2022-06-01 00:12:55 pixelkeen: if they tolerate and help me, you'll be just fine 😎 2022-06-01 03:13:58 https://wiki.alpinelinux.org/wiki/ALSA 2022-06-01 03:14:11 i followed this, and configured and enable/dstarted the alsas service 2022-06-01 03:14:21 but typing amixer or alsamixer returns that the host is down 2022-06-01 03:14:26 i did try rebooting 2022-06-01 03:14:52 when i start amixer with an arg pointing to my specific video card it sees it and all its values, cant change any of them though. it sees my hdm iaudio and normal intel hda 2022-06-01 03:15:33 and trying to run supertux or tux racer, just to test it anyways, it says couldnt initalize audioo device 2022-06-01 03:15:41 i am in the audio group, so is root 2022-06-01 03:19:42 pretty lost now ive been tackling this issue all day, i am just trying to get at laest alsa going right now so i have basic audio to work with 2022-06-01 03:28:46 what else acn i try...? all the googlign suggests it should be working 2022-06-01 03:48:54 pipewire is frequently recommended as a placement but its wiki page is scarce outdated and lacking 2022-06-01 03:48:59 i cant figure this audio out 2022-06-01 03:49:10 do i need friggin dbus for this or what 2022-06-01 04:13:36 9,99moose77199,99: you'll need dbus. I followed the wiki but installed pipewire-media-session and uncommented the pipewire-media-session bit at the bottom of /etc/pipewire/pipewire.conf rather than adding wireplumber. 2022-06-01 04:14:44 I think it was Bubblz in here that said it's not even necessary to use wireplumber or edit default conf, but I have not tried that method. 2022-06-01 04:19:18 From 7,99bubblz99,99 "pixelkeen: pipewire wiki instructions are outdated. no config file is needed to use wireplumber or pulseaudio plugin. you just need to use a custom launcher for pipewire. It can be found in /usr/libexec/pipewire-launcher" here's bubblz' solution posted earlier 2022-06-01 05:55:43 Hello, I noticed that many Alpine packages are built for Risc-v but there is not any information about that in the wiki. Is there an effort to port Alpine to risc-v? 2022-06-01 06:06:53 Answering my question: It looks liker there is, the packages are only in edge 2022-06-01 06:24:35 cclifforgg: https://gitlab.alpinelinux.org/alpine/aports/-/issues/12167 https://gitlab.alpinelinux.org/alpine/aports/-/issues/13269 2022-06-01 08:04:04 Hey there ! Someone could explain to me why tkremind (sub of remind) conflict with remind ? 2022-06-01 08:04:08 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/community/remind/APKBUILD 2022-06-01 08:06:08 It looks like the recipe is somehow wrong. remind already provide tkremind 2022-06-01 08:08:23 Okay I'm on it ! 2022-06-01 08:08:45 It seems like the Makefile in src already installs tkremind from the scripts directory to usr/bin 2022-06-01 08:09:06 yap it install the script from builddir instead of pkgdir 2022-06-01 08:10:51 Can I somehow query how a specific package was pulled in with apk? 2022-06-01 08:10:56 I.e., what things depend on it 2022-06-01 08:11:25 Perhaps: apk info --rdepends 2022-06-01 08:43:54 staceee: Have you tried running tkremind? 2022-06-01 08:46:03 rbq: I fixed my issue but now tkremind complaint about missing "mime" pkg 2022-06-01 08:46:56 You need tcl-lib, but that's in testing/, while remind is in community/ 2022-06-01 08:56:38 mhh 2022-06-01 08:58:31 is it something moveable to community ? 