugh. My fault. IPv6 + docker was not configured correctly. 2019-05-10 06:50:51 https://twitter.com/pid_eins/status/1126458346515705857 2019-05-10 06:50:55 troll detected 2019-05-10 06:51:21 This was discussed earlier 2019-05-10 06:51:36 Concusion is that it was a cheap shot 2019-05-10 06:51:43 oops, I've looked at the IRC logs but I've probably missed 2019-05-10 06:52:48 oh ncopa responded to him btw :P 2019-05-10 06:53:00 ACTION <3 2019-05-10 06:53:09 lol 2019-05-10 07:01:01 Laurent Bercot (skarnet) answer there +1 2019-05-10 07:20:47 I see there is an envar that can be set for setup-alpine for the bootloader. There is no command line option, but this just might solve my problem. The only question I have is why is grub the default for EFI (and some other systems like IBM mainframe?), but not for BIOS-based systems? 2019-05-10 07:53:33 syslinux is a dependency of xen. So I can't use grub if I want to install xen? 2019-05-10 07:53:40 ??? 2019-05-10 07:54:05 why can't grub be an option for xen, just like syslinux? 2019-05-10 07:54:27 (and, yes, apk is INSISTENT on this. There does not appear to be a workaround I can see) 2019-05-10 09:01:21 Ok. I've taken a new approach -- answers file for setup-alpine. It makes things go faster, and less for me to figure out. Making more progress, but still stuck. 2019-05-10 09:02:16 Now I do get the emergency shell, but I can't type anything into it. The error is that it can't find /dev/sda2. Which is sorta funny, given that it is the ONLY file system (/boot and / are on one fs together) 2019-05-10 09:02:41 I've googled to find out the reason it might not be mounting, but maybe someone here already knows. 2019-05-10 09:03:18 So, i've reached the initramfs stage of the boot, but it is failing because it can't mount the root fs 2019-05-10 09:03:54 It correctly generated the grub entries, from what I can see. And I'm using the plain vanilla entry, not the xen entries. 2019-05-10 09:05:16 I note that grub-install would need to be run separately. I didn't, adding the generated entries to the Devuan grub.cfg. I've also added devuan entry to the alpine config in case I need to recover devuan (there are other ways, but this is easier) 2019-05-10 09:06:16 so I could, in theory do grub-install from the alpine installation procedure. But I see no reason to do this, other than the remote possibility that grub2 has also been altered to suit alpine. 2019-05-10 09:06:59 (does not seem likely, but it is possible, given I know little at this point about alpine) 2019-05-10 09:18:06 Hello! 2019-05-10 09:18:50 I'm searching for an easy to use manual to rebuild a package. Please help! 2019-05-10 09:21:23 <_ikke_> lloydz: https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2019-05-10 09:21:38 <_ikke_> It mainly explains how to create a new package 2019-05-10 09:22:12 <_ikke_> but the part about setup your system and account is usefull 2019-05-10 09:22:14 yes. i know that. but i only want to rebuild - without any changes. 2019-05-10 09:22:36 .. changes on the source ;-) 2019-05-10 09:22:36 you can rebuild using the original APKBUILD for the package you want to rebuild 2019-05-10 09:22:56 <_ikke_> Yes, abuild just expects a few things to be setuop 2019-05-10 09:23:06 <_ikke_> and https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Setup_your_system_and_account mostly describes those things 2019-05-10 09:23:23 <_ikke_> You can ignore the git part 2019-05-10 09:24:08 <_ikke_> lloydz: Do you already have the APKBUILD? 2019-05-10 09:25:45 yes and no. i have a second installation in an virtual machine. now i'm standing at the git part step. 2019-05-10 09:26:18 <_ikke_> You don't need git to just build the package 2019-05-10 09:26:23 ..but i think i ask the internet for an easy to use - for dummys - howto for that job. 2019-05-10 09:27:26 i found a betadocumentation for alpine linux, but maybe someone knows more that i. 2019-05-10 09:27:42 <_ikke_> That's indeed still under development 2019-05-10 09:27:49 <_ikke_> For now, the wiki is the best source 2019-05-10 09:28:01 yes, you are right. 2019-05-10 09:30:14 .. i have to gamble with that. 2019-05-10 09:31:04 I wan't to rebuild the kernel to add and remove modules. 2019-05-10 09:32:52 But i don't want to lose any dependences. 2019-05-10 09:33:05 you're gonna want to set up a working abuild environment, pull these[1] files into a directory, and edit config-$flavor.$architecture 2019-05-10 09:33:07 [1]: https://git.alpinelinux.org/aports/tree/main/linux-vanilla 2019-05-10 09:35:06 ..or security features come with alpine. 2019-05-10 09:35:36 ok. I know the architecture ;-) thank you! 2019-05-10 09:35:41 <_ikke_> the apkbuild + files should contain everything 2019-05-10 10:16:38 how often does alpine release kernel updates 2019-05-10 10:17:58 cim209: you ask for major or minor releases 2019-05-10 10:18:30 oh just asking for a scenario in a production environment, sorry should've added that to my question 2019-05-10 10:18:35 we stick to LTS 2019-05-10 10:18:46 how often do you reboot on an LTS? 2019-05-10 10:19:10 I generally don't reboot unless a newer kernel has features I need or super important security patches, but that's just me 2019-05-10 10:19:38 when the new minor release appears it took about day or two to make new version for alpine 2019-05-10 10:19:44 danieli: what was your longest uptime before rebooting 2019-05-10 10:19:56 I don't know off the top of my head, a couple years 2019-05-10 10:20:02 not on alpine, but in genral 2019-05-10 10:20:06 s/genral/general/ 2019-05-10 10:20:45 I had a debian server with more than 1200 days uptime 2019-05-10 10:21:00 alpine's nld5 box has 395 days of uptime 2019-05-10 10:21:25 gbr1 had quite a bit too but iirc it was rebooted for some firewall stuff a while ago 2019-05-10 10:21:39 i'm running on ubuntu right now but thinking of switching 2019-05-10 10:21:45 for production servers 2019-05-10 10:21:58 just curious, what's your incentive? 2019-05-10 10:22:34 danieli: just asking cause ubuntu server reboots once a month 2019-05-10 10:22:55 it.. reboots by itself...? 2019-05-10 10:23:04 sounds like something is very off if that is the case 2019-05-10 10:23:10 no, i get nagged by the "system needs reboot" 2019-05-10 10:23:20 when opening a new ssh connection 2019-05-10 10:23:51 also my websites are in docker containers 2019-05-10 10:24:05 you can probably just disable the motd and it'll disappear 2019-05-10 10:24:19 cim209: are you sure it is not windows which impersonate itself as ubuntu :D 2019-05-10 10:24:24 iirc it says "*** System restart required ***" 2019-05-10 10:24:26 yeah but then i'll be experiencing cognitive dissonance lol 2019-05-10 10:24:27 is that the one? 2019-05-10 10:24:46 yeah that one 2019-05-10 10:25:10 there are mechanisms like kgraft/ksplice/kexec/whatever on some linux distributions so you in theory don't ever have to restart 2019-05-10 10:25:35 and i would recommend holding back the linux kernel and header packages and update them manually 2019-05-10 10:26:01 i do have livepatch enabled on my machines 2019-05-10 10:26:15 i don't see them working though 2019-05-10 10:26:18 looks like 5.2 will have live patching with the help of gcc 9 2019-05-10 10:26:35 canonical livepatch is not a silver bullet 2019-05-10 10:26:40 mps: cool 2019-05-10 10:26:42 mps: is that gonna be enabled by default? 2019-05-10 10:26:47 i would assume not 2019-05-10 10:27:17 don't know, yesterday read announce on phoronix 2019-05-10 10:27:32 i see something about it on phoronix from early april 2019-05-10 10:27:43 and something about it from late november last year 2019-05-10 10:28:39 as danieli says, it will not be silver bullet but could help in small changes, I hope 2019-05-10 10:29:52 i do like the idea of alpine being really minimal and secure by default 2019-05-10 10:30:03 all of my dependencies are installed in docker images 2019-05-10 10:30:14 doesn't mean you'll *never* have to reboot, there's more nuance to it than that 2019-05-10 10:30:20 probably it will need year or two to stabilize and become useful for long running servers 2019-05-10 10:30:41 yeah the reboot thing i'm fine with 2019-05-10 10:31:22 i also do not know of any corporate backers or companies offering paid support for alpine 2019-05-10 10:32:22 you'll have to assess your needs and measure them up to different distributions 2019-05-10 10:32:58 i have this script in bash https://demyx.sh 2019-05-10 10:33:04 danieli: why would they, alpine is simple ;) 2019-05-10 10:33:18 'simple' is relative 2019-05-10 10:34:04 i'm gonna make it compatible with all linux and put every dependencies in the image 2019-05-10 12:11:03 cim209: is that compose doing the work? 2019-05-10 12:11:26 yes] 2019-05-10 12:12:22 looks like it by scrolling through a tiny bit of the script 2019-05-10 12:12:27 some docker run 2019-05-10 12:15:35 i've got 30 unpushed commits 2019-05-10 14:25:30 I'm running alpine-3.9 under qemu as per the wiki. Trying to get XFCE or MATE desktop. Mostly works, but can't move mouse. I ran evtest on console, there are two entries for "VirtualIPS/2 VMware VMMouse". THe first one /dev/input/event2 even shows mouse movement and clicks. The second one shows no events. I've got xf86-input-evdev installed. Any idea why X doesn't seem to see mouse events? 2019-05-10 15:08:31 rfs613: have you tried with xf86-input-libinput? 2019-05-10 15:09:34 doggone: yep, have that installed too. 2019-05-10 15:10:43 the strange bit is that Xorg.log shows no obvious error. But it also doesn't show any indication of using /dev/input/eventN either... 2019-05-10 15:11:33 I've tried using "Xorg -config" to generate a config... modified that config with the 10-evdev.conf settings... ran with that config but still no dice. 2019-05-10 15:12:05 second, i'll try setting up a vm too 2019-05-10 15:12:40 it's alright... i was just doing this to help out someone else who asked yesterday... 2019-05-10 15:16:37 works just fine here. I just ran setup-xorg-base and apk add xfce4 2019-05-10 15:16:40 oh, from log file: "Hotplugging is on, devices using 'kbd', 'mouse' or 'vmmouse' will be disabled" 2019-05-10 15:16:59 the video driver is qxl 2019-05-10 15:17:13 oh, not xf86-modesetting? 2019-05-10 15:17:38 setup-xorg-base installs both 2019-05-10 15:19:21 oh hey, working. I re-ran setup-xorg-base. It installed hwdata-pci, pciutls-libs and pciutils packages. 2019-05-10 15:19:32 ah :) 2019-05-10 15:19:39 can't explain why it didn't install those the first time though 2019-05-10 15:19:42 anyhow thanks ;-) 2019-05-10 15:55:13 why does the alpine bug tracker have no obvious place to add comment? i finally figured out you have to "edit" the "include a note" on the edit. wtf 2019-05-10 15:55:32 (yes this should be a complaint about redmine, not alpine) 2019-05-10 15:55:36 Yeah, took me a sec too 2019-05-10 15:55:36 <_ikke_> :) 2019-05-10 15:55:59 this is like the most intentionally obtuse behavior i've ever encountered in a bug tracker 2019-05-10 15:56:36 <_ikke_> To edit the actual bug description, you have to click on a pencil icon that you will most likely miss if you are not aware of it 2019-05-10 15:57:43 yeah no, redmine isn't exactly eye candy nor super intuitive 2019-05-10 16:02:35 honestly bugzilla is the friendliest ux to me, maybe just because it's what i'm familiar with tho 2019-05-10 16:03:15 sorry if this is ot; just thinking about it because we still need a bug tracker for musl and i notice good and bad points when i use them with the idea of eventually choosing one 2019-05-10 16:10:55 <_ikke_> I don't have any issues with redmine personally because I'm used to that 2019-05-10 16:11:27 i don't like it, but i am used to it 2019-05-10 16:11:33 well.. for the most part. 2019-05-10 16:11:50 yeah, i can see it working fine if you're used to it 2019-05-10 16:11:56 but it's completely hostile to new bug reporters 2019-05-10 16:14:08 Well, Bugzilla seems just as hostile to me 2019-05-10 16:14:13 I'm super happy GNOME has moved to Gitlab which is the main thing I care about 2019-05-10 16:14:39 <_ikke_> I'm just sad that git(lab|hub) issue tracking is quite limited 2019-05-10 16:15:17 Yes, it doesn't have some nice stuff like more fine grained issue statuses 2019-05-10 16:15:25 But oh well, it works for me 2019-05-10 16:15:36 <_ikke_> It barely works for me 2019-05-10 16:15:46 <_ikke_> In a larger scale 2019-05-10 16:15:47 Oh? How come? 2019-05-10 16:15:49 <_ikke_> for small projects it's fine 2019-05-10 16:16:05 <_ikke_> Hard to keep track of issues / prs 2019-05-10 16:16:34 <_ikke_> See for example aports with 300+ open PRs 2019-05-10 16:16:49 Hum, at least im Void it worked pretty well, I can't really tell how that aspect with Bugzilla/Redmine better 2019-05-10 16:16:50 <_ikke_> The most recent PRs get the most focus 2019-05-10 16:17:02 i hate github's ui/ux. haven't used gitlab's as much 2019-05-10 16:17:05 Well, the 300+ PRs wouldn't be better on Patchwork, would they? 2019-05-10 16:17:10 <_ikke_> Nope 2019-05-10 16:17:52 <_ikke_> In redmine and I can imagine in bugzilla as well, you can add much more granular filters / sorting 2019-05-10 16:18:01 So it's more of a disproportionate amount of contributors compared to reviewers than the platform, I guess 2019-05-10 16:18:20 Yes, the filters are nicer on those two, labels aren't that nice 2019-05-10 16:18:28 cogitri: queue voidlinux/void-packages with 400+ PRs 2019-05-10 16:18:46 Heh, I guess a certain reviewer is missing :P 2019-05-10 16:20:22 i prefer working from the bottom of ticket/pr queues, i.e. fifo 2019-05-10 16:20:40 I very much like that GL/GH offer one platform for both bugreports and PRs 2019-05-10 16:21:20 At least I'd barely ever check an external bugreport thingie unless I'm pinged on there (or I happen to moderate it like I do here with Simon). 2019-05-10 16:54:52 cogitri, PRs are a lot less interesting to me because they're almost always not suited for direct merging 2019-05-10 16:55:14 commit messages are usually bad, commits not properly factored into independent changes, etc. 2019-05-10 16:55:39 i find git-am'able attachments easier to work with 2019-05-10 16:57:15 fwiw you can get .patch files from PRs 2019-05-10 16:57:19 just append .patch 2019-05-10 16:57:29 <_ikke_> that's what I do 2019-05-10 16:58:34 yes 2019-05-10 16:59:32 <_ikke_> curl .patch | git am 2019-05-10 17:06:01 Dalias: Yes, that has nothing to do whith the platform usually 2019-05-10 17:06:56 Maybe you less les low quality contributions because most contributors give up because the have to use smth like MLs for contributing 2019-05-10 17:10:12 Cogitri: other direction also, less contribution from users who are accustomed to ML 2019-05-10 17:15:52 Hm, true that, but I feel like more people dislike MLs these days than GH/GL 2019-05-10 17:18:04 _ikke_: to be fair alpinelinux turned off projects which are for sorting issues 2019-05-10 17:19:42 Cogitri: this is also my impression, but serious contributors use what is default for project 2019-05-10 17:47:29 i'd love a platform that integrates with mailing list well 2019-05-10 17:47:51 where you can use a web interface if you want to reply and comment on patches 2019-05-10 17:47:58 but it just bridges to the list 2019-05-10 17:48:42 doesn't sr.ht want to do that ? 2019-05-10 17:50:02 yes, sr.ht works in that direction but is not yet mature enough, imo, although it is fine 2019-05-10 17:50:22 <_ikke_> But don't you get the lowest common denominator effect? 2019-05-10 17:51:42 I like lowest common denominator 2019-05-10 17:52:32 <_ikke_> The worst of both worlds? 2019-05-10 17:53:49 or interoperability of more worlds 2019-05-10 19:45:01 just noticed that in my edge lxc build container time is off the two hours from host 2019-05-10 19:45:44 do we need ntp client in lxc guests? 2019-05-10 19:47:14 or setting TZ or /etc/localtime is enough 2019-05-10 19:50:38 i don't think linux has such a thing as "time namespaces", so container should not have its own version of the clock that would need setting 2019-05-10 19:50:59 compare time in UTC rather than local and see if it matches 2019-05-10 19:51:02 TZ=UTC date 2019-05-10 19:51:28 if that does then you're just seeing the result of having different timezones insde and outside the container 2019-05-10 19:51:50 timezone does not change the clock at all, it's just a matter of how it's interpreted for presentation 2019-05-10 19:53:20 dalias: thanks, I understand how timezone works, and UTC, I'm asking if it is enough to set to same as TZ as in host 2019-05-10 19:55:04 I read how it works for musl libc on musl site 2019-05-10 19:56:04 but for me it doesn't work, only setting link to /etc/localtime works 2019-05-10 19:57:09 meh, sorry, I messed set and setenv in tcsh, it work, ofc 2019-05-10 20:02:58 :) 2019-05-10 20:04:52 (must be that I'm tired after all day behind the screen and keyboard) 2019-05-10 20:18:38 I made modification to the init script (initramfs-init) in order to put in some debugging. I put in some statements to get debug info. But, somehow, instead of showing me the output and pausing for input (I inserted a "read" at the end of the instructions), it now boots all the way to the login prompt! (Oh, I had to add console=hvc0 on the cmdline at boot) 2019-05-10 20:19:31 What the heck happened? So, by inserting some debug statements in the init script, it bumped bytes around just enough so the system was no longer unable to find /dev/sda2 (my root fs)??? 2019-05-10 20:19:37 Really? 2019-05-10 20:20:43 Then I noted that not all the xen daemons had started (after login). When I tried them manually, the system again complained about a missing resource... and on it goes. 2019-05-10 20:21:55 I was hoping to use alpine as my dom0. But it seems to me, at least, alpine is not stable enough, or the tools not well-developed enough, to do this reliably. Would anyone here recommend waiting for a future release of alpine (I'm still rather interested) 2019-05-10 20:21:57 ? 2019-05-10 20:25:05 I wonder if, perhaps, certain system packages are being updated, more or less ongoing (makes sense), and those are being pulled into each of my attempts, further changing the variables. Idk. I'm a bit frustrated by all of this. 2019-05-10 20:25:11 systemdlete: i use alpine as xen guests in production, no big issues 2019-05-10 20:25:24 mps: Do you use it as dom0? 2019-05-10 20:25:59 also, what hardware are you on? 2019-05-10 20:26:32 Someone here a few days back mentioned that desktop boxes have some issues. 2019-05-10 20:26:34 no, as guest only, although I intended to test dom0 but didn't found time 2019-05-10 20:27:51 well, on desktop I use it about year and a half, on different architecture. I'm really satisfied 2019-05-10 20:28:33 and, I'm not alone :) there are a lot of with alpine on desktop 2019-05-10 20:28:38 which version(s) are you running 2019-05-10 20:28:57 3.9, stable 2019-05-10 20:28:59 I'm attempting 3.9.3 with xen 2019-05-10 20:29:36 did you install the xen version? Or did you install one of the generics and add xen afterward? 2019-05-10 20:29:49 (I'm trying to figure out what you and I did differently) 2019-05-10 20:29:54 desktop under VM could be problematic for inexperienced in alpine 2019-05-10 20:30:17 No, this is directly on the hardware. 2019-05-10 20:30:35 I mean, I run desktops native 2019-05-10 20:30:46 honestly, i wish it wasn't so, but alpine isn't exactly optimal as a daily driver 2019-05-10 20:30:58 danieli? 2019-05-10 20:31:06 just adding to what mps said 2019-05-10 20:31:08 "daily driver?" 2019-05-10 20:31:18 right, OS for day-to-day workstation use 2019-05-10 20:32:11 Sorry. I'm having trouble following you. I said "daily driver" meaning to inquire what you meant by that. You responded with "Right" -- what was i right about? 2019-05-10 20:32:12 eh, forgot to tell, I have one lxc with debian for some specific software I need which is not packaged in alpine 2019-05-10 20:32:31 an lxc running under xen on alpine? 2019-05-10 20:32:38 I wasn't aiming at your issue, don't worry 2019-05-10 20:32:46 no, lxc on alpine 2019-05-10 20:33:05 so you are not running xen then? 2019-05-10 20:33:16 ON THE HOST 2019-05-10 20:33:36 on cloud service xen guests as production servers 2019-05-10 20:33:54 I've successfully installed and run alpine under vbox, but no where else, until today on hw 2019-05-10 20:34:05 on desktops native alpine, and one lxc on it for debian 2019-05-10 20:34:11 I haven't ran alpine as a xen dom0 vm, so I have no clue about that specifically 2019-05-10 20:34:37 danieli, please explain what you mean by "alpine isn't exactly optimal" -- in what ways? 2019-05-10 20:34:44 also I didn't tried dom0 xen 2019-05-10 20:36:06 you didn't quote the most important part of what I said. I meant that it's unergonomic as a daily driver (day-to-day workstation OS with GUI) 2019-05-10 20:37:29 yeah, I got that. But I mean, specifically, what is wrong with it in your opinion? 2019-05-10 20:38:32 it depends, my 'gui' is awesome wm with about twenty st terms, one firefox, sometimes libreoffice and evince, so for me it comfortable 2019-05-10 20:39:12 (I was asking danieli. I take it that you, mps, like alpine; it works for you) 2019-05-10 20:39:16 and Xephir if need debian kicad (electronic design CAD) 2019-05-10 20:43:03 things break once in a while, especially larger projects that require a lot of patches, like chromium and firefox (because of rust) 2019-05-10 20:43:23 some utilities and software i need for work also isn't packaged for alpine, and it's not going to be trivial 2019-05-10 20:43:49 I like the idea of secure, lightweight, simple. Especially for a dom0; that's what I found so alluring about alpine. But it appears that maybe it is not quite ready for consumption yet. Or maybe I just need to install a more generic image and then add xen later. 2019-05-10 20:44:16 danieli: ofc, I didn't meant to say that is alpine appropriate for everyone 2019-05-10 20:44:32 it's quite all right for a server and container OS, but it's not for me when it comes to daily drivers 2019-05-10 20:46:46 I am also gussing that maybe the xen image is new this release or so, and maybe it needs some more work yet. I just wish someone could confirm this. If it were simply something I am doing, then I would expect much different results. Instead, I am getting different results depending on changes to the init script? The changes are not functional changes, just some debug statements and simple f 2019-05-10 20:46:46 ile operations. Those should NOT cause the error message and the emergency shell to be skipped all of a sudden. 2019-05-10 20:47:18 i personally don't know, but i do know that people use alpine both as a xen dom0 and domU 2019-05-10 20:47:26 This reminds me of the fortran compilers I used in college. Insert a debug statement in the code, and suddenly the bug disappeared. 2019-05-10 20:47:39 Which points to NOT a user error. 2019-05-10 20:48:19 ha, heisenbug 2019-05-10 20:49:30 yeah. Heisenbug. IMO, that points to some instability in a compiler, linker, or OS interface somewhere. 2019-05-10 20:50:32 the fact that Alpine is using musl could be suspect. Not that musl is an issue, necessarily, but musl is much newer "technology" than the older and venerable gnu compilers. Musl was another plus that lead me to Alpine. 2019-05-10 20:50:57 In theory, musl should be much more stable and reliable than the old libraries 2019-05-10 20:51:14 (I mean gnu libs, not compilers) 2019-05-10 20:54:39 I am running AMD am3 generation hardware. I've never seen an issue with it for any other distro I've installed. It's stable, fast, and quiet. 2019-05-10 20:55:37 But maybe xen or alpine are depending on some very new stuff, such as EFI, which neither of my boxes support. 2019-05-10 20:57:17 how old is your hardware, generation-wise? Are you using the latest hw, or something more dated, maybe up to 10 or 12 years old? In the pioneering days of Linux, every effort was made to ensure back-compat for older hardware. These days, it seems this may be becoming less of a consideration. 2019-05-10 20:57:50 Understandable, but still even hw a decade old is still useful, even if not quite as fast or not having the latest gadgetry yet. 2019-05-10 21:00:28 ... are you speaking to someone I'm not seeing? my ignorelist is empty 2019-05-10 21:00:42 wtf 2019-05-10 21:01:25 <_ikke_> I see mps, danieli and systemdlete being active 2019-05-10 21:01:52 I haven't seen mps in like 15-20 posts 2019-05-10 21:02:27 <_ikke_> 15 minutes ago 2019-05-10 21:03:02 it seemed like systemdlete was replying to someone 2019-05-10 21:03:13 ACTION becoming tired after all day behind screen 2019-05-10 21:25:01 mps: talk to me about it. working with an injured wrist with carpal tunnel >.< 2019-05-10 21:27:30 today started to feel something similar, I must start again my sport activities, I see 2019-05-10 23:27:46 hi 2019-05-10 23:28:13 i have a question about mkpasswd & chpasswd. i can't figure out how to make them work. 2019-05-10 23:28:24 # passwd test 2019-05-10 23:28:42 2019-05-10 23:28:51 grep ^test /etc/shadow 2019-05-10 23:29:03 mkbucc: your username is test?? 2019-05-10 23:29:19 sorry, didn't mean to put two '?'s :) 2019-05-10 23:29:25 correct 2019-05-10 23:29:31 grep ^test /etc/shadow | cut -d: -f2 2019-05-10 23:29:36 $6$/R335DLZUJeBjNN1$WLWyaU6m1zXT/5v1EmnPCeEnPUvJSDF2p41snmzdLseLKk4VOc81XMCz1noRPYRhBDP3ZXkAGixOFdMZgWJzH0 2019-05-10 23:29:51 $6$beRZRmvJaXdVYtZA$3FdNCN50R2uriEyMcYX4gR4zd1PKHi/070jQ168XP0zEpTQbx0yWCP.FIbWJtp0XUq5fVee7ksZLlyJNllJ9c. 2019-05-10 23:30:01 is the output from mkpasswd test 2019-05-10 23:30:13 so ... i manually reset the test users password to test 2019-05-10 23:30:20 then used mkpasswd 2019-05-10 23:30:33 and compared the output to shadow ... and they are not the sam.e 2019-05-10 23:30:45 mkbucc: No, because they have different salts 2019-05-10 23:30:48 That's by design 2019-05-10 23:31:36 mkbucc: a better way than grep is: getent shadow test 2019-05-10 23:31:45 :) 2019-05-10 23:32:03 so, how to i use mkpasswd + chpasswd -e to update a user's passwd? 2019-05-10 23:32:37 mkbucc: mkpasswd's output is fine - it doesn't have to match 2019-05-10 23:32:47 it's not meant to match, all the hashes are meant to be different due to the salt 2019-05-10 23:33:25 mkbucc: in other words, you're aiming at the wrong target when you're trying to get the exact same hash 2019-05-10 23:33:32 ok. 2019-05-10 23:33:44 let's see, i need to test 2019-05-10 23:33:48 haven't done that on alpine lately 2019-05-10 23:34:02 mkbucc: you can provide mkpasswd -S SALT to generate one that matches 2019-05-10 23:34:02 mkbucc: I mean, if you *WANTED* them the same, you *CAN* specify the salt 2019-05-10 23:34:07 mkpasswd -S beRZRmvJaXdVYtZA test 2019-05-10 23:34:07 $6$beRZRmvJaXdVYtZA$3FdNCN50R2uriEyMcYX4gR4zd1PKHi/070jQ168XP0zEpTQbx0yWCP.FIbWJtp0XUq5fVee7ksZLlyJNllJ9c. 2019-05-10 23:34:21 no matter how many times you run that, the output will be the same 2019-05-10 23:34:39 but, generally speaking, re-using salts defeats the entire purpose of salting 2019-05-10 23:34:43 er, it seems to work for me 2019-05-10 23:35:18 it had better work, or login is going to fail ;-) 2019-05-10 23:35:31 i know, i'm obviously missing something. 2019-05-10 23:35:32 :) 2019-05-10 23:35:54 my test was to generate passwd with mkpasswd, then echo "test:" | chpasswd -e 2019-05-10 23:36:02 then try to login as the user. failed every time. 2019-05-10 23:36:07 hm, let me try 2019-05-10 23:36:11 -e writes what is passed in right to shadow 2019-05-10 23:36:12 https://git.busybox.net/busybox/tree/loginutils/chpasswd.c#n90 2019-05-10 23:36:22 Well, I've never used chpasswd, so I can't speak to that 2019-05-10 23:36:52 presumably there's something about the CLI API of chpasswd you're using incorrectly 2019-05-10 23:37:27 if the echo is quoted like that the $ will be treated as a shell variable lookup 2019-05-10 23:37:38 you'll need to single quote if you use echo 2019-05-10 23:38:25 https://tpaste.us/RgOv?hl=true 2019-05-10 23:38:35 het werkt 2019-05-10 23:38:54 oops, i missed a ")" when copying. oh well. 2019-05-10 23:40:05 doh ... shell quoting. 2019-05-10 23:43:03 something like this is often useful to find out what is actually being input to the command echo "foo" | tee /dev/stderr | chpasswd -e 2019-05-10 23:46:11 echo "test:$(mkpasswd test)" | chpasswd -e would work fine 2019-05-10 23:46:23 lol 2019-05-10 23:46:30 ? 2019-05-10 23:46:32 nice. 2019-05-10 23:47:13 i like the brevity. but my use case is ansible, and so I don't want to store the clear text value. 2019-05-10 23:47:22 this was a big help, thanks for taking the time. 2019-05-10 23:48:00 aha 2019-05-10 23:48:03 mkbucc: you're welcome 2019-05-10 23:48:08 no worries ;) 2019-05-10 23:48:20 here is my favorite: printf '%s:%s\n' root "$(openssl rand 35 | openssl enc -a)" | chpasswd 2019-05-10 23:48:54 this is probably better though: passwd -l root 2019-05-10 23:50:38 pwd=$(openssl rand 35 -hex 32) 2019-05-10 23:51:27 hex it such a limited character space though 2019-05-10 23:53:44 root:d4dca807ae83cf9759fc0101ea8cd3158c1b0c2f05b023e2042147e97b970cea495216 vs root:vcVj6PIOMxZSjIn1jKrcHqXWSZF4D6nxtV+dFN0f0hYcang= 2019-05-10 23:57:05 true. though 32 character hex password, per Dashlane, would take three sesvigintillion years to crack. 2019-05-10 23:57:42 I use whatever the maximum for each service is 2019-05-10 23:57:54 some services are really nasty and truncate the entered password invisibly 2019-05-10 23:58:35 my entropy is usually 150-810 bits somewhere 2019-05-10 23:59:18 danieli: just generate a new rsa key and feed that in as your password ;-) 2019-05-10 23:59:30 heheh 2019-05-11 00:01:19 that was a joke, but the results from ssh-keygen aren't terrible. ecdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBDW/9HT4Ts6nnQsOefznFOt5e8C+bpdhH3vypsjeRT9tO0dNZ4+ntuaTgnf1vJuZA37ZRutZ5ScPbsJO2oX32Kc= 2019-05-11 02:18:49 hello 2019-05-11 02:18:50 hola 2019-05-11 02:28:16 can i use .DEB ? 2019-05-11 02:32:16 well you can install dpkg and then install a .deb, but I would not recommend it since Alpine used MUSL Libc and most Debian packages are compiled against GNU Libc. 2019-05-11 03:01:41 So I just download mate-polkit, how do I start it? 2019-05-11 06:07:09 should i use the extended iso for servers? 2019-05-11 06:07:19 or standard 2019-05-11 06:07:55 <_ikke_> standard should suffice 2019-05-11 06:09:01 <_ikke_> The only difference is that extended has more packages shipped on the iso 2019-05-11 06:13:26 hmm 2019-05-11 06:13:37 digital ocean doesn't seem to have an option 2019-05-11 11:48:58 yup, no alpine there, unless you install it yourself / use a custom image 2019-05-11 14:23:15 In an alpine container I'm trying to set `net.ipv6.conf.all.disable_ipv6 = 1` in /etc/sysctl.d/01-my.conf or /etc/sysctl.conf but the value doesn't stick on boot. Any idea what I might be doing wrong? The only file with settings is 00-alpine.conf but that should be overwritten by either file I mentioned. 2019-05-11 14:23:58 `/etc/sysctl/00-alpine.conf` that is 2019-05-11 14:39:34 <_ikke_> A container is not booted 2019-05-11 14:39:47 <_ikke_> adrian_1908: so there is nothing run on boot to read that value 2019-05-11 14:41:15 _ikke_: Hmm, but don't the init scripts do a similar job though? I'd suspect that config file isn't left in there for nothing, since containers are quite trimmed/customized. 2019-05-11 14:41:38 <_ikke_> adrian_1908: there is no init system 2019-05-11 14:41:39 if you're using docker, there isn't even an init system 2019-05-11 14:41:41 yes 2019-05-11 14:41:51 <_ikke_> a container is just a single process that is run 2019-05-11 14:41:53 Ok, learned something new today. 2019-05-11 14:41:54 <_ikke_> nothing more, nothing less 2019-05-11 14:42:06 <_ikke_> people assume containers are like virtual machines 2019-05-11 14:42:08 <_ikke_> they are not 2019-05-11 14:42:26 it's more similar to a zone or jail 2019-05-11 14:42:33 <_ikke_> or chroot 2019-05-11 14:42:35 or that 2019-05-11 14:42:46 I know they're a child process, I just thought they go through a lot of the motions at imitate normal bootup. 2019-05-11 14:42:57 <_ikke_> Nope, not for docker containers 2019-05-11 14:43:12 <_ikke_> lxc is a little bit more like that 2019-05-11 14:43:21 Ok then, will try another way. It works at runtime so I'll call it somehow. Thanks. 2019-05-11 14:43:54 <_ikke_> (ftr, there is nothing preventing you from using something like openrc as the first process which would then start other processes 2019-05-11 14:43:56 <_ikke_> ) 2019-05-11 14:44:54 then again, containers use the host kernel 2019-05-11 14:45:17 i can't remember how changing stuff like net.* with sysctl in a container would work, if at all 2019-05-11 14:50:27 You're right, having ipv6 on the host but not on in the container doesn't make sense. I think disabling it just remove ipv6 from the container veth interfaces, making it look like it worked. 2019-05-11 14:51:31 It's confusing though, that I can even touch sysctl inside the container and it'll update its state. 2019-05-11 14:51:41 it's not a real system 2019-05-11 16:43:51 How do I start mate-polkit? 2019-05-11 16:46:28 KH405: it should provide a .desktop file to be autostarted 2019-05-11 16:46:35 if not, uh, try `/usr/libexec/polkit-mate-authentication-agent-1` perhaps? 2019-05-11 16:53:46 When I try the second thing you posted I get 2019-05-11 16:54:18 Unable to determine the session we are in: GDBUS.Error 2019-05-11 16:54:43 The name org.freedesktop.ConsoleKit was n ot privded by any .service Files 2019-05-11 16:55:49 Where should the .desktop file be located ? 2019-05-11 16:56:00 i don't use mate so i'm not too familiar with it 2019-05-11 16:56:14 I guess you need ck2 installed then 2019-05-11 16:56:19 Or elogind w/ polkit-elogind 2019-05-11 16:56:27 mm 2019-05-11 16:56:32 i would assume ck2 2019-05-11 16:56:51 I have no clue what those packages are 2019-05-11 16:59:55 Anyway to check which one would be needed ? 2019-05-11 17:00:14 Aren't they suppose to be in 'dependencies'? 2019-05-11 17:01:08 Yes, I guess I missed that. 2019-05-11 17:05:16 So any other clues & 2019-05-11 17:07:28 Is everything it shows? 2019-05-11 17:08:38 Let me paste it, sorry i wasn't working remotely 2019-05-11 17:09:33 I'm not getting the same error while SSHed ... 2019-05-11 17:09:56 But yeah, this is the only error it gives me 2019-05-11 17:45:31 Hello. I've been trying to set up Alpine diskless on a USB flash drive. It boots and runs fine, but the problems arise when trying to set up lbu. /dev/sda1 contains Alpine, /dev/sda2 is a fat32 formatted partition covering the rest of the flash drive (around 20G). I'm unable to mount sdb2 as it's considered "busy," though this doesn't make sense to me. Only after I physically remove and reattach them, 2019-05-11 17:45:32 they show up and I'm able to mount both, and set them up for cache & lbu, however this is obviously not the way to handle it. I'm at a loss. If further details are needed I'll be happy to give them. 2019-05-11 18:08:51 KH405: Hey, mind running /usr/lib/polkit-1/polkitd as root? 2019-05-11 18:15:54 I'm geting some weird errors now 2019-05-11 18:17:01 It seems to start then it says: Finished loading, compiling and executing 2 rules 2019-05-11 18:17:30 Entering main event loop, connected to the system bus 2019-05-11 18:17:57 Lost the name org.freedesktop.policykit - exiting shutting down, exiting with code 0 2019-05-11 18:18:19 And I still can't access it with virt-manager 2019-05-11 19:17:56 FYI docker has --sysctl flag that can be used to disable IPv6 2019-05-11 19:19:03 I just used that recently when fixing prefix bug in my network setup 2019-05-11 19:27:51 KH405, sorry, but please provide a full log, otherwise it's hard to help 2019-05-11 19:28:01 Lost the name usually means another instance is open 2019-05-11 19:28:08 Try with --replace 2019-05-11 20:40:56 ncopa, the new firefox pkg doesn't fix the crash, does it? 2019-05-11 20:41:02 66.0.5 2019-05-11 20:43:06 dalias: I think ncopa is on vacation 2019-05-11 20:43:17 <_ikke_> Yes, officially starting from tomorrow 2019-05-11 20:43:56 iirc, fabled worked on firefox fixes, maybe ask him 2019-05-11 20:48:41 ah 2019-05-11 21:03:02 Cogitri: --replace worked, but I'm still not able to connect to it with virt-manager 2019-05-11 21:04:16 I get checkautorization: The name org.freedesktop.policykit1 was not provided by any .services file 2019-05-11 22:25:16 I keep seeing the error "Unexpected character in locale name '2E." in various programs 2019-05-11 22:25:27 is it from musl? 2019-05-11 22:25:29 Does the polkit needs to also be on the ''server'' with virt-manager? 2019-05-11 22:25:30 if so, why? 2019-05-11 22:26:08 danieli, i don't think so. what is your locale set to? 2019-05-11 22:26:16 haven't changed it, fresh install in a VM 2019-05-11 22:26:19 @ dalias 2019-05-11 22:26:33 what software prints it? 2019-05-11 22:26:33 I've seen it from various programs today, once from postgres when fiddling with it, once from urxvt now 2019-05-11 22:26:37 Xorg printed it earlier too 2019-05-11 22:26:59 is that the literal message (including quotation marks ? 2019-05-11 22:27:11 er 2019-05-11 22:27:12 panic: locale.c: 893: Unexpected character in locale name '2E. 2019-05-11 22:27:15 that's it, in verbatim 2019-05-11 22:27:18 s/in// 2019-05-11 22:27:37 I'm seeing it in perl source when googling 2019-05-11 22:27:45 it's possible that alpine perl is really broken for some reason 2019-05-11 22:27:54 looks like a perl issue 2019-05-11 22:27:59 _ikke_: google this, with the quotes: "Unexpected character in locale name" 2019-05-11 22:28:08 yeah, something is off, and it's probably perl 2019-05-11 22:29:59 Anybody using Alpine Linux for a commercial appliance, such as a router or firewall? What is the best approach to completely lock down the back end? PaX/RBAC or something else? 2019-05-11 22:37:11 danieli, from here it looks like perl is doing some nonsense of parsing the output of setlocale(LC_ALL, 0): https://github.com/perl11/cperl/blob/master/locale.c 2019-05-11 22:37:23 which is nonsense; the format of that string is implementation-defined 2019-05-11 22:37:50 yeah no, this looks stupid 2019-05-11 22:38:42 but i'm not so sure this is recent code (on gh) 2019-05-11 22:38:46 alpine should patch out whatever idiotic thing it's doing 2019-05-11 22:38:52 blame shows the last commit for that code was a year ago 2019-05-11 22:40:39 hm. this code really is that old, a year ago 2019-05-11 22:40:55 https://github.com/Perl/perl5/blame/v5.28.2/locale.c 2019-05-11 22:41:30 heh that's not really old for perl 2019-05-11 22:41:34 i am aware 2019-05-11 22:41:37 i wouldn't have been surprised if it hadn't been touched in 15 years 2019-05-11 22:41:44 but hey, it's perl, i'm not that surprised in general 2019-05-11 22:41:55 it's broken anyway, and i'm reinstalling the vm on 3.9 rather than edge 2019-05-11 22:42:02 i don't think there's any reason to believe it's a new regression 2019-05-11 22:42:07 looks just like nonportable crap in perl 2019-05-11 22:42:17 i had to patch postgres earlier because perl broke with strange errors, had to add --no-locales or something deep in an obscure makefile 2019-05-11 22:42:31 but urxvt has worked in the past? 2019-05-11 22:42:35 and so has postgres 2019-05-11 22:42:52 it only broke 5.26.x -> 5.28.2 2019-05-11 22:44:13 danieli: https://github.com/void-linux/void-packages/blob/da7eb611cbc7df8ab379082e0c86e4a871ce0750/srcpkgs/perl/template#L182 2019-05-11 22:44:14 Could be related to this ? 