2022-06-01 08:59:53 I'd say just include the move of tcl-lib to community in your MR (if you're going to create one with your fixed APKBUILD for remind), and see if it gets approved 2022-06-01 09:00:12 Otherwise tkremind would be broken 2022-06-01 09:00:47 okay 2022-06-01 09:01:56 I would recommend asking maintainer first 2022-06-01 09:07:18 That's probably better, tcllib is actually in unmaintained/ as well, so it has a complicated history...was in community/ under the name tcllib, got moved to unmaintained/, and then re-appeared in testing/ under a new name (tcl-lib) 2022-06-01 09:08:02 !34977 I marked is as draft for the moment 2022-06-01 09:13:51 latest release is 1.21, tck website says it's 1.20 and some lines lower it mentions "The latest version of tcllib, 1.17, includes [...]" 2022-06-01 09:15:15 ACTION sighs 2022-06-01 09:17:21 but it at least seems to be maintained, not like, e.g. agg 2022-06-01 09:39:19 rbq: --rdepends worked, thanks! 2022-06-01 09:41:41 Newbyte: You're welcome :) 2022-06-01 15:12:18 Hi 2022-06-01 15:13:29 I've got a problem running flatpak on alpine, complaining about 'F: Invalid proxy URI', then Glib-GIO:Error ... assertion failed: (data->error_info->best_error) 2022-06-01 15:33:06 do you have proxy_url in your environment 2022-06-01 15:34:32 no, fresh install and no proxy 2022-06-01 15:35:07 env | grep proxy 2022-06-01 15:36:23 no_proxy=localhost 2022-06-01 15:36:26 https_proxy= 2022-06-01 15:36:30 http_proxy= 2022-06-01 15:36:32 ftp_proxy= 2022-06-01 15:37:27 ahh, ok unset http_proxy works 2022-06-01 15:37:44 no idea where http_proxy is being set 2022-06-01 15:40:20 perhaps something in the new install scripts adds it somewhere to /etc/profile*, don't know 2022-06-01 15:57:46 try a frontend. unset http_proxy didn't help for me. also setting up a working proxy fixes the problem for some reason. 2022-06-01 16:12:07 you would have to start the frontend from the same place you `unset`, because it's only unset in the shell you ran it in 2022-06-01 16:12:28 as for what sets it in the first place, no idea 2022-06-01 16:12:57 if you could grep -r proxy /etc/profile* maybe that would have something 2022-06-01 16:41:59 psykose: unset doesn't fix the issue. installing flatpaks with ansible did work thought, that's why i'm assuming using a frontend would help 2022-06-01 16:43:02 if you actually set up a working proxy that'll also work for some reason 2022-06-01 16:44:34 can you post an `env` and subsequent `flatpak ..` in the same shell that fails? 2022-06-01 16:46:03 Hello, I have recently installed alpine on my raspberry pi b, I'd like to get graphics working, preferably sway, though I would also be content with xorg. I tried following the steps at https://wiki.alpinelinux.org/wiki/Sway but upon running sway, I get the following error: Unable to create backendn any DRM deviceackend. In all honesty im not too sure what this means. 2022-06-01 16:47:31 Also, this is my first time trying to use wayland instead of xorg. In the link provided, I wasn't sure about graphics drivers, so I tried with each one, though I don't know much when it comes to graphics and drivers 2022-06-01 16:49:47 Did you re-login after adding yourself to various groups? 2022-06-01 16:50:12 yes 2022-06-01 16:50:43 Can you post full sway log? 2022-06-01 16:52:32 I removed all the packed I installed after I tried a few hours ago. Ill redo everything and post it 2022-06-01 16:52:41 Where would I find the sway log? 2022-06-01 16:52:54 packages* 2022-06-01 16:53:13 psykose: http://0x0.st/oBlR.txt. env with the error from flatpak 2022-06-01 16:53:42 you can run it like dbus-launch sway > sway.log 2> sway.log.2 2022-06-01 16:53:59 also yes http_proxy, ftp_proxy and the rest are set in /etc/profile.d/proxy.sh 2022-06-01 16:54:36 jude: could you run `sway -dV &> log` and then post that log 2022-06-01 16:55:07 bubblz: well.. clearly it's set in that, so now try unsetting both http/https and doing it again 2022-06-01 16:55:14 now that i think about it, i think we had this issue before 2022-06-01 16:55:28 and this looks like a setup-proxy issue 2022-06-01 16:56:06 ah, i see the bug 2022-06-01 16:56:38 it works ofc. dumb me. i unset http_proxy and thought it was just a red herring 2022-06-01 16:57:16 setup-proxy gets called with nothing, this initialises proxyurl= as nothing. then it prompts you for a proxy 2022-06-01 16:57:24 if you just press enter, the response is '' 2022-06-01 16:57:40 then, if it checks if the response is 'none'. of course, this check fails.. 