2019-05-11 22:44:21 Void Linux also had locale problems on musl when it upgraded to 5.28 2019-05-11 22:44:34 north1: [ expletives ] 2019-05-11 22:44:52 thanks for the heads up, you're the best - I'll compile it with that fix and retry 2019-05-11 22:51:09 thats at best a workaround, not a correct change 2019-05-11 22:51:17 agered 2019-05-11 22:51:19 agreed* 2019-05-11 22:53:45 i just reinstalled that VM to downgrade to 3.9, and now i have to upgrade it to edge to test perl 2019-05-11 22:53:46 FML 2019-05-11 23:03:50 danieli: If you're going with that then it also needs every perl module that links to libperl to be rebuilt as well 2019-05-11 23:04:08 i figured.. might as well upgrade to latest perl then 2019-05-11 23:04:26 we've been rebuilding things built against perl-dev for ages now, bleeeehhhh 2019-05-11 23:08:33 Anybody knows if polkit has to be insatlled on the server machine while sshing with virt-manager? 2019-05-12 05:04:11 <_ikke_> danieli: and then? 2019-05-12 07:56:37 KH405_TV: Mind describing your setup? You have virt-manager on your client and libvirt on your server? 2019-05-12 07:59:12 You don't need polkit, but it kind of is the default: https://libvirt.org/auth.html 2019-05-12 13:45:19 _ikke_: hm? 2019-05-12 13:53:56 <_ikke_> danieli │ _ikke_: google this, with the quotes: "Unexpected character in locale name" 2019-05-12 13:54:31 _ikke_: our perl is broken and apparently has to be recompiled, together with everything linked against it 2019-05-12 14:07:00 <_ikke_> recompiled with a patch? 2019-05-12 14:09:37 _ikke_: correct 2019-05-12 14:10:36 <_ikke_> Is the patch ABI incompattible? 2019-05-12 14:11:18 north1: ^ ? 2019-05-12 14:11:44 this was void's fix https://github.com/void-linux/void-packages/blob/da7eb611cbc7df8ab379082e0c86e4a871ce0750/srcpkgs/perl/template#L182 2019-05-12 14:11:54 and this was apparently the result https://github.com/void-linux/void-packages/pull/1849 2019-05-12 14:12:48 the code that's parsing setlocale(LC_ALL,0) return value is utter nonsense and should be rempved 2019-05-12 14:13:38 agreed, but that's a bigger project, and a workaround patch is an immediate solution until it's fixed properly 2019-05-12 14:20:46 .... 2019-05-12 14:20:51 " * was meant by setting a locale of "". We handle this by not actually 2019-05-12 14:20:54 * ever setting to "" (unless querylocale exists), but to emulate what we 2019-05-12 14:20:57 * think should happen for "". 2019-05-12 14:21:00 that it utterly horribly wrong with musl 2019-05-12 14:21:07 and probably part of the problem here 2019-05-12 14:24:18 EVERYTHING inside this #ifdef is wrong and should be removed: 2019-05-12 14:24:19 https://github.com/perl11/cperl/blob/master/locale.c#L694 2019-05-12 14:25:18 guess we'll make an issue @ rt.perl.org to ask wtf the intention was 2019-05-12 14:25:44 if you remove all that crap, and in this line: 2019-05-12 14:25:45 https://github.com/perl11/cperl/blob/master/locale.c#L575 2019-05-12 14:25:51 remove && defined(DEBUGGING) 2019-05-12 14:26:01 it should use _NL_LOCALE_NAME, the correct way to do this 2019-05-12 14:26:15 holy crap, this issue at the perl5 issue tracker 2019-05-12 14:26:16 (they forced the correct way off under !DEBUGGING .....) 2019-05-12 14:26:21 "open" "19 years ago" 2019-05-12 14:26:23 ???? 2019-05-12 14:27:12 their claim that it "doesn't handle "" properly" is utterly wrong; they just have a wrong idea about what "" is supposed to do 2019-05-12 18:07:10 Cogitri: Yes, exactly, I can connect with virt-manager from my lapotp to my laptop, but I can't from my laptop to my server 2019-05-12 18:08:59 Have you tried using a different auth thingie for libvirt as described in the link I've posted here? 2019-05-12 18:26:12 No I haven't, let me give this a try 2019-05-12 18:31:11 I guess that having "none" as authentification method is a bad idea ? 2019-05-12 18:31:39 in general, yup 2019-05-12 18:32:05 Yup, you don't want any user to have access to your VMs usually 2019-05-12 18:33:09 So what would you suggest to fix my problem? 2019-05-12 18:33:36 Try SASL? 2019-05-12 18:34:41 UNIX socket permissions seems like it's by far the easiest to set up 2019-05-12 18:35:09 You could also just install polkit to your server, but that may not be what you want, it's kinda heavy 2019-05-12 18:41:25 Ohh Polkit has to be installed on the server? 2019-05-12 18:42:08 Yes. 2019-05-12 18:45:36 Just installed it on the server, and the error message hasn't changed 2019-05-12 18:47:48 Huh 2019-05-12 18:48:27 You definitely need it on the server if you want to use it though, since there's no need to authenticate for anything on the client 2019-05-12 18:48:35 (Well, other than SSH) 2019-05-12 18:48:43 dbus is running, yes? 2019-05-12 18:48:50 Correct 2019-05-12 18:48:56 Both on the server and on the client 2019-05-12 19:20:59 Is polkit the most secure way to identify ? 2019-05-12 19:21:09 Cause it works from laptop to laptop now 2019-05-12 19:21:15 But i can't ssh to my server 2019-05-12 19:24:15 Polkit certainly is the most fine-grain with its rules 2019-05-12 19:24:33 But UNIX sockets should be easier to use and work for your usecase just fine 2019-05-12 19:35:39 Anything else that I should try before rulling Polkit off the list? 2019-05-12 20:06:38 Do I have to manually start it on the server ? 2019-05-12 20:06:46 Cause I only installed it 2019-05-12 20:07:05 No, it's dbus activatable 2019-05-12 20:24:20 Then what's wrong :( ? 2019-05-12 20:25:39 What's the error it throws again? 2019-05-12 20:33:06 agent authentication unavailable, no polkit agent available to authenticate action org.libvirt.unix.manage 2019-05-12 20:33:41 Verify that libvirtd is running ... Which it it 2019-05-12 20:33:43 is * 2019-05-12 20:33:57 Then install one of the polkit-* thingies 2019-05-12 20:34:15 Although you'll need one that works on the CLI, I suppose? 2019-05-12 20:34:25 Yeah 2019-05-12 20:34:32 I have no gui installed on my server 2019-05-12 20:34:39 And don't plan to have one 2019-05-12 20:34:51 I though the polkit-* was a ''gui'' for polkit 2019-05-12 20:37:40 It's not a GUI, it's a frontend which may may not be a GUI 2019-05-12 20:38:37 https://www.freedesktop.org/software/polkit/docs/master/polkit.8.html 2019-05-12 20:39:05 Huh, it should work with pkttyagent which should be shipped with polkit 2019-05-12 20:39:37 Do I need to start that? 2019-05-12 20:39:51 No, libvirt should invoke it 2019-05-12 20:41:35 Any idea why it dosen't ? 2019-05-12 20:50:46 I mean, pkttyagent works 2019-05-12 20:51:04 I got version 0.115 2019-05-12 20:51:46 Hum, not quite sure about that. Maybe you could ask upstream about it? 2019-05-12 20:52:00 What do you mean? 2019-05-12 20:52:48 https://libvirt.org/contact.html 2019-05-12 21:27:50 hi Cogitri :) 2019-05-12 21:28:18 Hello there :) 2019-05-12 21:30:31 i still consider going for alpine on the next install on my server (instead of void). i guess i'll ping you in case that's gonna start ;) 2019-05-12 21:32:07 Ah, sure 2019-05-12 21:52:23 on alpine next doesn't work 'lxc-create --name=deb --template=debian --release=jessie' 2019-05-12 21:55:16 umh, forgot '--' in between, all is ok 2019-05-13 06:41:25 hello my friends 2019-05-13 06:52:31 i see you have cleaned up nicely the alpine dl page ;) 2019-05-13 07:01:30 hey, anyone succeeded in running alpine on odroid-c2? 2019-05-13 07:02:46 kahiru_: what is odroid-c2, i.e. what chipset 2019-05-13 07:03:14 amlogic s905 2019-05-13 07:04:17 it should work on amlogic, although I don't have such boards and because of that not sure 2019-05-13 07:30:43 I'd be happy to try it, but without any attached output it is hard to figure out whats going on if things dont work ootb 2019-05-13 07:35:07 kahiru_: do you have console, HDMI, VGA or serial 2019-05-13 07:35:35 I have HDMI, but last time I tried I couldn't get any output over it from the board, so that was that 2019-05-13 07:37:14 not sure if the u-boot is built for your board in alpine, you should look in /boot (iirc) of untared u-boot-xxx.tar.gz 2019-05-13 08:02:14 so everything is awesome in alpine? 2019-05-13 08:02:50 Haven't tried it in a while, does it have xzcomp compressex xfs now? 2019-05-13 08:03:02 i mean squashfs 2019-05-13 08:03:15 as boot system 2019-05-13 13:08:34 hi there, i'm a newbie to alpine-linux. can i please seek some assistance on configuring the kernal for an unrecognised sdio component ? 2019-05-13 16:22:25 hi 2019-05-13 16:22:27 i wonder if i should try to get my mouse working on alpine 2019-05-13 16:23:03 i wasn't able to solve the problem before though 2019-05-13 16:23:34 rain1: on X or console 2019-05-13 16:23:56 X 2019-05-13 16:24:31 Do you have xf86-input-libinput installed? 2019-05-13 16:24:33 did you tried xf86-input-libinput and udev 2019-05-13 16:24:47 And is udev loaded correctly? 2019-05-13 16:25:06 I think so but it just occured to me that I could try qemu alpine with USB passthrough to debug this easier 2019-05-13 16:26:27 Or just use magic SYSRQ to reset your system. If you do the two things we've mentioned it should work just fine 2019-05-13 16:40:27 im trying to run apk del npm but it never actually uninstalls it and i havent been able to find away to force it to uninstall 2019-05-13 16:41:37 does any one know a way to uninstall it? 2019-05-13 16:41:53 any packages that depend on it installed? 2019-05-13 16:42:25 no not taht i know of and at least when i try to uninstall other packages it tells me if it depends on something 2019-05-13 16:42:56 im currently trying to remove npm since it depnds on node so i can then uninstall node 2019-05-13 16:59:11 it seems to work in the VM using usb passthrough, i wasn't able to reproduce the problem of the mouse not working 2019-05-13 16:59:25 i guess i'll try installing alpine again and hopefully it'll work 2019-05-13 17:09:18 I'm having an issue while installing, I think that it has to do with when choosing a mirror but I am unsure. 2019-05-13 17:09:49 I'm getting that wget timed out and I believe that is the cause of the issue where I am trying to use a 'sys' install 2019-05-13 17:33:04 alaradia: check /etc/apk/world there may be two entries resolving to nprm 2019-05-13 17:33:14 *npm 2019-05-13 17:38:53 thank you i found it it was named nodejs-npm for some reason lets see if i can fix everything else now 2019-05-13 18:16:17 I found that mouse worked in VM but it does not work on here in the real install 2019-05-13 18:16:27 I have xf86-input-libinput and udev 2019-05-13 18:19:22 in Xorg.log i can see it loading the keyboard correctly but failing to open the /dev device for the USB and then unloading libinput 2019-05-13 18:40:31 i couldn't get it to work. i also tried sway but it hardlocked my system when I did sway -C 2019-05-13 18:46:49 rain1: does you mouse works on alpine at all, you can use evtest to check 2019-05-13 18:47:47 oh i didn't know about evtest 2019-05-13 18:47:58 ill make a note to try that next time 2019-05-13 19:38:20 is anybody still using a hardened grsec kernel with 3.9.4? I wonder how to set the configuration options to get it working 2019-05-13 19:45:26 hello friends 2019-05-13 19:45:29 privet 2019-05-13 19:45:32 you should try the leaked https://github.com/jameshilliard/linux-grsec/ - it's a bit more recent midasi 2019-05-13 19:46:14 how much better is the backdoor functionality of the new kernel compared to a 3.16 ? 2019-05-13 19:49:54 p4Wv1qn095FW: thanks... I will purchase an official license, but I'm not sure about the configuration options. I'm afraid Alpine with musl/libc might not work with all available options 2019-05-13 19:51:27 you can git clone the aports repo, checkout a revision where linux-hardened is not yet deleted and then look at the config there... 2019-05-13 19:54:45 I have one suggestion alpine devs, check out stali, a statically compiled linux 2019-05-13 19:55:12 yes, I've done that already and the config is actually quite weird... 2019-05-13 19:55:39 bsdnub: isn't stali dead? 2019-05-13 19:56:37 north1: you can download it, it works, small and fast 2019-05-14 01:53:40 stali practically died 2019-05-14 01:56:04 linux-hardened did too 2019-05-14 02:34:11 Am I missing something or does Alpine Linux not have a package for pip for Python 3? There's a py2-pip but no py3-pip. 2019-05-14 02:34:26 Also, how do I list all packages installed already? 2019-05-14 02:36:13 galaxie: apk list -I 2019-05-14 02:36:16 and pip comes with python3 2019-05-14 02:36:32 pip3/pip3.6 to be specific, in 3.9 2019-05-14 02:37:28 also, if you want to see only the packages that have been installed and not their deps, you can look at `/etc/apk/world` 2019-05-14 02:39:40 danieli: I guess I forgot to type pip3 instead of pip. 2019-05-14 02:39:47 pip[tab] 2019-05-14 02:39:54 or is it tab x2 2019-05-14 02:46:23 Huh. I tried using https://github.com/genuinetools/apk-file but apk-file search pip3 didn't bring anything relevant. Is there a working equivalent to apt-file? 2019-05-14 02:47:07 it looks like all that does is query http://pkgs.alpinelinux.org/contents 2019-05-14 02:47:27 also 2019-05-14 02:47:27 $ apk search pip3 2019-05-14 02:47:27 python3-3.6.8-r2 2019-05-14 02:47:47 and https://pkgs.alpinelinux.org/contents?file=pip3&path=&name=&branch=v3.9&arch=x86_64 2019-05-14 02:48:03 Huh. I used that exact command but somehow I didn't pay too much attention to it. 2019-05-14 02:48:30 if it's a fresh system, did you happen to run `apk update` or `apk upgrade -U ...` before searching? 2019-05-14 02:48:59 I did. I saw that entry. Does it require you to have Internet access? 2019-05-14 02:49:24 genuinetools/apk-file queries that API, so yes, it requires internet access 2019-05-14 02:49:57 you should however be able to do `apk add cmd:pip3` and `apk search pip3` without internet (as long as you have a mirror accessible to pull index files from) 2019-05-14 03:47:07 setup-alpine data or sys 2019-05-14 03:48:41 cim209: depends on what kind of install you want 2019-05-14 03:49:06 setting up alpine as the host OS to run docker containers 2019-05-14 03:49:21 i should go with sys mode right 2019-05-14 03:49:26 you most likely want sys 2019-05-14 03:49:34 https://wiki.alpinelinux.org/wiki/Installation#Basics 2019-05-14 03:49:39 yeah i'm in there 2019-05-14 03:49:42 reading 2019-05-14 03:50:35 data mode is practically diskless but with a writable data partition that persists 2019-05-14 03:53:56 setting this up first time on a kvm 2019-05-14 03:56:52 wow installation was quick 2019-05-14 04:04:14 do you guys install sudo or just switch to root 2019-05-14 04:04:31 i'd recommend using root as little as possible in general 2019-05-14 04:04:41 and a strict sudo configuration 2019-05-14 04:07:51 indeed 2019-05-14 06:15:46 when running virt-manager, getting an error during vm creation that it failedto reach the end of file (not sure which file they mean), and there is a segfault recorded in dmesg saying that libvirtd segfaulted with error 6 in ld_musl-x86_64.so 2019-05-14 06:16:55 Googling, I discovered that other people have reported similar issues in docker and virsh, at least in the past. Just to be sure, I did apk update and apk upgrade --available and rebooted, so I have the latest from the repos. 2019-05-14 06:21:48 there is #5193, but that seems to be for running remotely. I am simply running on one box, and I don't even have a single VM running yet. 2019-05-14 06:28:05 systemdlete: Post the entire log, the error 6 in libc can mean about anything 2019-05-14 06:28:33 sure. You want the entire dmesg output? 2019-05-14 06:28:53 The stacktrace of it would be nice to have 2019-05-14 06:29:09 Yeah, I think I can get that from inside the tool (virt-manager) 2019-05-14 06:29:46 Thank you. Please give me just a few minutes... that box is slow atm because I limited memory to 1G to allow for VMs. 2019-05-14 06:45:10 traceback is at https://pastebin.com/kWrnkazZ I was creating a VM with a raw file img and booting from an ISO image 2019-05-14 06:45:47 dmesg log is at https://pastebin.com/pV01zqW2 2019-05-14 06:45:57 (thanks again for looking into this) 2019-05-14 06:53:07 Cogitri: Being this is only a testbox, I can paste any logs, etc. if that will help 2019-05-14 06:56:34 Cogitri: I've been thinking of doing an ltrace or strace on the program. 2019-05-14 06:59:56 I'm a bit curious as to why it uses py2 stuff when virt-manager is py3 now 2019-05-14 07:00:22 And pastebin seems to be down? 2019-05-14 07:00:29 what? 2019-05-14 07:00:53 yipes... well, very slow 2019-05-14 07:01:12 but it is up (over here) 2019-05-14 07:02:35 Is this as simple as hacking the shebang at the top of the script? 2019-05-14 07:02:45 (i could try that) 2019-05-14 07:06:48 I doubt it is. 2019-05-14 07:07:37 :( 2019-05-14 07:07:57 Pastebin works again now. Not quite sure why it fails creating/installing the xml though, could you ask upstream about that? 2019-05-14 07:08:41 you can also apk add tpaste 2019-05-14 07:08:45 should be a bit faster 2019-05-14 07:08:55 thanks clandmeter. 2019-05-14 07:08:58 will do 2019-05-14 07:09:02 its in house paste 2019-05-14 07:09:54 bigger problem is that I am also seeing intermittent issues with internet, but only when I boot alpine dom0. The test box is multiboot, and devuan, e.g., does not have network weirdness. 2019-05-14 07:10:49 So I rebooted devuan, and copied the files up to pastebin. (I cross-mounted the alpine partition) 2019-05-14 07:11:20 you seem to be having *a lot* of strange issues specifically with alpine 2019-05-14 07:11:35 what is "house paste?" I see tpaste package. 2019-05-14 07:12:04 tpaste is great, but I often hit the limit when using it. dpaste is what I use when that happens 2019-05-14 07:12:21 in house, made by alpine. 2019-05-14 07:12:26 dpaste. Thanks tcely 2019-05-14 07:12:27 hosted by alpine 2019-05-14 07:12:30 hit which limit? 2019-05-14 07:12:36 ah! thank you for the clarity 2019-05-14 07:12:37 it has a size limit 2019-05-14 07:12:38 danieli: size usually 2019-05-14 07:12:40 ah right 2019-05-14 07:12:41 gotcha 2019-05-14 07:12:42 that makes sense 2019-05-14 07:13:16 i do not hit it that often so its fine for me. 2019-05-14 07:14:08 I just use wgetpaste with github gists 2019-05-14 07:14:12 Cogitri: upstream, as in the xen devs? 2019-05-14 07:14:30 i like how i can paste small binaries to tpaste 2019-05-14 07:14:37 I tried the ubuntu and fedora paste solutions and did not like them much. I recently found out I can clone my gists, so that's useful for big things. 2019-05-14 07:16:47 clandmeter: there was a question earlier about drone and git am: does that use git am -3 ? 2019-05-14 07:17:04 probably not 2019-05-14 07:19:18 Cogitri: I poked around in the home directories and found no xml file -- I am guessing you mean xml file maybe generated by virt-manager? 2019-05-14 07:19:43 If there is, I don't see one 2019-05-14 07:26:55 systemdlete: Upstream as in libvirt 2019-05-14 07:27:10 Libvirt creates xml files in /var/lib/libvirt IIRC 2019-05-14 07:27:19 Oh. Nvm, I mis-read the broken line. Right. Got it now. 2019-05-14 07:27:21 ok 2019-05-14 07:27:28 For the VM definitions (what components it has etc.) 2019-05-14 07:27:33 even though it is a user-level prog? 2019-05-14 07:28:32 (It wasn't setuid root when I looked last) 2019-05-14 07:29:52 It does via the libvirtd services 2019-05-14 07:32:47 ok, thanks. Got it. 2019-05-14 07:46:11 good morning 2019-05-14 07:46:23 Today I am supposed to receive my new notebook 2019-05-14 07:46:45 telmich: I hope that goes well 2019-05-14 07:46:55 How crazy would I be if I tried running xorg + i3 + emacs with alpine on it on a scale from 1-10? (1 = totally makes sense, 10 = how stupid can you be) 2019-05-14 07:48:05 Why wouldn't it be a good idea? 2019-05-14 07:48:16 Seems perfectly fine to me, although I dislike using just WMs 2019-05-14 07:48:36 you don't need any OS other than emacs :) 2019-05-14 07:48:50 Heh 2019-05-14 07:48:52 anyway, i'm running Xorg -> i3 -> my stuff 2019-05-14 07:49:16 when i log into tty1, it fires up X, which starts i3 2019-05-14 07:49:36 at one point, i ran xmonad with xfce4-panel 2019-05-14 07:50:28 I was tempted by xmonad, but also hold back due to haskell - but that's more a personal fear/preference 2019-05-14 07:50:45 that's the reason i enjoy using and fiddling with it :) 2019-05-14 07:51:06 ok... mu is also available in testing 2019-05-14 07:54:05 I'll try to run my cdist manifest for my notebook against a VM... if that actually installs 85% of what I usually install, then I'll give alpine a try on the notebook later on 2019-05-14 07:56:04 <_ikke_> telmich: I switched from xmonad to awesomewm, which is quite similar but uses lua as a config language 2019-05-14 07:56:19 _ikke_: oh, interesting to know! 2019-05-14 07:56:30 yeah, it's neat 2019-05-14 07:56:47 i should give awesome a spin, it's been a long time - and i need to practice lua :P 2019-05-14 08:02:14 ok, first package I don't see is magit 2019-05-14 08:02:43 -> adding it to my queue of things that I want to fix if using alpine later today 2019-05-14 08:02:53 ohh 2019-05-14 08:05:53 I have to say that using the @ syntax for repos works quite nicely in cdist 2019-05-14 08:06:50 next one is xmbc - but I guess my config is just too old and I should have changed to kodi long time ago... *fix* 2019-05-14 08:07:10 xbmc even 2019-05-14 08:07:54 next in list is synergy, which I also don't see in testing - will skip it for the moment 2019-05-14 08:08:35 (or is it also already "outdated"?) 2019-05-14 08:16:46 What was the apk command to show which package contains a specific file? I am looking at apk list, but seem to not see it 2019-05-14 08:17:00 pkgs.alpinelinux.org 2019-05-14 08:17:31 someone posted a link to a github repo of some go code querying pkgs.a.o for you in CLI, but to my knowledge, that's it 2019-05-14 08:18:05 danieli: you mean apk-file ? 2019-05-14 08:18:14 i think so, i don't remember :) 2019-05-14 08:18:56 it is in testing 2019-05-14 08:19:06 aha, didn't know it was a package 2019-05-14 08:19:06 neat 2019-05-14 08:20:58 Info options: 2019-05-14 08:20:58 -W, --who-owns Print the package owning the specified file 2019-05-14 08:21:25 that's for files installed on your system 2019-05-14 08:21:49 i guess the original answer was ambiguous 2019-05-14 08:21:55 original question, sorry 2019-05-14 08:26:05 if you look for bins you can use cmd:xxx 2019-05-14 08:26:11 those are in the index 2019-05-14 08:27:56 and so: for libs, pc: for pkgconfig files 2019-05-14 08:29:39 How can I generate a loadmap for an alpine binary executable for gdb stack analysis? 2019-05-14 08:29:50 Do I have to build it myself? 2019-05-14 08:30:08 (been a while since I've done this) 2019-05-14 08:32:19 which pkg? 2019-05-14 08:35:53 any package 2019-05-14 08:35:57 any binary 2019-05-14 08:36:09 I mean, just generally speaking, how is it done (on Alpine) 2019-05-14 08:36:35 (oh, sorry. maybe you were not replying to me) 2019-05-14 08:37:37 we have -dbg subpkg support. it would contain symbols. if thats what you are after. 2019-05-14 08:38:54 and where are those? 2019-05-14 08:39:24 on the mirrors, as with all packages 2019-05-14 08:39:29 some pkgs have it enabled, some not. 2019-05-14 08:39:52 so I look for libvirtd-dbg and it is not available. 2019-05-14 08:40:15 is that a an actual pkg? 2019-05-14 08:40:18 or binary? 2019-05-14 08:40:21 idk 2019-05-14 08:40:28 it is a binary, it night not be a pkg 2019-05-14 08:40:47 check which pkg it belongs to 2019-05-14 08:41:02 then check for $pkgname-dbg 2019-05-14 08:41:02 `apk search libvirt` :) 2019-05-14 08:41:08 it'\s libvirt-daemon 2019-05-14 08:41:09 it's* 2019-05-14 08:41:14 if it doesnt exist, build it yourself 2019-05-14 08:41:44 https://pkgs.alpinelinux.org/package/edge/main/x86_64/libvirt and expand 'Sub Packages' on the right side - I'm not seeing a -dbg package for it 2019-05-14 08:41:45 ^ 2019-05-14 08:41:50 i think you need to look at the origin pkgname 2019-05-14 08:42:01 right thats the one 2019-05-14 08:42:09 yup :) 2019-05-14 08:42:23 so if it doesnt list it in subpkgs then its not added. 2019-05-14 08:43:18 libvirt-daemon, but there is no libvirt-daemon-dbg, sadly 2019-05-14 08:43:37 ok thanks guys. 2019-05-14 08:43:39 there are no -dbg packages for libvirt at all, by the looks of it 2019-05-14 08:43:40 now I know 2019-05-14 08:43:57 dbg is only for origin 2019-05-14 08:44:22 it will contain symbols of all subpkgs afaik 2019-05-14 08:44:30 that would make sense 2019-05-14 08:44:35 i don't remember off the top of my head 2019-05-14 08:44:51 you will have to build it yourself 2019-05-14 08:47:43 Do python packages even have -dbg stuff? 2019-05-14 08:48:08 you can try https://git.alpinelinux.org/docker-abuild to quickly build something. 2019-05-14 08:48:18 <_ikke_> if there is no compiled extension, I don't think so 2019-05-14 08:48:55 thanks, but all I need is stack backtrace, and I can get it with the core file and a gdb command 2019-05-14 08:49:15 so, for now, no worries. But I'll look at the link 2019-05-14 08:49:18 later. 2019-05-14 08:51:02 Hi, stupid question :) Alpine on RPI... means FS is all FAT ? 2019-05-14 08:52:39 iirc only the boot partition is fat32 2019-05-14 08:53:15 so a small boot partition should be enough ± 100MB 2019-05-14 08:53:45 from their website: "The Raspberry Pi's bootloader, built into the GPU and non-updateable, only has support for reading from FAT filesystems (both FAT16 and FAT32)" 2019-05-14 08:54:27 bootloader needs fat 2019-05-14 08:54:38 by default rpi will run from ram 2019-05-14 08:54:55 there are some hacks to make a disk install 2019-05-14 08:55:32 Run from ram is no problem 2019-05-14 08:55:54 but i want to store some data also.... second partition ext4 preffered 2019-05-14 08:56:06 so boot ± 100MB, rest ext4 2019-05-14 08:56:31 second part you can mount wherever you like 2019-05-14 08:56:40 ok Thnx 2019-05-14 09:43:50 if apk add ... says "1 error: " and same goes for apk -v add ..., how do I get the reason for the error? 2019-05-14 09:44:46 the error was already shown before 2019-05-14 09:44:54 you can fix it by apk fix 2019-05-14 09:46:00 clandmeter: apk fix is good; shown before is true, but in my case the problem is that I automatically configure the system and the config management might have seen the error, but I never did 2019-05-14 09:46:57 Maybe a second line "Try using apk fix to fix the error." would be helpful for people coming to alpine 2019-05-14 09:47:01 so check the log of config management 2019-05-14 09:47:25 apk does not log errors 2019-05-14 09:48:10 telmich: you can add a suggestion to our bug tracker. it has an apk-tools section. 2019-05-14 09:48:11 I understand that, but why does it not generate/show an error when I try to add again? 2019-05-14 09:48:19 clandmeter: will do, thanks 2019-05-14 09:48:28 its most probably an upgrade bug 2019-05-14 09:48:40 trying to overwrite another apk file 2019-05-14 09:48:51 the file probably moved from 1 pkg to another 2019-05-14 09:49:41 and this happens more often when you jump a version 2019-05-14 09:49:51 skip a version i mean 2019-05-14 09:50:37 clandmeter: it was actually https://bugs.alpinelinux.org/issues/10453 that triggered the problem and was "stuck" 2019-05-14 10:22:48 So from trying so far, there are about 20 packages that are not in alpine that I usually install to my notebook, out of which 4 are "critical" 2019-05-14 10:23:22 (those 4 being vym openconnect magit shotwell) 2019-05-14 10:23:24 i've been procrastinating porting packages i need to alpine, some of them will be extremely difficult 2019-05-14 10:23:41 so my daily drivers will still be other distros 2019-05-14 10:24:31 I think packaging vym and magit should not be hard, not sure about openconnect and shotwell is just temporarily broken 2019-05-14 10:24:45 I've ported most stuff to Alpine now, I just need my stuff to be merged now :P 2019-05-14 10:25:10 Will fix shotwell once I'm home, telmich 2019-05-14 10:25:20 Cogitri: you are great, thanks a lot! 2019-05-14 10:26:22 looking at arch's vym, it requires some qt5 stuff, zip/unzip, and a tool to manage desktop entries 2019-05-14 10:26:44 telmich: Glad to be of help :) 2019-05-14 10:27:01 vym is *really* awesome 2019-05-14 10:27:10 it's by far the best mind mapper I have ever found 2019-05-14 10:27:28 i occasionally use graphviz to be honest 2019-05-14 10:27:44 danieli: I'll check out graphviz 2019-05-14 10:27:57 it's more of a general graphing tool 2019-05-14 10:28:14 http://viz-js.com/ 2019-05-14 10:28:47 ACTION would've needed that a long time ago already 2019-05-14 10:28:52 hmmm 2019-05-14 10:29:07 not exactly convinced, it seems to be a very different approach 2019-05-14 10:29:20 Hi, I'm trying to install alpine to PCengines APU that has only serial port. I've found this https://wiki.alpinelinux.org/wiki/Enable_Serial_Console_on_Boot but when I dd the iso image to USB stick, I can't write changes to it. What am I missing? 2019-05-14 10:29:41 jvl: so nice! are you based in .ch? 2019-05-14 10:29:42 telmich: indeed, i'm not trying to sell it to you :) 2019-05-14 10:29:43 <_ikke_> jvl: the iso image is isofs, which is read-only by design 2019-05-14 10:29:56 jvl: I actually planned to do the same this weekend! 2019-05-14 10:30:20 also came across this the other day, haven't tried it but it looks cool https://github.com/ivanceras/svgbob 2019-05-14 10:30:22 <_ikke_> jvl: if you already have an alpine system running, you can use setup-bootable to create a bootable USB stick which is writable 2019-05-14 10:31:15 telmich: nope, .cz ;) 2019-05-14 10:31:37 _ikke_: ah thanks. I could do this in qemu and write the image afterwards. thanks! 2019-05-14 10:31:51 danieli: I am actually quite open for selling new stuff, so I had to check it out whether it could replace vym 2019-05-14 10:32:19 jvl: do you have an internal ssd in it? 2019-05-14 10:32:38 if so, I'd boot from usb => write to the ssd 2019-05-14 10:32:39 graphviz is quite versatile, i mostly use it in an automated fashion, occasionally as a sort of mind map more concrete thoughts 2019-05-14 10:32:50 for more* 2019-05-14 10:35:41 btw, related tweet: https://twitter.com/NicoSchottelius/status/1128203335662751744 2019-05-14 10:43:28 leonardus: yes I do 2019-05-14 10:43:41 telmich: yes I do 2019-05-14 10:43:45 leonardus: sorry, typo 2019-05-14 10:43:53 so my new notebook has just arrived... time to check it out 2019-05-14 10:44:07 telmich: what kind? 2019-05-14 10:44:24 danieli: good question! some kind of hp, the only one "portable" with 32gb ram 2019-05-14 10:48:09 hp x360 1040 g5 it is 2019-05-14 10:48:23 ah, an elitebook 2019-05-14 10:49:30 my first hp, had thinpads for about 15y or so, so I am curious 2019-05-14 10:50:12 i end up hating laptops in general if i get something consumer grade rather than business grade, build quality is often terrible 2019-05-14 10:50:20 guess i'm a bit spoiled 2019-05-14 10:50:46 so.. guys.. if I am crazy to go with alpine on my notebook, do you think you can bear with me / help me getting stuff running that doesn't work? 2019-05-14 10:51:04 Sure 2019-05-14 10:51:10 I happen to have a Spectre X360 too 2019-05-14 10:51:28 The one with the Vega M configuration though 2019-05-14 10:51:35 i wanted to run alpine as a daily driver, but it would ruin my workflow :( 2019-05-14 10:52:25 So the 15-ch009ng 2019-05-14 10:52:34 plus, some of the things i need are pretty darn difficult to port and get running properly 2019-05-14 10:52:35 dd if=alpine-standard-3.9.4-x86_64.iso of=/dev/sdb bs=1M it is then 2019-05-14 10:52:45 go for it, it'll be fun 2019-05-14 10:53:04 i distro-hop once in a while, but i have a couple go-tos for daily drivers 2019-05-14 10:53:15 danieli: I think I might even use Alpine as my work distro because CLion works on it \o/ 2019-05-14 10:53:57 vim, emacs, vscode here 2019-05-14 10:54:40 Yeah, I use CLion because it's the best IDE for Rust imho 2019-05-14 10:55:10 And I don't like Electron :c 2019-05-14 10:55:28 i don't like electron either, but i don't like java either 2019-05-14 10:55:38 that sentence was very poorly phrased, sorry 2019-05-14 10:55:58 being distracted does that :) 2019-05-14 10:58:34 Java isn't nice, but it's nice how cross-platform it is 2019-05-14 10:58:52 So Electron is the greater evil for me I guess 2019-05-14 11:00:07 additionally, i'd imagine most vscode installs are proprietary software 2019-05-14 11:00:55 hm.. i digress 2019-05-14 12:38:23 when trying to mount the cryptsetup formatted ext4 filesystem on /dev/mapper/root I get the message "Invalid argument", but no error in dmesg 2019-05-14 12:38:32 Any clue on what the cause could be? 2019-05-14 12:40:05 I did run cryptsetup luksFormat before and luksOpen'ed it 2019-05-14 12:43:56 dm_crypt is loaded in the kernel 2019-05-14 12:44:38 hmpf, using -t ext4 worked 2019-05-14 13:31:44 ahhhhhhhhhhhhhhhhh 2019-05-14 13:31:50 the hp restore just killed my installation 2019-05-14 13:31:59 "restoring" 2019-05-14 13:38:02 hello! Im using alpine linux in raspberry pi and i want to use the data mode, as described in https://wiki.alpinelinux.org/wiki/Installation 2019-05-14 13:38:55 But i cant set the second partition in data mode and im afraid the lbu misbehave with vfat because filesize limitations 2019-05-14 13:39:01 So what is recommended in this case? 2019-05-14 14:02:30 how / when in the install process is the uefi entry created? 2019-05-14 14:15:43 Is the encryption guide still up-to-date? I updated the initramfs and added the kernel parameter, but there is no password prompt for / on bootup 2019-05-14 14:19:02 Do you happen to use a NVMe drive (you most likely do)? Have you added `nvme` to your mkinitfs.conf? 2019-05-14 14:19:09 That screwed me over the first time I installed Alpine on my laptop 2019-05-14 14:22:45 argh 2019-05-14 14:22:49 Yes and no 2019-05-14 14:22:52 thanks 2019-05-14 14:23:12 No problem :) 2019-05-14 14:26:58 the resolution is crazy.... 2019-05-14 14:27:05 a character on the console is < 2mm 2019-05-14 14:27:29 Yeah, reading the TTY on 4K is hard 2019-05-14 14:27:42 I can recommend setting a bigger consolefont 2019-05-14 14:27:43 what do you use for setting the efi boot entry in the bios? 2019-05-14 14:28:08 I have or not seen that there is no efibootmgr and the bios does not jump to syslinux.efi automatically 2019-05-14 14:28:48 I haven't changed anything in the bios, efibootmgr did everything for me 2019-05-14 14:29:11 it boots!!! 2019-05-14 14:29:23 Nice! :) 2019-05-14 14:29:44 next step: accessing via ssh to configure via cdist 2019-05-14 14:30:24 ssh nico@2a0a:e5c0:2:5:2279:18ff:fefa:fb95 works, yeah! 2019-05-14 14:30:58 it took two or three tries to get alpine to boot on my laptop too 2019-05-14 14:31:45 After that hurdle with the initramfs it worked for me 2019-05-14 14:31:59 *updatingdns* 2019-05-14 14:32:22 what is pretty cool that wpa_supplicant keeps the installation config 2019-05-14 14:32:27 a bit different to arch 2019-05-14 14:32:59 ssh root@snow.place6.ungleich.ch => now it's ready to run cdist over it... let's see how many errors I get after fixing my types today 2019-05-14 14:35:15 "ssh root@" :( 2019-05-14 14:35:59 sure 2019-05-14 14:36:03 pubkey only 2019-05-14 14:39:00 what is quite nice with alpine is ash as a default 2019-05-14 14:39:13 with bash as default, cdist is about 20-30% slower (a feeling, not measured) 2019-05-14 14:41:12 hmm... my postfix types somehow fail on arch with getent errors... even though there should be no getent involved 2019-05-14 14:41:20 I can't really find benchmarks for that, having a bash vs ash vs dash one would be kinda cool 2019-05-14 14:42:56 i don't like using root in general, pubkey only or not 2019-05-14 14:43:02 easier to make fatal mistakes 2019-05-14 14:43:53 danieli: I agree to that. However in my case I am using an automated configuration management tool -> I do need root access for remotely configuring it 2019-05-14 14:44:12 i make ansible use sudo or some other mechanism 2019-05-14 14:44:20 But that's in the end the same thing 2019-05-14 14:51:52 there is no usermod in alpine; is there any default way of changing users automated in alpine? i.e. setting primary group / uid / etc. ? 2019-05-14 14:55:09 <_ikke_> THe shadowpackage contains usermod 2019-05-14 14:55:30 ahh, ok. so I'll modify __user in cdist to require shadow 2019-05-14 15:12:26 yeah, x started! 2019-05-14 15:12:37 But I don't have keyboard or mouse support at the moment :-) 2019-05-14 15:15:27 Install one of the xf86-input and make sure udev works 2019-05-14 15:15:47 xf86-input I have - where do I read about udev best? 2019-05-14 15:16:06 even though... it seems there is a lot of stuf in /dev/input already, so this should be fine [tm] 2019-05-14 15:16:54 Make sure udev, udev-postmount and udev-trigger are all enabled 2019-05-14 15:17:51 started, but still no keyboard / mouse... checking the xorg log file 2019-05-14 15:19:52 Hm 2019-05-14 15:21:42 [ 2910.064] (II) No input driver specified, ignoring this device. 2019-05-14 15:21:57 I get this for every device 2019-05-14 15:22:11 I guess that explains it 2019-05-14 15:23:24 ah, missing the xf86-input-keyboard 2019-05-14 15:26:27 I'd just go for xf86-input-libinput which is maintained and works best for most stuff 2019-05-14 15:35:15 yeah! omg! That was too easy[tm] 2019-05-14 15:35:31 telmich: cdist is neat. Used it 6 years ago. Now I've just hacked together my own poor mans config mgmt system in shell (and some inline python for templating): https://github.com/3ju/cfg 2019-05-14 15:38:58 eu: oh! That is cool! I'll certainly check it out! cdist is never finished and new inspirations are always welcome 2019-05-14 15:38:59 i'm gonna have to rewrite my bash script to use only docker commands instead of docker-compose 2019-05-14 15:39:09 using alpine as a host, requires you to install that nasty python/pip and its dependencies just to use docker-compose 2019-05-14 15:44:19 heeeeeeeeeey 2019-05-14 15:44:26 I am writing from my new notebook on alpine! 2019-05-14 15:44:31 I can type! 2019-05-14 15:44:34 And I can read! 2019-05-14 15:44:38 congrats 2019-05-14 15:44:50 Quite impressive with a 3840x2160 resolution 2019-05-14 15:45:16 next in queue: emacs + mu and then I should be set 2019-05-14 15:46:08 Don't know about the situation nowadays, but 5 years ago emacs was not working on musl due to some insane glibc specific session saving 2019-05-14 15:47:40 I see that nowadays emacs is unpatched in alpine \o/ 2019-05-14 15:50:22 hmm... mu exists, but no mu4e 2019-05-14 15:50:43 and for some reason emacs starts in the console, even when $DISPLAY is setup 2019-05-14 15:50:56 cim209: you can run docker-compose as a docker container. No need to install on the host. 2019-05-14 15:51:31 tcely: don't you have to give it super priv? 2019-05-14 15:51:53 Not that I recall 2019-05-14 15:52:19 how would you send docker-compose commands from container to host? 2019-05-14 15:52:32 docker.sock? 2019-05-14 15:53:19 is it correct or a bug that firefox cannot be installed? When trying to install from testing, I get an error for so:libicuuc.so.64 (missing): 2019-05-14 15:54:00 <_ikke_> telmich: might be transitional 2019-05-14 15:54:21 ok 2019-05-14 15:54:50 downloading my mails with mbsync already, 100k+ takes a bit of time via imap 2019-05-14 15:56:15 cim209: http://tpaste.us/nMwE 2019-05-14 15:57:28 guessing that's the entrypoint script 2019-05-14 15:58:27 That's the host side wrapper 2019-05-14 15:59:05 If I need/want to have the usual build tools, is there some virtual package bringing in make, gcc, etc.? 2019-05-14 15:59:08 It can share the socket or TCP with the container to issue commands 2019-05-14 15:59:23 tel 2019-05-14 15:59:51 telmich: alpine-sdk or build-base 2019-05-14 16:10:16 telmich: you need emacs-x11 or emacs-gtk2 2019-05-14 16:17:17 p4Wv1qn095FW: thanks! 2019-05-14 16:17:48 hmm 2019-05-14 16:17:55 xrandr does not detect the external usb-c screen 2019-05-14 16:19:08 tcely: lol, i just installed docker-compose in an alpine base, it works when i mount docker.sock 2019-05-14 16:20:12 also gotta mount the directory where the yml is stored 2019-05-14 16:20:38 sweet, now i can still use docker-compose 2019-05-14 16:24:42 So I am thinking about building my first alpine package (mu4e) - I have the pkgbuild from ARCH, what's the smoothest / easiest way to get started in alpine? 2019-05-14 16:25:44 <_ikke_> telmich: Probably read this through: https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2019-05-14 16:26:08 <_ikke_> PKGBUILDS are similar to APKBUILDS, but Arch and Alpine differ in some policies 2019-05-14 16:28:10 ok! Trying to build it in the first place and then create a package out of it 2019-05-14 16:29:23 _ikke_: I released atools up to version 6 2019-05-14 16:29:29 Also has scdoc documentation 2019-05-14 16:30:28 <_ikke_> north1: https://pkgs.alpinelinux.org/package/edge/testing/x86_64/atools :-) 2019-05-14 16:31:16 <_ikke_> How do I build scdoc? 2019-05-14 16:34:12 Build the manpage? 2019-05-14 16:34:46 scdoc < alint.5.scd > $pkgdir/usr/share/man/man5/alint.5 2019-05-14 16:36:37 <_ikke_> Wonder if some things in https://github.com/alpinelinux/aports/pull/7855/files should be merged into the main package? 2019-05-14 16:39:25 <_ikke_> north1: done 2019-05-14 16:40:46 _ikke_: nice 2019-05-14 16:42:31 the link to https://pkgs.alpinelinux.org/packages?name=newapkbuild&branch=edge&repo=&arch=x86_64&maintainer= is broken from https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2019-05-14 16:43:51 <_ikke_> It's part of abuild 2019-05-14 16:44:23 <_ikke_> telmich: would this be a good replacement you think? https://pkgs.alpinelinux.org/contents?file=newapkbuild&path=&name=&branch=edge 2019-05-14 16:45:40 <_ikke_> Hmm, it uses a macro to generate the link 2019-05-14 16:45:58 omg... I cannot see the mouse pointer because of the resolution to be able to c&p the link 2019-05-14 16:49:04 autsch. there is a inet_protocols = ipv4 in the standard postfix config 2019-05-14 16:51:32 so besides a tiny bug in the natural scrolling (enabling it via xinput does not enable it), I am 85% set with the new notebook. hacking magit, checking out suspend and then everything should be fine [tm] 2019-05-14 16:55:40 telmich: what distro were you previously using on laptops? 2019-05-14 16:56:14 eu: I am currently using devuan, arch and alpine 2019-05-14 17:00:49 tcely: https://hub.docker.com/r/demyx/docker-compose 2019-05-14 17:07:33 Is it not recommended/possible to mix edge and stable packages? There's this one package I need but it's only available in edge. 2019-05-14 17:11:14 Hmmm. Appears there's going to be more than one package. 2019-05-14 17:12:21 https://www.nico.schottelius.org/blog/alpine-linux-on-the-hp-x360-1040-g5-notebook/ 2019-05-14 17:12:30 I just started to publish my notes 2019-05-14 17:21:40 for the broken xbacklight support, should I open a ticket somewhere or post to lkml? I assume that this is the usual xorg-or-kernel-too-old problem 2019-05-14 17:22:02 actually... the kernel has the backlight control, so probably xorg or xorg misconfig 2019-05-14 17:22:43 <_ikke_> galaxie: correct, they often don't work together due to different dependencies 2019-05-14 17:24:31 _ikke_: So why is it only in edge then? geos/geos-dev. Is it because it needs more testing? Or what? 2019-05-14 17:28:45 telmich: I use https://github.com/haikarainen/light for backlight control. I have a pending PR to alpine here: https://github.com/alpinelinux/aports/pull/7878 2019-05-14 17:29:02 Also, does anyone know if ash supports pipefail? https://docs.docker.com/v17.09/engine/userguide/eng-image/dockerfile_best-practices/#using-pipes 2019-05-14 17:29:18 <_ikke_> galaxie: It's still in testing, so it would not be included in stable releases 2019-05-14 17:29:35 <_ikke_> galaxie: and apparently no one is maintaining 2019-05-14 17:29:36 <_ikke_> it 2019-05-14 17:29:43 _ikke_: How long does the testing phase last typically? Oh. 2019-05-14 17:29:54 <_ikke_> galaxie: No predefined time 2019-05-14 17:30:03 <_ikke_> Mostly up to the maintainer 2019-05-14 17:31:20 <_ikke_> galaxie: did you test it? 2019-05-14 17:32:22 _ikke_: Test what? The packages? No. I'm on stable. 2019-05-14 17:33:04 <_ikke_> galaxie: I you could test it in edge just to verify it works, I can probably get it into community, and then it will be part of the next stable release 2019-05-14 17:34:56 _ikke_: I'm afraid I wouldn't really know how to test it. I haven't had much experience in PostgreSQL/PostGIS. 2019-05-14 17:35:21 <_ikke_> me neither 2019-05-14 17:37:00 galaxie: No pipefail in posix shell. There are ugly hacks to emulate it though. 2019-05-14 17:39:53 telmich: created a mu4e subpackage of mu for you: https://github.com/alpinelinux/aports/pull/7881 2019-05-14 17:46:54 eu: wow, that is very cool! 2019-05-14 17:49:50 I've never used emacs, but I was able to add (require 'mu4e), M-x mu4e, and it asked me to create a ~/Maildir 2019-05-14 17:54:36 eu: interesting, also I have https://github.com/haikarainen/light APKBUILD and using it for more that a year but didn't posted or pushed although I thought to add it after finish works needed for 3.10 2019-05-14 17:54:46 it works quite fine 2019-05-14 17:55:11 great minds think alike 2019-05-14 17:55:25 hehe 2019-05-14 17:56:00 did you opt for SUID or udev setup in your APKBUILD? 2019-05-14 17:56:07 I have about 50 apk's in my private repo 2019-05-14 17:56:11 udev 2019-05-14 17:56:19 good 2019-05-14 17:56:29 I see that Arch uses SUID 2019-05-14 17:56:34 shame on them 2019-05-14 17:56:55 eh, suid is last choice, imo 2019-05-14 17:57:56 eu, will look at your PR 2019-05-14 18:00:38 hmm, looks like it could be pushed, but need udev.rules file 2019-05-14 18:04:51 mps: the rules file is there 2019-05-14 18:05:02 updatream provides it 2019-05-14 18:05:36 s/updatream/upstream/ 2019-05-14 18:05:50 I don't see file in APKBUILD 2019-05-14 18:06:15 maybe is due my inexperience with github 2019-05-14 18:06:39 its in $src, make installs it from the upstream tarball 2019-05-14 18:06:50 given the correct ./configure flags 2019-05-14 18:07:31 ah, so, nice. will try to build it later 2019-05-14 18:08:22 oh, I see that I have outdated version 2019-05-14 18:10:05 so xbacklight works with the intel driver 2019-05-14 18:11:23 telmich: I think so, but not with arm machines, ime 2019-05-14 18:29:56 cim209: the wrapper I linked to earlier gets installed as docker-compose and takes care of running the container for me. 2019-05-14 18:30:20 my image works 2019-05-14 18:30:27 exec, scale, up, down 2019-05-14 18:55:03 mps i also have a private repo. would be a shame if people would do duplicate work just because these are a bit hidden. 2019-05-14 19:03:29 p4Wv1qn095FW: my experience is that a lot of people do duplicate or similar work in free/open software 2019-05-14 19:04:09 although I agree, would be better if we coordinate our work somehow 2019-05-14 19:04:58 something like 'wish list' and 'intent to package' on bugs.a.o or somewhere else 2019-05-14 19:12:08 a list of packages is unofficial repos 2019-05-14 19:12:44 is alpine rolling? 2019-05-14 19:12:47 at least for my repo, the packages in there are solid and i use them regularly 2019-05-14 19:12:51 cim209: edge is rolling 2019-05-14 19:12:57 ah ok 2019-05-14 19:12:59 just checking 2019-05-14 19:13:05 which one is LTS? 2019-05-14 19:13:15 so it goes beyond "intent to package" at least for my repo 2019-05-14 19:14:04 i also noticed jirutka has "private" repo which is a nice setup, it even builds pkgs using CI services iiuc 2019-05-14 19:18:33 mps: I have the feeling that arm is the poor child that does not gen enough love... I browsed etcd yesterday and for some reason, their arm support is just experimental 2019-05-14 19:19:38 'etcd' key/value daemon? 2019-05-14 19:20:09 ack 2019-05-14 19:20:20 I am considering to migrate away from consul to etcd + something else 2019-05-14 19:20:33 I have the feeling that the consul crowd and I are not really compatible 2019-05-14 19:21:11 right, arm on servers is not yet supported well, which is big shame 2019-05-14 19:21:33 it really is 2019-05-14 19:22:04 telmich: what are you using consul for ? 2019-05-14 19:22:23 koollman: finding servers and I wanted to use it for our distributed nft based firewall 2019-05-14 19:23:02 and what's the incompatibility you're encountering with the rest of the crowd ? :) 2019-05-14 19:25:15 For reference, this is our ticket that we work on: https://redmine.ungleich.ch/issues/6681 2019-05-14 19:26:14 The problem I have is that the consul crowd does not really think in ipv6 world or in other words: the security model of consul relies on running your own CA 2019-05-14 19:26:34 While one can do that, it easily involves adding another tool (like vault) to manage your CA 2019-05-14 19:27:02 And when it comes to certificates, we use letsencrypt on any system, because every system has IPv6 and get can a valid certificates 2019-05-14 19:28:11 So consul has 2 use cases: knowing which servers should be alive and register in service discovery that we use for prometheus; the second case is the distributed firewall 2019-05-14 19:28:40 Thus I was thinking whether we use one tool (etcd) for the distributed k/v store and something else (maybe self written) for service/server discovery 2019-05-14 19:34:38 eu: built light from your PR and testing it on aarch64, it works :) 2019-05-14 19:34:59 under awesome wm 2019-05-14 19:35:36 it gives me some warnings by building it on armv7, though 2019-05-14 20:28:10 mps: are you thinking about the %lu format errors etc? 2019-05-14 20:28:19 s/about/of/ 2019-05-14 20:30:56 eu, yes 2019-05-14 20:31:14 but I tried it on armv7 also and it works 2019-05-14 20:37:31 eu: i can push it as it is now if you agree 2019-05-14 20:38:09 although I have fear of s390x and ppc64le ;) 2019-05-14 20:43:45 ACTION feels fear of ppc64le is entirely justified 2019-05-14 20:44:16 LuaJIT ppc assembly is a giant project 2019-05-14 20:44:27 please don't stone me to death, but what is wrong with ppc64le? 2019-05-14 20:46:19 power pc 64 bit little endian 2019-05-14 20:47:36 I know what it stands for, I just don't know it enough to have any idea if it is any good or not 2019-05-14 20:48:17 I'm sure it is fine. I just hate it right now. https://gist.github.com/tcely/5b9a145cedfd5ed60ebf15b1d85fb97b 2019-05-14 20:51:10 mps: ah, you have commit bit? fire away 2019-05-14 20:52:11 ok 2019-05-14 20:58:45 it passed ppc64le, nice. waiting for s390x 2019-05-15 03:47:40 hello 2019-05-15 03:50:52 I've been running alpine aarch64 on a raspberry pi model 3a+ for about a day now, and have run into segfault/oops issues when running iptables (both latest stable and edge). is anyone aware of similar issues? 2019-05-15 03:53:42 my wifi is working on it so I can link relevant info when necessary/requested 2019-05-15 08:12:44 has anyone here compiled eternal terminal before? 2019-05-15 08:16:30 uriah: what version of alpine? 2019-05-15 08:16:39 oh right, both 2019-05-15 08:16:44 you've upgraded recently, yes? 2019-05-15 08:18:06 danieli: yes 2019-05-15 08:18:16 well i'm trying to install it on a docker image 2019-05-15 08:18:32 but i'm looking at the package page, the required libs aren't in there 2019-05-15 08:18:36 only some are 2019-05-15 08:18:41 what do you mean? 2019-05-15 08:19:14 let me paste the error 2019-05-15 08:19:52 and uh.. you mind explaining what your setup is, exactly? it doesn't sound like it's simply alpine on an rpi3 2019-05-15 08:20:12 based on 3.9.4 docker image 2019-05-15 08:20:13 https://kopy.io/dULyD 2019-05-15 08:24:33 huh 2019-05-15 08:24:37 those libs are in edge 2019-05-15 08:30:11 Hi! have someone expirience with tpm 2.0 and alpine linux? 2019-05-15 08:31:48 oops, the other guy was talking about an rpi 2019-05-15 08:31:52 sorry, i had just woken up :) 2019-05-15 08:32:32 i don't know of any eternalterminal package, so you'll probably have to package it yourself 2019-05-15 08:42:43 finally it's compiling 2019-05-15 08:45:35 I did some fighting with TPM and came to the conclusion it was a tad too complicated to continue with 2019-05-15 08:58:41 ugh 2019-05-15 08:58:46 build failed at 64% 2019-05-15 09:32:19 yes, fighting is the right word. packages like tcsd or tpm-tools are missing 2019-05-15 09:34:17 does someone know what this error is https://github.com/MisterTea/EternalTerminal/issues/188 2019-05-15 09:34:50 the tpm module is working, dmesg says: 2019-05-15 09:34:53 [ 3.772319] tpm_tis 00:02: 2.0 TPM (device-id 0x1A, rev-id 16) 2019-05-15 09:35:54 cim209: ah, backtrace 2019-05-15 09:36:01 'd patch the code if i were you' 2019-05-15 09:36:10 darn irc client 2019-05-15 09:36:13 .. and the kernel modules are automaticly loading: 2019-05-15 09:36:15 danieli: what does it mean? 2019-05-15 09:36:15 tpm_tis 16384 0 2019-05-15 09:36:15 tpm_tis_core 20480 1 tpm_tis 2019-05-15 09:36:15 rng_core 16384 2 ccp,tpm 2019-05-15 09:36:15 tpm 57344 2 tpm_tis,tpm_tis_core 2019-05-15 09:36:16 i'd patch the code if i were you( 2019-05-15 09:36:44 danieli: i don't know how 2019-05-15 09:36:55 cim209: undefined reference to `backtrace_symbols' 2019-05-15 09:37:44 actually, you could try libexecinfo-dev 2019-05-15 09:38:03 i do have it in the lib list 2019-05-15 10:02:12 lloydz: tpm-tools you can find here: https://github.com/aports-ugly/aports/tree/master/ugly 2019-05-15 10:02:18 https://github.com/aports-ugly/aports/tree/master/ugly/tpm-tools 2019-05-15 10:02:27 together with trousers 2019-05-15 10:03:20 you'll have to build them yourself. 2019-05-15 10:09:12 i also have a repo full of too ugly aports, lol 2019-05-15 10:09:47 i'm considering writing a wrapper around abuild à la arch's yay (aur helper) 2019-05-15 10:12:02 A wrapper around `abuild -r && apk add `? 2019-05-15 10:13:50 in aur, every package is a separate repo, so not quite 2019-05-15 10:13:56 there's more to it than that 2019-05-15 10:15:42 p4Wv1qn095FW: thank you. i got it. i have to learn about aports. 2019-05-15 10:21:21 'abuild -ri' 2019-05-15 10:21:33 mps and jirutka also have user repos 2019-05-15 10:21:59 we should really do something about those. would possibly elliminate lots of duplicate efforts 2019-05-15 10:22:02 p4Wv1qn095FW: my is not on GH 2019-05-15 10:22:17 mine's also not on github, it's on a private git server 2019-05-15 10:22:36 oh, they're private? that's sad 2019-05-15 10:22:36 danieli: mine to 2019-05-15 10:23:16 they're for personal use and i don't care to maintain them for everyone else, it'd take more effort than i have to waste on them 2019-05-15 10:23:36 they're "ugly" aports, and there aren't much more than 60-70 there 2019-05-15 10:26:03 p4Wv1qn095FW: I've seen ugly (it's your iirc) and even downloaded kicad but didn't found enough time to make pkg 2019-05-15 10:29:47 p4Wv1qn095FW: btw, hamlib is already in aports 2019-05-15 10:31:55 I maintain my own private packages as well, mainly because I've never bothered to figure out getting packages to the official repos... 2019-05-15 10:33:48 I just upstream everything 2019-05-15 10:34:14 the whole point of having our own aports is that upstream wouldn't accept it 2019-05-15 10:38:48 some of my packages is not yet in a shape to be pushed to aports, for some I'm not sure if they needed for general use (example chromebook vboot and cgpt tools) and some are niche tools (gnu forth for example) 2019-05-15 10:39:17 and games, gnugo, gnuchess and xboard 2019-05-15 10:40:00 and a few ham (radio amateur) tools and software 2019-05-15 10:46:08 bye. 2019-05-15 10:49:44 mps: i didn't mean to put real maintainer effort into it. just that these should be available if someone else also needs these and they can decide if they want to use those, roll their own, or even take full responsibility. 2019-05-15 10:50:56 also i think this "I've never bothered to figure out getting packages to the official repos..." and this "not yet in a shape to be pushed to aports" seem to be some inherent barriers to sharing work, maybe worth consider change something upstream. 2019-05-15 10:51:45 these ugly/personal things might come with a disclaimer, i use them, but i do not provide support. 2019-05-15 10:52:16 some of my private aports use bundler/npm/pip for instance, adn they would not be accepted into upstream aports. i simply don't care to put the effort into getting them there either, i quickly hacked them together for myself 2019-05-15 10:52:29 s/adn/and/ 2019-05-15 10:53:13 <_ikke_> I feel that in some cases, we probably do need to use something like that, if only to avoid conflicting dependencies 2019-05-15 10:53:47 putting them on github would be possible, sure, but they're ugly and i hate publishing unfinished or half-assed work 2019-05-15 10:54:17 <_ikke_> It would allow others to continue on it though 2019-05-15 10:54:26 the cult of done manifesto migh help overcome these feelings: https://medium.com/@bre/the-cult-of-done-manifesto-724ca1c2ff13 2019-05-15 10:54:35 what _ikke_ sys 2019-05-15 10:54:38 s/sys/says/ 2019-05-15 10:54:56 it would reduce waste on redundant effort. 2019-05-15 10:55:08 it's not my problem 2019-05-15 10:55:33 but i'll read it, perhaps i'll change my mind 2019-05-15 10:55:35 it is, if you are doing something that someone else did already but you did not know about it. 2019-05-15 10:55:47 you are doing double work then. 2019-05-15 10:55:49 I could and maybe I even wish to push some of my private aports but need a time 2019-05-15 10:55:52 which is a waste 2019-05-15 10:55:54 that might be, but i don't really care 2019-05-15 10:56:02 it's not too much wasted effort 2019-05-15 10:56:19 lucky you 2019-05-15 10:56:43 and, good idea would be to put some of them somewhere to the net for those who need it to play with them 2019-05-15 10:58:11 right, i read it, and i don't agree with several of the points 2019-05-15 10:59:47 perfect, that means theres points that you agree with :) 2019-05-15 11:14:42 p4Wv1qn095FW: and in ugly, irssi-otr is in upstream irssi already 2019-05-15 11:19:49 possible, that's from someone else, i share this repo with others. 2019-05-15 11:20:28 i think the wireguard stuff might also be obsolete in there. 2019-05-15 11:21:52 there are more obsoletes 2019-05-15 11:50:54 danieli: you sounds like a very good friend of mine (re publishing unfinished things) 2019-05-15 11:51:12 danieli: like, about 100% overlap, it is very interesting to read you 2019-05-15 11:51:25 aw, you're telling me i'm not unique? :( 2019-05-15 11:52:34 no, he's saying you're not a very good friend of his :( 2019-05-15 11:52:45 oh guys! 2019-05-15 11:53:08 <_ikke_> :D 2019-05-15 11:53:16 danieli could sure be a good, unique friend 2019-05-15 11:53:36 But what I actually wanted to say is that I always advise my friend to release-early-release-often 2019-05-15 11:53:53 Because it helps others to jump onto the unfinished work and make it pretty 2019-05-15 11:54:22 I think danieli, you should not underestimate which ground work you do by getting something working already - it helps others a lot to be able to start from that and "just make it pretty" 2019-05-15 11:55:15 my point is that i don't need pretty, i need it to work - and if anyone else wants to avoid duplicating my work, it doesn't matter to me in this case 2019-05-15 14:45:54 remote code execution in kernels below 5.0.8 with rds loaded: https://security-tracker.debian.org/tracker/CVE-2019-11815 2019-05-15 14:46:20 just a heads up - it's rare anyone uses RDS anyway 2019-05-15 14:53:28 i posted a bunch of cves in -devel earlier 2019-05-15 14:53:37 there are some vulns in intel processors too, once again 2019-05-15 15:49:35 "some vulns" is generous. From what I've seen Intel as a design is broken and getting worse. https://twitter.com/IanColdwater/status/1128395135702585347?s=19 2019-05-15 15:50:45 what is RDS? 2019-05-15 15:50:48 yup, i've kept up with it, but just barely 2019-05-15 15:51:00 <_ikke_> remote desktop sesison? 2019-05-15 15:51:04 <_ikke_> session* 2019-05-15 15:51:10 no, the rdp stuff is a separate issue 2019-05-15 15:51:15 <_ikke_> ok 2019-05-15 15:51:19 oh right, yeah, RDS is separate 2019-05-15 15:51:36 MS has patched it all the way back to XP/2003 to avoid another WannaCry 2019-05-15 15:52:07 for the record, RDS = remote desktop services (remote desktop stuff collectively) 2019-05-15 15:52:29 huh? 2019-05-15 15:52:37 why would RDS have anything to do with kernel? 2019-05-15 15:52:47 "with RDS loaded" ?? 2019-05-15 15:53:00 the RDS stuff has nothing to do with the intel issues, it's separate 2019-05-15 15:53:15 MDS is microarchitectural data sampling 2019-05-15 15:53:19 remote code execution in kernels below 5.0.8 with rds loaded: https://security-tracker.debian.org/tracker/CVE-2019-11815 2019-05-15 15:53:22 There is a lot of confusion around this. MS RDS had a flaw and kernel RDS module did too. 2019-05-15 15:53:23 was looking at that 2019-05-15 15:53:35 oh that 2019-05-15 15:53:41 and the tracker is idiotic as usual 2019-05-15 15:53:54 no indication of what RDS is, whether it's enabled by default, who is affected, etc 2019-05-15 15:53:59 lol, yeah, useless tracker and a crapload of vulns disclosed in general over the past couple of days 2019-05-15 15:54:16 any vuln announcement needs A 100% FUCKING CLEAR EXPLANATION of who is affected 2019-05-15 15:54:17 the linux rds is reliable datagram sockets 2019-05-15 15:54:30 eew 2019-05-15 15:54:35 I've seen no less than three stories mix rdp and Kern RDS so far 2019-05-15 15:54:36 <_ikke_> isn't that an oxymoron? :P 2019-05-15 15:54:51 tcely: me too, but it's easy to get it wrong 2019-05-15 15:54:51 ;-) 2019-05-15 15:54:57 so basically some SOCK_SEQPACKET hack for tcp 2019-05-15 15:55:03 RDS, RDS, MDS 2019-05-15 15:55:25 also, there's a list of versions in the debian tracker 2019-05-15 15:55:37 a whole section even 2019-05-15 15:58:04 Hmm. I didn't see Huawei reported this net/RDS problem before. 2019-05-15 15:58:40 huawei? why would they be one of the first to react? 2019-05-15 15:59:23 hm, this got off topic 2019-05-15 15:59:26 sorry. 2019-05-15 16:00:53 # CONFIG_RDS is not set 2019-05-15 16:00:57 I presume they must use net/rds for something on android devices, but I have no idea why they discovered the problem first. 2019-05-15 16:01:05 yay it's a kconfig thing you can omit 2019-05-15 16:01:28 why does alpine kernel lack /proc/config.gz ? :( 2019-05-15 16:01:48 try /boot/config-vanilla 2019-05-15 16:02:06 vanilla says CONFIG_RDS=m 2019-05-15 16:02:20 _TCP, _RDMA, and _DEBUG are not set 2019-05-15 16:03:14 /etc/modprobe.d/blacklist.conf: 2019-05-15 16:03:17 +blacklist rds 2019-05-15 16:03:24 fixed 2019-05-15 16:04:00 i should probably just find |sed to make a huge blacklist of all the nonstandard useless modules 2019-05-15 16:04:57 (it wouldn't even be loaded unless there's an evil local user or bad local sw, but i honestly don't trust chromium or firefox not to start using some nonsense for some sort of "accelerated http" crap) 2019-05-15 16:05:56 at what point does module whitelist become easier? 2019-05-15 16:06:57 :-) 2019-05-15 16:07:25 the problem is that i kinda want random hardware i connect to a usb port to work without editing config 2019-05-15 16:08:12 what i don't want is dynamic loading of obscure functionality i didn't know existed based on software request 2019-05-15 16:08:31 because plugging in a device is a physical action i'm aware of and expect something to happen 2019-05-15 16:08:51 <_ikke_> Don't you need root permissions to load modules? 2019-05-15 16:08:58 running an upgraded version of some sw and having new shit get pulled into the kernel is not 2019-05-15 16:09:17 _ikke_, kernel will auto-probe for it when the right socket calls are made to request such a socket 2019-05-15 16:09:27 <_ikke_> ah, ok 2019-05-15 16:09:46 so if firefox or chrome started using this, it would definitely get pulled in 2019-05-15 16:12:05 I wonder if seecomp / apparmor / selinux can be used to deny that behavior 2019-05-15 16:14:28 i don't think it can be done globally ("never let this process do anything that would cause a new module to be loaded") and you probably don't wanat that with a highly modular kernel 2019-05-15 16:14:51 for instance, some very core functionality might be modular, and if you get unlucky and the first process to try to use it is a sandboxed one, it would break badly 2019-05-15 16:15:16 you can definitely use seccomp to block certain known syscalls that would pull in unwanted modular functionality 2019-05-15 16:15:24 also blocking them from accessing it if it's already loaded 2019-05-15 16:38:19 danieli: yeah, both 3.9.4 and edge, fresh install. sorry for delay, was asleep. waiting on coffee now :) 2019-05-15 18:07:00 dalias: modprobe configs 2019-05-15 18:54:53 hello there, I have some problem with using LVM with the root partition 2019-05-15 18:55:55 Well, define "some problem" 2019-05-15 18:57:40 after booting up the kernel, it couldn't find the lvm path for the /sysroot and fallback to emergency mode. dm-mod,dm_mod and lvm modules are to be loaded by the bootloader, but dm-mod doesn't seems to loaded. after executing the lvm lvs on the emergeny mode, the modules is loaded. i can mount the FS and after an "exit" the init system just start fine 2019-05-15 18:58:33 *sorry, there couple of misspells... such a long day :) 2019-05-15 19:01:58 one more thing what I forgot to mention, the same image works fine with KVM and VMware player, I got this problem on ESXi VMware hypervisor. 2019-05-15 19:02:39 Hm, is the disk setup the same for all of them? 2019-05-15 19:02:50 I guess they're all configured as SATA? 2019-05-15 19:06:19 on KVM it's configured as normal 'hda' disk, it's iSCSI ont the VMw player, but I have no information about the ESXi. it's a black box, but using the same bootloader parameters, kernel and initrd just with normal partition it works fine 2019-05-15 19:17:09 is the dpaste package what I should use in alpine for facilitating uploading a dmesg log? 2019-05-15 19:24:53 vinibali: Huh, that's odd indeed. Maybe setting `root=` as kernelcmdline fixes it? At least it did for me with ZFS 2019-05-15 19:25:21 uriah: You can use dpaste (or other services like ix.io via curl) 2019-05-15 19:27:12 ok ix.io with curl it is then, ty 2019-05-15 19:31:21 http://ix.io/1J9G 2019-05-15 19:31:44 this is on aarch64 with 512MB of ram 2019-05-15 19:40:34 hey, i got an amd graphics card and the screen freezes at "loading hardware drivers", do i maybe need to enable kernel modesetting? 2019-05-15 19:42:42 Not likely. Do you have the appropriate firmware packages installed? 2019-05-15 19:43:48 root rootfstype and rootflags are defined in my case: root=/dev/vg0/lv_root rootfstype=btrfs rootflags=compress=zstd,noatime,nobarrier 2019-05-15 19:43:55 SpaceToast: seems like your assumption of needing beefier hardware was correct, looks like it tries to allocate to a nonexistent memory address 2019-05-15 19:57:09 trying to install https://bitbucket.org/EionRobb/purple-rocketchat/src/default/ but failing at cc -fPIC -O2 -g -pipe -Wall -DROCKETCHAT_PLUGIN_VERSION='"0.9.2019.05.15.hg.r90.826990b48f41"' -shared -o librocketchat.so librocketchat.c -Wl,-z,relro `pkg-config purple glib-2.0 json-glib-1.0 --libs --cflags` -Ipurple2compat -g -ggdb -lmarkdown 2019-05-15 19:57:29 /usr/lib/gcc/x86_64-alpine-linux-musl/8.3.0/../../../../x86_64-alpine-linux-musl/bin/ld: cannot find -lmarkdown \n collect2: error: ld returned 1 exit status \n make: *** [Makefile:93: librocketchat.so] Error 1 2019-05-15 19:57:49 have installed discount which should be lmarkdown? 2019-05-15 19:58:05 Well, does it provide a markdown.so? 2019-05-15 19:58:44 Cogitri, i just tried installing linux-firmware-amd, linux-firmware-amd-ucode and linux-firmware-amdgpu but it did not help 2019-05-15 19:59:16 oh maybe discount-dev 2019-05-15 19:59:28 contains /usr/lib/libmarkdown.so 2019-05-15 20:00:51 realies: That seems more approach. 2019-05-15 20:00:59 fassl: Have you regenerated your initramfs after installing those? 2019-05-15 20:01:19 nope, at it atm 2019-05-15 20:26:54 Cogitri, do i need to do something else before regenerating it or use some flags for mkinitfs? /lib/firmware is empty in the initramfs 2019-05-15 20:27:43 Huh, that should be full of the firmware 2019-05-15 20:27:54 Maybe try deleting the old one 2019-05-15 20:31:17 i have a backup initramfs which does contain a bunch of files there, these new ones have none 2019-05-15 20:31:56 i am issuing: mkinitfs -c /etc/mkinitfs/mkinitfs.conf -b / 4.19.29-0-vanilla 2019-05-15 20:44:04 i installed linux-firmware and in /lib/modules are way more than in the initramfs, but at least some like isci_firmware.bin, ql2200_fw.bin end up in the initramfs 2019-05-15 20:52:42 Cogitri, not sure if installing whole linux-firmware or enabling kms helped, it does not freeze now 2019-05-15 20:53:04 Cogitri, thanks for the help! 2019-05-15 20:56:09 Has anyone noticed TCP networking problems when using java with musl ? 2019-05-15 20:56:25 Specifically, alpine linux, alpine java and musl in a docker container 2019-05-15 20:56:44 jcdutton: what kind of problems have you noticed? 2019-05-15 20:56:45 when swapping out musl for glibc, everything works again 2019-05-15 20:57:00 netstat -ntp shows Send/Recv-Q backing up 2019-05-15 20:58:15 More details here: https://issues.apache.org/jira/browse/IGNITE-11842 2019-05-15 21:00:05 fassl: Glad it works now :) 2019-05-16 08:54:36 is there alpine ash-completion? 2019-05-16 09:10:34 cim209: you probably want bash or even zsh 2019-05-16 09:10:53 yeah i'm using zsh 2019-05-16 09:11:25 i ended up using an executable script named s and put it in the local bin lol 2019-05-16 09:12:47 so yeah, yesterday i was able to compile eternal terminal 2019-05-16 09:12:51 https://github.com/MisterTea/EternalTerminal/issues/188#issuecomment-492882321 2019-05-16 09:16:29 cim209: what are advantages of using eternal terminal instead of mosh and tmux combo 2019-05-16 09:17:16 mps: mosh doesn't allow you to scroll up through the window, instead it scrolls through history 2019-05-16 09:17:26 never tried tmux 2019-05-16 09:17:30 tmux has scrolling 2019-05-16 09:17:38 yeah never tried that 2019-05-16 10:22:52 _ikke_: iirc you can push to community - mind having a look at my erlang and elixir patch in pw? (4869, 4870) 2019-05-16 10:28:14 hello. 2019-05-16 10:28:47 I've got reports of a behaviour/build change for one of the projects that uses alpine:3.9 docker images: https://github.com/nginxinc/docker-nginx/issues/327 2019-05-16 10:28:58 I believe it could be due to https://github.com/docker-library/official-images/commit/889772fcf5b4bdbd5ea502656956c99b671c1970 2019-05-16 10:29:01 Hello there 2019-05-16 10:30:01 I guess I could set PATH the way I'd like it to be, but I wonder if that's a regression that should be fixed in alpine:3.9 instead? 2019-05-16 11:57:49 <_ikke_> danieli: Sure, will look at it later when I have time 2019-05-16 11:58:34 _ikke_: no worries - also, now that i think about it, some other stuff like ejabberd/cloudi/tsung might have to be rebuilt as well 2019-05-16 19:11:44 I have successfully managed to repair iptables in aarch64 on my pi3a+ 2019-05-16 19:13:40 fwiw, i had to recompile it natively using abuild ; I tried updating the kernel first and it failed, then rebuilt iptables *after* making sure linux-rpi-dev was installed and now it doesn't cause kernel panics 2019-05-16 19:14:51 since this is just irc, do you think I should file a bug report with my findings? 2019-05-16 19:15:10 (might need to do so tomorrow, busy day today) 2019-05-16 22:02:21 How do I check the last time apk's index has been updated? 2019-05-17 04:04:05 Hey, I have a couple of questions about Alpine. I am working on an embedded project (RPi Zero) and I want to know if there is a way to run in diskless mode without loading everything into RAM. 2019-05-17 04:04:59 I.e. boot the system, set it up how I want, create an APKOVL, and then lead that (along with an APKCACHE) on a diskless system without loading everything in the APKOVL into RAM. 2019-05-17 04:05:10 load* that 2019-05-17 04:06:18 The reason for this is because the Pi Zero is resource constrained, and 512MB of RAM fills up fast. 2019-05-17 04:06:32 Would it make more sense to do this with a traditional sys install? 2019-05-17 04:08:23 Different question: I am having problems with an APKOVL. I create it on an Alpine system running under Virtualbox, and then mount the VDI offline and copy that tar.gz to the host. 2019-05-17 04:09:02 But when I try to extract it on the host, it is missing two files (.ash_history and .xinitrc, which I created under the root home directory). 2019-05-17 04:10:25 When I try to extract it on the Alpine VM, those files are in the tar. When I try on the host, they aren't there. As far as I can tell, everything else is preserved, except the files I added to the APKOVL with `apk include` 2019-05-17 04:10:54 Sorry, *lbu include. 2019-05-17 04:12:08 I have also manually edited /etc/apk/protected_paths.d/lbu.list, with the same result. 2019-05-17 04:17:02 I am not sure why the tar would be missing files on one machine but not another. AFAIK, tar is a pretty standard posix util... I don't see how it would be possible to have an incompatibility between two tar versions. 2019-05-17 04:17:26 Also: the host machine is Debian based. Nothing too exotic. 2019-05-17 06:08:44 not sure I understand - diskless mode means loading everything into memory (a ramdisk), you need *somewhere* to store it 2019-05-17 10:31:36 is there a list of packages that is going to be installed from the extended image? 2019-05-17 10:38:25 afaik there is no extra pkgs installed from extended, extended is only in case you want local packages instead of downloading them over the network 2019-05-17 10:38:44 cim209: I'll find them, one sec 2019-05-17 10:39:15 cim209: https://git.alpinelinux.org/aports/tree/scripts/mkimg.standard.sh#n34 2019-05-17 10:40:06 lol nano is installed 2019-05-17 10:40:16 hmm sudo too 2019-05-17 10:40:41 busybox magick 2019-05-17 10:40:54 hmmm maybe not 2019-05-17 10:41:11 irssi.. really 2019-05-17 10:41:13 it's called 'extended' for a reason :P 2019-05-17 10:41:18 does extended install these or only provide them in case someone wants to install them? 2019-05-17 10:41:25 but even that isn't as much bloat as some other distros' images come with 2019-05-17 10:41:28 p4Wv1qn095FW: it installs them 2019-05-17 10:41:36 wow. thx TIL 2019-05-17 10:42:05 p4Wv1qn095FW: lol 2019-05-17 10:42:08 that's why i always go with standard on regular boxes and virt (slimmed down kernel) for VMs 2019-05-17 10:43:12 but, ehm. this is strange, i think to remember to having a standard image also having wpa_supplicant available 2019-05-17 10:44:29 it has it available in the ISO, yes, but it doesn't install it 2019-05-17 10:44:47 run '7z l alpine-standard-3.9.4-x86_64.iso' and look for 'apks/x86_64' 2019-05-17 10:45:13 aah. 2019-05-17 10:45:59 so the extended label is really just about installing stuff. hmmmm. strange but ok. 2019-05-17 10:46:03 thx 2019-05-17 10:46:45 it seems so 2019-05-17 10:47:12 i find it odd that those apks are included in the standard ISO at all if they aren't installed 2019-05-17 10:47:39 where is the mirror file 2019-05-17 10:47:46 can't seem to find it in docs 2019-05-17 10:50:37 mirror file..? 2019-05-17 10:50:51 yeah 2019-05-17 10:51:43 nvm 2019-05-17 10:51:49 it's the actual repository file 2019-05-17 10:51:59 it's generated on the fly 2019-05-17 10:53:16 https://git.alpinelinux.org/alpine-conf/tree/setup-apkrepos.in 2019-05-17 10:53:24 pulls from here https://mirrors.alpinelinux.org/mirrors.txt 2019-05-17 10:53:41 crazy how the netherlands have relaly high speeds 2019-05-17 10:58:24 <_ikke_> We're a big relay 2019-05-17 11:03:18 <_ikke_> AMS-IX is one of the larger nodes 2019-05-17 11:04:06 yeah, that IX is one of the larger ones 2019-05-17 15:34:55 There is no disown on alpine, am i correct? 2019-05-17 15:35:33 Disown isn't a shell built-in? 2019-05-17 15:35:44 <_ikke_> yes, I think so 2019-05-17 15:35:51 <_ikke_> It's not a command, it can't be 2019-05-17 15:35:53 I assume you're talking about the disown command that frees background jobs 2019-05-17 15:36:25 correct 2019-05-17 15:37:00 <_ikke_> So you need a shell that supports it 2019-05-17 15:37:13 I think ash (alpine system shell) has it 2019-05-17 15:37:36 <_ikke_> north1: afaik ash is not short of Alpine system shell 2019-05-17 15:37:40 <_ikke_> s/of/for 2019-05-17 15:37:55 _ikke_: it isn't i know 2019-05-17 15:38:09 I meant Ash is the system shell used by Alpine Linux 2019-05-17 15:38:26 But now let's rewrite history so that Ash actually means Alpine System Shell 2019-05-17 15:38:44 ash: disown: not found, but same with zsh 2019-05-17 15:39:39 _ikke_: almquist shell 2019-05-17 15:39:43 <_ikke_> right 2019-05-17 15:39:45 zsh neither 2019-05-17 15:39:58 danieli: shhhhh 2019-05-17 15:40:01 <_ikke_> disown ~ 2019-05-17 15:40:03 <_ikke_> disown: no current job 2019-05-17 15:40:06 <_ikke_> on zsh 2019-05-17 15:40:21 yash has it, and it's in the docs for zsh 2019-05-17 15:40:24 Please tow the revised history lines danieli 2019-05-17 15:41:23 my bad, to many open windows ... and a long day 2019-05-17 15:42:16 was logged in with the wrong user 2019-05-17 15:42:22 works with zsh, sorry 2019-05-17 15:48:34 i think disown and nohup do the same thing 2019-05-17 15:48:52 "Both disown and nohup are used to allow you to exit a running shell session without stopping running jobs. ... nohup is defined by POSIX while disown is not." 2019-05-17 15:49:18 so you should prefer nohup 2019-05-17 15:49:41 the difference for me is, nohup can only be applied at the beginning of the task 2019-05-17 15:50:02 disown can be applied at any time, but yeah nohup should be preferred 2019-05-17 15:50:18 you can run either one after backgrounding the task afaik 2019-05-17 15:52:24 i don't know how to do it with nohup 2019-05-17 18:08:35 Just in case... https://make-linux-fast-again.com/ 2019-05-17 18:13:46 hahahahhaha 2019-05-17 18:14:10 security people are going to develop (even more of) an eye twitch 2019-05-17 18:14:23 danieli: For home users... 2019-05-17 18:14:32 still 2019-05-17 18:14:36 Why? 2019-05-17 18:15:01 it's not like non-business users aren't at risk 2019-05-17 18:15:28 Nobody has access to my machine, only I do. 2019-05-17 18:15:39 that's true until it isn't 2019-05-17 18:16:17 If you know the candle is fire, then the meal was cooked a long time ago? 2019-05-17 18:16:30 I'm sorry? 2019-05-17 18:16:36 that didn't really make sense in my head 2019-05-17 18:17:23 Sorry, quote from Stargate SG1. One of the main guys is named "Daniel Jackson". Your quote "that's true until it isn't" sounded pretty equal to me :-) 2019-05-17 18:18:05 I never watched Stargate much, occasionally as white noise - more Star Trek here 2019-05-17 18:18:42 my point is that threat modelling is hard, and that even home users can be at risk security-wise 2019-05-17 18:19:00 only you have access to your computer.. until someone malicious does too. 2019-05-17 18:19:45 True. But 20% of performance degrading sounds more like a Intel marketing feature to me to buy newer CPUs 2019-05-17 18:20:03 danieli: Did you watch Discovery Season 1 & 2? 2019-05-17 18:20:07 that might be, and Apple has done just that in the past 2019-05-17 18:20:22 Nobody has access to my machine, only I do. <-- nope 2019-05-17 18:20:35 honestly, I just watched roughly the first half of S1 of Discovery 2019-05-17 18:20:48 dalias: nah... you need to watch S2. Forget S1 2019-05-17 18:21:12 wrong highlight 2019-05-17 18:21:17 oops 2019-05-17 18:21:27 my IRC client gave both of you the wrong color, my mistake :) 2019-05-17 18:21:38 oooops. 2019-05-17 18:21:41 er.. same color. 2019-05-17 18:21:42 I need sleep. 2019-05-17 18:22:16 danieli: S2 fixed some stuff and makes real fans pretty happy. 2019-05-17 18:22:27 I am however looking forward to the next season of The Expanse 2019-05-17 18:22:30 it was great 2019-05-17 18:22:53 Yeah, gonna watch this too. So little time :-( 2019-05-17 18:23:18 velix, every bottom-feeding adtech/tracking/analytics/etc producer partnered with every website you visit does 2019-05-17 18:23:25 so.... 2019-05-17 18:23:55 depends on what is meant by 'has access' 2019-05-17 18:24:01 dalias: But I mean, it doesn't have harmful access to the CPU to use Spectre etc.? 2019-05-17 18:24:09 CPU & memory 2019-05-17 18:24:09 YES IT DOES 2019-05-17 18:24:13 mmh. 2019-05-17 18:24:17 that's the whole point of spectre and why it's so evil 2019-05-17 18:24:22 yeah, it can be exploited via JS 2019-05-17 18:24:33 But it's a postgresql server ;) 2019-05-17 18:24:35 spectre completely crosses all hardware privilege domains 2019-05-17 18:25:37 because speculative execution is fundamentally unsafe without _extreme_ levels of context isolation no cpu vendor has ever implemented (and probably never will, since once you go to those lengths, non-speculative is a HUGE win in performance and power efficiency) 2019-05-17 18:25:49 okay, gonna read about it 2019-05-18 00:41:19 Anyone know of a way to run Calibre under Alpine other than using a chroot? 2019-05-18 02:49:40 Where do I find the Calibre you're asking about? 2019-05-18 02:51:01 https://calibre-ebook.com/ 2019-05-18 02:51:36 Any good alternatives for Alpine would be appreciated as well 2019-05-18 02:56:40 Did you try gcompat already? 2019-05-18 04:04:59 Yeah, sort of; I may or may not have experimentally run `patchelf --add-needed /lib/libgcompat.so.0 /lib/ld-musl-x86_64.so.1` 2019-05-18 04:05:23 ... I do some dumb things when I'm bored 2019-05-18 04:07:38 Either way, it doesn't work with gcompat (and I'm going to just casually undo that stupid decision) 2019-05-18 04:19:40 pixelherodev: I think you'd be best off asking calibre to support musl or python3 2019-05-18 04:20:02 python2 + PyQt5 is the stumbling block I ran into trying to compile from source 2019-05-18 04:20:52 Or "better" yet: adding that to my (lengthy) list of things to do... 2019-05-18 04:32:02 Right below "learn how to prioritize" and "actually get shit done" 2019-05-18 04:33:44 that's usually tricky 2019-05-18 07:30:20 hi there 2019-05-18 07:30:34 I just see a video, of Alpine on the Raspberry Pi 2019-05-18 07:31:47 why do you use a - character, instead of simple "setup" or "alpine" when you KNOW that at this time in point, people are working in a potentially strange keyboard layout? Thanks a lot 2019-05-18 07:35:03 ShalokShalom: tab-complete works 2019-05-18 07:35:30 aha 2019-05-18 07:35:46 still seems a bit counterintuitive 2019-05-18 07:36:03 especially considering Pi users might not know about tab completion 2019-05-18 07:36:15 is there ny benefit? 2019-05-18 07:36:55 don't know to be fair, and there is a keyboard without '-' ? 2019-05-18 07:37:21 it is not visible 2019-05-18 07:37:28 you have it somewhere on your keyboard 2019-05-18 07:37:44 try to find - on the swedish layout just right now :D 2019-05-18 07:37:56 if the people works with strange layouts they probably know default one, i think 2019-05-18 07:38:02 like out of your head and while you are in the command line 2019-05-18 07:38:09 why should they? 2019-05-18 07:38:29 strange layouts ARE the default way in other countrys :D 2019-05-18 07:38:59 hmm, never had such issue so don't know, really 2019-05-18 07:40:21 which type of layout do you use? 2019-05-18 07:41:05 four, english, serbian, serbian latin, russian and sometimes dutch 2019-05-18 07:41:17 even sanskrit sometimes 2019-05-18 07:42:36 Pi users tend to be new to the scence 2019-05-18 07:42:51 i see no reason why making their lives hard makes any sense 2019-05-18 07:43:55 this naming is not intended to make someone lives hard, it is made at the beginning of development 2019-05-18 07:44:29 although I think it will not hurt to add aliases to these commands 2019-05-18 07:44:35 so, do you see any reason to change it? 2019-05-18 07:44:39 thats perfect 2019-05-18 07:44:56 while they will type the command that you show in the line above 2019-05-18 07:45:12 why do you type in Sanskrit sometimes? :D 2019-05-18 07:45:30 four languages is already a hard task to pull off, right? 2019-05-18 07:46:14 well, because I like old languages and trying to learn, but when I have time, which is less and less 2019-05-18 07:47:06 no, it is not hard to learn languages. when you leard few others are a lot easier 2019-05-18 07:47:18 Ah I see 2019-05-18 07:47:24 s/leard/learn/ 2019-05-18 07:47:25 Yeah, I can see that point 2019-05-18 07:47:41 While I guess this also differs between all these languages 2019-05-18 07:47:54 some are completely different, like that one of the native american 2019-05-18 07:48:16 the US army used them in the cold war to encrypt messages, since their language is so fundamentally different 2019-05-18 07:48:26 I didn't tried Chinese :) 2019-05-18 07:48:35 nice to meet you :) 2019-05-18 07:48:44 i would not try it xS 2019-05-18 07:49:14 that was Navajo lang, it differs a lot from other american natives langs 2019-05-18 07:49:42 this video series shows how complex Japanese is quite nicely 2019-05-18 07:49:43 https://www.youtube.com/watch?v=ZefYEaNUJ7w 2019-05-18 07:49:52 yeah, Navajo 2019-05-18 07:50:10 heh, I learnt some japanese words and terms 2019-05-18 07:50:47 to me it doesn't look hard, but don't have big reason to learn it deeply 2019-05-18 07:51:44 i see 2019-05-18 07:52:01 as martial artist (japanese martial) I had to learn some terms and phrase 2019-05-18 07:52:15 the developer of a emulator, byuu tries it since a lot of years and says its super hard 2019-05-18 07:52:23 while you might have a talent for it 2019-05-18 07:52:23 and I like haiku (japanese poetry) 2019-05-18 07:52:27 I see 2019-05-18 07:52:37 yeah, I remember they told us some in Judo as well 2019-05-18 07:52:45 while I was a kid, so forgot everything 2019-05-18 07:56:30 do you know where this script lives? and the message in the output, which hints that command? 2019-05-18 07:56:51 Github seems just your mirror 2019-05-18 07:57:07 And none of your repos seems to hint where the original source is 2019-05-18 07:57:08 Why? 2019-05-18 07:57:28 primary repo is git.alpinelinux.org github is a mirror 2019-05-18 07:57:52 Ah, I see. 2019-05-18 07:58:11 Adding the original repo into the Github description sounds sane? 2019-05-18 07:58:13 there is cgit web interface 2019-05-18 07:58:19 So other people can find it 2019-05-18 07:58:21 Oh... 2019-05-18 07:59:12 if you visit http://alpinelinux.org you will see all what you need 2019-05-18 07:59:42 yeah, thats cool 2019-05-18 07:59:50 would you add these changes? 2019-05-18 08:00:01 you have more oversight over your repos 2019-05-18 08:00:57 well, changes are discussed and patches are needed before something is changed, look wiki there 2019-05-18 08:02:01 fine, sure 2019-05-18 08:02:26 so, what can I do? 2019-05-18 08:02:36 ShalokShalom: sorry, I have to go to other city to meet some friends from distant country, bye 2019-05-18 08:02:46 sure sure 2019-05-18 08:02:48 much fun 2019-05-18 08:03:30 thanks, same to you and enjoy usage of alpine :) 2019-05-18 08:04:12 lets see :) 2019-05-18 08:06:22 ah; i have another question: 2019-05-18 08:07:27 if anybody else is here: why does Alpine provide no translations over libconv or a similar library, in order to replace the character encoding in glibc 2019-05-18 08:45:54 Hi all, has anybody ever achieved to run C++ grpc-based service on alpine? I'm able to build and run the service, however, as soon as I try to call Write-method on unique_ptr > writer(stub->put(&clientContext, &res)); /*...*/ writer->Write(req); the method returns false immediately (broken stream). Connecting using netcat / telnet, however, works; compiling the same sources on ubuntu-image and issuing t 2019-05-18 08:46:10 Might this issues connected to libmusl? 2019-05-18 09:09:03 ShalokShalom: I'm quite used to the Scandi layout (we for the most part have the same layout) but I stick to US most of the time 2019-05-18 09:09:07 did you ever find that hyphen? :P 2019-05-18 15:37:02 how do I setup the serial console speed when trying to boot the install media? 2019-05-18 15:41:43 press tab in syslinux when offered to boot, and then write the name of the kernel syslinux wants to boot and appens serial=ttyS0,115200 2019-05-18 15:42:43 oh. not serial -> console=ttyS0,115200 2019-05-18 15:43:10 yes, the latter, and you can have multiple console= options 2019-05-18 15:43:15 it's pretty much a universal thing (not alpine-specific) 2019-05-18 15:43:33 details: https://www.kernel.org/doc/html/latest/admin-guide/serial-console.html 2019-05-18 16:05:44 greetings, Im trying to install xvfb-run script which says it's an edge package, but when I run "apk add xvfb-run", I get "missing world[xvfb-run]" https://pkgs.alpinelinux.org/package/edge/testing/x86/xvfb-run 2019-05-18 16:05:58 /code # cat /etc/apk/repositories 2019-05-18 16:05:59 http://dl-cdn.alpinelinux.org/alpine/v3.9/main 2019-05-18 16:05:59 http://dl-cdn.alpinelinux.org/alpine/v3.9/community 2019-05-18 16:06:00 @edge http://nl.alpinelinux.org/alpine/edge/community 2019-05-18 16:06:02 @edge http://nl.alpinelinux.org/alpine/edge/main 2019-05-18 16:06:07 Is there some flag or other that retries downloading packages if the network is down? 2019-05-18 16:06:44 holler: if you want to install a command you can do apk add cmd:xvfb-run 2019-05-18 16:07:05 holler: oh, the package is in the testing repo. 2019-05-18 16:07:34 how can I add that? 2019-05-18 16:07:59 just change main/community for testing 2019-05-18 16:08:02 i think testing is only built for edge though 2019-05-18 16:09:14 fetch http://nl.alpinelinux.org/alpine/edge/testing/x86_64/APKINDEX.tar.gz 2019-05-18 16:09:15 ERROR: unsatisfiable constraints: 2019-05-18 16:09:15 xvfb-run-1.20.4-r0: 2019-05-18 16:09:17 masked in: @edge 2019-05-18 16:09:18 satisfies: world[xvfb-run] 2019-05-18 16:10:30 careful with mixing edge and stable 2019-05-18 16:10:39 anyway, do apk add xvfb-run@edge 2019-05-18 16:11:14 galaxie: not that I can think of, but writing a wrapper to do that should be fairly easy 2019-05-18 16:12:18 /bin/sh: fmt: not found 2019-05-18 16:12:21 north1: correct, testing is only available in edge 2019-05-18 16:12:25 isn't that a unix command? fmt 2019-05-18 16:12:40 it's in coreutils for sure 2019-05-18 16:12:57 trying to set up selenium in an alpine docker container :) 2019-05-18 16:13:04 do I need to add coreutils? 2019-05-18 16:13:20 i don't know 2019-05-18 16:13:25 yep 2019-05-18 16:13:27 i'm saying that fmt is in coreutils as far as i can see 2019-05-18 16:13:28 that worked 2019-05-18 16:15:06 does selenium even run on alpine? 2019-05-18 16:15:46 I got it to run using selenium-standalone 2019-05-18 16:16:28 but tbh I havent actually got it up and connected yet 2019-05-18 16:16:31 the process ran 2019-05-18 16:17:07 Im hoping to encapsulate a node e2e test repo into a docker container and run it to run tests 2019-05-18 16:36:04 Black screen, no POST at all, works with other distribution, Raspberry Pi 2019-05-18 16:36:05 https://i.imgur.com/mrpqrad.png 2019-05-18 16:36:37 which image on which model of raspberry pi? 2019-05-18 16:36:55 That one from the homepage with Rasperry Pi 3+ 2019-05-18 16:37:10 Is there more than one image? 2019-05-18 16:37:19 armv7, armhf, aarch64 2019-05-18 16:37:22 there are three 2019-05-18 16:37:42 ARMv7 2019-05-18 16:37:57 Sorry, I have downloaded it a few days ago, didnt remember anymore :) 2019-05-18 16:38:46 try the arm64 one 2019-05-18 16:38:48 er, aarch64 2019-05-18 16:39:48 ok 2019-05-18 16:39:58 didnt use that more memory? 2019-05-18 16:40:46 i don't know about that, but i know it's supposed to work on pi3s (they have 64-bit CPUs but most distros for the RPi3 runs in 32-bit mode) 2019-05-18 16:41:06 I see 2019-05-18 16:41:11 Yes, it can 2019-05-18 16:41:21 try it, see if it works 2019-05-18 16:41:22 The most distros released 32 Bit ISOs before 64 Bit ones 2019-05-18 16:41:24 Sure 2019-05-18 16:45:42 Oh, the ISO is broken 2019-05-18 16:45:48 This is why I prefer torrents 2019-05-18 16:45:53 So, the tar.gz I mean 2019-05-18 16:46:02 Different SHA256 2019-05-18 16:46:05 Thanks 2019-05-18 16:52:23 well, same thing 2019-05-18 16:52:33 aha, i assumed you had checked the download 2019-05-18 16:52:33 with correct summed aarch64 2019-05-18 16:52:43 has anyone a PI to check? 2019-05-18 16:52:53 i have a few but i think they're all occupied 2019-05-18 16:53:12 ok, I see 2019-05-18 16:53:26 someone else might 2019-05-18 16:53:35 I thought a 50mb download from a dedicated server would not break 2019-05-18 16:53:40 ok, I see, thanks 2019-05-18 16:53:50 dedicated server = not sourceforge :D 2019-05-18 16:54:21 which mirror did you dl from? 2019-05-18 16:54:30 and what was the exact filename+? 2019-05-18 16:55:56 https://elixirforum.com/t/is-elixir-good-for-beginner-programmer/1675/12?u=shalokshalom 2019-05-18 16:56:00 oh, sorry 2019-05-18 16:56:14 http://dl-cdn.alpinelinux.org/alpine/v3.9/releases/armv7/alpine-rpi-3.9.4-armv7.tar.gz 2019-05-18 16:56:15 this one 2019-05-18 16:57:43 elixir is great but it's a bit different from most other things 2019-05-18 16:58:02 and all right, thanks, I'll check it 2019-05-18 16:58:41 checksum is valid for me 2019-05-18 16:58:45 i just posted this in an accident xD 2019-05-18 16:58:56 i figured, but still :) 2019-05-18 16:58:57 it was still in my clipboard for some reason 2019-05-18 16:59:02 ok, I see 2019-05-18 16:59:06 then look at Elchemy also 2019-05-18 16:59:16 its Elm > Elixir :) 2019-05-18 16:59:26 elm was never my thing 2019-05-18 16:59:27 nice community 2019-05-18 16:59:30 I see 2019-05-18 16:59:34 i'm more of an erlanger, working on getting better at elixir 2019-05-18 16:59:44 You like type inference? 2019-05-18 16:59:54 Erlang is also nice :) 2019-05-18 17:00:05 All these languages are so elegant 2019-05-18 17:00:06 imho 2019-05-18 17:00:09 the type system is probably erlang's #1 downside 2019-05-18 17:00:19 well, Elchemy provides that 2019-05-18 17:00:29 It has nothing to do with Elm besides that 2019-05-18 17:00:39 it just uses its syntax and tools, thats all 2019-05-18 17:00:52 i'm fine with thorough testing and with dialyzer 2019-05-18 17:01:00 yeah, this is interesting 2019-05-18 17:01:17 so far as I understand, is this some kind of type checker for dynamic types? 2019-05-18 17:01:31 while you miss Type Driven Development with that 2019-05-18 17:02:11 i looked over elchemy, it doesn't look like something i want to get involved with 2019-05-18 17:02:13 domain driven design is also pretty save with inference, due to the fact that the whole thing is fundamental 2019-05-18 17:02:19 I see (Y) 2019-05-18 17:02:42 I love "Learn you some Erlang" its such an awesome book 2019-05-18 17:02:45 dialyzer is a static analysis tool (discrepancy analyzer for erlang) 2019-05-18 17:02:50 One of the most helpful tutorials 2019-05-18 17:02:54 Ah, I see :) 2019-05-18 17:05:21 Manjaro can run on this Pi 2019-05-18 17:05:29 While its a horrible distribution xD 2019-05-18 17:06:17 a lot of different OSes have been ported to the pi 2019-05-18 17:06:26 Yes... 2019-05-18 17:06:35 Just not my favorite one ^.^ 2019-05-18 17:06:42 So Alpine is next 2019-05-18 17:06:50 Let`s look at freeBSD 2019-05-18 17:06:59 Maybe they provide something useful 2019-05-18 17:07:31 which one is your favorite then? 2019-05-18 17:08:27 KaOS, the one that packages focused on one DE and this is Plasma 2019-05-18 17:08:34 http://kaosx.us/ 2019-05-18 17:08:50 never heard of it 2019-05-18 17:09:03 All the others think they can dance on 10 weddings at the very same time 2019-05-18 17:09:18 You can never package 4 different DEs on one of the same base 2019-05-18 17:09:29 You will always prefer one :) 2019-05-18 17:09:36 let's agree to disagree 2019-05-18 17:09:42 And others split soooo much 2019-05-18 17:09:44 it's difficult for smaller projects though 2019-05-18 17:09:48 Well, have you tried? 2019-05-18 17:09:56 I distro jumped from distro to distro 2019-05-18 17:10:10 I've tried a thing or two 2019-05-18 17:10:31 Some offer pretty stable distributions, so long as you stay on their main DE, which is mostly GNOME or another GTK based one. 2019-05-18 17:11:21 How can you imagine that hundreds of packages can behave completely the same on other thousands of packages, especially when you split all and everything into million pieces :D 2019-05-18 17:12:27 I have never experienced a constantly stable and fresh Plasma until I visited this little distro. 2019-05-18 17:58:28 yep, it seems to be an Alpine issue 2019-05-18 17:58:32 works now with Arch 2019-05-18 17:58:46 thanks a lot 2019-05-18 17:58:54 for the support 2019-05-18 17:58:57 :D 2019-05-18 18:14:42 does Alpine support ext4 on the Pi? 2019-05-18 18:14:51 since your documents say fat32 only 2019-05-18 18:15:09 which wonders me, since ext4 is probably three times faster 2019-05-18 18:26:12 ext4 is supported 2019-05-18 18:26:20 only the boot partitoin is fat 2019-05-18 18:27:33 thanks a lot 2019-05-18 18:28:10 ShalokShalom: it's an rpi limitation, not alpine 2019-05-18 18:28:22 How that? 2019-05-18 18:28:32 /b 3 2019-05-18 18:28:39 You mean for the boot? Yes 2019-05-18 18:28:46 its bootloader can only boot off fat32/vfat 2019-05-18 18:28:47 yes 2019-05-18 18:28:54 I just forgot that Alpine is RAM only by default 2019-05-18 18:29:05 While its said, its not booting for gods sake 2019-05-18 18:29:08 :D 2019-05-18 18:29:22 They dont boot at all 2019-05-18 18:29:31 Yellow light is not lighting at all 2019-05-18 18:29:36 So no process at all 2019-05-18 18:29:48 I'm gonna test 2019-05-18 18:30:53 Thanks :D 2019-05-18 18:30:58 3+ 2019-05-18 18:32:50 you mean 3b+? 2019-05-18 18:34:56 Yes 2019-05-18 18:35:08 Any other data I can provide? 2019-05-18 18:35:23 i'm just going to test if it works for me 2019-05-18 18:35:27 or if the image is broken 2019-05-18 18:35:56 Yeah, nice 2019-05-18 18:36:12 I use the kdepartitionmanager maybe that plays a role 2019-05-18 18:36:30 no clue, I haven't used KDE in years 2019-05-18 18:36:32 While I used FAT32, so .. 2019-05-18 18:36:41 Its gparted under it 2019-05-18 18:47:34 And did you find the issue? 2019-05-18 18:47:47 i'm a little preoccupied, i'll boot it in a bit when i have time 2019-05-18 18:48:05 i imaged the card a little while ago 2019-05-18 18:54:35 sure sure 2019-05-18 18:54:41 I mean with the current ISO 2019-05-18 18:54:48 maybe it broke recently 2019-05-18 18:54:54 iso? 2019-05-18 18:54:59 the rpi image isn't iso, it's a tarball 2019-05-18 18:55:01 except its the same tar.gz of course 2019-05-18 18:55:03 I know 2019-05-18 18:55:05 and it booted perfectly here 2019-05-18 18:55:08 used too 2019-05-18 18:55:08 on an rpi3b+ 2019-05-18 18:55:10 ok ,fine 2019-05-18 18:55:13 the current one? 2019-05-18 18:55:35 alpine-rpi-3.9.4-aarch64.tar.gz 2019-05-18 18:56:43 thanks a lot 2019-05-18 19:37:24 ... and another alpine system up and running 2019-05-18 19:39:14 honestly, it's so annoying to install a ""traditional"" Alpine install on the raspberry pi I'm gonna go with something else 2019-05-18 19:39:39 what do you mean? 2019-05-18 19:55:03 It seems that "ip -o link show" does not show anything on alpine, whereas it usually shows interfaces in one line - do you consider this a bug? 2019-05-18 19:56:10 <_ikke_> telmich: busybox ip? 2019-05-18 19:59:53 _ikke_: yes 2019-05-18 20:00:49 <_ikke_> Works with the iproute2 version 2019-05-18 20:01:07 <_ikke_> So something funky / not implemented in the busybox ip commandlet 2019-05-18 20:01:25 <_ikke_> apk add iproute2 2019-05-18 20:01:55 _ikke_: sure, that works. It will break cdist on standard installations for interface discovery at the moment - I've to see for a workaround then 2019-05-18 20:40:35 Has anyone had success getting a wireguard interface to work with wg-quick on alpine? 2019-05-18 20:44:42 rmsvc: yes, many. but also you can use /etc/network/interface 2019-05-18 20:47:43 Hmm, I'll keep debugging. I'll try the way it shows on the wiki, too. Thanks. 2019-05-18 20:47:56 coming back to my serial question: when I bootup the APU, I see the ISOLINUX message and boot:, but it does not accept any input 2019-05-18 20:48:11 i.e. pressing tab or enter or a does not change the prompt 2019-05-18 20:48:28 rmsvc: yes, I did 2019-05-18 20:49:01 rmsvc: what is the output that you get? 2019-05-18 20:49:16 telmich: I'm not sure, but I think standard iso doesn't support serial console 2019-05-18 20:49:55 mps: oh. which image should I be using in that case? 2019-05-18 20:50:54 well, not sure. I know that virt images have support for serial console and maybe some arm tarballs 2019-05-18 20:52:08 i actually installed alpine on an apu with the standard image and in the way i said earlier here 2019-05-18 20:52:17 last time I tried standard x86 and x86_64 iso's it didn't worked but that was about year ago 2019-05-18 20:52:26 just last week 2019-05-18 20:52:48 p4Wv1qn095FW: it have serial console? 2019-05-18 20:52:52 yes 2019-05-18 20:53:09 oh, that's a good news 2019-05-18 20:53:22 telmich: sorry, I was wrong 2019-05-18 20:53:48 like i said above: on the syslinux prompt write whatever the name of the menu entry is in syslinux (you can find out with pressing tab) and then write console=ttyS0,115200 2019-05-18 20:54:04 i think it is something "vanilla" 2019-05-18 20:54:43 and you have to be quick, i think syslinux only waits for 3 secs before booting automatically 2019-05-18 21:07:25 telmich: http://ix.io/1Jpk 2019-05-18 21:15:19 rmsvc: why do you think something is wrong? 2019-05-18 21:15:28 rmsvc: from the output, it looks rather normal to me 2019-05-18 21:19:44 telmich: everything looks normal to me too and the interface is setup correctly, but once it's up I have no connection to the internet. I think it might be more of an iptables or ipv6 issue than a wireguard issue 2019-05-18 21:40:16 rmsvc: ahh..that rings a bell 2019-05-18 21:41:03 So after commenting out the net settings in /etc/sysctl.d/00-alpine.conf, everything seems to be working now. 2019-05-18 21:41:05 I've seen that under macos w/ wireguard; in your case: can you ping ip addresses inside the vpn / on the other side? 2019-05-18 23:08:30 hello everyone 2019-05-18 23:08:41 I have a question about zfs on ALpine 2019-05-18 23:08:54 I can't make it run on latest stable alpine 2019-05-18 23:09:23 it seems the kernel modules for the package zfs-vanilla are from another kernel 2019-05-18 23:09:45 and I have no clue on hot to get the correct kernel modules/drivers for zfs 2019-05-18 23:12:38 the package is set for 4.19.41-r0 kernel 2019-05-18 23:12:44 but in 3.9 the kernel is different 2019-05-19 00:44:13 are you using linux-vanilla as your kernel? 2019-05-19 05:47:23 as iptables is now deprecated and nftables has taken it's place, is there any plans to add the iptables-nft bridge to alpine as in many distros have done? 2019-05-19 08:41:45 ShalokShalom: nothing relevant to your issue, i was just ranting 2019-05-19 10:04:31 henlo, here for a quick support question 2019-05-19 10:05:05 soo... i made the dumb decision to use alpine as a daily driver... 2019-05-19 10:05:26 how do i get icons and themes properly working with xfce? 2019-05-19 10:05:46 ive had this issue before with gentoo but its happining again 2019-05-19 10:15:08 Not necessarily a stupid decision, I use it as my daily driver with GNOME :P 2019-05-19 10:16:33 Hm, do you have adwaita-icon-theme, hicolor-icon-theme and gnome-themes-standard installed? 2019-05-19 10:23:27 yeah, you gotta install icons to get icons 2019-05-19 10:24:44 Although it'd most likely be a good idea to have xfce depend on these by default 2019-05-19 10:25:17 I'm using alpine 3.9.4 on aarch64, custom kernel, and I have a funny bug with openssh.when calling `ssh user@1.1.1.1` ie using an ip addr it succeeds but if I use `ssh@host.com` to the same host using a hostname, it segfaults 2019-05-19 10:26:02 well, i know that. im trying to rice it up with adapta theme and papirus icons. there is no packages for these 2 so i used the installer. it extracted to the right locations but the icons are dysfunctional and the theme doesnt apply to all of xfces menus. I tried downloading a package from arch repos and extracting it myself and it has no effect. 2019-05-19 10:27:01 xcko: that's funny.. what does gdb say? 2019-05-19 10:27:48 Do you have librsvg installed? Maybe some icons are only available in svg 2019-05-19 10:37:13 i could try installing that, thanks! 2019-05-19 10:38:32 Cogitri: one of the reasons I switched from debian was a so called 'Debian dependency hell' 2019-05-19 10:39:42 Well, depending on the default icon and themeset doesn't sound like a "dependency hell" to me, but rather like sane behaviour 2019-05-19 10:39:42 xcko: I'm using alpine aarch64 with custom kernel (4.4.179) without problem, ssh works without problem 2019-05-19 10:40:25 what is 'sane' depends on the ones POV 2019-05-19 10:40:56 danieli: if I run `gdb ssh` then run user@foo I get a SIGILL, but it happens no matter what foo is - off hand you know how to enable coredumps? 2019-05-19 10:41:04 mps: which version of alpine? 2019-05-19 10:41:28 for example. why xfce have to depend on adwaita-icon-theme if I install local one I prefer 2019-05-19 10:41:31 mps: Very much so, but IMHO if I install a desktop it should work :) 2019-05-19 10:41:52 and I don't like to have 'bloat' from adwaita-icons 2019-05-19 10:42:22 FWIW, I plan to make a gnome metapackage which contains the default deps too. If you don't want that, don't install the metapackage and install the components yourself. How does that sound? 2019-05-19 10:42:36 xcko: last stable, 3.9.4 but it worked with previous versions also 2019-05-19 10:43:55 mps: dropbear or openssh? 2019-05-19 10:44:54 xcko: openssh 2019-05-19 10:45:50 Cogitri: that approach sounds fine, I even wanted to say that (for xfce) but you were faster 2019-05-19 10:46:01 Ah :) 2019-05-19 10:46:46 danieli: got a coredump and here is the backtrace https://0x0.st/zmA2.txt 2019-05-19 10:47:29 i'm not against making alpine (DE) easy install, but I'm also to retain it's first characteristic, i.e. "Small" 2019-05-19 10:47:42 mps: hrm. maybe it's a config option then, I'll experiment 2019-05-19 10:48:47 wow it is I moved my config and ssh works fine, thanks mps lets see if I can figure out which option it is 2019-05-19 10:48:54 xcko: and I use alpine on different arm boxes with different kernels, mostly upstream, and never had problem you describe 2019-05-19 10:49:06 yeah very weird I agree 2019-05-19 10:49:39 xcko: if you found bug please report it to bugs.a.o 2019-05-19 10:54:54 something to do with 'VerifyHostKeyDNS yes' yeah I'll report a bug. Thanks for your help mps (and danieli) 2019-05-19 10:55:24 sorry, I was a little busy - and thanks, that'd be great. I'll check it out asap 2019-05-19 10:55:33 mind posting the ticket ID when it's made? 2019-05-19 10:56:20 yeah np I'm gonna see if I can isolate the bug further for reproducing, then I'll make a report 2019-05-19 11:27:54 danieli: bug #10467 2019-05-19 11:30:27 do bugs get assigned to Copa by default? I didn't intend to do that 2019-05-19 11:31:01 fixed 2019-05-19 11:33:24 it's a bug with redmine apparently, i don't know why it happens 2019-05-19 11:33:29 and thanks, i'll have a look 2019-05-19 11:48:17 it is 'VerifyHostKeyDNS' and not 'HostKeyVerifyDNS' as is in bug report 2019-05-19 11:49:10 and, yes it segfaults 2019-05-19 12:00:58 xcko: danieli: here is little about ssh segfault https://bridge.grumpy-troll.org/2018/03/alpine-arm/ 2019-05-19 12:01:17 author thinks it is upstream bug 2019-05-19 12:01:26 so it's a known issue that's been there for a long time 2019-05-19 12:01:44 woops sorry for the typo 2019-05-19 12:01:46 looks like it is 2019-05-19 12:02:00 openssh-server is version 7.9 in alpine 3.9 2019-05-19 12:02:05 and 8.0 in edge 2019-05-19 12:04:00 heh, it is time to look at openwrt openssh pkg 2019-05-19 12:10:14 on that grumpy-troll link I also had a 700 / and needed to chmod 755 / - no idea how that happened, surprised to see someone else get hit by that 2019-05-19 14:06:23 xcko: danieli: I found one calloc function in openssh which segfault when the VerifyHostKeyDNS is used 2019-05-19 14:07:34 and found this https://github.com/openwrt/packages/issues/2059 but according to this it is fixed, and it is not, obviously 2019-05-19 14:36:02 Literally seen just a minute ago: https://twitter.com/NicoSchottelius/status/1130119696387379207 2019-05-19 14:36:28 eyyyy 2019-05-19 15:20:22 Hey how can I make a bootable iso from the minimal fs download 2019-05-19 15:42:35 has anyone tried running alpine on espresso bin? 2019-05-19 15:42:54 I am soon giving an IPv6 workshop and I am considering to use alpine + espresso bin as a tool for the workshop 2019-05-19 16:12:53 i hope ipv6 will pick up when i retire. cant even read one ipv6 address. 2019-05-19 16:13:05 even localhost 2019-05-19 16:13:12 yeah im dumb 2019-05-19 16:19:11 move to programming :-D 2019-05-19 16:20:34 <_ikke_> 8::! 2019-05-19 16:20:37 <_ikke_> ::1 2019-05-19 16:20:39 <_ikke_> localhost 2019-05-19 16:22:17 tmhoang: it's as easy as "::" 2019-05-19 16:22:38 which is muuuuuuuuuch shorter than ipv4 ;-) 2019-05-19 20:40:37 What I don't understand about IPv6 is why should :: represent localhost? 2019-05-19 20:47:32 <_ikke_> It's shorthand for 0000:0000:0000:0000:0000:0000:0000:0001 2019-05-19 20:49:00 <_ikke_> But that's a bit much to type each time :-) 2019-05-19 20:49:31 yea that seems a little inefficient, but allows for a lot of addresses I guess 2019-05-19 20:51:26 So can I just run mkisofs on the extracted alpine minimal filesystem download to create a bootable iso? 2019-05-19 20:51:40 I need a method that can be completely automated 2019-05-19 20:52:43 <_ikke_> aports contains a script to create iso's 2019-05-19 20:53:08 <_ikke_> https://git.alpinelinux.org/aports/tree/scripts/. 2019-05-19 20:53:12 <_ikke_> https://git.alpinelinux.org/aports/tree/scripts/ 2019-05-19 20:55:27 cool so I can alter what I need of the minimal fs and then run mkimage.sh probably? 2019-05-19 20:56:00 <_ikke_> I've never used it myself, so I don't know exactly how you should use it 2019-05-19 20:56:14 <_ikke_> https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage 2019-05-19 20:56:25 sweet thank you 2019-05-19 23:00:53 Just to clarify: :: is the IPv6 version of 0.0.0.0, and ::1 is localhost 2019-05-19 23:01:37 http://teachmeipv6.com/IPv6-Essentials-Cheat-Sheet.pdf 2019-05-19 23:07:32 That's a nice sheet 👍 2019-05-19 23:40:25 See that makes more sense to me 2019-05-19 23:40:46 :: being shorthand of 0000:0000:0000:0000:0000:0000:0000:0000 and ::1 being 0000:0000:0000:0000:0000:0000:0000:0001 2019-05-19 23:41:23 each : being 0000:0000:0000:0000 ? 2019-05-19 23:41:42 no 2019-05-19 23:42:08 :: is a shorthand for "fill in as many 0000 blocks as possible" 2019-05-19 23:43:05 you can have it anywhere in an ipv6 address, but only once 2019-05-19 23:43:40 2::1 would be 0002:0000:0000:0000:0000:0000:0000:0001, for example 2019-05-19 23:48:36 ahhh gotcha 2019-05-19 23:49:01 So I'm guessing this requires alpine to run this howto since it uses apk: https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage 2019-05-19 23:49:15 is alpine-sdk available for other distro's? 2019-05-19 23:49:43 alpine-sdk build-base apk-tools alpine-conf 2019-05-19 23:50:32 are the only requirements that I can see may not be available in other distros? I'm guessing build-base might correlate to build-essential in other distros? 2019-05-20 00:19:04 hey, does anyone know how to download packages from another computer and install them via say a usb key 2019-05-20 00:19:08 i need this for drivers 2019-05-20 00:20:32 i can't find a download link on the alpine package repo site 2019-05-20 00:20:51 apk add --allow-untrusted 2019-05-20 00:21:04 ah 2019-05-20 00:21:14 and i found the link on the site, i think 2019-05-20 00:21:27 you just click the commit hash and then the download link right? 2019-05-20 00:21:50 curl / wget the apk from a mirror, e.g. dl-cdn.alpinelinux.org/alpine/ then what p4Wv1qn095FW said 2019-05-20 00:32:02 @p4Wv1qn095FW, this can be run from the live USB? 2019-05-20 01:01:32 there is apk fetch too if you have an internet connected alpine linux system 2019-05-20 01:02:38 http://tpaste.us/jX9v 2019-05-20 01:06:17 This could be very handy for something that has a lot of dependencies. For example: apk fetch -o /mnt/thumb_drive -R bash vim less git-perl 2019-05-20 01:06:35 Otherwise, you need to figure out which packages you need and fetch them yourself. 2019-05-20 01:35:58 is there a way to list only installed packages with apk? 2019-05-20 01:36:28 apk list -I 2019-05-20 01:36:33 apk list -I 2019-05-20 01:37:23 thanks 2019-05-20 02:08:57 Hmmm running into issues building anything but min fs 2019-05-20 02:09:27 oh wait, I'm on the virtual thought not standard 2019-05-20 02:10:51 is there a way to install everything from standard if I installed from virtual image 2019-05-20 02:11:46 or better yet, just include a specific home directory in the min fs build 2019-05-20 02:24:25 Ahhh, it actually builds from the apk repos and not from anything on the filesystem except etc I believe 2019-05-20 02:24:28 for min 2019-05-20 03:37:32 the minrootfs build creates a tar.gz, how can I get it to create an iso? 2019-05-20 04:08:06 no mount -o loop ? 2019-05-20 04:11:40 mkisofs ? 2019-05-20 04:51:55 Is there a way I can create my own bootable iso with my own selection of apks or patches? Links to useful documentation would also be appreciated. Thank you 2019-05-20 04:52:32 <_ikke_> https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage 2019-05-20 04:53:57 @_ikke_ Thank you kindly 2019-05-20 11:18:39 Is there a way to go from 3.9.4 back to 3.9.2 ? probably not ?!? 2019-05-20 11:22:14 klesk: no 2019-05-20 11:23:05 ok, thanks 2019-05-20 14:56:08 good day 2019-05-20 14:56:22 In case you haven't seen it already, we just posted a list of distributions w/o sytemd: https://ungleich.ch/en-us/cms/blog/2019/05/20/linux-distros-without-systemd/ (also on HN/reddit) 2019-05-20 15:23:06 Hi all, is there a way to make a persistent vlan with a custom interface name? 2019-05-20 15:24:06 eKKiM: custom interface name, as in completely arbitrary name? 2019-05-20 15:24:16 Yes 2019-05-20 15:24:20 "# ip link add name lan-voip link lan type vlan id 10" does the trick but doesn't survive reboots 2019-05-20 15:24:52 persistent vlan, sure, but i don't know about arbitrary interface name with busybox ifconfig 2019-05-20 15:25:15 eKKiM: add it in pre-up or post-up in 'interfacees' 2019-05-20 15:25:30 that could work 2019-05-20 15:25:44 Hmm that could work indeed 2019-05-20 15:26:03 heh, that works ;) 2019-05-20 15:26:38 i have that for some interfaces 2019-05-20 15:31:44 "# ip link add name lan-voip link lan type vlan id 10" does the trick but doesn't survive rebootsroot 2019-05-20 15:31:50 woops sry 2019-05-20 15:41:33 I have issues with the combination of nameif 2019-05-20 15:41:52 nameif: can't change ifname lan-test to lan: File exists 2019-05-20 15:42:28 nameif works only on boot when mdev take action, iirc 2019-05-20 15:43:34 i added "nameif -s" to /etc/network/if-pre-up.d/nameif 2019-05-20 15:43:44 probably a bad idea? 2019-05-20 15:47:42 But this seems the way to go. Tyvm for the tip! 2019-05-20 16:27:46 I created a file /etc/ld-musl-x86_64.path with a bunch of paths I need ldconfig to crawl, and ran "/sbin/ldconfig", but it doesn't appear to have read that file. Am I missing something obvious? 2019-05-20 16:34:09 TML: ldconfig is shell script, you can add paths there if you need 2019-05-20 17:01:23 mps: So it is - is that the best practice, though? 2019-05-20 17:05:31 love how all the options to ldconfig are no-ops o.O 2019-05-20 17:06:09 TML: I'm not sure, but musl is different than glibc based systems, and probably it will work. you can add lib paths on cli when invoking ldconfig 2019-05-20 17:09:07 sure, I get that it's different - just trying to learn the musl tools 2019-05-20 17:11:22 "scanelf -qS -f /etc/ld-musl-x86_64.path" seems to have worked 2019-05-20 17:18:02 Hmm – so why isn't there an Abuild for the scanelf package? 2019-05-20 17:20:06 scanelf is a subpkg of pax-utils 2019-05-20 17:20:12 TML: because it's built by a difffernet APKBUILD 2019-05-20 17:20:44 ah - I have not encountered that before, didn't realize a given APKBUILD could make multiple packages 2019-05-20 17:20:47 thanks 2019-05-20 17:21:16 TML: a differnt example are APKBUILDs for libraries, which produce the library itself, and the development headers 2019-05-20 17:21:26 and sometimes included tools 2019-05-20 17:21:30 makes sense 2019-05-20 17:22:07 is there a tool of some sort that will tell me which APKBUILD generated a given package? "apk info" didn't seem to have any hints 2019-05-20 17:22:26 (pkgs.alpinelinux.org conveniently has a subpackages box on the packages details page) 2019-05-20 17:23:15 TML: apk info --who-owns 2019-05-20 17:24:04 c705: I tried that, it's not quite what I'm asking for, though/usr/bin/scanelf is owned by scanelf-1.2.3-r0 2019-05-20 17:24:24 jn__: oh, I see an "Origin" entry there - that will work 2019-05-20 17:40:47 installed fresh alpine, then 'nginx' and 'nginx-mod-http-lua' packages. nginx fails to start because it wants the "Resty" version? (cannot find certain shared libraries); any ideas? is the package failing to pull dependencies? 2019-05-20 17:42:35 logs: http://paste.debian.net/plainh/a5783a8b 2019-05-20 17:50:34 Thalheim: try apk add lua-resty-http? 2019-05-20 17:50:42 Or apk search *resty* 2019-05-20 17:59:59 c705: no luck; installed all of those 2019-05-20 18:03:08 greetings, I'm trying to debug an openrc file: https://hastebin.com/iqebihayam.makefile ; On restart, nomad seems to detatch and not actually be killed. I have to manually pkill the stragglers to complete the restart 2019-05-20 18:04:02 <_ikke_> maldridge: Is nomad actually starting on the foreground, or is it double forkin itself (to daemonize) 2019-05-20 18:04:41 it starts in the foreground, attatched to a terminal 2019-05-20 18:05:09 <_ikke_> And if you kill it, the other processes are gone too? 2019-05-20 18:05:24 sudo pkill nomad does remove the ghost processes 2019-05-20 18:16:04 I suspect that something obvious is wrong with my file, but I lack the experience with openrc to see it 2019-05-20 18:16:40 <_ikke_> maldridge: I suspect that process doing something strange 2019-05-20 18:16:50 <_ikke_> normally killing the parent should also kill the children 2019-05-20 18:17:23 and I think it should, but nomad defers the kill until it can clean up references, so I think the restart is happening before that can finish and the process is deadlocking 2019-05-20 19:22:45 Someone has an explanation for this: https://serverfault.com/questions/968136/arbitrary-vlan-interface-name-undocumented-configuration 2019-05-20 19:22:52 ? 2019-05-20 19:27:03 it's a bit different with busybox 2019-05-20 20:49:54 Hey when trying to make a custom standard iso and adding openjdk8 I get this on script run: openjdk8-jre: unable to select package (or its dependencies) 2019-05-20 20:51:24 I'm guessing it's using main instead of community repo but how do I get apk to pull from both when running this custom mkimg.$PROFILE.sh 2019-05-20 20:52:18 in genrootfs it looks like it uses the system repo's on whatever it is building on 2019-05-20 20:52:25 repositories_file=/etc/apk/repositories 2019-05-20 20:56:00 ooh nope it's only fetching from main for apk adds: fetch http://dl-cdn.alpinelinux.org/alpine/v3.9/main/x86_64/APKINDEX.tar.gz 2019-05-20 20:59:43 *sigh* i just found polkitd running consuming 1GB of ram 2019-05-20 20:59:51 <_ikke_> yikes 2019-05-20 20:59:55 is there any way the polkit dep could be extricated from alpine? 2019-05-20 20:59:58 polkit is pure evil 2019-05-20 21:00:14 only consolekit2 and xfce-power-manager depend on it 2019-05-20 21:00:29 (and xfce depends on xfce-power-manager despite me not caring about having power-manager) 2019-05-20 21:00:44 if consolekit2 could be fixed not to need polkit, nobody would need polkit installed 2019-05-20 21:00:59 and we could rid ourselves of yet another fdo/poettering-fans hellware package 2019-05-20 21:03:20 dalias: today I talked on #alpine-devel to make 'slim' DM without consolekit/polkit/mozjs dependencies 2019-05-20 21:03:33 DM? 2019-05-20 21:03:41 display manager? 2019-05-20 21:03:43 Display Manager 2019-05-20 21:03:54 i'm not even using any DM, just startx 2019-05-20 21:03:59 dalias: since xfce4 is just a metapackage you could just install its dependencies sans xfce4-power-manager 2019-05-20 21:04:07 DM is like 300 megs of wasted ram to replace /bin/login 2019-05-20 21:04:09 ahhh nvm on my question... --extra-repository 2019-05-20 21:04:17 eu, yeah, i know 2019-05-20 21:04:23 that's why consolekit2 was the one i focused on 2019-05-20 21:04:24 ah, so startx depends on polkit, or some xfce pkg's 2019-05-20 21:04:26 xfce is just an annoyance 2019-05-20 21:04:35 but would be nice with a sane default for those that do not know dalias :) 2019-05-20 21:04:40 mps, no, consolekit2 depends on it 2019-05-20 21:04:56 yeap and that worked perfect 2019-05-20 21:05:05 and bluez-libs (*libs* only, even) depends on consolekit2 2019-05-20 21:05:22 and lots of stuff that has capacity to use bt stuff depends on bluez-libs 2019-05-20 21:05:28 even if you don't have bluez installed 2019-05-20 21:05:34 well, I've built slim without consolekit deps, and it works quite fine, no polkit 2019-05-20 21:06:06 now I just need to figure out how to include a specific user and their home directory in the iso 2019-05-20 21:06:55 consolekit depends on polkit, and polkit depends on mozjs. Javascript for system service !? 2019-05-20 21:07:04 AppXprt: Did you folow the wiki guide and write your own mkimg.$PROFILENAME.sh script? I'm curious because I'm about to attempt doing the same thing myself 2019-05-20 21:07:11 yes 2019-05-20 21:07:22 it works well once you figure everything out how it's working 2019-05-20 21:07:28 AppXprt: Would you mind sharing with me? I could use an example 2019-05-20 21:08:02 what are you trying to build? standard or what? 2019-05-20 21:08:19 and what output format do you want? presumably iso? 2019-05-20 21:08:40 AppXprt: I think so. Standard + a few extra apks & kern mods > iso 2019-05-20 21:09:07 ok so go take your mkimg.standard.sh and copy it to a new one 2019-05-20 21:09:16 I see the example on the wiki, but it's not 100% clear to me yet whats going on 2019-05-20 21:09:29 yea the documentation is lacking severely 2019-05-20 21:09:39 then edit it and change the function names to your custom profile name 2019-05-20 21:09:42 same as $PROFILENAME 2019-05-20 21:09:44 AppXprt: So basically I just patch mkimage.standard.sh and roll with that? 2019-05-20 21:09:50 well sorta 2019-05-20 21:10:14 Or do I copy it to mkimg.$PROFILENAME.sh and edit 2019-05-20 21:10:17 you don't execute those scripts 2019-05-20 21:10:29 mkimg.customprofile.sh and edit 2019-05-20 21:10:43 but you don't execute those scripts 2019-05-20 21:10:44 they won't do anything 2019-05-20 21:10:48 you execute like this: 2019-05-20 21:10:54 Ok, thanks. I'll figure it out after some hacking 2019-05-20 21:11:01 sh mkimage.sh --arch x86_64 --outdir /root/iso/ --repository http://dl-cdn.alpinelinux.org/alpine/v3.9/main/ --profile appxprte --extra-repository http://dl-cdn.alpinelinux.org/alpin 2019-05-20 21:11:01 e/v3.9/community/ 2019-05-20 21:11:36 in the standard profile there are other defined too like extended 2019-05-20 21:11:49 you'll see where the apk's are being added in there 2019-05-20 21:12:08 you can do the same for standard, but I just went with altering the extended profile in the standard profile file 2019-05-20 21:12:41 but the extended profile and virt profile will call standard_profile at the top of those functions, so you have to change that to whatever you change your standard profile function name to 2019-05-20 21:13:01 mps, polkit used to be a rootkit (backdoors to let random desktop shit run commands as root) using an XML-based DSL 2019-05-20 21:13:03 profile_appxprt() { 2019-05-20 21:13:10 they decided the XML-based DSL was too hideous 2019-05-20 21:13:16 profile_appxprte() { 2019-05-20 21:13:16 profile_appxprt 2019-05-20 21:13:16 so now it's a rootkit based on javascript 2019-05-20 21:13:30 profile appxprt extended (appxprte) calls appxprt (standard) 2019-05-20 21:14:03 i don't even know how polkitd got running -- what starts it? 2019-05-20 21:14:07 that would be profile_standard in the unaltered version 2019-05-20 21:14:29 dalias: Most likely dbus 2019-05-20 21:14:30 uhg some dbus crap i think 2019-05-20 21:14:31 AppXprt: I don't care for extended, just need to add ansible,python,cryptsetup to the standard image to 2019-05-20 21:15:26 So I'll use the example on the wiki, and call profile_standard at the top of my function, along with the other stuff 2019-05-20 21:15:32 dalias: s/crap/helpfully starting the applications required for my stuff to work/ 2019-05-20 21:16:00 no, starting stuff i absolutely DO NOT WANT 2019-05-20 21:16:14 there is also something about copying genapkovl-dhcp.sh to your own matching your custom profile as an apk overlay, but I just ended up adding my apk's in the custom profile like they do 2019-05-20 21:16:46 also I made some minor changes / additions to genrootfs.sh to include the community repo 2019-05-20 21:16:59 what i would want, if some application actually depended on this crap being running 2019-05-20 21:17:09 AppXprt: I'll figure it out I'm sure. Thak you for your help 2019-05-20 21:17:11 is a hard error message "Y can't run without daemon X running" 2019-05-20 21:17:27 dalias: Then make your stuff not depend on polkit/blacklist the polkit dbus services/remove that service 2019-05-20 21:17:31 then i could make an educated choice whether i want to run X, patch Y not to need X, or just rm Y for being awful crap that needs X 2019-05-20 21:17:38 yea np, I like this way it's just took a little understanding of how they were doing things 2019-05-20 21:17:43 how do you blacklist dbus services? 2019-05-20 21:17:48 can i just blacklist * ? :) 2019-05-20 21:17:51 Which like 1/1000th of all users want? 2019-05-20 21:17:51 I'll be here if you need any help with building iso 2019-05-20 21:17:59 because i really don't want dbus starting *anything* 2019-05-20 21:18:15 AppXprt: Yeah, ability to custom iso's is a huge pullze piece I've been missing. My goal is mininmal image managed by Ansible 2019-05-20 21:18:57 nice, well keep in mind that the minimal fs profile outputs a .tar.gz intsead of an iso, but I think you could combine that profile with standard to get min fs iso 2019-05-20 21:19:15 AppXprt: My thoughts exactly 2019-05-20 21:20:08 dalias: maybe don't use 'Desktop Environment' but some simple Window manager 2019-05-20 21:20:08 let me know if you do this because I might go that route, but as it stands now I made a 450MB iso that should fit my needs for an android build ENV... I'm curious if it would all work ok on min fs though 2019-05-20 21:20:57 mps, that's not the problem, and i'm not using any heavy DE crap 2019-05-20 21:21:04 AppXprt: I run from RAM - the idea is I boot system and then whatever environment I want, I do ansible-playbook dev_env.yml...or something 2019-05-20 21:21:10 ahhh wait I gotta add nodejs, npm and cordova... now that's a problem... nodejs and npm is no problem as it's provided by apk, but cordova has to be installed via npm intsall -g cordova@8.1.2 2019-05-20 21:21:17 I use Docker often these days if I'm testing 2019-05-20 21:21:39 dalias: you mentioned xfce so I thought you use it as DE 2019-05-20 21:21:45 I may switch to Docker instead of custom ESXi infrastructure automation 2019-05-20 21:21:53 Docker has changed my life in such a good way, I can never go back to VMs 2019-05-20 21:22:17 It's like God mode for creating static environments 2019-05-20 21:22:27 mps, i use the panel and wm but i don't use any DE features of it 2019-05-20 21:22:31 I wonder how to do a post install through another package manager before creating iso, like through npm? 2019-05-20 21:22:39 like control over device mounting, etc. 2019-05-20 21:23:16 I'll check it out, I'm building this custom iso for an automated android build environment 2019-05-20 21:23:18 but the fundamental problem here is dbus being able to start stuff i don't want 2019-05-20 21:23:29 I do everything 'by hand' from terminal, udevil for mounting devices 2019-05-20 21:23:35 which could happen whenever installing a package without heavy auditing 2019-05-20 21:23:50 I like the idea of being able to even build the ISO for each env on the fly if I can automate it smoothly 2019-05-20 21:23:56 so i'd rather just block it globally 2019-05-20 21:24:19 rm -rf /etc/dbus-1/* && chattr +i /etc/dbus-1 would work.... 2019-05-20 21:24:31 imagine, cups and gimp depends on dbus 2019-05-20 21:24:31 but it'd give lots of spurious apk errors 2019-05-20 21:24:37 i don't use cups 2019-05-20 21:24:58 and i do use gimp but purely as an image editor with no integration (and don't want integration) with anything else 2019-05-20 21:25:03 <_ikke_> Why does gimp depend on dbus? 2019-05-20 21:25:08 i don't want it trying to talk to other things i have running 2019-05-20 21:25:16 I don't think you'd want to delete all of those files but only the service's dir 2019-05-20 21:25:19 IDK I love the power behind ESXi though 2019-05-20 21:25:32 cogitri, well delete all services and prevent addition of any new ones 2019-05-20 21:25:53 but i'd rather just a clean config that can be added on top to block loading of services 2019-05-20 21:25:58 yes, I think it is enough to disable dbus start script 2019-05-20 21:25:59 Yes, but doesn't /etc/dbus-1 contain some more config? 2019-05-20 21:26:12 mps, but then stuff spams gratuitous errors that dbus isn't running 2019-05-20 21:26:22 even when it doesn't need to do anything with it 2019-05-20 21:26:31 mps: If nothing uses dbus for ipc, yes 2019-05-20 21:26:37 AppXprt: I'm planning on writing a Dockerfile that does just that. Stay tuned, I'll share when I'm finished 2019-05-20 21:27:01 note: i don't want any ipc happening between desktop apps, at all 2019-05-20 21:27:10 dalias: that is second issue, and pkg's should be built without dbus requirements, but some packages can't be built without dbus 2019-05-20 21:27:17 *nod* 2019-05-20 21:27:38 My idea is docker run .. -e PROFILE=/path/to/profile.sh docker-make-alpine-iso 2019-05-20 21:27:38 if i wanted OLE/COM hell i'd be using windows 2019-05-20 21:28:16 (and i wouldn't have written musl and alpine would still be using uclibc and ... :-P) 2019-05-20 21:29:27 dalias: I'm on your side of POV. but distro's trying to be more user friendly and less 'power users friendly' 2019-05-20 21:29:50 i don't see how this makes anything "user friendly" 2019-05-20 21:29:59 dalias: Are you Rich Felker? 2019-05-20 21:30:02 yes 2019-05-20 21:30:04 sent you a message... also my image is booting with the installed apk's 2019-05-20 21:30:22 dalias: I want to personally thank you for writing musl 2019-05-20 21:30:34 there are great ways to do friendly desktop behavior without dbus/ipc/giving desktop apps root backdoors 2019-05-20 21:30:39 one of reasons for me to switch to alpine from debian was 'Debian dependency hell' 2019-05-20 21:30:53 c705, thanks :) 2019-05-20 21:31:29 like for removable devices, just plain automount and a having the file manager inotify on /media... 2019-05-20 21:32:05 (that's one example; pretty much all the other stuff has nice ways to do it too that don't massively violate unix access control model) 2019-05-20 21:33:01 I'm trying wherever and whenever I can to make package to have less dependencies 2019-05-20 21:33:08 thanks :) 2019-05-20 21:33:51 i do like devel libs being built with full deps tho or having a way to get them all 2019-05-20 21:34:08 some python lib packages are missing some optional features iirc and makes it so you can't run python apps that need them 2019-05-20 21:34:29 (this is different from deps that impose running a daemon as root by default as soon as you install them) 2019-05-20 21:35:00 during last stable release we didn't upgrade polkit because it started to depends on mozjs, but looks like for new release it will be upgraded. I'm not happy about that but accepted that as something which is inevitable 2019-05-20 21:44:55 _ikke_: on https://www.gimp.org/source/ dbus is not mentioned 2019-05-20 21:45:50 mps: Yup, the newer polkit is required for lots of stuff :/ 2019-05-20 21:46:16 so my method only adds the packages to the cd iso, what about installing them for availability on boot of the iso? 2019-05-20 21:46:23 BTW, is there a reason we disable nls on so many pkgs? I mean Alpine tries to be minimal, but isn't that too much since we already have -lang packages? 2019-05-20 21:47:30 Cogitri: I know, I know, we discussed that for the previous stable release (v3.9) but then decided to no upgrade in hope that maybe someone will make patch for it to not depend on mozjs 2019-05-20 21:48:10 That'd require a rework of like half of polkit and every package that install polkit rules 2019-05-20 21:48:27 And doesn't really have any perspective of being adopted upstream, so yeah 2019-05-20 21:49:07 I have then looked at some requests to remove mozjs deps but polkit devs didn't cared about these requests 2019-05-20 21:49:49 Yup :c 2019-05-20 21:50:04 Well, I'm not quite sure what would've been a bettet choice (lua?l 2019-05-20 21:50:36 imo, lua will be better 2019-05-20 21:51:39 _ikke_: just rebuilt gimp without dbus deps 2019-05-20 21:52:01 so, looks like it is not needed for gimp 2019-05-20 21:52:57 will look later or tomorrow if it works without dbus 2019-05-20 21:54:23 https://unix.stackexchange.com/questions/370895/install-packages-into-a-custom-alpine-linux-iso 2019-05-20 21:54:40 sorry wrong window 2019-05-20 21:59:20 why does the genapkovl scripts need a hostname? 2019-05-20 22:09:44 the overlay works if you add it with rootfs_apks="" or maybe it was just the rootfs_apks that did it 2019-05-20 23:52:37 Hey everyone, I'm testing a new alpine installation on an old laptop of mine 2019-05-20 23:53:36 I'm trying to set some passwords, but the password strength meter is on and won't let me use some easy to remember ones 2019-05-20 23:53:45 How can I turn it off? 2019-05-20 23:55:40 for root? 2019-05-20 23:56:06 iirc if you repeat the same rejected password again it will accept it or something; i forget 2019-05-20 23:56:50 imo there's little sense in enforcing login password strength unless physical attack is part of your threat model or you intend to allow remote password-based login (this is 2019, remote should be ssh pubkey ONLY!) 2019-05-21 00:01:51 Hah, yep that worked great thanks. 2019-05-21 00:03:56 I agree, very few people are going to try to physically brute force your local account 2019-05-21 00:04:47 And it doesn't even matter because you could boot externally and then chroot in anyway 2019-05-21 00:09:22 right 2019-05-21 00:09:35 it really just matters if you allow remote login using password auth 2019-05-21 00:09:44 or su with password 2019-05-21 00:26:53 anyone know how to add a specific user to remastered iso? 2019-05-21 00:40:39 and default network conf 2019-05-21 01:25:31 Thalheim: that nginx / openresty confusion might be caused by pulling resty modules from GitHub while building nginx 2019-05-21 01:45:17 tcely: yes, probably, but it suggests that this has not been tested 2019-05-21 01:45:33 which, when upgrades break a system, is ... frustrating 2019-05-21 01:47:28 Thalheim: indeed. Did you file an issue for this on bugs.a.o? 2019-05-21 02:14:29 tcely: I can, but tbh it'd be less of a waste of time to fix the package myself and share it, rather than file a report and wait for someone else to do so 2019-05-21 04:54:14 Hey, if you want to report and fix. I won't say no. ;-) 2019-05-21 05:24:04 so when building custom iso's, how can I add a specific user and their home directory + files, and also need to execute a script on every boot 2019-05-21 05:24:37 or install an npm package into the iso filesystem like is done with the apk's 2019-05-21 05:30:34 ahhh damn he left, I figured out c705's problem 2019-05-21 05:36:10 waha I figured out some stuff but still need help 2019-05-21 05:36:12 anyone? 2019-05-21 05:36:58 <_ikke_> maybe some kind of post install script? 2019-05-21 05:38:20 yea how can I do that 2019-05-21 05:38:40 or rather where should I place it 2019-05-21 05:39:49 <_ikke_> Not sure if it's the best way, but one option would be to place it in /etc/local.d and to enable the local service 2019-05-21 05:39:54 <_ikke_> the script would remove itself once run 2019-05-21 05:42:01 ok cool I'll try that 2019-05-21 07:03:05 the boot speed of alpine is pretty similar to tiny core linux 2019-05-21 08:26:05 who cares 2019-05-21 08:30:36 tmhoang: cares those who uses short lived VM's and containers 2019-05-21 08:31:05 <_ikke_> containers don't boot :) 2019-05-21 08:31:09 <_ikke_> but it counts for vms 2019-05-21 08:32:39 _ikke_: strictly speaking it is true, but containers also can start some system daemons and this could be important 2019-05-21 08:33:11 <_ikke_> Sure, but that's more user space where alpine has less influence 2019-05-21 08:33:18 <_ikke_> Or the specific distro 2019-05-21 08:33:20 you cannot do much about a proces starting 2019-05-21 08:34:59 again true, but compare speed of systemd and openrc boot speed 2019-05-21 08:35:30 you are running openrc in a container? 2019-05-21 08:35:38 <_ikke_> We are :-) 2019-05-21 08:35:39 ofc :) 2019-05-21 08:36:06 lol 2019-05-21 08:36:29 that what happens when you work too much on a docker project.. 2019-05-21 08:36:50 <_ikke_> hehe 2019-05-21 08:37:06 ACTION hides back in his docker cave 2019-05-21 08:39:12 I'm moving some containers from debian to alpine, and speed of starting is quite impressive. also stopping 2019-05-21 09:09:13 that's actually exactly why I care... short lived vm's 2019-05-21 09:10:35 specifically, automated build environments... see why it matters? 2019-05-21 09:11:24 <_ikke_> A lot of people say, what does it matter, we have enough disk space and ram 2019-05-21 09:11:35 <_ikke_> but they forget these kinds of workloads 2019-05-21 09:11:38 <_ikke_> which really add up 2019-05-21 09:12:10 that kind of attitude will end up with node_modules... 2019-05-21 09:13:00 _ikke_: you can post that pick again now ;-) 2019-05-21 09:13:20 s/pick/pic 2019-05-21 09:13:57 <_ikke_> http://devhumor.com/content/uploads/images/July2017/npm_package.png 2019-05-21 09:13:59 <_ikke_> this one? 2019-05-21 09:15:17 https://cdn-images-1.medium.com/max/1600/1*tLmdca91O3WXf_ZTiy8JFA.jpeg 2019-05-21 09:16:00 <_ikke_> ah yeah, that one as well 2019-05-21 09:16:31 but similar :) 2019-05-21 09:20:12 if you wanted to even build custom iso's on the fly it could be automated and run pretty quickly actually... I've ran through a ton of iso builds tonight and each one takes less than 60 seconds each 2019-05-21 09:20:44 okay maybe under 90 sec... 2019-05-21 09:21:21 it all depends on what you're doing, the image you select and size of everything + your internet speed for downloading apk's 2019-05-21 09:21:38 speed of CPU's, amount of RAM 2019-05-21 09:21:51 shit... bus speed 2019-05-21 09:22:09 even 2019-05-21 10:13:32 Hi all. Hoping someone can help. I tried to install alpine today behind our corporate firewall, however "apk" doesn't want know about it. I set the proxy details and they are working fine, tested via wget downloading the APKINDEX.tar.gz "apk update" just sits there for ages saying - fetch http://dl-3.../APKINDEX.tar.gz and eventually returns with ..permission denied ...No such file or directory 2019-05-21 10:19:30 you can wget the APKINDEX? 2019-05-21 10:20:20 <_ikke_> magoo: How did you set the proxy? 2019-05-21 10:20:30 yes wget works fine 2019-05-21 10:20:40 <_ikke_> magoo: https://github.com/gliderlabs/docker-alpine/issues/171 ? 2019-05-21 10:21:39 saw that ticket... I tried every combination. with and without the http:// 2019-05-21 10:23:05 like I mentioned wget works fine... if I change my username or password and try to wget the apkindex it will fail with proxy authentication 2019-05-21 10:23:18 <_ikke_> apk is using libfetch 2019-05-21 10:24:14 I saw someone mention libfetch and to set the HTTP_PROXY_AUTH... which I also tried. 2019-05-21 10:24:41 <_ikke_> https://git.alpinelinux.org/alpine-conf/tree/setup-proxy.in 2019-05-21 10:26:02 <_ikke_> That seems to suggest setting http(s)_proxy should be enough 2019-05-21 10:27:20 I thought so too, but just does not work... I used the setup-proxy and saw it wrote those details to /etc/profile.d/proxy.sh which I sourced. 2019-05-21 10:27:57 <_ikke_> Looking at https://git.alpinelinux.org/apk-tools/tree/libfetch/http.c now 2019-05-21 10:28:46 I work for supporting people to run containers but I never use containers myself. too old for that 2019-05-21 10:29:13 Someone mentioned I would have to find it that some function resets the http_proxy when using apk... not sure how true it is 2019-05-21 10:31:47 not sure if its related https://bugs.alpinelinux.org/issues/6219 2019-05-21 10:37:07 tunnel it? 2019-05-21 10:39:21 tunnel it to where? 2019-05-21 10:41:44 through the proxy? 2019-05-21 10:43:01 instead of using the apk proxy settings / functions, send the traffic through a tunnel utilizing the proxy instead 2019-05-21 10:43:29 It is a while since I looked at Alpine and I wish to know what the current status is for zfs on raspberry pi3 (aarch64). Is it possible to create a ZFS root and run everything from ZFS? Is the ZFS module ready compiled or is it built upon install? 2019-05-21 10:43:56 how would I go about that? 2019-05-21 10:44:22 via cntlm? 2019-05-21 10:44:28 first you would setup a tunnel to use the proxy and test it and then you have to apk traffic to go over that tunnel 2019-05-21 10:44:34 engblom: i have no clue how it works on raspberry pis, but zfs-vanilla is built for aarch64 (64-bit image you can run on a RPi 3) 2019-05-21 10:45:02 IDK it's been forever since I've setup and used tunnel, but should work, they would know better / quicker in a networking or the official linux chan 2019-05-21 10:45:23 also, it's precompiled against the current kernel version in the branch you're on (3.8, 3.9, edge, etc) 2019-05-21 10:46:11 I only have control over our server, so if it has to be done on the network side first via IT, well its not going to happen :/ 2019-05-21 10:46:22 danieli: Thanks! 2019-05-21 10:46:26 nah you can do this all on the server 2019-05-21 10:46:46 if your server can use the proxy then a tunnel on your server can too..... 2019-05-21 10:47:19 engblom: for what it's worth, raspberry pi's bootloader can only boot off fat32, but you could probably load zfs and pivot to it later in the boot process 2019-05-21 10:47:31 *probably* 2019-05-21 10:47:34 any links to references or which software on alpine I can use? 2019-05-21 10:48:02 is it a socks proxy? 2019-05-21 10:48:08 https://www.google.com/search?ei=NNfjXP7vIu2OggfM_amoBQ&q=linux+tunnel+all+traffic+through+proxy&oq=linux+tunnel+all+traffic+through+proxy&gs_l=psy-ab.3..0i22i30.10943.12996..13259...0.0..0.63.393.7......0....1..gws-wiz.......0i71j35i39.Oao88v8CyPI 2019-05-21 10:48:09 no 2019-05-21 10:48:18 danieli: I am looking at the download page and I notice the generic arm uses uboot. If rpi images would also use uboot, it would be possible to have /boot on zfs too. 2019-05-21 10:48:43 engblom: you'd have to use the rpi image, the generic arm image will not work on raspberry pi 2019-05-21 10:48:49 you sure? 2019-05-21 10:49:05 magoo: https://www.linuxjournal.com/content/use-ssh-create-http-proxy 2019-05-21 10:49:06 yes 2019-05-21 10:49:19 engblom: rpi uses some custom binary bootloader, afaik 2019-05-21 10:50:08 https://www.ssh.com/ssh/tunneling/example 2019-05-21 10:50:10 mps: I know, but some distros have created a uboot system that boots from the fat32 partition and then chained with a normal boot loader like grub which then boots from a normal linux partition 2019-05-21 10:50:53 mps: This removes the burden of keeping the fat32 parition updated with current kernel etc. 2019-05-21 10:51:02 that will slowdown boot and complicate it 2019-05-21 10:51:35 mps: Not much as uboot is not big. It is just to tell the system were the real bootloader is. 2019-05-21 10:51:40 and maybe make it more fragile 2019-05-21 10:51:41 https://daniel.haxx.se/docs/sshproxy.html 2019-05-21 10:51:56 you'll have to combine some methods 2019-05-21 10:52:05 I don't think that ssh tunnel will work as thats normally used from one machine through to the next? 2019-05-21 10:52:59 engblom: so, some distro's boot rpi with u-boot as first stage bootloader, iiuc 2019-05-21 10:53:33 should be a way to use a tunnel through the proxy that lets you pass apk traffic through it ie through the proxy as well 2019-05-21 10:53:36 mps: Yes, and it is actually less fragile as then there needs to not be any extra scripts handling the kernel upgrade for rpi. 2019-05-21 10:54:15 you could always setup a machine on the outside as the endpoint, but probably breaking your enterprise policy 2019-05-21 10:54:35 if the u-boot can be used as first stage then it could load kernel and rpi binary is not needed at all? 2019-05-21 10:55:45 yeh definitely cant do that, I'd have no job left :) 2019-05-21 10:56:02 do you have link or hint where i can look for rpi u-boot info 2019-05-21 10:56:32 so figure out a way to tunnel traffic to the proxy 2019-05-21 10:56:58 the proxy is a machine with an endpoint that you can connect to and pass traffic to..... 2019-05-21 10:57:07 all you need really, figure out a way to utilize it 2019-05-21 10:57:18 errr with a gateway 2019-05-21 10:57:30 I will see if I can... I wonder why apk isn't using the proxy details :/ 2019-05-21 10:57:30 it is an endpoint 2019-05-21 10:58:20 <_ikke_> magoo: did you actually verify it's not using the proxy? 2019-05-21 10:58:33 read this: https://unix.stackexchange.com/questions/10428/simple-way-to-create-a-tunnel-from-one-local-port-to-another 2019-05-21 10:58:54 no I just assumed that was the case... is there a way to check? 2019-05-21 10:59:15 <_ikke_> strace / tcpdump I guess 2019-05-21 10:59:39 netstat -t -n 2019-05-21 10:59:54 while using apk 2019-05-21 11:00:05 mps: I do not know how uboot is exactly loaded in those distros, but I know they can handle the kernel like on any other system during upgrade and that that /boot does not need to be on fat32 anymore. Somehow uboot + maybe something else is on the fat32 which then chains to the normal boot loader used by that particular distro. 2019-05-21 11:00:40 ok I will try tomorrow 2019-05-21 11:01:03 mps: Now without uboot + normal boot loader, the fat32 partition need to be handled by special scripts during each kernel upgrade, which makes it more fragile. 2019-05-21 11:01:21 engblom: RPI firmware (on SOC) looks for fat partition and binary file with hardcoded name on it 2019-05-21 11:02:14 mps: Yes, and there uboot comes in and then loads the real boot loader (grub/lilo/syslinux/whatever) and the kernel from the real /boot (on ext4/xfs/zfs/whatever) 2019-05-21 11:03:03 so it is a chain but this chain simplyfies it as the system will look the same regardless if it is pc or arm or any special case like rpi. 2019-05-21 11:03:08 look here: https://www.raspberrypi.org/documentation/hardware/raspberrypi/bootmodes/bootflow.md 2019-05-21 11:03:37 engblom: u-boot is _real_ bootloader 2019-05-21 11:04:02 to boot linux you don't need anything else 2019-05-21 11:07:36 mps: By chaining uboot with grub you can have the kernel on a better filesystem than fat32. You can build up normal linux filesystem hiarchy with kernel. 2019-05-21 11:08:19 with u-boot I have kernel on ext4, no FAT's 2019-05-21 11:08:48 Does u-boot have support for zfs? 2019-05-21 11:08:58 u-boot supports more FS's, waiting for f2fs 2019-05-21 11:09:32 engblom: I suspect it support zfs, but ext2/3/4 are there 2019-05-21 11:09:33 While I really hate grub (and most stuff that comes from GNU), it got the advantage that it supports almost any file system including zfs. 2019-05-21 11:10:16 aha, you want to have zfs on rpi, interesting 2019-05-21 11:10:46 Yes, I would want a zfs root system with just enough stuff on the fat32 to chain the boot into the zfs system. 2019-05-21 11:11:04 fat32 does not allow me to have snapshots for the kernel. 2019-05-21 11:11:47 I have used zfs both with debian and gentoo on rpi. 2019-05-21 11:12:09 And I am now trying to figure out how it would be possible to do with alpine. 2019-05-21 11:12:55 and what FS is your boot partition on these distro's 2019-05-21 11:12:56 The 1Gb of RAM is enough for zfs as long as you just run what rpi is intended for (small programs doing IO and some small services). 2019-05-21 11:13:32 mps: Some of them require me to put the kernel on fat32, some of the systems I have tried will chain uboot with grub. 2019-05-21 11:14:00 All of them have the fat32 system, but not all of them have the kernel there nor the config files for what arguments should be called. 2019-05-21 11:14:37 I think SuSE was the first to chain uboot + grub. But they do not have zfs support. 2019-05-21 11:15:16 On debian and gentoo the zfs module is compiled on each upgrade, which is a bit too much for an rpi. 2019-05-21 11:16:48 Ubuntu is also using this uboot -> grub -> loading kernel chain. 2019-05-21 11:17:10 And Fedora too 2019-05-21 11:19:15 I don't understand, sorry. they don't have FAT partition and uses u-boot directly from the rpi firmware? 2019-05-21 11:20:00 They have the FAT partition with uboot. This uboot then boots the normal grub (the same they use for the rest of the arm platforms, unifying the boot system) 2019-05-21 11:22:24 ah, so. I understood you that you want to remove FAT part from rpi media 2019-05-21 11:23:03 so, FAT must stay there as first thing and rpi binary bootloader? right? 2019-05-21 11:24:17 auto eth0 \n iface eth0 inet dhcp should bring up network with dhcp on boot right? 2019-05-21 11:24:27 in interfaces file 2019-05-21 11:25:02 Yes, fat is needed (because the rpi firmware is hardcoded to boot from there). But if you have uboot, you do not need to have the kernel anymore on the FAT partition, instead you can have a normal boot loader supporting several kernel versions and a real /boot on any file system (even zfs). 2019-05-21 11:27:06 anyone smarter than me able to build this on alpine https://github.com/tmate-io/tmate-ssh-server 2019-05-21 11:28:02 engblom: can't follow all this because I actually never worked with rpi's, only 'normal' ARM boxes 2019-05-21 11:28:22 mps: https://www.suse.com/media/article/UEFI_on_Top_of_U-Boot.pdf 2019-05-21 11:29:15 I work with u-boot long time and I know how it could be chained to different 'things' 2019-05-21 11:36:18 mps: I think the full chain is uboot (from FAT) -> efi parition containing grub -> grub. After this the boot is similar for all platforms and the kernels are in /boot. Thus the FAT parition can remain untouched during system upgrade (as no config files there need to be changed for each upgrade nor any kernel files needs to be placed there) 2019-05-21 11:37:27 mps: The the real linux system and upgrade process/scripts can be identical for all arm platforms 2019-05-21 11:38:13 One would not need to detect if the upgrade process is running on rpi and then do exceptional handling of the special FAT partition and the stuff there. 2019-05-21 11:38:53 and /boot can be a directory on /. 2019-05-21 11:46:18 engblom: does rpi have limitation on boot partition size 2019-05-21 11:47:00 mps: I do not know. I am not an expert on rpi boot, I have just noticed a few things while trying several rpi linux distros. 2019-05-21 11:47:57 mps: All I want is a system allowing me to easy update everything (including kernel) with just the package manager and that everything is running on zfs (also the kernel). 2019-05-21 11:49:03 I cannot help here because I do not have rpi's. I prefer more open boards 2019-05-21 11:49:41 mps: I have achieved this with some other distros, but they have not had binary zfs modules, so the kernel update takes almost a day compiling the zfs and spi modules during 'apt-get dist-upgrade' or whatever package manager those distros are using. 2019-05-21 11:50:09 Also the compiling is very memory intensive and 1Gb of ram is not really enough for that. 2019-05-21 11:51:37 well, alpine needs less memory than other (glibc) distros 2019-05-21 11:52:26 Yes, and I have been playing with alpine on amd64 some years ago and I liked it. I would happily run it on rpi too, especially if it is possible to install to an zfs root. 2019-05-21 11:52:33 I was surprised when run first time alpine on arm boxes 2019-05-21 11:53:26 before that I had debian and raspbian. what a difference ;) 2019-05-21 11:53:46 don't worry, it will all be over soon and your linux installations are replaced by nodejs based operating systems :D ... 2019-05-21 11:53:55 I have several times been tempted to switch my normal desktop to alpine, but alpine is missing several really important packages. 2019-05-21 11:53:59 (dear God no) 2019-05-21 11:58:39 (the packages missing are at least the following: thunderbird, openscad-nightly, sbcl, wxmaxima and probably a few more) 2019-05-21 12:03:30 i switched to alpine on all my workstations (x86_64, arm64 and arm32) and members of my family 2019-05-21 12:31:35 engblom: request them on bugs.a.o and someone might just package them 2019-05-21 12:52:24 there's a lot of stuff missing that i'd need to use alpine as a daily driver too 2019-05-21 12:52:42 i packaged about half of it in a private aports repo, some of them would never make it upstream 2019-05-21 12:52:53 others have been too much effort, more than i cared to put into it 2019-05-21 14:02:29 Hi everyone. I installed git and tried to add hunks to the stage interactively, but this does not work anyhow: "git add -p" results in "git: 'add--interactive' is not a git command. See 'git --help'." 2019-05-21 14:02:37 Am I missing a neccessary package? 2019-05-21 14:15:31 Nevermind, it's apparently a part of git-perl... Not the most obvious name, though^^ 2019-05-21 14:55:12 misterunknown: based on the description that name seems entirely reasonable 2019-05-21 14:55:32 What would be a better name? 2019-05-21 15:27:16 tcely: The description is indeed correct, but the "reverse search" is not so easy if you don't know that this particular subcommand depends on perl. Maybe the description could contain the names of the three features which the package contains, but doesn't have to. I found the solution finally via pkgs.alpinelinux.org in a reasonable time. 2019-05-21 15:31:25 misterunknown: would a install_if="git perl" help in your case? 2019-05-21 15:39:37 https://github.com/alpinelinux/aports/pull/7982 2019-05-21 17:26:01 If I change the visudo file, is there a way to make the new config current without a reboot? 2019-05-21 17:28:39 I don't think that ever needs a reboot 2019-05-21 17:29:06 You have to log out and log in if you've only just joined the group that you've added to the sudoers file 2019-05-21 17:33:30 yes, needs a relog for the group to take effect 2019-05-21 17:33:42 changes to /etc/sudoers should be instantaneous 2019-05-21 17:50:28 Cogitri, danieli, ah yes... an option I didn't consider. Thanks :) 2019-05-21 17:50:42 :) 2019-05-21 18:27:32 Okay so I was able to build a successful live android / cordova build environment that includes bash, sudo, nano, node, npm, openJDK, Android Studio and Cordova... Studio and Cordova are stored in ISO only so doesn't load to ram during live session... boots in like 8 seconds and runs stable at 456MB Ram out of 4GB... 2019-05-21 18:28:11 just wanted to see if it could be done and how small 2019-05-21 18:28:17 Neat 2019-05-21 18:33:22 aports mkimage and the profiles are awesome, but a little confusing and could use a little more documentation so maybe I can contribute when I have some time 2019-05-21 18:36:57 AppXprt: nice work 2019-05-21 18:38:54 only thing is I have to access cordova through /media/cdrom/ so I don't load to RAM by copying everything to live filesystem, but should be alright since this is going to be ISO's on a fs instead of a physical optical media, right? 2019-05-21 18:39:24 will you share your work somewhere 2019-05-21 18:39:40 IDK it's just an experiment and not sure about the legality of it 2019-05-21 18:40:01 storing Android Studio in an ISO? 2019-05-21 18:40:37 other than that, I wouldn't mind 2019-05-21 18:41:10 ok, however you think is appropriate. but I think you can describe or put script/s without problem, but IANAL so don't know at the end 2019-05-21 18:41:52 I think everything else other than that and maybe cordova should be fine to store and user could always download and store Android Studio because it gets pulled from user home dir on ISO build machine 2019-05-21 18:42:21 I wouldn't mind sharing in that way at all and letting user build Android Studio into it themselves 2019-05-21 18:42:58 Honestly I thought I would get some backlash / ridicule for this 2019-05-21 18:43:56 ok, np, however you think 2019-05-21 19:01:09 mps you really need this? 2019-05-21 19:02:17 no, but what you done could be useful for someone with similar needs 2019-05-21 19:02:48 LoL 2019-05-21 19:02:59 and don't worry about backlash 2019-05-21 19:04:31 for example, short guide or script how to setup android develepment with cordova on alpine could be useful for androd devs which uses alpine 2019-05-21 19:05:28 you can even create wiki page on wiki.alpinelinux.org 2019-05-21 19:06:58 is android studio open source yet? 2019-05-21 19:07:55 that reminds me that we need to upgrade android-tools 2019-05-21 19:32:27 yea I see what you're saying, I would like to contribute to to Alpine in any way possible 2019-05-21 19:32:42 I like what you all have done for sure 2019-05-21 19:36:29 AppXprt: that is :) 2019-05-21 20:17:28 Did something change in mpv packaging recently? My syncplay stopped working. 2019-05-21 20:19:05 <_ikke_> It did have some upgrades 2019-05-21 20:19:20 <_ikke_> Are you on edge? 2019-05-21 20:19:32 tcely: from yesterday, if you're interested in a follow-up https://bugs.alpinelinux.org/issues/10478 2019-05-21 20:20:23 _ikke_: Yeah, edge x86_64 2019-05-21 20:22:27 <_ikke_> SInce the beginning of this year, there were 2 changes. One is a rebuild against ffmpeg 4.1.1, and the other as to enable wayland 2019-05-21 20:23:08 Hmm, any major version bump that would change the accepted options or the control protocol? 2019-05-21 20:23:45 <_ikke_> On october from 0.27.2 to 0.29.1 2019-05-21 20:24:11 Syncplay is unfortunately rather unhelpful and doesn't show me the error message, just says the mpv is too old. But as I looked in the code it does that universally when the output is not what it expects. 2019-05-21 20:24:49 Hmm, I believe I've used it since then. Weird. 2019-05-21 20:24:52 <_ikke_> CcxWrk: One think you can try is to build on older version of the package and see if that would fix it? 2019-05-21 20:27:45 I guess. Is there a command for pulling all the build deps? 2019-05-21 20:28:35 <_ikke_> abuild can build the entire package in one go, if you have setup your system for it 2019-05-21 20:28:39 ^ 2019-05-21 20:28:53 or `abuild deps` only for the build deps, if you're in the package's directory in the aports tree 2019-05-21 20:33:49 Thalheim: thanks 2019-05-21 20:42:20 pr7848 2019-05-21 20:44:08 ah nice 2019-05-21 21:07:26 _ikke_: Managed to dig out the actual error and it seems that --script=... option is no longer recongized. Was Lua support removed or something? I see nothing relevant in upstream changelog. 2019-05-21 21:11:03 The manpage does have the option but actually using it gives me: Error parsing option script (option not found) 2019-05-21 21:18:14 Also urxvt started crashing on me: panic: locale.c: 893: Unexpected character in locale name '2E. 2019-05-21 21:18:49 yeeeah, i noticed that too 2019-05-21 21:18:53 apparently it's a perl bug with locales 2019-05-21 21:18:56 Apparently it doesn't like the locale.encoding syntax for LANG anymore. LANG=C is fine but LANG=C.UTF-8 is not 2019-05-21 21:19:16 yeah, perl changed, so we have to recompile perl and everything built against it 2019-05-21 21:19:23 it has some really screwy code for parsing locale strings 2019-05-21 22:45:00 Trying to build custom iso via wiki post: https://pastebin.com/raw/peEhgC7m I think im pissing package on build machine, any ideas? 2019-05-21 22:45:10 Err..missing* 2019-05-21 22:46:31 Build profile: https://pastebin.com/raw/xTHhBPDM 2019-05-21 22:49:16 fwiw s390-tools is only built for s390x, not ppc64le 2019-05-21 22:49:28 (it's probably not relevant, just thought i'd point it out) 2019-05-21 22:50:14 judging by the output there, it sounds like openssl isn't getting a file argument it expects 2019-05-21 22:50:35 Indeed 2019-05-21 22:50:51 right, happens when it tries to sign it 2019-05-21 22:51:03 i'll skim through the source real quick 2019-05-21 22:51:07 The only dir left in temp after crash is mkimage.x 2019-05-21 22:51:40 >> mkimage-x86_64: Building custom 2019-05-21 22:51:40 >> mkimage-x86_64: --> kernel x86_64 2019-05-21 22:51:47 happens, runs, then crash 2019-05-21 22:53:25 i see what happens, the build scripts are calling update-kernel from the alpine-conf package, which does the signing 2019-05-21 22:54:53 line 5-17 https://git.alpinelinux.org/aports/tree/scripts/mkimg.base.sh#n5 and line 183-196 + line 315-318 in https://git.alpinelinux.org/alpine-conf/tree/update-kernel.in#n183 2019-05-21 22:55:03 not sure why the file it expects isn't being passed though 2019-05-21 22:55:19 Let me apk update on my machine before build. Im using Docker alpine image to build 2019-05-21 22:57:22 i'm not sure if it does it automatically, but perhaps you could test it: do you have the abuild environment variables set? 2019-05-21 22:57:35 look at that line 183 in the second file 2019-05-21 22:57:46 (i might fall asleep soon, but someone else can probably help you :) 2019-05-21 22:58:44 Thank you for that - this explains the problem 2019-05-21 23:01:21 Im trying to automate the build process so I can build as non-root user 2019-05-21 23:02:01 Problem now is that it wants to copy key into /etc/apk/keys which requires root 2019-05-21 23:04:02 OK, I have derpy hack - let us try that 2019-05-21 23:23:13 Ah, progress! Thank you for your help danieli 2019-05-22 00:02:09 i did apk update and then apk upgrade, the result was just OK 926 MiB in 149 packages. did it update anything? 2019-05-22 00:02:41 cim2095: That means nothing was upgraded 2019-05-22 00:03:11 ah got it 2019-05-22 00:36:12 that would make sense 2019-05-22 00:36:33 cause alpine is the host OS for docker and there's only docker installed 2019-05-22 00:36:46 everything else is inside my docker images 2019-05-22 00:57:21 Is the virtuhardened kernrel as used with alpine 3.7 abandoned and no longer updated? 2019-05-22 00:57:35 4.9.65-1-virthardened #2-Alpine SMP Mon Nov 27 15:53:30 GMT 2017 2019-05-22 01:04:14 ah https://git.alpinelinux.org/aports/commit/?id=94a65a421705eb0152c2a6cdeb0bffd269c58e97 2019-05-22 01:04:19 Well then. 2019-05-22 01:05:42 JTL: grsecurity put up a payment wall for the patches 2019-05-22 01:05:47 I'm aware 2019-05-22 01:05:58 Moving over to regular kernel now 2019-05-22 01:06:37 (This was just a external infrastucture box used for external monitoring my home/datacenter servers, and I haven't touched it in a while) 2019-05-22 01:06:39 THanks 2019-05-22 03:07:25 so i can't seem to download ovmf 2019-05-22 03:07:36 with a simple `apk add` like the xen dom0 wiki article states 2019-05-22 03:07:49 its in a different repo, i think? how do you access a repo besides world 2019-05-22 03:08:54 <_ikke_> jacob____: you need to enable the community repp in /etc/apk/repositories 2019-05-22 03:09:14 thanks 2019-05-22 03:09:37 <_ikke_> It's only available for x68_64 though 2019-05-22 03:09:46 that's fine, that's my platform 2019-05-22 03:10:52 On the topic, when may one expect to see packages move from testing to community? 2019-05-22 03:11:17 should i make a wiki edit to warn people that they will need to add the community repo first on the xen dom0 guide? 2019-05-22 03:13:23 <_ikke_> jacob____: Yes, feel free 2019-05-22 03:13:46 <_ikke_> When that wiki page was written, community most likely did not exist yet 2019-05-22 03:15:26 <_ikke_> c705: that's up to the maintainer. In principle packages don't need to be in testing for a long period, but it should be at least verified that the package is properly working 2019-05-22 03:42:41 did anyone ever get powershell workingon alpine? 2019-05-22 03:43:18 I only see docker packages for it 2019-05-22 03:47:46 ok, so i've rebooted 2019-05-22 03:47:56 how do i swap grub for syslinux 2019-05-22 03:48:02 i apk del'd grub 2019-05-22 03:48:31 and i still have grubx64.efi and the grub folder in /boot 2019-05-22 03:48:46 <_ikke_> those are not managed by a package 2019-05-22 03:49:16 is there a program or script to uninstall them 2019-05-22 03:49:23 <_ikke_> rm 2019-05-22 03:49:28 hmmmmmm 2019-05-22 03:51:06 <_ikke_> Then you need to make sure you properly install syslinux 2019-05-22 03:52:01 any idea on coreclr / dotnet core? 2019-05-22 03:52:06 "temporary error (try again later)" 2019-05-22 03:52:07 uh oh 2019-05-22 03:58:14 okay now whenever i install anything i get temporary error (try again later) 2019-05-22 04:00:17 nvm, network 2019-05-22 04:07:42 isn't it a bit of a problem if the installer installs grub2 by default but xen requires syslinux? 2019-05-22 04:37:38 AppXprt: I have plans about making aports for dotnet core and ps 2019-05-22 04:40:05 PLS... when? 2019-05-22 04:50:16 soonish .... :P 2019-05-22 05:18:56 Please oh please 2019-05-22 05:20:05 what all has to be done? adapt from glibc to musl? 2019-05-22 05:33:38 I have to reluctantly switch to tinycore until dotnet and ps work on alpine 2019-05-22 06:27:25 tcely: Well, this would be really nice, thank you! (Sorry for the late reply) 2019-05-22 06:29:33 re: pr7982 you are welcome 2019-05-22 08:06:52 tcely: with this change to git-perl Alpine starting to 'feel' like Debian to some degree :) 2019-05-22 08:07:33 mps: why? 2019-05-22 08:07:58 'Debian dependency hell' 2019-05-22 08:08:18 i dont follow 2019-05-22 08:09:29 it is term in debian community which describes when one simple package pulls a lot of other packages 2019-05-22 08:09:50 i dont understand how that relates to this change. 2019-05-22 08:10:23 did you notice ';)' at the end of my message 2019-05-22 08:10:52 no i only saw":)" 2019-05-22 08:10:53 I wrote it as small sarcasm 2019-05-22 08:11:51 meh, I'm just drinking first morning coffee, and shouldn't write anything in this state of brain :) 2019-05-22 08:12:18 yes, it was ':)' and not ';)' 2019-05-22 08:12:21 it makes sense to pull in git-perl when git and perl are installed. 2019-05-22 08:12:57 I don't mind, and in most cases it is ok, I agree 2019-05-22 08:12:58 we dont want to pull in perl when installing git, but when you explicitly or by dep pull in perl why not add perl features to git? 2019-05-22 08:13:46 and, I do that by hand wherever install git 2019-05-22 09:58:57 Hello! What is the ABI level for Alpine S390x? z196 or is it newer? 2019-05-22 10:04:15 tmhoang: ^ 2019-05-22 11:01:33 blueCmd: z196, correct 2019-05-22 11:01:42 z196 or newer 2019-05-22 11:01:55 I should update things to newer cpu in new gcc8 ... but I'm dumb 2019-05-22 11:03:32 all update to gcc 9.1 ! 2019-05-22 11:56:13 tmhoang: I have a z114 so I prefer if people keep it at z196 :) 2019-05-22 11:56:41 but thanks, I'll add that to a blog post I'm writing 2019-05-22 11:57:34 blueCmd: sooner or later the base would be z12 instead of z196 2019-05-22 11:57:55 any specific reason why? 2019-05-22 11:58:09 is there any particular instruction set that is needed for something? 2019-05-22 11:58:58 should be but I have to check 2019-05-22 11:59:07 I mean, z114 is only 5-6 years old - it sucks a bit to have ABI change that fast 2019-05-22 12:22:50 blueCmd: sorry I was afk. I think z12 would enjoy support time as much as z196. 2019-05-22 12:23:03 so nothing special 2019-05-22 12:23:27 <[0xAF]> Hi there. I'm stuck with one SBC with Vortex86 DX2 CPU. This CPU has sort of i686, but is missing CMOV instruction, so everything has to be compiled for i585 at max. So far the Alpine was the only distro which I was able to run out-of-the-box (i.e. w/o compiling everything from sources). I'm using Alpine edge i386 and I need a recent browser - FF or Chrome, preferably Chrome, but it's missing in the repo. Is there a reason for that? 2019-05-22 12:45:57 [0xAF]: you're looking for chromium, not chrome 2019-05-22 12:50:15 <[0xAF]> danieli: Yeah, that's what I meant, sorry for typing Chrome instead of Chromium. But still Firefox and Chromium are missing from the i386 repo. 2019-05-22 12:54:35 [0xAF]: firefox is missing because rust is missing for x86 (i3860 2019-05-22 12:57:39 [0xAF]: looks like chromium is built only for x86_64, armv7, and aarch64 2019-05-22 12:59:32 <[0xAF]> mps: thanks for the info. 2019-05-22 12:59:54 <[0xAF]> danieli: yeah, hence I'm asking my question... 2019-05-22 13:00:43 there is a massive patchset for chromium, i'm guessing it simply refuses to build on x86 and that fixing the build has been too much work 2019-05-22 13:01:52 chromuim is hard to maintain especially on musl based systems 2019-05-22 13:02:21 I tried few months ago but give up 2019-05-22 13:02:38 blueCmd: out of curiosity, are you Alpine running in z/VM or LPAR ? 2019-05-22 13:02:42 <[0xAF]> huh, I see... 2019-05-22 13:03:12 yup, maintaining browsers is just a world of hurt 2019-05-22 13:03:22 i don't blame people for not putting much effort into it 2019-05-22 13:03:42 FF is pretty good 2019-05-22 13:03:50 <[0xAF]> Any suggestion for a browser (in the i386 repo), with some recent JS engine/support? I need to run a VueJS based app in it. 2019-05-22 13:03:51 well firefox is not hard 2019-05-22 13:04:06 firefox used to be decent, but not anymore due to rust 2019-05-22 13:04:21 only missing rust part and rust is not easy 2019-05-22 13:04:25 to be fair, rust is the problem there, not firefox itself 2019-05-22 13:04:33 danieli: true 2019-05-22 13:04:52 <[0xAF]> danieli: FF would be quite good candidate, I'm developing on it, but I didn't know that rust part is messing up the i386 support... 2019-05-22 13:05:10 [0xAF]: you're mistaken, i wasn't referring to i386 support 2019-05-22 13:05:19 rust is a blocker for firefox in general, not just for specific architectures 2019-05-22 13:05:33 [0xAF]: rust upstream don't care for musl port 2019-05-22 13:05:46 <[0xAF]> danieli: ahh, i see... 2019-05-22 13:05:56 i've hardly touched i386 in 10+ years 2019-05-22 13:06:04 perhaps even more than that 2019-05-22 13:07:14 <[0xAF]> danieli: I'm stuck with x86 32bit SBC... even worse, it's missing some instrunctions to be able to run i686 code ... I dislike that too, but ... no choice for me... 2019-05-22 13:08:12 <[0xAF]> so, FF now uses rust for it's internals or only for some peripherial features? can it be compiled w/o the rust stuff? 2019-05-22 13:08:17 internals 2019-05-22 13:08:21 it can not 2019-05-22 13:08:21 <[0xAF]> i see 2019-05-22 13:08:54 it's used for both the browser engine and other features integral to firefox 2019-05-22 13:09:34 danieli: Rust actually isn't that hard, but right now Rust PRs have been waiting for some months to be merged 2019-05-22 13:09:49 fair enoguh, that's no big surprise to me either 2019-05-22 13:09:52 enough* 2019-05-22 13:10:18 So yeah, I won't plunge effort into that until the other stuff is merged 2019-05-22 13:10:35 why has it been waiting for that long to be merged? 2019-05-22 13:10:41 i'm sure there's a good reason 2019-05-22 13:10:50 <[0xAF]> Hmm, no luck for me then... I'll be AFK for 10 mins, in case someone come with an idea for recent browser on i386... 2019-05-22 13:11:25 Cogitri: did you tried to apply PR and build rust and did you had success? I tried but without luck 2019-05-22 13:11:29 danieli: Because they've waited for the maintainer to review it, but he never did (due to inactivity or smth) 2019-05-22 13:11:53 Cogitri: okay, I gotta say, that sounds a bit ridiculous to me 2019-05-22 13:12:07 or does nobody else want to be the maintainer? 2019-05-22 13:13:47 I'd take over maintainership and bring it up to speed, but right now I'm waiting for 30 other PRs to be merged, so motivation is a bit low rn 2019-05-22 13:14:06 yeah.. i feel the same way 2019-05-22 13:14:24 I will work on some commits after I eat my lunch 2019-05-22 13:16:07 Thank you :) 2019-05-22 13:16:38 Didn't mean to sound offensive there, I just feel like Alpine's review policy is a bit too restrictive for its manpower 2019-05-22 13:18:27 tmhoang: LPAR, I am not running it yet though 2019-05-22 13:19:07 Cogitri: no worries, i'm just getting jaded 2019-05-22 13:19:23 tmhoang: the reason I bring it up is that it does not really compute to me. for all other CPU platforms I have not seen this kind of rushing to upgrade -march as s90x seems to have, and I am not sure why 2019-05-22 13:20:08 imagine if you would require -march=some_new_intel_cpu, you would not unless you were a very specialized OS 2019-05-22 13:20:25 and z196 I understand as that gives you rust and golang, but why z12? 2019-05-22 13:21:34 I am writing a blog post about the state of zLinux for hobbyists and I would like to understand the motivating factor, if there is one 2019-05-22 13:28:55 blueCmd: running LPAR would be a little bit of issue (incorrecting tty to have output at booting). Im trying to make that happen in this 3.10 release 2019-05-22 13:28:58 z/VM would be fine 2019-05-22 13:30:16 I'm not saying we are going to eol z196 soon. What I'm saying what z196 has support time, z12 would be the same 2019-05-22 13:30:50 so Alpine is not doing something crazy. it just follows what IBM intends to do on other supported distros (RHEL SUSE Ubuntu) 2019-05-22 13:32:30 I'm not sure about rust golang arguments you mentioned though 2019-05-22 13:32:50 I can comeback to you about this discussion later, currently on some training ... 2019-05-22 13:44:37 <[0xAF]> Thanks for the info guys. It seems I'm going to try to compile everything from source with glibc and try to compile chromium. I'm heading to Gentoo again... 2019-05-22 15:25:22 the busybox xargs doesn't support -L1, is there any alternative, thx 2019-05-22 15:25:41 docker images --format "{{.Repository}}:{{.Tag}}" | grep -v "" | sort | uniq | xargs -L1 docker pull 2019-05-22 15:26:05 https://pkgs.alpinelinux.org/contents?file=xargs&path=&name=&branch=edge&arch=x86_64 2019-05-22 15:30:32 thanks TBK[m], however i got error with: # apk add xargs 2019-05-22 15:30:33 ERROR: unsatisfiable constraints: 2019-05-22 15:30:35 xargs (missing): 2019-05-22 15:30:37 required by: world[xargs] 2019-05-22 15:33:18 f# apk add wget 2019-05-22 15:33:20 OK: 1251 MiB in 359 packages 2019-05-22 15:41:56 i see, it is in findutils pkg, thanks! 2019-05-22 15:48:24 appll: np 2019-05-22 16:01:58 so I can run glibc programs in a chroot? 2019-05-22 16:03:38 AppXprt: better yet in lxc container 2019-05-22 16:04:01 although you can run them in chroot also 2019-05-22 16:04:51 I extracted a tinycore iso to alpine, but when I chroot to it, I can't use the tce package manager even after getting network / dns operational 2019-05-22 16:05:49 I run debian without problem 2019-05-22 16:06:01 maybe I am running the chroot wrong 2019-05-22 16:06:34 do you mount /proc /sys and /dev under chroot dir 2019-05-22 16:07:18 oh no I forgot that part 2019-05-22 16:07:58 been a while since I've used chroot really 2019-05-22 16:10:26 Heh 2019-05-22 16:13:45 proc, sys, dev? 2019-05-22 16:14:04 still not working, using: for i in proc sys dev; do mount -o bind $i /folder/$i; done 2019-05-22 16:14:27 /folder being the chroot dir 2019-05-22 16:14:43 mount -o bind /$i 2019-05-22 16:15:18 proc and bind -o, no. mount -t proc none /edge/proc 2019-05-22 16:15:48 egde is your chroot dir 2019-05-22 16:16:06 mm 2019-05-22 16:16:14 different types 2019-05-22 16:17:18 proc as proc, sys as sysfs, dev as bind 2019-05-22 16:18:19 here is one of my chroot scripts http://tpaste.us/mZXL 2019-05-22 16:19:17 this one is for alpine chroot, but have somewhere similar for debian 2019-05-22 16:26:14 what does this line do exactly? mount -o bind /home/mps/sdcard/sdcard/aports edge/home/mps/aports 2019-05-22 16:26:26 link old system aports to chroot system? 2019-05-22 16:26:58 i'm pretty sure you can omit that one 2019-05-22 16:31:06 yes, this is my path to bind mount aports three under chroot, you don't need it 2019-05-22 16:31:16 argh still no tce-load with your script mps 2019-05-22 16:31:40 I think it's a tinycore issue from it not being "booted" ? 2019-05-22 16:31:48 /usr/bin/tce-setup: line 8: can't open /etc/sysconfig/tcuser: no such file 2019-05-22 16:32:00 I'll check with them, thanks for the help 2019-05-22 16:33:47 uhm, I don't have any experience with tinycore, but similar scripts work for different distro in my experience 2019-05-22 16:35:16 yea it works after creating those missing directories the tce-setup was reporting 2019-05-22 16:37:02 nice to hear :) 2019-05-22 17:55:56 damn I need powershell so bad on alpine so I don't have to use tinycore 2019-05-22 17:56:28 what is powershell? windows one? 2019-05-22 17:57:11 yes, it's a microsoft thing 2019-05-22 17:57:23 they open sourced it and made it run on debian a while back, runs on a couple more distros now 2019-05-22 17:57:35 it runs on a lot of distro's now 2019-05-22 17:57:48 I can get it working in tiny, but would rather use alpine 2019-05-22 17:57:59 i think i have a half-finished apkbuild somewhere 2019-05-22 17:58:02 it's for VM automation, esxi / xen 2019-05-22 17:58:12 powershell is quite versatile 2019-05-22 17:58:14 or what I'm using for I mean 2019-05-22 17:58:19 figured as much 2019-05-22 17:58:29 considered looking at libvirt? 2019-05-22 17:58:45 I'll read about it right now 2019-05-22 17:59:23 Hmmm yea that would work, but probably more than I need really 2019-05-22 18:00:35 ACTION will not use ms 2019-05-22 18:01:07 ok ok I will try libvirt 2019-05-22 18:01:11 :-D 2019-05-22 18:01:30 something else to learn 2019-05-22 18:01:39 zsh is quite powerful 2019-05-22 18:02:10 mps: the use case here is different 2019-05-22 18:02:23 although I use tcsh for interactive work and sh (ash) for scripts 2019-05-22 18:03:11 danieli: understand, but I express my strong preferences and opinion 2019-05-22 18:03:46 fair enough 2019-05-22 18:03:52 for what it's worth, it's MIT licensed 2019-05-22 18:05:05 anything from company known for it's not friendly behavior to free source should be escaped 2019-05-22 18:05:10 agreed 2019-05-22 18:05:14 here is what I'm doing, I'm exporting build packages and need to build them, but want to use immutable infrastructure so boot -> build -> die... I don't want to use containers because of security, I want the builds fully separated... I need a REST API that accepts the configs (node.js Restify) then builds the alpine ISO with the exported build config files included, spins up the build VM on either ESXi or XEN with the ISO atta 2019-05-22 18:05:14 ched (powershell / libvirt / script), boot -> build -> die 2019-05-22 18:05:22 i'm not super strict about it now though 2019-05-22 18:05:42 it just isn't a realistic point of view in the industries i do work 2019-05-22 18:05:46 MS has been doing a lot better under satya nadella 2019-05-22 18:05:57 A LOT of open source 2019-05-22 18:06:06 AppXprt: new tactics only, imo 2019-05-22 18:06:21 better, yes, but i'm not sure i'd call it great 2019-05-22 18:06:31 Extend, Embrace, Extinguish 2019-05-22 18:06:31 perhaps not even good with everything taken into account 2019-05-22 18:06:33 IDK, they have really been trying to change their attitude I think 2019-05-22 18:06:45 embracing linux much more 2019-05-22 18:06:56 and all of OSS 2019-05-22 18:07:23 any idea's on the technique / method I should be using? 2019-05-22 18:07:34 i'm not in the mood to die on this hill right now, but normally, i'd go on a long and fiery rant 2019-05-22 18:07:50 I mean I am open to being 100% FOSS 2019-05-22 18:08:04 I'm very open to idea's and opinions 2019-05-22 18:08:54 I have the alpine iso customization and iso build taken care of now, working on API and controlling the infrastructure 2019-05-22 18:09:16 If they want to be FOSS they should release source of their OS first. All other is fake 2019-05-22 18:09:20 I already have a powershell script that buils ESXi VM's based on ZERO user input 2019-05-22 18:10:08 mps: can we please refrain from fighting a holy war? 2019-05-22 18:10:16 which is why I was going with powershell, but it only took a night to make that script and my team members want me to switch to XEN XCP instead of ESXI and I assume I should switch to libvirt if I do that as well 2019-05-22 18:10:36 AppXprt: you can use some of the true programming langs (scripting) if you need some deep customization, python perl etc. 2019-05-22 18:10:49 I know python pretty well 2019-05-22 18:10:52 danieli: agree, and sorry 2019-05-22 18:10:54 "I need help with X" "but have you tried Y/Z"? 2019-05-22 18:11:06 this is pure CLI setup, no GUI's 2019-05-22 18:11:12 s/"?/?"/ 2019-05-22 18:12:03 mps: you don't happen to know a good way to convert a variable to lowercase only with posix sh? 2019-05-22 18:12:14 it's fairly hard to find stuff that isn't bash-specific online 2019-05-22 18:12:28 I'm open to building multiple ways of doing it and find out which is the quickest... Powershell build them VM's fairly quickly on linux dotnet core, ash is building the alpine iso through mkimg profiles 2019-05-22 18:12:42 the* 2019-05-22 18:14:00 danieli: convert the contents ? 2019-05-22 18:14:05 north1: ? 2019-05-22 18:14:11 oh, yes 2019-05-22 18:14:17 something something substitution 2019-05-22 18:14:23 convert the contents of the variable or the variable name itself ? 2019-05-22 18:14:24 i'm not sure there's a good way to do that with busybox ash 2019-05-22 18:14:27 the contents 2019-05-22 18:14:53 I was going to use nodejs restify to build the rest API that accepts the build config files and then have that API machine (alpine) build the iso with the build files included in it, then execute the infrastructure automation script to build the generic VM based on ZERO user input and just attach that build ISO which would have build scripts included that build on the injected build files and then shutdown 2019-05-22 18:15:21 tr? 2019-05-22 18:15:37 danieli: i think tr [:upper:] [:lower:] 2019-05-22 18:15:56 mps: north1: yeah, that's the only way i can think of doing it, but i'll just use $_pkgname instead 2019-05-22 18:15:59 thanks 2019-05-22 18:16:32 Any idea's on if this an alright method? Build's are going to be in a queue system that notifies when build is complete and available, so doesn't really matter how long it takes in the end... 2019-05-22 18:17:09 I know there are multiple ways to do this 2019-05-22 18:18:41 containers vs vm, but want separation... ESXi vs XEN, need Free and Flexible (Xen?), API vs ???, PHP vs Node (node for security), transfer build files vs build iso including them, etc....... 2019-05-22 18:19:04 any input at all is welcome 2019-05-22 18:19:14 i don't really understand exactly what you are asking 2019-05-22 18:19:35 just some input on how this should properly be done 2019-05-22 18:20:07 I have an idea on how to limit issues / hang ups like building the iso with build files included instead of transferring over network 2019-05-22 18:21:00 which should be more secure and eliminate some issues like intermittent connection / network / server issues 2019-05-22 18:21:59 So it's looking like XEN + Node API + Build ISO + libvirt controller 2019-05-22 18:22:57 is it okay to run alpine as a node.js rest api? should I use nginx with it? 2019-05-22 18:23:11 whether it's alpine or not makes no difference 2019-05-22 18:23:21 php and node both run fine on alpine 2019-05-22 18:23:31 so does apache and nginx 2019-05-22 18:23:35 an alpine runs fine as a webserver? 2019-05-22 18:23:56 if you install a httpd, yes, whether it's alpine or not hardly makes any difference 2019-05-22 18:24:08 it won't factor into your application design much 2019-05-22 18:24:26 but is probably more secure since alpine uses a hardened kernel? 2019-05-22 18:24:40 functionally, it makes no difference 2019-05-22 18:24:43 and alpine doesn't use a hardened kernel 2019-05-22 18:24:51 we abandoned it a good while ago 2019-05-22 18:25:06 ahhh ok 2019-05-22 18:26:47 Thanks 2019-05-22 18:33:15 something is broken with gimp package 2019-05-22 18:33:22 and rerunning fix doesn't help 2019-05-22 18:33:24 ERROR: Failed to create usr/lib/gimp/2.0/plug-ins/file-jpeg/file-jpeg: No such file or directory 2019-05-22 18:33:29 many lines like that 2019-05-22 18:33:42 the file-jpeg/file-jpeg at end is suspicious 2019-05-22 18:35:30 dalias: on stable or edge 2019-05-22 18:36:27 edge 2019-05-22 18:36:40 /usr/lib/gimp/2.0/plug-ins/ exists 2019-05-22 18:36:52 it's like it's trying to mkdir foo/foo under that without first mkdir'ing foo 2019-05-22 18:37:15 I have this ' parse_iso_codes: error parsing '/usr/share/xml/iso-codes/iso_639.xml': No such file or directory' on stable 2019-05-22 18:37:29 ? 2019-05-22 18:37:45 could it be some other tooling that's broken? apk itself? 2019-05-22 18:37:51 can't check on edge, don't have GUI edge yet 2019-05-22 18:38:49 probably abuild package() or some of the subpackage functions 2019-05-22 18:41:23 ' apk-file file-jpeg' shows only gimp and full path on all arch's except x86_64 2019-05-22 18:42:59 ok i just sed'd the error messages to mkdir -p them all 2019-05-22 18:43:04 now apk fix is working 2019-05-22 18:43:05 *sigh* 2019-05-22 18:43:18 nope.. 2019-05-22 18:43:26 this file exist on stable 2019-05-22 18:43:47 i think the last component was supposed to be a file not a dir 2019-05-22 18:43:52 so taking it off 2019-05-22 18:43:54 looks like it missing on edge and only on x86_64 2019-05-22 18:44:09 yes, sorry, file 2019-05-22 18:44:18 weird 2019-05-22 18:44:54 'file /usr/lib/gimp/2.0/plug-ins/file-jpeg/file-jpeg' gives 'ELF 64-bit LSB pie executable, x86-64, version 1 (SYSV), dynamically linked, ...etc' 2019-05-22 18:46:17 heh, yesterday I rebuilt it without dbus deps but deleted it after that 2019-05-22 18:47:58 will try now again, it will take time on my notebook 2019-05-22 18:48:53 good excuse to leave keyboard and go out a little 2019-05-22 19:00:46 ok forget powershell, I'm just going to switch to libvirt 2019-05-22 19:11:59 danieli: try, sed, awk, printf +case. There are a lot of ways you could change case. 2019-05-22 19:12:09 tcely: i already got it sorted 2019-05-22 19:12:16 but thanks 2019-05-22 19:12:36 s/try/tr/ 2019-05-22 19:12:42 Welcome 2019-05-22 19:12:44 i am aware that i could use different external tools for it, but i was wondering if there was a good way to do it in pure posix sh using a substitution or something 2019-05-22 19:13:32 Reading character at a time and using case / printf 2019-05-22 21:32:35 I know bash has such a substitution, but busybox ash I wouldn't be so sure of 2019-05-22 21:33:51 ? 2019-05-22 21:34:21 my much-outdated sh_tricks.html might have a solution to whatever danieli is trying to do 2019-05-22 21:34:41 it's all right dalias, i got it sorted 2019-05-22 21:40:17 TBB_: I'd just juse tr myself. Avoiding the exec call is not going to be worth the trouble most times 2019-05-22 21:42:18 agreed; I'd basically just use bash since I find it better suited for shell scripting. 2019-05-22 21:43:04 I also did some digging that seems to confirm that the ^^ and ,, substitutions used in bash to do case conversion don't exist in the busybox shell 2019-05-22 21:43:29 yes, i also figured that 2019-05-22 21:44:00 i posted a remark about most of the 'shell scriping' posts online are bash-specific 2019-05-22 21:44:55 now that my mostly-bash work has ended after six years, I find myself drifting back towards Tcl ... 2019-05-22 21:47:54 I don't even exactly know what it is about tcl that got me so deep into it but my rationalization is basically me doubting any other scripting language having the expressive power of it 2019-05-22 21:48:52 but then, if a language is not even on the TIOBE top 50, is there really any use for it in the market... 2019-05-22 21:50:44 (please nobody say Perl) 2019-05-22 21:52:36 Heh. I did a lot of work in Perl for a while. Lately, it's more Lua, but I see a few things that TCL is great at. 2019-05-22 21:52:41 TBB: obligatory https://www.xkcd.com/224/ 2019-05-22 21:53:42 mps: :-D 2019-05-22 21:55:23 *facepalm* 2019-05-22 21:57:22 long time ago I used tcl, and after that perl 2019-05-22 21:58:38 these days I'm using tcl for scripts for managing radio repeaters and expect (which is tcl) for some automation 2019-05-22 21:58:56 for me it was the other way around; when I started in my first job, the first thing that happened was, I was given a book on Perl and sent home for two weeks 2019-05-22 21:59:36 why? so that I could work the entire Easter while everyone else was on vacation... that left a trauma :D 2019-05-22 22:00:17 Tcl I figured out because of a Tcl scriptable IRC client, and was later happy to find the tcl shell in Cisco products 2019-05-22 22:11:57 junos has c and bourne shells 2019-05-22 22:13:19 some other vendors drop you straight into different shells too 2019-05-22 22:16:22 mikrotik routeros has its own funny scripting language 2019-05-22 22:16:36 in a way it remindsme of TCL meets some cisco weirdness 2019-05-22 22:17:24 yeah, mikrotik's is a bit odd but it's not terrible 2019-05-22 22:17:50 mikrotik is comparatively weird 2019-05-23 00:08:14 Is there a way to always install -doc packages without explicitly requesting them? 2019-05-23 00:09:33 apk add docs 2019-05-23 00:11:34 <3 2019-05-23 07:53:08 Hi. probably a frequently asked question, but I can't find anything on how to build an alpine image from scratch without using dockerhub. debian does that with debootstrap . is there such a thing with alpine ? 2019-05-23 07:57:04 https://github.com/alpinelinux/docker-alpine/blob/master/docs/build.adoc 2019-05-23 07:58:28 <_ikke_> car: this should also work: https://wiki.alpinelinux.org/wiki/Installing_Alpine_Linux_in_a_chroot#Install_the_alpine_base_installation_onto_the_chroot 2019-05-23 07:59:40 Ah cool. I'm gonna take a look at both things. I think that helps me a lot. Thank you both very much. 2019-05-23 08:01:50 Crazy. why didn't i find this. exactly what i was looking for. thank you so much again. 2019-05-23 14:19:06 I'm not sure where else to ask so I'll ask it here; I'm trying to set up Alpine on a USB flash drive with persistence. It boots fine, but the problem is it doesn't mount itself properly. Instead of mounting /dev/sdX1 on which Alpine resides, it mounts /dev/sdX, making /dev/sdX2 (on which I want to store apkovl's) inaccessible 2019-05-23 14:19:35 I may just be fantastically retarded, but I can't figure it out. Google is of no help, either 2019-05-23 14:52:34 lockhead2: Have you tried explicitly setting your rootfs with `ROOT=/dev/sdX1` on your kernel cmdline? 2019-05-23 14:53:16 s/ROOT/root/ 2019-05-23 15:38:06 I apologise for the spam, I have a proper account on here now 2019-05-23 16:44:17 is there any way to make scripts go through/retry when alpine package install is failing? 2019-05-23 16:44:26 it always gets stuck when downloading from CDN 2019-05-23 16:47:33 <_ikke_> Not a good CDN if it's constantly failing 2019-05-23 16:47:42 <_ikke_> You mean fastly (dl-cdn)? 2019-05-23 16:55:02 _ikke_: it is dl-cdn, yes 2019-05-23 16:55:12 _ikke_: it's the default in the alpine docker images 2019-05-23 17:02:57 <_ikke_> yes 2019-05-23 17:03:02 <_ikke_> I wonder why it's failing so often for you 2019-05-23 17:07:09 _ikke_: i have no idea. i am running fedora, ubuntu, alpine and arch with some autoamtion 2019-05-23 17:07:20 ubuntu and alpine are usually the ones that fail to download 2019-05-23 20:50:41 i know lists.alpinelinux.org uses hypermail for web archives but is there a way to grab a mbox file from it? i only know how for gnu mailman 2019-05-23 20:52:37 <_ikke_> opal: I don't think so 2019-05-23 20:53:40 ri 2019-05-23 20:53:41 rip* 2019-05-23 20:53:59 same issue with suckless MLs, and ive brought it to their attention. shame 2019-05-23 20:54:12 i guess the proper place to complain would be hypermail itself 2019-05-23 20:54:51 <_ikke_> opal: We are working on migrating to a different system, but I'm not sure if it would support it 2019-05-23 20:54:59 http://www.hypermail-project.org/ seems that there really isnt a place to ask other than github, and i dont have an account there 2019-05-23 21:07:58 i emailed the maintainer so i'll see if i get a reply about that. i dont want to scrape the archives and hit every html page 2019-05-24 04:06:52 in the man-pages package, I have unknown characters in `man 1p fold` regardless if I set LANG=C or LANG=en_US.UTF-8 or which font I use. Can anyone replicate? 2019-05-24 04:19:58 what's $TERM set to? 2019-05-24 04:26:20 /usr/share/man/man1/fold.1.gz is owned by coreutils-doc-8.31-r0 <-- looks fine when I display it 2019-05-24 05:24:12 yeah not the one from coreutils docs, /usr/share/man/man1/fold.1p, I'll check with another term 2019-05-24 05:24:35 occurs with xterm TERM=xterm 2019-05-24 07:13:28 Anyone can explain this: https://serverfault.com/questions/968136/arbitrary-vlan-interface-name-undocumented-configuration 2019-05-24 07:15:42 eKKiM_: explain what exactly? that vlan-id wasn't documented beyond the name? 2019-05-24 07:17:43 vlan-id is not documented at all.. anywhere.. 2019-05-24 07:18:43 i don't even get how it could work that way 2019-05-24 07:19:34 eKKiM_: oh. that part is easy. many who delve into linux networking stack don't value documenting things they take for granted. 2019-05-24 07:20:40 eKKiM_: Are you familiar with paper clips? If you use them to keep papers together on your desk everyday, would you consider writing instructions on how to use them? 2019-05-24 07:20:42 But even when i look into the source of the vlan package it makes no sense (for me) 2019-05-24 07:21:14 <_ikke_> ifupdown allows you to extend properties 2019-05-24 07:21:19 <_ikke_> ie, define your own 2019-05-24 07:22:15 _ikke_: Can you point me to the location where the vlan-id property is defined? i can't seem to find it MS 2019-05-24 07:22:17 :S* 2019-05-24 07:23:41 tcely: if all other (more common) ways to use that paper clip is documented i would think this vlan-id option would be documented too 2019-05-24 07:24:10 at this moment it feels wrong for me to use the vlan-id option because i have a feeling it could be gone any new update 2019-05-24 07:25:46 to be honest, I'm more used to seeing eth0.VLANID interfaces 2019-05-24 07:27:37 tcely: me too. Just discussing it with a friend. we would like the have arbitrary names for some vlan interfaces to get more readability in log files 2019-05-24 12:42:01 Hi everyone. I have a weird problem with apache2 processes within LXC containers. I see segfauls in the dmesg all the time, like this: https://termbin.com/seuw 2019-05-24 12:42:17 Has anyone an idea how to solve that? 2019-05-24 12:42:50 what alpine version are you on, and what's the version of the apache2 package? 2019-05-24 12:42:51 I'm not even sure how to debug, because this happens across many (all) containers, regardless of the alpine linux version (3.8, 3.9, edge) 2019-05-24 12:42:55 that's not good 2019-05-24 12:42:57 i see 2019-05-24 12:43:27 The host system has this kernel: Linux themis 4.19.41-0-vanilla 2019-05-24 12:43:38 Reboot didn't help. 2019-05-24 12:43:59 The host system runs on alpine 3.9 2019-05-24 12:44:29 Currently I try to strace various apache2 processes and hope one of them dies, so maybe I can see something... 2019-05-24 13:23:16 danieli: I found out that this problem only happened with debian containers. I stopped all of them and the messages stopped also. 2019-05-24 13:23:27 I'm further investigating... 2019-05-24 15:48:52 https://git.alpinelinux.org/aports/log/community/easy-rsa/APKBUILD 2019-05-24 15:49:02 is adam dobrawy on irc? 2019-05-24 15:51:38 that easy-rsa version is bugged 2019-05-24 15:51:41 our automation crashed 2019-05-24 15:52:17 <_ikke_> His github nick is ad-m, but I haven't seen him here 2019-05-24 15:52:33 <_ikke_> gunix: Bugged in what way? 2019-05-24 15:55:58 _ikke_: https://github.com/OpenVPN/easy-rsa/issues/302 2019-05-24 15:56:07 quote: "Not missing anything. There is a bug and we will try to look into it. I’d suggest using v3.0.5 temporarily." 2019-05-24 15:56:38 <_ikke_> Ok, so that version is just bugged 2019-05-24 15:57:38 _ikke_: yea. so yesterday our new deployments started to fail and we were like "wat? we didn't touch that area of the code in like 1 month" 2019-05-24 15:58:07 <_ikke_> So you run your deployment system on edge? 2019-05-24 15:58:37 gunix: edge is development so it fan fail 2019-05-24 15:58:47 s/fan/can/ 2019-05-24 16:00:04 although, is not excuse 2019-05-24 16:00:27 _ikke_: i am used to having edge since some stuff is only there 2019-05-24 16:01:14 _ikke_: like this: https://gitlab.com/gun1x/alpine-wireguard/blob/master/Dockerfile#L2 2019-05-24 16:01:58 but anyway, it's also my fault i had big folder with 5 subfolders of containers, all on edge, all automated to deploy together, without separate projects and versioning of containers 2019-05-24 16:02:23 but than again, it is a test deployment so i want it to fail with new software. it's not prod. i want to see errors as soon as possile 2019-05-24 16:02:28 <_ikke_> gunix: It would help to provide feedback to the maintainers of the packages in testing so that they can be moved to community 2019-05-24 16:02:33 <_ikke_> then it would be part of a stable release 2019-05-24 16:02:47 <_ikke_> gunix: right, in that case it makes sense 2019-05-24 16:03:08 _ikke_: i cried on #wireguard about that being only in testing and i think i should have wrote a proper tutorial on wireguard + alpine 2019-05-24 16:03:22 but i burning out since i work too much 2019-05-24 16:03:24 :-D 2019-05-24 16:03:24 <_ikke_> clandmeter is working on moving it to community 2019-05-24 16:03:31 sounds awesome 2019-05-24 16:04:08 <_ikke_> Do you know if there are any significant changes between 3.0.5 and 3.0.6 of easy-rsa? 2019-05-24 16:06:12 <_ikke_> Hmm, so they add new functionality in patch releases, right 2019-05-24 16:06:38 I was suprised by missing 'gpm' package in stable and then suprised again that is in 'testing' :D 2019-05-24 16:07:17 <_ikke_> gunix: I suspect this is the PR that caused the issue? https://github.com/OpenVPN/easy-rsa/pull/63/files 2019-05-24 16:10:30 easy-rsa has a test suite which we properly should use - http://tpaste.us/NK6m 2019-05-24 16:11:31 <_ikke_> if it exists, definitely 2019-05-24 16:12:05 _ikke_: its already in community 2019-05-24 16:12:28 so should be included in 3.10 2019-05-24 16:13:03 <_ikke_> clandmeter: right 2019-05-24 16:13:38 <_ikke_> TBK[m]: But I suspect that the test suite would have caught it before they made the 3.0.6 release if they had tests for this 2019-05-24 16:14:02 <_ikke_> TBK[m]: The test suite passes indeed 2019-05-24 16:15:05 <_ikke_> gunix: Still here? 2019-05-24 16:15:09 _ikke_: yea sorry 2019-05-24 16:15:12 <_ikke_> np 2019-05-24 16:15:14 _ikke_: there was one major change 2019-05-24 16:15:36 _ikke_: they also moved the required location for openssl-easyrsa.cnf from /usr/share/easy-rsa to /usr/share/easy-rsa/pki 2019-05-24 16:15:54 we had the automation put the file in /usr/share/easy-rsa. now it crashes 2019-05-24 16:16:07 <_ikke_> Someone should hit easy-rsa with a semver bat 2019-05-24 16:17:47 https://bpaste.net/show/258db601de42 2019-05-24 16:17:51 this is what we are using 2019-05-24 16:19:24 <_ikke_> I have to go soon, but do you have a way to reproduce the issue? 2019-05-24 16:20:12 _ikke_: that automation has some private keys in there 2019-05-24 16:20:14 :-D 2019-05-24 16:20:41 _ikke_: but there is a bug already on easy-rsa 2019-05-24 16:21:04 <_ikke_> Yes, reading through it 2019-05-24 16:26:39 going back to 3.9 fixed the problem for me 2019-05-24 16:28:32 but that is not a permanent fix obviously. we will do a "stable" container for now, that doesn't get auto-updated, and rewrite the automation when we have time 2019-05-24 16:31:21 gunix: with 'edge' you can always expect some issues, it is testing development branch 2019-05-24 16:40:30 mps: i don't think this issue was in any way the responsability of alpine. 2019-05-24 16:40:55 mps: i think the release policy of easy-rsa is bad. this was their STABLE version. 2019-05-24 16:41:21 they also pushed a new version that is not backwards compatible by adding a minor version bump :-D 2019-05-24 16:44:32 gunix: right, but you encountered issue in alpine 2019-05-24 16:45:12 mps: because i didn't try in fedora, arch or other distros 2019-05-24 16:45:14 pacman -Ss easy-rsa 2019-05-24 16:45:16 extra/easy-rsa 3.0.6-1 [installed] 2019-05-24 16:45:52 and that is good, it will be fixed for next stable, probably 2019-05-24 16:46:16 mps: also fedora: https://bpaste.net/show/0fef0ccb9534 2019-05-24 16:46:47 all distros will be affected by this 2019-05-24 16:47:50 we are accustomed to cope with careful and less careful upstream devs 2019-05-24 16:48:46 at the end, everyone can make bugs 2019-05-24 16:49:11 mps: me too, but the upstream devs should understand they release the package for linux, not for every small communities to fix all the shitty bugs :-D 2019-05-24 16:49:48 mps: there are lots of way to do this today. automated build, CI/CD, unit tests, integration tests. and openvpn is a really big project. 2019-05-24 16:53:15 bugs slips, even with CI/CD. only 'real' test can show bugs and sometimes not from the start 2019-05-24 17:02:23 gunix, mps, and sometimes, it is just better to do a more thorough test for release. 2019-05-24 17:03:03 e.g., https://github.com/gavinhoward/bc/blob/master/RELEASE.md and https://github.com/gavinhoward/bc/blob/master/release.sh 2019-05-24 17:04:40 gdh: true, personally I would like of every distro have some kind of 'quality assurance' team or procedure, but the problem is usually not enough man power 2019-05-24 17:05:31 mps, I agree on both points. 2019-05-24 17:07:15 I'd rather sacrifice 'featurism' in favor of quality, but real life forces us to make 'new shiny featurefull' distro 2019-05-24 17:07:45 Yeah... 2019-05-24 17:08:48 mps: if you automate package checks, you don't need such a big team 2019-05-24 17:09:00 but distros usually rely on the comunity to point out the bugs 2019-05-24 17:09:34 suse has some cool testing now. fedora started too. even so, archlinux is awesome :)) 2019-05-24 17:10:44 automated tools help ofc, but they can't find corner cases bugs, usually, or bugs not anticipated during making tests 2019-05-24 17:11:49 debian is good at this but mostly because their users reports a lot of bugs 2019-05-24 17:12:20 and slow release cycle ;) 2019-05-24 17:20:51 mps: so ... debian is good at testing because it tests 3 years after everybody else? 2019-05-24 17:21:40 they will just get this version of easy-rsa 3.0.6 after everybody else, so they will get this bug 2 years later 2019-05-24 17:23:24 gunix: long freeze time also helps them 2019-05-24 18:05:43 mps: but if you are running something strongswan on debian, you will have an older version probably vulnerable to DDOS 2019-05-24 18:24:49 gunix: debian has older versions for most of their packages however they backport most security fixes 2019-05-25 02:15:30 Hi, just playing around with Alpine just wondering if it is possible to cross compile to ARM when using Alpine, I don't see a arm-none-eabi-gcc package anywhere? 2019-05-25 13:52:12 hi, could any one tell me how to prepare the bootloader parts for my arm devices? 2019-05-25 13:52:53 i could download the generatic arm version , but which is a tarball, how can i prepare the bootloader? 2019-05-25 13:54:51 https://wiki.alpinelinux.org/wiki/Partitioning_and_Bootmanagers this link is empty page :[ 2019-05-25 13:55:33 yunfan, I have no experience with writing tarballs to SD-cards for RPi or similar devices, only iso images 2019-05-25 13:55:50 Unless inside the tarball is an iso 2019-05-25 13:57:56 yunfan: which board 2019-05-25 13:59:07 in tarball there is /boot dir from which you can flash u-boot for your board, of course if your board is supported 2019-05-25 13:59:24 u-boot dir, sorry 2019-05-25 14:01:46 mps: mine is pcuidno3-nano i saw there is relevant dtb file in the tarball 2019-05-25 14:02:39 mps: but the sd card i am going to use is a blank one just bought online 2019-05-25 14:02:45 I don't have experience with this board, but generic instruction would help 2019-05-25 14:03:02 so how should i prepare the bootloader for it? i dont know how to do with uboot 2019-05-25 14:03:35 you have to make partition table and two partitions, one for boot and one for root filesystem 2019-05-25 14:04:02 ok, this is doable for me, what's next? mps 2019-05-25 14:04:12 of course, you can have more partitions if you need or want, but these two are minimum 2019-05-25 14:05:04 ehm, I'm right now out to drink coffee, but I think I have short guide somewhere 2019-05-25 14:05:14 where is the relevant uboot should i get? 2019-05-25 14:05:49 should i download a special one or just a common one for the generic arm supporting target? 2019-05-25 14:07:45 if there is no u-boot for your board in tarball you have to build it from u-boot source or use u-boot from some other distro which have it 2019-05-25 14:10:43 mps: but i had download the generic arm version of alpine 2019-05-25 14:11:03 yunfan: here https://wiki.alpinelinux.org/wiki/DIY_Fully_working_Alpine_Linux_for_Allwinner_and_Other_ARM_SOCs you can read how to make bootable Alpine sdcard 2019-05-25 14:11:23 it is little outdated but is good starting point 2019-05-25 14:11:54 mps: thanks very much ! 2019-05-25 14:11:57 reading 2019-05-25 14:12:06 you are welcome 2019-05-25 14:12:27 yep, i like alpine for these sbc boards 2019-05-25 14:13:16 i used to install arm debian for my board, and it had gone fly after only run for 1 days and doing some magnet downloading 2019-05-25 14:13:18 I run it on arm's, from very small to 'big' ones 2019-05-25 14:13:30 i guess there were too much sdcard writing during that day 2019-05-25 14:13:57 well, it depends for what is used 2019-05-25 14:14:15 and i'd tried alpine on my rpi zero with wifi version, the diskless mode i think is very suitable for this purpose 2019-05-25 14:25:50 mps: can i use the prebuilt binary file from archlinux's ? like this https://archlinuxarm.org/platforms/armv7/allwinner/pcduino3 2019-05-25 14:27:40 yunfan: if you mean u-boot, then yes, usually it works without problem 2019-05-25 14:28:24 ok 2019-05-25 14:28:41 i think I have similar board 2019-05-25 14:29:30 i have few allwinner a20 2019-05-25 14:31:07 ah no, it is cubie board, but still allwinner 2019-05-25 14:33:07 mps: i had 2 cubieboard 7 from my leader as gift 2019-05-25 14:33:25 mps: they had sata ports too like my pcduino3 nano 2019-05-25 14:33:32 but i havnt got any dtb files 2019-05-25 14:34:26 mps: also the currently tarbal of alpine had the dtb files of my pcduino3-nano, but i havnt saw any boot.scr files in it, sounds like not the same as in the article you gave 2019-05-25 14:35:11 one note, be aware that you can have a looooong boot time with alpine kernel 4.19.x, 10 to 20 minutes on some board 2019-05-25 14:35:35 that's really a looooong time 2019-05-25 14:35:38 alpine uses extlinux.conf for u-boot 2019-05-25 14:35:50 similar to syslinux boot 2019-05-25 14:35:53 ok, now i am trying 2019-05-25 14:37:07 I patched kernels for for these small devices by which I reverted crng changes introduced in kernel 4.19 2019-05-25 14:39:01 btw, archlinux alarm guides are fine, small and simple. can be used as a base guide how to install alpine 2019-05-25 15:22:33 mps: just tried alpine for my boards, the uboot binary files might had some problem 2019-05-25 15:22:56 which will said crc failed, and it will use vga as the out while i am using hdmi screen 2019-05-25 15:23:30 so after it boots into alpine(i saw the extlinux menu had been displayed and enter), i had no screen out :[ 2019-05-25 15:28:55 crc error on allwinner are harmless in my experience, it is bug in firmware iirc 2019-05-25 15:30:32 and you have to add some parameters to extlinux.conf for hdmi, I don't have them right now 2019-05-25 15:30:40 okay, then it should be my misusing uboot binary of pcduino3 ? 2019-05-25 15:30:46 try to find them on archlinux alarm 2019-05-25 15:30:48 while my boards is pcduino3-nano 2019-05-25 15:31:07 mps: ah, got it 2019-05-25 15:31:11 will try 2019-05-25 15:38:39 by checking the uart, i found it will use pcduino3.dtb not pcduino3-nano.dtb :L 2019-05-25 15:39:18 and now it hangs at Staring kernel, i guess that begins with the loooooong time you mentioned before? 2019-05-25 15:41:14 mps: how can i edit the grub.cfg to force it use pcduino3-nano.dtb ? 2019-05-25 15:52:08 yunfan: you mean extlinux.conf and not grub.conf? 2019-05-25 15:53:29 you can add 'FDT /sun7i-a20-lamobo-4.18.3-2-sunxi.dtb', not that dtb, it is just my near board example 2019-05-25 15:58:19 or you can use 'FDTDIR /boot/dtbs-4.14.80-1-bpi' by which u-boot will find proper dtb for your board if it in the dir you added as parameter to FDTDIR 2019-05-25 16:08:00 mps: aha, after setting the FDT options, it just works, thanks very much for the kindness help 2019-05-25 16:10:33 i guess i could use the same alpine tarball with the dtb file i found from cubieboard7's specific debian 2019-05-25 16:22:25 yunfan: nice to hear that it works :) 2019-05-25 16:23:25 would you tell me board name and version, I would like to add support for it on Alpine u-boot package 2019-05-25 17:37:42 greetings, I'm trying to install a slightly older version of chromium/chromium-chromedriver in my Dockerfile but when I do the following I get "ERROR: unsatifiable constraints" 2019-05-25 17:37:44 RUN apk update \ 2019-05-25 17:37:44 && apk upgrade \ 2019-05-25 17:37:44 && apk add --no-cache \ 2019-05-25 17:37:46 chromium=72.0.3626.121-r0 \ 2019-05-25 17:37:48 chromium-chromedriver=72.0.3626.121-r0 \ 2019-05-25 17:37:56 https://pkgs.alpinelinux.org/package/v3.9/community/x86_64/chromium 2019-05-25 17:38:12 Please, use a paste service next time 2019-05-25 17:38:17 sorry 2019-05-25 17:38:40 That is in community, did you enable it in /etc/apk/repositories ? 2019-05-25 17:39:01 <_ikke_> What is the exact error you get? 2019-05-25 17:39:12 http://dpaste.com/0D86BZ0 2019-05-25 17:39:40 error: http://dpaste.com/1PXCEJJ 2019-05-25 17:39:58 breaks: world[chromium-chromedriver=72.0.3626.121-r0] 2019-05-25 17:51:48 _ikke_: /etc/apk/repositories already has community and edge 2019-05-25 17:52:00 looks like the docker base image comes with that 2019-05-25 17:52:08 https://medium.com/@stschindler/the-problem-with-docker-and-alpines-package-pinning-18346593e891 2019-05-25 19:53:44 Greetings, someone knows something about the fall of the server http://dl-cdn.alpinelinux.org 2019-05-25 19:54:59 Hm? 2019-05-25 19:57:10 I'm trying to download packages and I'm getting the following message: ERROR: http://dl-cdn.alpinelinux.org/alpine/v3.5/community: network error (check Internet connection and firewall) 2019-05-25 19:57:59 wfm 2019-05-25 20:00:19 sorry my main language is not English, I do not know what wfm means 2019-05-25 20:03:35 Works for me 2019-05-25 20:04:33 It's a CDN, so it might use a different server depending on your region 2019-05-25 20:06:17 Btw. 3.5 is EOL 2019-05-25 20:10:45 ok 2019-05-26 03:21:11 Hi 2019-05-26 03:21:18 hi 2019-05-26 03:25:40 I've seen alpine is most used docker linux image which most of the docker images extended from, how can we rely on the packages alpine used to make sure in doesn't have any vulnerability. Is there any official governance from the community? 2019-05-26 04:01:48 mps: aha, there's another 2 querstions, the first one is that when it boots into the login ui, the board dont take my input from usb. while on uboot, it accept 2019-05-26 04:02:43 mps: another is when i back to use usb2tll, it disconnetinued to output from alpine boot 2019-05-26 04:10:01 and i had append console parameters in extlinux.conf 2019-05-26 08:51:52 yunfan: do you use serial console or hdmi and usb keyboard for boot and login 2019-05-26 09:08:18 mps: yes 2019-05-26 09:09:38 yes? which one 2019-05-26 09:10:56 mps: i had tried both way 2019-05-26 09:11:07 first i tried hdmi and usb keyboard 2019-05-26 09:11:43 but then i met the problem which dont take input from my usb keyboard when login, while in the uboot, it could accept input from the same keyboard 2019-05-26 09:11:52 then i think i could try serial console 2019-05-26 09:12:28 but i found that after the kernel load, it dont output anymore content to my serial wire 2019-05-26 09:12:45 also it used to output content while uboot doing its staff 2019-05-26 09:13:16 I didn't worked much with hdmi/usb keyboard, mostly using arm32 boards in so called headless mode 2019-05-26 09:13:34 i could accept headless 2019-05-26 09:13:54 by headless I mean serial console 2019-05-26 09:14:06 yep thats also what i am mean 2019-05-26 09:15:10 i have in extelinux.conf that parameter ' console=${console}' and it usually works 2019-05-26 09:15:35 i just put console=ttyS0,9600 2019-05-26 09:16:06 should i try ttyAMA0? 2019-05-26 09:16:17 but, also look in your root filesystem /etc/inittab do you have line 'ttyS0::respawn:/sbin/getty -L ttyS0 115200 vt100' 2019-05-26 09:16:17 i remember my rpi zero w need that 2019-05-26 09:16:34 ok trying 2019-05-26 09:16:50 yes, some board use ttyAMA0, mostly arm64 and qemu 2019-05-26 09:17:56 by root filesystem, you mean the /boot/initramfs-vanilla? 2019-05-26 09:18:21 but, you have interaction with u-boot over serial cable, and that mean cable and your terminal works 2019-05-26 09:18:59 root FS usually second partition, if you followed arch linux guides 2019-05-26 09:19:22 i only had one :[ 2019-05-26 09:19:37 and, of course, it could be issue in initramfs-vanilla 2019-05-26 09:19:59 i prefer the diskless mode on sdcard 2019-05-26 09:20:34 change card change os, reminds me those old familly computer 2019-05-26 09:21:30 aha, I wrote some guides how to install alpine arm under qemu, few weeks ago, and there are sections how to unpack initramfs, change it and repack again 2019-05-26 09:26:40 mps: is it work if i patch the /etc/inittab file in alpine.apkovl.tar.gz? 2019-05-26 09:27:30 not sure, last time I tried this is about 5-6 months ago 2019-05-26 09:28:16 i think i prefer a common base from official plus my own customized files 2019-05-26 09:28:59 and I work in sys mode (not diskless) and my tests are mostly in this mode 2019-05-26 09:29:57 ok i will find how to do these, and if i got it work , i think i could take some time to complete the wiki entry 2019-05-26 09:30:09 remember that kernel 4.19.x could cause long boot timeout 2019-05-26 09:30:26 yep 2019-05-26 09:32:00 few times I thought setup doesn't work till I left it in boot state and come back later. I was surprised seeing login prompt :) 2019-05-26 09:32:50 so what happened during this looooong time 2019-05-26 09:33:37 kernel wait for enough entropy for RNG 2019-05-26 09:34:14 which means if i sent much keystrokes, it will speedup the booting? 2019-05-26 09:34:23 btw, haveged (soft RNG) could help in such cases 2019-05-26 09:34:55 keystrokes help with usb keyboards 2019-05-26 09:35:46 in my experience they are nearly useless with serial consoles 2019-05-26 09:36:32 i am checking the initramfs, its interesting that there is no usbhid driver, am i missed something? 2019-05-26 09:37:47 don't know really because I mostly build kernels and use it without initramfs 2019-05-26 09:38:03 ok 2019-05-26 09:41:28 you can look at my guide how to install alpine arm under qemu here http://arvanta.net/mps/install-aarch64-under-qemu.txt 2019-05-26 09:42:17 it could help you with setting initramfs and extlinux.conf, although guide is far from good 2019-05-26 09:55:34 mps: after remastering the initramfs, it still hang on staring kenrel, i guess i met the problem you mentioned 2019-05-26 09:57:57 could be that the initramfs doesn't detect your root FS correctly 2019-05-26 09:58:32 i dont think so, because when i use hdmi output it could boot to login prompt 2019-05-26 09:58:42 aha, true 2019-05-26 09:59:03 is that because alpine use kms mode? 2019-05-26 09:59:59 when i use qemu on server, i often met these things if i choose curses mode 2019-05-26 10:00:53 I don't know for sure although could be right this 2019-05-26 10:03:30 maybe it is time for me to install again alpine on one of the free boards and make notes how it works 2019-05-26 10:04:32 ha, try it 2019-05-26 10:04:58 but still i found there is no usbhid loads in the APPEND parameter values 2019-05-26 10:06:10 last time I tried usb keyboard worked, but not sure if something is changed in meantime 2019-05-26 10:06:30 do you use latest stable, v3.9.4 2019-05-26 10:06:43 of course 2019-05-26 10:07:11 I think that I worked with v3.8 2019-05-26 10:07:34 which platform ? 2019-05-26 10:07:41 artok: pcduino3-nano 2019-05-26 10:07:56 also i had 2 cubieboard7 needs to be installed with this too 2019-05-26 10:08:04 after i had done pcduino 2019-05-26 10:13:03 I have to find free micro SD to try again installation and see what are issues and how it could be done 2019-05-26 10:13:55 thanks very much for this 2019-05-26 10:17:29 np, I thought some time to do this 2019-05-26 14:52:45 can't run chromium inside docker alpine, it closes instantly https://pastebin.com/raw/ZS6yKXL9 2019-05-26 15:03:57 ah there must be some other flag that forces it to linger when in headless mode 2019-05-26 21:16:21 yunfan: are you here? did you booted your board with Alpine kernel from tarball or you built your own 2019-05-26 21:22:31 When using surf, a webkit2gtk based browser, I get '(process:1135): libsoup-WARNING **: 14:04:49.868: soup-tld: There is no public-suffix data available.' spammed to output - I think it's related to libpsl not having any builtin or runtime support 2019-05-26 21:22:36 anyone else get that? 2019-05-26 21:23:02 I get it on luakit doesn't seem to have any negative side-effects 2019-05-26 21:23:46 Mind trying with pr7891? 2019-05-26 21:24:40 yeah, but reading the mainpage for libpsl, which seems to handle the public-suffix data for libsoup, it sounds like it'd be nice to have https://github.com/rockdaboot/libpsl 2019-05-26 21:25:02 but yes, no noticed side-effects beyond log spam 2019-05-26 21:34:36 hi all, I'm trying to run a haskell application in a docker container running alpine. When the application tries to obtain network resources I get an error: ConnectionFailure Network.BSD.getProtocolByName: does not exist (no such protocol name: tcp) 2019-05-26 21:34:52 googling for this there are some references to missing resources in /etc 2019-05-26 21:35:09 I've tried adding alpine-sdk, nss and a few other packages but I still get this error 2019-05-26 21:35:23 does anyone know what could be missing that might be causing this issue? 2019-05-26 21:35:47 Cogitri: Building libsoup now, I'll let you know 2019-05-26 21:36:13 anyone uses alpine linux-vanilla armhf/armv7 4.19.x on arm32 allwinner board? in my case kernel does not boot, u-boot loads it but nothing happens after that 2019-05-26 21:36:32 Anyway, linking it with spl sounds like a good idea, will look into that 2019-05-26 21:37:15 Unode: usually people switch to another distro when alpine doesn't work for them 2019-05-26 21:37:37 adu: that's not really a solution is it? 2019-05-26 21:38:40 libpsl is already in community, it just doesn't seem to have any runtime or builtin capabilities for dealing with the pub-suf list. libpsl-dev I think comes with a tool for checking it 2019-05-26 21:39:30 Unode: have you tried $(sudo printf "tcp\t6\tTCP\n" > /etc/protocols)? 2019-05-26 21:40:11 adu: it's already present 2019-05-26 21:40:22 yeah, but what's in it? 2019-05-26 21:40:51 mps I'm using a custom kernel on aarch64. ime if u-boot loads the kernel but there is no output from the kernel over serial it means you're missing a critical driver. I'd make sure your initramfs is loading all your serial drivers to start with 2019-05-26 21:41:40 among other protocols, tcp 6 TCP # transmission control protocol 2019-05-26 21:41:49 xcko: I have no problem with custom kernels. I want to see/find reason why alpine kernel doesn't work 2019-05-26 21:43:27 Unode: then I'm not sure why Haskell would have trouble finding that protocol 2019-05-26 21:43:31 qemu and aarch64 works, but allwinner boards don't 2019-05-26 21:44:11 mps: ah. well I can't test but if I were you, I'd check the device drivers are enabled in the vanilla kernel for allwinner stuff. basically anything CONFIG_SUNXI is for allwinner 2019-05-26 21:44:37 adu: ok I think I found the answer here: https://vadosware.io/post/static-binaries-for-haskell-a-convoluted-approach/ 2019-05-26 21:45:11 I'm afraid the binary I'm trying to run was built on Ubuntu and triggers the same situation mentioned in this link 2019-05-26 21:45:16 xcko: yes, i'm looking through vimdiff vanilla config and one which works 2019-05-26 21:45:39 adu: thanks for the help. The /etc/protocols pointer led me to that page 2019-05-26 21:45:49 and vimdiff is not good for this, will try to find better tool 2019-05-26 21:45:49 Unode: huh, ok 2019-05-26 21:48:22 Cogitri: got the same warnings on output with libsoup-2.66.2-r0 from your pr 2019-05-26 21:51:02 mps: yeah, I've struggled with that before. ended up going with running menuconfig from both kernel trees and tracking down likely config options. maybe with some clever shell scripting you could narrow it down a bit. if you find a good tool I'd be curious to know for next time 2019-05-26 21:54:48 xcko: diff and vimdiff is all I have in mind for now :( 2019-05-26 21:55:50 I will try drivers one by one, which will take some time 2019-05-26 22:00:50 xcko: Thanks for confirming :) 2019-05-26 22:27:07 mps: maybe you could experiment with diff `vanilla.config custom.config | grep -- '+CONFIG.*=[ym]' ` which would give you places where a config option was added to the custom config that wasn't in the vanilla 2019-05-26 22:27:17 Cogitri: no problem, thanks for keeping it updated! 2019-05-26 22:34:08 xcko: yes, could help. thanks. Although I think I found issue, i.e. some sunxi drivers are modules and should be given as modules=xx params 2019-05-27 00:19:59 hi. the "virtual" version says "Similar to standard. Slimmed down kernel. Optimized for virtual systems" - where can i read about what specifically "slimmed down" means, and what specifically "optimized" means? 2019-05-27 00:32:00 the most specific answer would be to have a look at https://github.com/alpinelinux/aports/blob/master/main/linux-vanilla/config-virt.x86_64 2019-05-27 00:32:14 or https://github.com/alpinelinux/aports/blob/master/main/linux-vanilla/config-virt.x86 or https://github.com/alpinelinux/aports/blob/master/main/linux-vanilla/config-virt.aarch64 2019-05-27 00:39:10 oh, cool, thanks. so given x86_64, for example, i could compare config-virt and config-vanilla, and that will tell me how the kernels differ? 2019-05-27 00:39:30 does standard = vanilla? 2019-05-27 00:42:51 yes 2019-05-27 00:44:08 ok. are kernels provided as packages that can be switched between? e.g. could i use the standard installer and then switch to the virtual kernel later? 2019-05-27 01:05:06 I haven't tested it but yes, linux-vanilla and linux-virt are the names of the packages 2019-05-27 01:07:03 ok, thanks. i'll experiment 2019-05-27 01:40:54 mps: i booted with the kernel from tarball 2019-05-27 06:40:12 yunfan: thanks for info. Is your board Allwinner and if it is which version, i.e. series A10, A20 or something else 2019-05-27 06:49:08 mps: the board is pcduino3-nano with A20 inside 2019-05-27 06:53:12 yunfan: thanks again for info. trying to install alpine on cubieboard1 with A10 SOC but kernel doesn't boot 2019-05-27 07:30:41 mps: so its the kernel's problem? 2019-05-27 07:30:52 mps: maybe i should use alpine 3.8 ? 2019-05-27 07:31:29 yes, I think. I will try later to boot bananapi A20 to see if there is a difference 2019-05-27 07:32:59 I can boot these boards with custom made kernels without problem, but I want to fix alpine vanilla kernel to work with these boards 2019-05-27 07:34:23 mps: wow, if you had the chance to fix the vanilla kernel, can you add support for cubieboard7? i could got its dtb file fromt he official image for debian 2019-05-27 07:34:30 but the kenerl is 3.x 2019-05-27 07:36:34 I will try what I can do, but I don't have much different boards so I can't test them 2019-05-27 07:37:17 your help will be appreciated if you have boards 2019-05-27 07:40:03 and, Alpine u-boot need some fixes, I think. For example, u-boot in alpine for cubie 1 doesn't work, I think it is to big to fit in boot sectors of partition table 2019-05-27 08:05:04 mps: well i found their boot partition is much smaller than alpine's 2019-05-27 08:10:59 I've built u-boot for cubie 1 from upstream and it is a just little smaller but it fits and works properly while alpine u-boot doesn't 2019-05-27 11:50:27 ddevault: What is the status of urxvt/perl locale error? I saw you commenting that something went wrong without specifying. 2019-05-27 11:52:07 fix is known, but deliberated. seems it needs full rebuild of all pkgs that link against libperl, which is something that people want to avoid iiuc 2019-05-27 11:54:04 maybe not everything should be rebuilt, just some pkg's 2019-05-27 11:54:36 Does the configure option change the ABI? 2019-05-27 11:55:06 no, afaik 2019-05-27 11:55:35 right solution will be patching perl, imo 2019-05-27 11:56:57 why is the cflags option not the right solution? 2019-05-27 11:58:19 p4Wv1qn095FW: dalias know better answer than I, you should ask him when he comes online 2019-05-27 12:55:06 p4wv1qn095fw, that's not a fix. it's a workaround that goes back to perl's old brokenness of handling locale wrong 2019-05-27 12:55:55 i don't see why it would change abi or require rebuilding all packages, but it might, because perl is awful :) 2019-05-27 12:56:50 but i already described the right fix to ncopa; he just ran out of time to test it before taking break for weekend 2019-05-27 13:03:52 hi dalias 2019-05-27 13:08:54 hi 2019-05-27 13:09:32 are you back and wanting to work on this further? i think the cleanest fix for now would be doing at the top of the file 2019-05-27 13:10:56 #if !defined(HAVE_QUERYLOCALE) && defined(_NL_LOCALE_NAME) / #define querylocale(mask,loc) nl_langinfo_l(mask2cat(mask),loc) / #define HAVE_QUERYLOCALE 1 / #endif 2019-05-27 13:11:17 where mask2cat is a function they may already have or that you'd have to provide 2019-05-27 13:11:44 this will ensure that the same code paths (the only correct ones) that get used with bsd querylocale get used on musl and glibc too 2019-05-27 13:12:07 and that the wrong fallback code only gets used on systems completely lacking any way to do it right 2019-05-27 13:12:53 really !defined(HAVE_QUERYLOCALE) && !defined(_NL_LOCALE_NAME) should forcibly disable perl's thread-local locale support, not use the buggy fallback they have now 2019-05-27 13:13:17 erm above, that should be: 2019-05-27 13:13:39 #if !defined(HAVE_QUERYLOCALE) && defined(_NL_LOCALE_NAME) / #define querylocale(mask,loc) nl_langinfo_l(_NL_LOCALE_NAME(mask2cat(mask)),loc) / #define HAVE_QUERYLOCALE 1 / #endif 2019-05-27 13:47:17 exit 2019-05-27 14:12:12 actually converting back from a mask is gratuitous mess so maybe it's best to just add the #elif inline 2019-05-27 14:15:04 ncopa, in addition to the changes i suggested last time, just removing lines 584-602 and replacing them with return temp_name; should finish fixing things 2019-05-27 14:15:39 hey I am trying to get php 7.3 installed I follwed this but still get php 7.2 files 2019-05-27 14:15:43 https://github.com/codecasts/php-alpine 2019-05-27 14:16:49 dalias: ok. so as i understand, i have 2 options: either do the define querylocale + implement mask2cat, or do what you suggested last time and do the return temp_name. 2019-05-27 14:28:37 how do I tell which version of a package is installed? 2019-05-27 14:29:27 `apk info`? 2019-05-27 14:29:58 apk info just says it sinstalled no version numbers 2019-05-27 14:30:39 vizo: apk version pkgname 2019-05-27 14:31:26 ah thanks 2019-05-27 14:33:37 I am tryin g to get php7-7.3.2-r1 installed but I cant seem to figure out how 2019-05-27 14:34:23 yit depends on the alpine version vizo or you manually compile it 2019-05-27 14:34:33 7.3 is in edge not stable, iirc 2019-05-27 14:34:39 vizo:^ 2019-05-27 14:34:51 it is? 2019-05-27 14:35:31 AL 3.9 = php 7.2.18 2019-05-27 14:35:57 compile it manually then vizo, you can take a look at php's official dockerfiles to see how they do it 2019-05-27 14:36:36 php 7.3 is being worked on for edge - https://github.com/alpinelinux/aports/pull/5863 2019-05-27 14:44:32 so would this get tophp 7.3 https://github.com/codecasts/php-alpine 2019-05-27 14:44:50 to php 7.3 2019-05-27 14:45:34 oh that for docker isnt it? 2019-05-27 15:16:03 ncopa, yes, that sounds right 2019-05-27 15:16:21 ideally i think the code paths for querylocale and NL_LOCALE_NAME should be unified because it's the same thing with different names 2019-05-27 15:17:28 i came up with this: 2019-05-27 15:17:37 but the gratuitous use of mask vs cat number makes it more work than it should be 2019-05-27 15:17:42 http://tpaste.us/vbDo 2019-05-27 15:17:51 (and they converted a cat to a mask just above the use of mask) 2019-05-27 15:17:58 i did realize what you just said 2019-05-27 15:18:26 there are a couple of more uses for querylocale 2019-05-27 15:18:27 some of your cases are missing _MASK 2019-05-27 15:18:48 so we cannot simply drop use of querylocale 2019-05-27 15:18:57 but I came up with this alternative: 2019-05-27 15:19:09 http://tpaste.us/Q1x0 2019-05-27 15:19:18 which is significantly less intrusive 2019-05-27 15:19:50 im about to test it 2019-05-27 15:22:10 hum 2019-05-27 15:22:11 nope 2019-05-27 15:22:14 didnt fix it 2019-05-27 15:24:51 what happens now? 2019-05-27 15:25:13 oh your alternative is incomplete 2019-05-27 15:25:25 there are more #ifndef HAS_QUERYLOCALE 2019-05-27 15:25:26 yup 2019-05-27 15:25:31 right 2019-05-27 15:25:36 that whole #ifndef should be #if 0 2019-05-27 15:25:36 there is another place as well 2019-05-27 15:26:44 i was originally thinking of replace all querylocale's with my_querylocale(category, ....) 2019-05-27 15:27:43 yes that would work too 2019-05-27 15:29:32 and that is probably better than: http://tpaste.us/Kg8o 2019-05-27 15:44:55 this works though: http://tpaste.us/oaYk 2019-05-27 15:45:17 but those if(n)defs HAS_QUERYLOCALE is somewhat tricky to get right 2019-05-27 16:03:27 looks ok 2019-05-27 16:03:36 it if works i'd go with that for now 2019-05-27 16:03:44 yup 2019-05-27 16:04:04 i was thinking implement const char * getlocalename_l (int category, locale_t locobj); 2019-05-27 16:04:22 but it gets complicated because the mask is calculated from index 2019-05-27 16:04:28 its all messy 2019-05-27 16:04:39 so i think the http://tpaste.us/oaYk works 2019-05-27 16:05:50 fwiw i would consider just providing querylocale in musl, but the bsd semantics are odd enough that i wouldn't want to do it without a sign that posix won't want to do something conflicting 2019-05-27 16:06:12 (use of mask instead of a category, undefinedness of mask has more than one bit set, etc.) 2019-05-27 16:06:13 might be worth wait for this: http://austingroupbugs.net/view.php?id=1220 2019-05-27 16:06:31 i like the nl_langinfo approach better 2019-05-27 16:07:20 i think we should report it to perl though 2019-05-27 16:07:29 yes 2019-05-27 16:07:37 and add a testcase to their testsuite 2019-05-27 16:07:53 i don't know what we'll hit with their bogus claims that NL_LOCALE_NAME doesn't work tho... 2019-05-27 16:08:33 i suppose just starting defining the problem in their bugtracker is a start 2019-05-27 16:08:36 but whatever problems they hit with it on glibc, they *can't* be anywhere near as wrong as the hacks they're doing parsing setlocale() return 2019-05-27 16:09:07 setlocale return is just irrelevant because it's for the global locale not the thread-local one 2019-05-27 16:09:35 we should make testcase for it 2019-05-27 16:12:23 im going with http://tpaste.us/oaYk for alpine for now 2019-05-27 16:16:31 ok sounds good 2019-05-28 00:01:17 given a package [say iproute2], how can i list its dependencies? 2019-05-28 00:01:35 e.g. i want to see what other packages will be dragged in, before doing apk add 2019-05-28 00:03:30 apk info -R 2019-05-28 00:03:51 see the output of apk info --help for more switches 2019-05-28 00:05:57 Hey guys, I have a sys installation of alpine that I would like to boot directly into x. I'm trying to modify some arch linux .bash_profile lines, but the environment variables are different. How can I accomplish what I want? thanks! 2019-05-28 00:07:06 directly into X as your user without having to log in into a terminal ? 2019-05-28 00:07:14 north1: ah, thanks 2019-05-28 00:10:18 yes, I'm going to use it as a no-internet kiosk, sort of. 2019-05-28 00:12:15 i get this: http://dpaste.com/1ZDTWCD.txt 2019-05-28 00:12:27 those seem to be files / shared objects though, not packages? 2019-05-28 00:13:05 it depends on those libraries, some packages will provide that so: 2019-05-28 00:13:45 i'd have to then track down the various packages providing each of the shared objects? 2019-05-28 00:14:07 if you apk search so:libxtables.so.12 it will return iptables 2019-05-28 00:14:08 lunaphyte: no 2019-05-28 00:14:35 this is not slackware 2019-05-28 00:14:57 i think i might be confused 2019-05-28 00:15:33 -R says "List packages that the PACKAGE depends on" 2019-05-28 00:15:45 but the output is not a list of packages, right? 2019-05-28 00:16:00 yes 2019-05-28 00:16:00 Is there an envronment variable that tells what tty you're on? 2019-05-28 00:16:30 scrawler: you can run the `tty` binary 2019-05-28 00:16:48 i guess -R should say something like "List files that the PACKAGE depends on"? 2019-05-28 00:17:05 lunaphyte: They are not necessarily files 2019-05-28 00:17:10 thanks. 2019-05-28 00:17:21 north1: what else are they? 2019-05-28 00:17:37 they can be straight packages or other stuff like so: (shared librares) cmd: (binaries) 2019-05-28 00:17:52 but yes in the end those so: and cmd: will be mapped to a package that has them 2019-05-28 00:18:59 i guess it would be nice if it told me the packages 2019-05-28 00:19:37 you can use apk search and it will find those 2019-05-28 00:19:51 yeah 2019-05-28 00:20:08 how can i see why iproute2 depends on iptables? 2019-05-28 00:21:27 if it is because of a so: or a cmd: then it is easy enough, for the others you might need to consult upstream documentation 2019-05-28 00:23:29 i can see it's because of a dependency on libxtables.so.12, but after installing, it's not clear to me. libxtables.so.12 isn't listed in ldd output 2019-05-28 00:27:42 /sbin/tc links to libxtables.so.12 2019-05-28 00:29:41 oh, i'm dumb. thanks. i wasn't considering i need to inspect more than just ip 2019-05-28 00:32:29 bummer for that dependency 2019-05-28 04:58:49 Hi all, is there any way to override the abuild behaviour where it wont create a package that installs files to /opt? 2019-05-28 05:00:25 <_ikke_> options="!hfs" 2019-05-28 05:02:09 <_c705> Or if you run mkimage.sh --outdir /whatever/you/want 2019-05-28 05:30:16 _ikke_: Sorry was that for me? 2019-05-28 05:31:53 <_c705> Yes 2019-05-28 05:32:06 Didnt seem to help 2019-05-28 07:44:59 hi folks, how can i enable a boot log? i don't mean dmesg, but a log of the boot-process after the kernel. 2019-05-28 07:45:13 basically i wanna see which services failed at boot and why. 2019-05-28 07:47:38 <_ikke_> https://wiki.archlinux.org/index.php/OpenRC#Boot_logs I guess 2019-05-28 07:48:06 <_ikke_> alpine doesn't have /etc/openrc though 2019-05-28 07:50:49 n.b. openrc bootlogs don't appear to log the steps before the root disk is available 2019-05-28 08:31:37 too bad. well i look into the archlinux article. maybe there is a way to enable it 2019-05-28 08:37:20 ouch... there is a rc.conf. i think it's what i searched. 2019-05-28 13:37:41 Is there a way to login without executing /etc/profile? 2019-05-28 13:37:41 I editted it and now it hangs on log in ... 2019-05-28 13:39:17 nm - I had jsut set the color to black accidentally >< 2019-05-28 15:43:38 I'm using docker base image python:3.7-alpine3.9 and when adding some development packages, getting some package conflicts for readline-7 vs readline-8. What is the best way to remove the out of date packages that depend on readline-7? 2019-05-28 15:44:44 I recommend compiling all from binaries for the docker image 2019-05-28 15:44:55 from sources to binaries that is 2019-05-28 15:45:51 have python3.7, add pip and buildtools, pip install all requirements and then just remove build tools 2019-05-28 15:45:58 multi stage build 2019-05-28 15:46:12 more work but... 2019-05-28 15:52:32 yeah I am happy to build as much from source as possible. I guess the problem is that I have to move from v3 packages to edge packages to obtain librdkafka-dev 2019-05-28 15:55:03 https://pastebin.com/tf6gLXaS 2019-05-28 15:55:38 unsatisfiable constraints: readline-8.0.0 (needed by python 3.7) and readline-7.0.003 (from .python-rundeps) 2019-05-28 15:58:11 oh librdkafka was added in a recent minor release! I don't need to swap up to edge packages anymore 2019-05-28 15:58:35 but do yall publish edge builds? like python:3.7-alpine-edge ? 2019-05-28 15:59:25 <_ikke_> We do have regular edge snapshots, but I'm not sure if docker images are created out of that 2019-05-28 15:59:48 <_ikke_> There are 2019-05-28 16:00:01 <_ikke_> 20190508 is the latest 2019-05-28 16:00:33 <_ikke_> That's not a python image though 2019-05-28 16:00:39 <_ikke_> just vanilla alpinelinux 2019-05-28 16:01:16 <_ikke_> alpine:edge 2019-05-28 16:01:21 yeah that would be a fine base image though. `apk add python3` is easy enough 2019-05-28 16:03:16 hmmm, i spoke too soon. alpine 3.9 still has librdkafka 0.11, not 1.0 2019-05-28 16:03:34 why did my `apk add librdkafka>1.0.0-r0` line not fail? it installed 0.11 :( 2019-05-28 16:07:31 <_ikke_> zetuno: lol, 2019-05-28 16:07:40 <_ikke_> try cat 1.0.0-r0 2019-05-28 16:08:16 <_ikke_> (I was confused for a second myself) 2019-05-28 16:12:49 huh? no file named 1.0.0-r0? 2019-05-28 16:13:20 also, I was wrong `apk install python3` is apparently not sufficient, looking at the https://github.com/docker-library/python/blob/master/3.7/alpine3.9/Dockerfile but thats fine, I will fork this and move it to edge 2019-05-28 16:13:32 <_ikke_> Did you type exactly apk add librdkafka>1.0.0-r0? 2019-05-28 16:14:23 --no-cache, and with more packages on the same line, but yeah 2019-05-28 16:14:38 <_ikke_> Do you see the > / 2019-05-28 16:14:40 <_ikke_> ? 2019-05-28 16:15:16 oh , yeah okay 2019-05-28 16:15:18 ha 2019-05-28 16:15:24 <_ikke_> hehe :) 2019-05-28 16:15:57 is there a package-at-least-version specifier I can use? should I just peg it with = ? 2019-05-28 16:16:14 <_ikke_> you can use it, but you have to escape it 2019-05-28 16:16:24 <_ikke_> apk add 'librdkafka>1.0.0-r0' 2019-05-28 16:16:35 <_ikke_> or 2019-05-28 16:16:38 <_ikke_> apk add 'librdkafka>=1.0.0-r0' 2019-05-29 05:46:46 anybody in the room experienced with getting Alpine to netboot? 2019-05-29 05:52:22 from what I can see it's not trying to mount alpine_dev (no entries in the logs on my nfs server to indicate any attempt) 2019-05-29 06:07:40 scratch that problem was apparently apline_repo being an https address, dunno why I need to specify alpine_dev still being that it doesn't actually mount it :/ 2019-05-29 06:10:17 or it does but my nfs server isn't reporting the mount even though it's in debug mode o.0 2019-05-29 06:17:33 hmm, nope not sure what's going on here we're back to where we were... that being the boot dumping to an emergency shell after attempting to retrieve the index from alpine_repo 2019-05-29 06:28:42 aaaand working again with no change *scratches head* 2019-05-29 06:31:32 definitely not mounting alpine_dev... 2019-05-29 06:32:43 which TBH is preferable, just weird that it's marked as required. 2019-05-29 06:33:01 also it appparently works without the param entirely, I consider that a win though 2019-05-29 09:33:35 Are Skype, Spotify, and Steam available for use on alpine? 2019-05-29 09:33:43 hi! how can i install google-chrome? 2019-05-29 09:34:17 No 2019-05-29 09:34:27 They're all proprietary applications, which is one reason to not package it, and they require glibc to run which isn't available on Alpine 2019-05-29 09:34:41 BOPOH: you can't, but you can install chromium 2019-05-29 09:35:33 ok, where i can see error logs? chromium gives me "Segmentation fault" error 2019-05-29 09:35:44 periffis: to my knowledge, none of them are available in alpine 2019-05-29 09:36:20 i have heard mentions of running spotify in a container/chroot, but it is unsupported if you decide to try it 2019-05-29 09:36:22 PureTryOut[m]: Why is that a requirement? Another question that eats away at me is what the real differences between musl and libc are. 2019-05-29 09:36:29 docker FROM php:5.6-fpm-alpine 2019-05-29 09:36:43 you'll have worse luck with skypeforlinux and steam 2019-05-29 09:36:57 periffis: because those packages are built by their developers with glibc. You can't just link them to a different libc without recompiling 2019-05-29 09:37:04 Steam runs fine in a chroot btw 2019-05-29 09:37:24 Do note that Flatpak is available in Alpine, and stuff can be installed through there as well 2019-05-29 09:37:41 PureTryOut[m]: Ah, so it's only if we can only use binaries. Right right right. 2019-05-29 09:38:03 Does the linux kernel not depend upon a certain version of libc? 2019-05-29 09:38:15 it does not 2019-05-29 09:38:35 it's also not about versions, they're entirely different projects (glibc/musl/uclibc etc.) 2019-05-29 09:39:41 I believe the kernel has it's own internal libc 2019-05-29 09:40:11 How the hell does that work? 2019-05-29 09:40:36 Also, is musl just built from the ground up, then? 2019-05-29 09:44:35 libc is an API to the syscalls the kernel provides. The implementation of libc decides which function calls are available. If a program depends on a glibc specific function that musl does not provide, then you have a tough time compiling that program on a musl system. You can use gcompat for some pre-built binaries. 2019-05-29 09:46:04 Myself, if I find something that needs glibc, I'm likely to use a container. 2019-05-29 09:48:54 How does THAT work, then? Containers have their own libc library somewhere? 2019-05-29 09:49:07 I'll be back soon. 2019-05-29 09:50:12 Containers basically run a stripped down version of a Linux distro. This can be the same as the host distro, but also a completely different one, including distros with glibc. So you could run Alpine Linux on your host, and run a container with e.g. Debian and it's programs 2019-05-29 12:04:40 flatpak and appimage could use more loving on Alpine, but they're definitely the way to go... no more "not packaged for my distro" problems 2019-05-29 12:06:44 you can also use singularity as alternative to docker. 2019-05-29 12:10:28 TBB: Pretty sure AppImage isn't going to be a thing on Alpine because it needs glibc IIRC 2019-05-29 12:10:29 there are plenty of tools, but the question is, what are the application makers using to package their software? some package things using snap, some use flatpak... it's difficult to choose, and of course the complexity of your systems will increase as you run multiple package managers 2019-05-29 12:10:38 Flatpak works just fine for me though 2019-05-29 12:11:10 for me it didn't work out of the box, but then, you're expected to do some work to run things on Alpine anyway, it's not an ubuntu after all 2019-05-29 12:11:33 Well, IMHO it should be as easy to setup as possible :) 2019-05-29 12:12:13 So maybe you could tell me what "didn't work out of the box"? Then I can look into fixing it 2019-05-29 12:14:54 oh, I wish I could remember the exact details. It was probably about running one of the big browsers; I usually write notes about attempts like these, but those specific notes are on a laptop I no longer am in possession of 2019-05-29 12:17:19 this was also over half a year ago, things may have improved since 2019-05-29 13:29:12 hi 2019-05-29 14:56:08 I'm looking to use apk-tools / abuild to build a the glibc version of Linux From Scratch, is the scripts/bootstrap.sh just used to build the toolchain + the base system apk's? I've looked through the irclogs trying to find some more information about it. But just wanted to confirm on that. Also is there any specific code inside abuild or apk that will prevent / make it harder to build glibc apk's? From the looks of it abuild looks fine. 2019-05-29 15:01:04 question, is LFS an absolute requirement for you or are you just planning to have a minimal Linux+glibc built? 2019-05-29 15:02:41 LFS is not a requirement par say but I'm borrowing just the knowledge from that book to build a minimal linux+glibc. 2019-05-29 15:03:56 okay. I just stumbled upon a tool called buildroot for building embedded linux systems for various hw platforms 2019-05-29 15:05:51 Yea, I've used buildroot and Yocto Project. Both great projects to build root filesystems. 2019-05-29 15:06:22 buildroot builds the toolchain then you can build the rootfs from that single tool. 2019-05-29 15:11:42 yup. it looks really handy so far, been getting used to it for a week and it's pretty similar to what I've done in the past, only in a not package manager oriented fashion 2019-05-29 15:12:08 yocto is a bit on a heavy side though :) 2019-05-29 15:12:28 Yea yocto is super heavy and the configuration language is a bit high level. 2019-05-29 15:14:03 I remember touching yocto at some point in 2011, launching my first build - and having the rest of the day free as my laptop was completely consumed by the build :D ... 2019-05-29 15:19:45 but looking at bootstrap.sh (one of those scripts I haven't yet had a look at before) your assumption is correct 2019-05-29 15:24:01 abuild I've been reading through seriously for a couple of years and I don't think there's anything that would prevent it from being used to build glibc... but that's one of just packages of which you never know 2019-05-29 15:48:30 Is it possible to get APK to just download a package and all of it's dependencies without installing them? I'm looking to set up my own (very minimal) APK mirror and it's a pain downloading packages and their dependencies by hand 2019-05-29 15:49:58 apk fetch 2019-05-29 15:50:30 How can I force a specific version of a package to be installed? In my repos there are 2 available, but for some reason apk only installs the older version 2019-05-29 15:51:02 jackwilsdon: specifically, 'apk fetch -R ' the -R will grab deps too 2019-05-29 15:51:20 ACTION sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/JuhmjGKIaVyDugVSIqYpSdiB > 2019-05-29 15:51:26 I don't understand that ^ 2019-05-29 15:51:56 That's perfect, thanks TBK[m] and programmerq! (no idea how I missed fetch in the help) 2019-05-29 15:52:28 PureTryOut[m]▸ that shows two separate packages 2019-05-29 15:53:54 programmerq: one is just the -dev version of the other 2019-05-29 15:53:54 the package that has -dev in its name has the necessary headers for compiling or similar 2019-05-29 15:54:03 But there is a newer available as shown, but it won't install 2019-05-29 15:54:04 yup, still two different packages. 2019-05-29 15:54:42 I realize gtk+2.0-maemo and gtk+2.0-maemo-dev are different... 2019-05-29 15:54:44 your paste didn't include the fail to install part of the error 2019-05-29 15:54:45 I just want them both to be upgraded to 2.24.23 2019-05-29 15:54:56 *2.24.32 2019-05-29 16:00:35 the fail to install? 2019-05-29 16:02:32 So when I run `apk fetch -R [packages...]`, how do I know which arches the packages belong in? Some seem to be noarch and others x86_64 2019-05-29 16:57:59 Is there a history for this channel? I asked a question earlier but my client disconnected and I didn't see if anyone replied 2019-05-29 16:58:37 jackwilsdon: nobody replied - but here are the logs https://dev.alpinelinux.org/irclogs/ 2019-05-29 16:59:21 Ah, thanks :) 2019-05-29 17:04:24 Is it documented anywhere how to create a custom APKINDEX with noarch and x86_64 packages? I seem to be having odd issues 2019-05-29 19:09:32 Hello, I am unable to do `modprobe fbcon` and I don't know how to fix this. x86, Alpine 3.9, kernel 4.19.41-0-vanilla 2019-05-29 19:09:42 what's the output? 2019-05-29 19:10:38 modprobe: FATAL: Module fbcon not found in directory /lib/modules/4.19.41-0-vanilla 2019-05-29 19:12:35 we seem to compile linux with CONFIG_DRM_VMWGFX_FBCON=y 2019-05-29 19:14:07 that's the vmwaregraphics driver for 3d hw accel iirc 2019-05-29 19:14:15 So there's no need to load the module 2019-05-29 19:14:23 s/vmwaregraphics/vmware graphics/ 2019-05-29 19:14:29 oh, I had nomodeset enabled, can this be the cause? 2019-05-29 19:22:59 nope 2019-05-29 19:37:16 upgrading to edge didn't help either 2019-05-29 19:38:17 and Xorg -configure will fatal withou fbcon 2019-05-29 19:51:11 oh, ok, it appears that this error is not important at all 2019-05-29 20:23:13 Halamix2, you dont need do -configure, X org auto detection usually working good enough and if wanna enforce some setting then creating .conf files with them in /etc/X11/xorg.conf.d/ 2019-05-29 20:58:50 Is it right that Alpine Linux 3.19 doesn't have the loopback interface configured properly by default? 2019-05-29 20:59:02 i.e. you can't ping localhost, the loopback does not have 127.0.0.1 assigned 2019-05-29 20:59:29 Sorry v3.9 not 3.19 2019-05-29 21:00:10 wfm on edge at least 🤔 2019-05-29 21:00:19 Or maybe NetworkManager sets that up for me 2019-05-29 21:00:20 Need to enable loopback rc service 2019-05-29 21:00:29 Oh interesting, is that not enabled by default? 2019-05-29 21:09:46 isn't like "networking" service reading "/etc/network/interfaces" where "setup-alpine" put "auto lo ...." ? sounds like default :P 2019-05-29 21:11:52 I'll take a look at what setup-alpine does, thanks! 2019-05-29 21:14:28 MY_R, it does look like alpine-setup adds "auto lo" and "iface lo inet loopback" to /etc/network/interfaces, which correctly configures the loopback correctly. 2019-05-29 21:15:03 Somewhat confused by the fact that there's even a /etc/init.d/loopback now though, what purpose does it serve that /etc/network/interfaces doesn't achieve? 2019-05-29 21:15:20 more like "setup-interfaces" which is executed by setup-alpine 2019-05-29 21:15:38 Yeah sorry, that's the file I'm looking at: https://github.com/alpinelinux/alpine-conf/blob/c2275905471687b0cf2470edc20d25f2192b8250/setup-interfaces.in#L552-L571 2019-05-29 21:16:13 I'm doing a network boot for a diskless server, so I guess I should be using an overlay to provide a basic /etc/network/interfaces 2019-05-29 21:20:20 Is it possible to create an overlay (i.e. an apkvol) that deletes a file on a system? Or would I need some form of script for that 2019-05-29 21:20:40 s/apkvol/apkovl 2019-05-29 21:43:48 hi, does anyone use multipath for san? 2019-05-29 23:51:03 PureTryOut[m]: Does that turn into performance losses in any real way if I've got the memory for it? 2019-05-30 08:45:10 is there a third-party alpine linux repo that offers links with gui support? 2019-05-30 09:20:12 winny: some time ago I tried to build links with X but it wasn't much useful for me so I didn't posted patches for gui version to Alpine 2019-05-30 09:21:02 ah bummer 2019-05-30 09:21:18 i didn't appreciate how many deps the graphical version of links has 2019-05-30 09:23:14 in nowadays web with JS everywhere it is not much useful to have gui browser without JS 2019-05-30 09:24:12 and links doesn't support html5, afaik 2019-05-30 09:24:40 yeah, that's pretty true too 2019-05-30 09:28:08 tbh it seems like there are enough dependencies that alpine might want to do what debian does, and ships a links package, and a xlinks package 2019-05-30 09:28:49 periffis: hi, sorry, what do you mean performance losses? What are you talking about? 2019-05-30 09:32:22 winny: maybe we can do something about this after finishing next stable release 2019-05-30 09:32:59 you can fill issue (wishlist) at bugs.a.o 2019-05-30 09:33:02 Maybe, I think i have a working links now. :) 2019-05-30 09:53:24 heyho, anyone willing to help me with a weird openvpn issue? 2019-05-30 09:54:06 when I run /usr/sbin/openvpn --config /etc/openvpn/openvpn.conf the interface is created and a client connection works (my alpine machine gets an ip from my openvpn remote server) 2019-05-30 09:54:41 but when I try to use rc-update add openvpn and then service openvpn start it just says " start-stop-daemon: failed to start `/usr/sbin/openvpn' 2019-05-30 09:55:31 the error message is not really helpful 2019-05-30 09:57:35 I suppose you could check your logs for the actual command that it launches 2019-05-30 10:00:07 @Cogitri any idea where I could find openr logs? 2019-05-30 10:01:27 `/var/log` 2019-05-30 10:01:33 Well, If you have a logging daemon setup 2019-05-30 10:05:01 @Cogitri well that gave me an idea to checkout dmsg first, seems like I get an Invalid user from port 2019-05-30 10:05:58 Ah, good 2019-05-30 10:06:04 @Cogitri this is weird, but I guess that is because I am trying to use username, pass auth from a textfile instead using a cert and key file, and the script by gentoo community probably expects that 2019-05-30 10:06:25 that would explain why manual run works and why the default openvpn script fails 2019-05-30 10:07:38 thanks for help 2019-05-30 11:44:04 How dodgy is it to add the edge repositories to Alpine 3.9? 2019-05-30 11:45:13 jackwilsdon: depends on what you're gonna do. I only use them if it's just to add a few packages. 2019-05-30 11:45:45 you'll likely run into some bugs 2019-05-30 11:45:54 alpine_fan_01 : clandmeter, another user including me says that default openvpn script fail to run as server mode. 2019-05-30 11:46:03 Makes sense, I was just curious about downloading the latest version of Docker from edge 2019-05-30 11:46:30 (although the latest Docker on 3.9 is from January this year so it's not exactly outdated) 2019-05-30 11:47:18 openvpn.init 2019-05-30 11:47:26 tmhoang: please report it on bugs.a.o 2019-05-30 11:47:37 ok 2019-05-30 11:47:53 jackwilsdon: you probably won't run into problems if you just apk add docker@edge 2019-05-30 11:48:19 Will try it out, thanks :) 2019-05-30 11:48:55 take a look at this: https://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management#Repository_pinning 2019-05-30 11:53:39 is there any webbrowser for x86? 2019-05-30 11:55:16 Looks like Firefox is still available but it's not on apk 2019-05-30 11:55:20 https://ftp.mozilla.org/pub/firefox/releases/67.0/linux-i686/ 2019-05-30 11:55:30 You'd have to download it and just put it somewhere and run it 2019-05-30 11:56:14 Looks like midori is available on the community repositories too 2019-05-30 11:56:30 Halamix2, https://pkgs.alpinelinux.org/package/v3.9/community/x86/midori 2019-05-30 11:57:10 thanks 2019-05-30 11:57:34 Midori does have some quirks IIRC but it does work well enough to be used 2019-05-30 12:12:14 jackwilsdon: Pretty sure that won't run due to it being compiled against glibc and not musl 2019-05-30 12:22:07 Oh interesting, you mean it won't run _at all_? It seems to depend on musl on 3.9 but not on edge :/ 2019-05-30 12:22:08 Weird 2019-05-30 12:22:44 I mean that the FF tarball from Mozilla won't work, sorry if I wasn't clear 2019-05-30 12:22:49 Ahh sorry was confused 2019-05-30 12:22:51 Midori will run 2019-05-30 12:22:55 Yeah that makes sense 2019-05-30 12:23:00 Didn't think about that 2019-05-30 13:00:07 Is there a CLI tool to find out information about an apk file without me having to extract it? 2019-05-30 13:00:29 tar? 2019-05-30 13:00:59 >without me having to extract it 2019-05-30 13:01:25 I want to read the arch from the .PKGINFO but I've got a bunch of files and was wondering if the APK tool had a way to do it 2019-05-30 13:04:02 apkinfo() { tar -xOf "$1" .PKGINFO ; } 2019-05-30 13:04:29 for f in *.apk; do echo -ne "$f: "; tar xf $f .PKGINFO -O | grep arch; done 2019-05-30 13:04:32 Yep just wrote that :P 2019-05-30 13:04:51 Cheers tcely 2019-05-30 17:35:59 #ansible 2019-05-30 20:33:27 Can someone sell me on putting alpine on the pi instead of raspbian? 2019-05-30 20:33:39 I guess it's only if I want to use it, but I have smallbrain, I've decided. 2019-05-30 20:33:43 <_ikke_> apk 2019-05-30 20:33:45 <_ikke_> :P 2019-05-30 20:35:57 _ikke_: hhahahhahah 2019-05-30 20:36:35 if it was easier to do 'regular' sys installs with alpine on rpi, i'd be using it all the time 2019-05-30 20:37:27 <_ikke_> What is preventing us from providing that? 2019-05-30 20:41:46 its pretty fast on the rpi ^^ 2019-05-30 20:42:21 _ikke_: setup-disk probably need learn to recognize mmcblk devices as target for sys install 2019-05-30 20:43:28 <_ikke_> other distro's just provide a tarball you can extract on the SD card 2019-05-30 20:44:08 that's what I do, and flash boot loader, u-boot 2019-05-30 20:45:34 Armbian provide gziped complete FS, i.e. boot loader, partition table and rest. just gunzip and `dd` image to card 2019-05-30 20:46:32 but they are target these devices only 2019-05-30 20:50:47 _ikke_: ask the wiki 2019-05-30 21:13:59 Hi 2019-05-30 21:14:09 which is the status of the firefox package? 2019-05-30 21:14:38 i see it is only on edge and not on the normal repos 2019-05-30 21:14:58 is it possible to have it packaged for ARM64? 2019-05-30 21:16:34 <_ikke_> afaik, the problem is Rust 2019-05-30 21:17:00 <_ikke_> which is only available for x86_64 on alpine atm 2019-05-30 21:17:17 Yup, Rust is x86_64 only right now 2019-05-30 21:18:27 ok, so, i assume it is an upstream issue for rust to support ARM64, isn't it? 2019-05-30 21:18:43 That should be fixed in the near future though 2019-05-30 21:18:46 <_ikke_> In combination with Musl, it is 2019-05-30 21:18:50 No, Rust supports aarch64 just fine 2019-05-30 21:19:16 With LLVM7, yes 2019-05-30 21:19:32 <_ikke_> ah, LLVM8 should fix these issues? 2019-05-30 21:19:45 Yup 2019-05-30 21:20:08 That's why I'm so keen on getting it merged so I can work on Rust 1.35 with support for other arches :) 2019-05-30 21:20:37 <_ikke_> heh 2019-05-30 21:20:47 then the next question is: which is the status of llvm8? 2019-05-30 21:21:32 <_ikke_> https://github.com/alpinelinux/aports/pull/7500 2019-05-30 21:23:13 oh, that looks like inminent! 2019-05-30 21:23:27 thank you all for the clarifications 2019-05-30 22:13:59 When I run apk add docs, everything runs fine until Executing mdocml-apropros-1.14-3-r0.trigger . Not sure what it's doing, but it manages to exhaust my total available VM on this relatively low memory system.(~800 MiB hardware memory, no swap) 2019-05-30 22:15:27 i'll drop in swap and hope for the best, but that's pretty impressive on that trigger's part 2019-05-30 22:18:48 winny, just wait until "makewhatis" do the job 2019-05-30 22:19:16 i think it's completely locked up, nothing refreshes on screen, but yes, i'll try again with some swap 2019-05-30 22:19:32 winny, juuuuust wait :) 2019-05-30 22:19:40 can take some time when doing it first time 2019-05-30 22:21:42 winny, it doing this: http://mandoc.bsd.lv/man/makewhatis.8.html 2019-05-30 22:21:56 I mean: fluxbox, conky, Xorg itself are locked up, and when i switch to a virtual console it takes several minutes 2019-05-30 22:22:01 so i need more VM 2019-05-30 22:28:06 yep that did the trick, funny how that out of all things brought this box to its knees 2019-05-31 03:12:42 Hello folks. I'm trying to build Dockerized firefox running alpine on my main Alpine OS. I use firefox with ALSA sound on my Desktop fine with sound, but I can't get sound through the container, even when using --privileged. I'll post my Dockerfile in a second. Anyone have any ideas? 2019-05-31 03:13:26 https://pastebin.com/raw/T5UJ8EyX 2019-05-31 07:05:48 So I just lost electricity and my server rebooted an I had a message that my ZFS repaired some blocks, should I worry or everything is normal ? 2019-05-31 07:11:11 KH405: although I don't use ZFS, some filesystems report changes like that after running fsck during boot, it's nothing to worry about. If it failed to repair some blocks it would tell you 2019-05-31 07:12:43 Aight, thanks man! 2019-05-31 07:12:53 First time I had a power loss since I have my server 2019-05-31 07:24:38 No worries! yeah power failures aren't fun 2019-05-31 07:35:37 mps: what's the final conclusion? 2019-05-31 08:17:21 yunfan: final conclusion? about what? 2019-05-31 09:03:20 mps: you mentioned you will try to test for the boards 2019-05-31 09:04:19 aha, yes I tested it on two different boards. A10 and A20 and kernel doesn't work 2019-05-31 09:04:43 u-boot works, load kernel but that's all 2019-05-31 09:05:55 didn't had enough time to check why, probably some modules are missing in modloop or in kernel boot parameters 2019-05-31 09:22:01 anyone using alpine on desktop / laptop ? Is there an initiative already to facilitate post-install configuration ? (dotfiles repos, ansible playbooks, ...) 2019-05-31 09:22:24 I use Alpine on my laptop, yes 2019-05-31 09:23:04 But why would you need post-install config? At least I basically just put NetworkManager and GNOME on the thing and it just werks 2019-05-31 09:23:22 And with other DEs I imagine it to be about the same and with WMs you have to setup your own thing anyway 2019-05-31 09:24:01 yeah ? I'm not using gnome, i like i3, which is pretty minimal. And i need fonts, etc. The list is not "1 or 2 things" 2019-05-31 09:25:50 but yes.. that's probably too specific 2019-05-31 09:25:59 Yes, but that falls under "with WMs you have to setup your own thing anyway", doesn't it? 2019-05-31 09:26:29 If you want preconfigured WMs you can just grab existing dotfiles from users of other distros, no? 2019-05-31 09:27:36 is alpine linux a gentoo fork? 2019-05-31 09:28:15 No 2019-05-31 09:28:37 dmcpoll: no 2019-05-31 09:29:04 so you folks have openrc but portage stuffs are missing? 2019-05-31 09:30:17 dmcpoll: it is not present 2019-05-31 09:53:52 mm maybe an alternative to doing https://wiki.alpinelinux.org/wiki/Setting_up_ZFS_on_LUKS will be using ZFS's native encryption 2019-05-31 09:53:53 dmcpoll: Alpine is binary distribution not like Gentoo 2019-05-31 09:54:17 tghttps://www.phoronix.com/scan.php?page=news_item&px=ZFS-On-Linux-0.8-Released 2019-05-31 09:54:21 err https://www.phoronix.com/scan.php?page=news_item&px=ZFS-On-Linux-0.8-Released 2019-05-31 09:54:51 though i wonder if that will be a problem 2019-05-31 09:54:53 https://www.phoronix.com/scan.php?page=news_item&px=NixOS-Linux-5.0-ZFS-FPU-Drop 2019-05-31 09:55:32 dngray: Set that up yesterday :) 2019-05-31 09:55:35 Works great (if you don't typo on the password creation like I did lol) 2019-05-31 09:55:37 my guess would be the alpinelinux kernel wouldn't support it 2019-05-31 09:55:56 Cogitri: did you use dmcrypt or did you use zfs 0.8's encryption 2019-05-31 09:56:29 Asking the password in the initramfs doesn't work yet though, so you'll have to manually call `zfs load-keys -a` and then `mount -t zfs rpool/ROOT/alpine /sysroot` 2019-05-31 09:56:32 some people suggested proxmox, but yeah this is just a small home server 2019-05-31 09:56:34 I use ZFS' native encryption 2019-05-31 09:56:39 maybe 2-3 VMs 2019-05-31 09:56:47 Why wouldn't it? 2019-05-31 09:57:13 Cogitri: what performance do you get read/write 2019-05-31 09:57:36 by support it i meat the patch 2019-05-31 09:57:41 *meant 2019-05-31 09:57:59 Well, I have an NVMe SSD, so I don't think that those numbers will help you? 2019-05-31 09:58:04 "the patch"? 2019-05-31 09:58:09 You mean the SIMD one? 2019-05-31 09:58:13 yeah 2019-05-31 09:58:42 Those aren't in, no 2019-05-31 09:59:01 https://github.com/NixOS/nixpkgs/pull/61076 2019-05-31 09:59:24 It doesn't mean that it doesn't work, it just slower 2019-05-31 09:59:34 yeah quite a bit slower 2019-05-31 09:59:50 A NixOS developer reports that the functions no longer exported by Linux 5.0+ and previously used by ZoL for AVX/AES-NI support end up dropping the ZFS data-set encryption performance to 200MB/s where as pre-5.0 kernels ran around 1.2GB/s. 2019-05-31 10:00:05 i'm sure you'd notice that on your SSDs 2019-05-31 10:00:28 Cogitri: what storage does your server have 2019-05-31 10:00:44 i was thinking of buying 6x6TB https://www.wd.com/products/internal-storage/wd-red.html 2019-05-31 10:01:07 Ah, on my server I use FreeNAS :) 2019-05-31 10:01:31 That thing has 2x6TB SeaGate Ironwolf and 2x2TB WD Red 2019-05-31 10:01:35 well a NAS is sort of secondary 2019-05-31 10:01:44 Can definitely recommend WD Red over the Ironwolfs 2019-05-31 10:01:46 the server is just a machine that does a bunch of things 2019-05-31 10:02:17 https://www.vueville.com/home-security/cctv/nvr/seagate-ironwolf-vs-wd-red-nas-hard-drives/ 2019-05-31 10:02:49 huh i thought they were 7200rpm 2019-05-31 10:04:14 apparently not. 2019-05-31 10:04:30 Cogitri: maybe i should go with WD blacks 2019-05-31 10:05:13 might just do that 2019-05-31 10:05:31 5yr warranty vs 3 too 2019-05-31 10:07:31 might be worth doing that if i run VMs off it 2019-05-31 10:08:15 expensive though :) 2019-05-31 10:08:39 red aren't really that much chaper 2019-05-31 11:14:13 to run VMs i would use SSD 2019-05-31 11:27:19 dngray: Also, the Blacks are pretty loud and power hungry 2019-05-31 11:28:48 Cogitri: that sentence sounds a bit wrong 2019-05-31 11:29:16 danieli: I brought that to attention already 2019-05-31 11:29:19 s/Blacks/WD Black HDDs/ then 2019-05-31 11:37:14 context matters, no matter how much postmodernists disagree :) 2019-05-31 11:47:49 does alpine support selinux? 2019-05-31 11:48:07 (and kindly a reference to alpine's selinux/hardened page?) 2019-05-31 11:48:26 is there something like adelie's gcompat here? 2019-05-31 11:49:37 alpine used grsecurity for a time, but I don't think there's another officially supported mechanism like that now 2019-05-31 11:49:54 gcompat, yes 2019-05-31 11:50:16 dmcpoll: as someone who backported selinux to one of debian release which didn't had it, I think it is not worth efforts 2019-05-31 11:50:18 it was an unofficial port of grsec, not actually grsec 2019-05-31 11:50:41 nothing prevents you from using selinux, apparmor or something else like that, it just has no official system wide support 2019-05-31 11:51:24 so you don't have specific packages for selinux policy modules and such? 2019-05-31 11:52:18 that'd be correct 2019-05-31 11:52:36 danieli, well, it did quack like a duck 2019-05-31 11:53:48 you've already been using openrc, so why don't you just borrow all the selinux modules from gentoo as well? 2019-05-31 11:53:57 mallards and pekins are both ducks, but still different types 2019-05-31 11:54:25 dmcpoll: because someone needs to maintain them 2019-05-31 11:56:56 mps: selinux is more than security. it could also be about namespaces, say polyinstantiated directories 2019-05-31 11:57:23 and much more 2019-05-31 11:58:02 I prefer Apparmor for its simplicity 2019-05-31 12:01:07 Does Xfce work for anyone after the recent upgrade to 4.14_pre1? It just boots to a black screen for me 2019-05-31 12:22:59 greetings! 2019-05-31 12:23:14 > Cogitri ╡ [07:32:52] dngray: Also, the Blacks are pretty loud and power hungry 2019-05-31 12:23:20 ah yeah i hadn't thought of that 2019-05-31 12:23:26 the blacks i do have are offline 2019-05-31 12:29:35 Cogitri: might get another solar panel :) 2019-05-31 12:29:39 that way i won't feel bad 2019-05-31 12:32:20 As long as you have good batteries to power the thing :P 2019-05-31 12:33:33 it's on a UPS as well 2019-05-31 12:33:44 the power isn't that much 2019-05-31 12:33:48 https://documents.westerndigital.com/content/dam/doc-library/en_us/assets/public/western-digital/product/internal-drives/wd-red-hdd/data-sheet-western-digital-wd-red-hdd-2879-800002.pdf 2019-05-31 12:33:54 https://www.wd.com/content/dam/wdc/website/downloadable_assets/eng/spec_data_sheet/2879-771434.pdf 2019-05-31 12:34:17 if you look at the 6TB idle it's 0.8 2019-05-31 12:34:24 vs 1.0 2019-05-31 12:34:49 and for read it's 8.8 on the red and 9.1 on the black 2019-05-31 12:35:25 oops that's the 8TB one 2019-05-31 12:35:32 4.8 RW on the red 2019-05-31 12:35:49 vs 9.1 on the black yeah 2019-05-31 12:36:00 but the black is 7200rpm vs 5400rpm so that makes sense 2019-05-31 12:36:59 idle is quite different 3.1W on the 6TB red vs 7.1W on the 6TB black 2019-05-31 12:37:41 i read somewhere not to use VMs stored on on a zpool 2019-05-31 12:37:57 not sure if it was this channel or somewhere else someone said about that 2019-05-31 12:39:35 Oh? 2019-05-31 12:39:38 I run a Fedora 30 VM on my pool, it's super comfy on FreeNAS 2019-05-31 12:39:49 oh okay 2019-05-31 12:39:52 (Note: My root pol, so It's on a SSD) 2019-05-31 12:40:00 yeah 2019-05-31 12:40:10 they said not to do it on plain old hdds, use a ssd for the VMs 2019-05-31 12:41:31 Yes, otherwise your HDD will have a hard time spinning down :) 2019-05-31 12:41:38 then again this time 2019-05-31 12:41:46 im thinking i won't build a server out of consumer parts 2019-05-31 12:41:56 although it's only for small business, and its onsite 2019-05-31 12:42:10 by small business i mean 3 people practically 2019-05-31 12:42:23 im thinking i might try to get a HP ProLiant ML350 2019-05-31 12:42:48 with ECC RAM for the ZFS 2019-05-31 12:42:59 somewhere i read ECC RAM is a good idea 2019-05-31 12:43:19 ECC is never a bad idea :) 2019-05-31 12:43:24 I personally wouldn't with HP, but sure 2019-05-31 12:43:41 I had a Microserver Gen8 and that worked reasonably well 2019-05-31 12:43:44 i'm not sure if there's really a lot to gain from going with HP 2019-05-31 12:44:42 i was thinking of going with one of those new ryzen 3 cpus 2019-05-31 12:44:48 I personally don't like HP a lot (even though I have a Spectre X360 as my daily driver) 2019-05-31 12:44:58 Yup, great idea 2019-05-31 12:45:05 epyc is way too overkill for me 2019-05-31 12:45:07 Anything but the APUs support ECC, so that's good :) 2019-05-31 12:45:09 as nice as it would be 2019-05-31 12:45:16 I run a Ryzen 2600 in my NAS thingie 2019-05-31 12:45:23 yeah 2019-05-31 12:45:43 nice new 7nm AMD cpu would be nice 2019-05-31 12:45:58 im kind of a bit sick of intel and their hyper threading vulnerabilities of late 2019-05-31 12:46:20 i have to wonder whether AMD has some hidden vulnerabilities down there 2019-05-31 12:46:50 Cogitri: maybe a Ryzen 5 3600 2019-05-31 12:47:12 the 3600X will probably be overkill 2019-05-31 12:47:45 i suppose i could get that 2019-05-31 12:47:52 will be waiting for reviews 2019-05-31 12:47:57 but i think both of those only take DDR4 2019-05-31 12:48:34 i guess it's really more up to the board 2019-05-31 12:49:06 i have someone else that looks at reviews for me so i don't have to :) 2019-05-31 12:52:24 Cogitri: thinking of going with the Fractal Design Define XL R2 https://www.fractal-design.com/home/product/cases/define-series/define-xl-r2-black-pearl 2019-05-31 12:54:34 it's refreshing to see something without stupid lights 2019-05-31 12:57:11 i have to admit i would have preferred https://www.newegg.com/p/2AM-0020-00063 but so hard to get 2019-05-31 12:58:39 I have mine in a BeQuiet Silentbase 800 2019-05-31 12:58:47 But I suppose both have plenty airflow 2019-05-31 12:58:47 we have #alpine-offtopic channel for discussion not related to alpine 2019-05-31 12:59:03 Overclocking a server doesn't seem wise :) 2019-05-31 12:59:16 good point we moved on from ZFS on alpine :P 2019-05-31 13:00:33 dngray: np, but for people who are here to help alpine specific question could be annoying to follow unrelated talks 2019-05-31 13:27:40 looks like selinux on gentoo/musl is hell a lot of trouble. so how does the situation look like here on alpine? 2019-05-31 13:30:44 looks like no one tried it on Alpine 2019-05-31 13:33:05 mps: so is it possible if i use alpine 3.8 or lower version that could be boot success? 2019-05-31 13:33:28 i had spent almost a week on this issue :L 2019-05-31 13:35:30 yunfan: I tried earlier releases but i think they also didn't worked. that was long ago so I forgot exact results 2019-05-31 13:36:39 I simply built kernel with needed drivers and replaced linux-vanilla. that is how I like to use alpine on arm's 2019-05-31 13:37:32 I usually need some patches and out of tree drivers, specific kernel settings etc 2019-05-31 13:39:17 I started to test linux-vanilla just to find what is the reason it doesn't boot on some boards. 2019-05-31 13:41:50 mps: of course i could built, but i dont know which options should i choice, i am worried about some option could blowed myself :D 2019-05-31 13:42:54 dmcpoll: to my knowledge, it hasn't been a priority for the project itself, and i can't remember hearing about anyone else fiddling with it either 2019-05-31 13:43:48 danieli: yes, looks like no one think it is worth effort 2019-05-31 13:45:21 yunfan: look at sunxi kernel site https://sunxi.org/Main_Page 2019-05-31 13:45:38 dngray: I very much like my define xl r2 2019-05-31 13:45:55 there are a lot infos, guides and advices about sunxi boards 2019-05-31 13:45:58 TomJepp: they seem like decent quality cases 2019-05-31 13:46:00 also... 'moved on from zfs'... what to? :) 2019-05-31 13:46:14 oh the conversation moved from that 2019-05-31 13:46:21 we were talking about zfs on alpinelinux 2019-05-31 13:46:29 oh, right :p 2019-05-31 13:46:34 and whether it would be better to use dmcrypt or the encyrption option in 0.8 2019-05-31 13:46:38 I thought you meant there was some new hotness filesystem 2019-05-31 13:46:48 btrfs, strata? 2019-05-31 13:46:54 btrfs makes me sad 2019-05-31 13:47:04 yeah makes me sad too i wish it was better 2019-05-31 13:47:20 https://bcachefs.org/ is hte one I have my eye on at the moment 2019-05-31 13:47:29 Btrfs isn't the new hotness :P 2019-05-31 13:47:30 i wonder what happened to that stratis filesystem redhat was doing 2019-05-31 13:47:34 <_ikke_> Someone said they just lost their data after a hard shtudown 2019-05-31 13:47:41 red hat dropped btrfs, but suse said they're committed to it 2019-05-31 13:47:42 It's uber cool in principle but died twice on me :c 2019-05-31 13:47:58 Cogitri: I corrupted btrfs many times by daring to use compression on a single disk 2019-05-31 13:48:01 ive never used btrfs in raid, i do use it on my archlinux laptop though 2019-05-31 13:49:18 I really would like an in-kernel FS as powerful as ZFS with the same level of reliability 2019-05-31 13:49:26 I think bcachefs is the most likely to get there 2019-05-31 13:49:51 oh i think stratis is just a management tool 2019-05-31 13:49:53 not a filesystem 2019-05-31 13:49:58 https://stratis-storage.github.io/ 2019-05-31 13:50:05 https://www.phoronix.com/scan.php?page=news_item&px=Stratis-1.0-Released 2019-05-31 13:50:18 ah, that's RH's stuff that heavily uses LVM? 2019-05-31 13:51:05 i don't know that it does 2019-05-31 13:51:09 Oh well, ZFS works fine for me and the new encryption is pretty amazing 2019-05-31 13:51:26 i am tempted to use the new encryption as opposed to dmcrypt 2019-05-31 13:52:43 mps: to be honest, i could follow document to do menuconfig, but patching is a job that i cant understand 2019-05-31 13:53:01 https://utcc.utoronto.ca/~cks/space/blog/linux/ZFSEncryptionNotReady mmm i dunno what to make of that though 2019-05-31 13:53:27 maybe i will just go with dmcrypt 2019-05-31 13:53:27 Works well for me 2019-05-31 13:53:30 tried and true 2019-05-31 13:53:46 until you hit a bug ;) 2019-05-31 13:53:57 I'll make a write-up on how to do it in the Wiki latet this day, maybe you want to wait until that? 2019-05-31 13:54:00 yunfan: if you don't need special drivers or patches you don't need to do patching etc.. 2019-05-31 13:54:11 Cogitri: oh awesome i'd appreciate that 2019-05-31 13:54:15 btw do you use anything like raidz2? 2019-05-31 13:54:19 mps: that's the problem, i dont know if i need 2019-05-31 13:54:25 i was thinking of using 6x6TB disks in raid-z2 2019-05-31 13:54:37 yunfan: it is not hard task 2019-05-31 13:54:52 actually arch for arm could work on my board, but it failed to use the builtin ethernet 2019-05-31 13:55:00 i was considering if i could use a usb wifi 2019-05-31 13:55:25 https://wiki.alpinelinux.org/wiki/Setting_up_ZFS_on_LUKS there is this article but that's only mirror 2019-05-31 13:55:51 the idea was i would boot from a alpinelinux usb stick (diskless) 2019-05-31 13:55:53 mps: do you have any model name that recommended to use about usbwifi or use2rj45? 2019-05-31 13:55:53 yunfan: working with ARM require some hacking at the end 2019-05-31 13:56:02 and then mount my pool, and then run some VMs on it 2019-05-31 13:56:06 or off a separate ssd 2019-05-31 13:56:28 Cogitri: it does say at the bottom 2019-05-31 13:56:30 " Todo: Multiple Disk ZFS on LUKS?" 2019-05-31 13:56:42 so maybe someone planned to do that 2019-05-31 13:57:05 mps: i am just user :L 2019-05-31 13:57:07 yunfan: I didn't used usb wifi's, just one two simple usb ethernet adapters 2019-05-31 13:57:29 mps: usb2rj45 is ok to me 2019-05-31 13:57:34 azarus wrote that he's sometimes in here i'm sure of it 2019-05-31 13:57:41 and is apparently on freenode at the moment 2019-05-31 13:58:27 I hope that I will have more time after stable release (3.10) to work on boot Alpine on sunxi boards out of the box :) 2019-05-31 13:59:12 yunfan: in my experience, whatever usb device I put it worked 2019-05-31 14:00:02 mps: ok, i will find a usable alpine version using binary search :D 2019-05-31 14:06:17 apparently though raid 10 with btrfs is safe 2019-05-31 14:06:22 https://www.phoronix.com/scan.php?page=news_item&px=Btrfs-RAID-56-Is-Bad 2019-05-31 14:07:10 yunfan: if you will be here I will inform you of the progress, if there will be some 2019-05-31 14:08:59 dngray: Well, "safe" :P 2019-05-31 14:10:28 though raid 6 is double parity which is what i did 2019-05-31 14:10:40 with my 6 disks, more than 2 would have to fail 2019-05-31 14:20:34 https://old.reddit.com/r/freenas/comments/a90z7h/raidz2_vs_raid10_for_10tb_drives/ecgrk4v/ 2019-05-31 14:20:38 might stay with raidz2 then 2019-05-31 14:21:15 https://old.reddit.com/r/freenas/comments/a90z7h/raidz2_vs_raid10_for_10tb_drives/ech55s5/ 2019-05-31 14:21:19 although that's a good point 2019-05-31 17:19:29 Is there a flag I can pass to apk to just get the version number of an installed package? apk version -q [pkg] gives me just the package name and I can't work out a flag for apk info 2019-05-31 19:38:12 can alpine:edge be used as a daily driver for the laptop? 2019-05-31 19:38:38 how is it compare to arch? 2019-05-31 19:39:17 Alpine generally has less software available in its repos 2019-05-31 19:39:30 I guess just try it 2019-05-31 19:39:56 Cogitri: an e-fried tried it and he is extatic about it 2019-05-31 19:41:03 i am depending on AUR for some stuff... 2019-05-31 19:41:09 mostly because i am lazy 2019-05-31 19:43:00 <_ikke_> edge can be a but unstable though 2019-05-31 19:43:13 <_ikke_> Packages are updated / changed constantly 2019-05-31 19:44:21 I guess it can be compared to Arch testing 2019-05-31 19:44:28 It works fine for me though (well, other than QtWebengine :P)