2022-06-01 16:57:48 then it exports proxy_url='' to the proxy.sh 2022-06-01 16:58:08 if you answer literally 'none' it would do nothing, but by default it gives you an empty proxy 2022-06-01 16:58:21 i'll go fix it in setup-* but the actual rollout will take a while 2022-06-01 17:01:07 jude: i think you need kms graphics for wayland to work. you could try something similar to https://wiki.alpinelinux.org/wiki/Radeon_Video. this(https://forum.qt.io/topic/107764/drm-kms-configuration-for-raspberry-pi-4) could give some pointer as to what driver is needed 2022-06-01 17:04:52 It seems be something like that. I can't get ssh working rn because of network issues, but the important part of the log seems to be such: 2022-06-01 17:04:52 (10 seconds later) 2022-06-01 17:04:52 Waiting for a DRM card device 2022-06-01 17:04:52 Found 0 GPUs, cannot create backend 2022-06-01 17:04:54 Failed to open any DRM device 2022-06-01 17:04:56 Unable to create backend 2022-06-01 17:05:34 sure, you have no gpu 2022-06-01 17:05:40 from what i'm reading, you need to enable it in config.txt 2022-06-01 17:05:41 dtoverlay=vc4-kms-v3d 2022-06-01 17:05:50 then you should see a /dev/dri/card* 2022-06-01 17:06:10 check if `vc4` is loaded in `lsmod` after adding that overlay and rebooting 2022-06-01 17:06:29 then for the driver.. 2022-06-01 17:06:58 mesa-dri-gallium has vc4_dri 2022-06-01 17:07:08 as usual everything is just in gallium 2022-06-01 17:07:20 do all of that, verify the card exists in /dev/, then try the say logs again 2022-06-01 17:07:36 sway* 2022-06-01 17:08:06 config.txt is in the first partition of the sd card right? 2022-06-01 17:08:27 yes, i think it should already be mounted as /boot or something, but you would know more than me 2022-06-01 17:08:56 I did a sys install so I think its in /media/mmcblk0p1 2022-06-01 17:08:58 Seems right 2022-06-01 17:14:48 I added it to usrconfig.txt and that worked 2022-06-01 17:18:16 In addition, tty text is now much smaller on my screen, seems screen resolution has been detected, but it don't look too good. is there a way to make it restore the old screen resolution? 2022-06-01 17:18:36 7,99bubblz:99,99 deleting the proxy var export script did the trick for me. 2022-06-01 17:19:45 psykose: sway launches for me now, thanks very much 2022-06-01 17:20:50 Oops missed where you said it worked, bubblz. 2022-06-01 17:24:06 bubblz: this was broken right before the 3.16 release, just a small oversight 2022-06-01 17:24:37 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/merge_requests/69 2022-06-01 17:25:51 jude: to set the framebuffer resolution (i assume pre-graphical), you can use fbset 2022-06-01 17:26:12 e.g. stuff from here https://stackoverflow.com/questions/34904763/linux-framebuffer-set-resolution-correctly 2022-06-01 17:26:14 i assume it works 2022-06-01 17:27:08 if you mean resolution in sway, then that's the usual sway-output stuff 2022-06-01 17:28:06 It is both, so ill look more into using sway, since its my first time on wayland 2022-06-01 17:48:05 for sway you usually have a section like `output DP-1 subpixel bgr mode 2560x1440@164Hz pos 0 0`, and what settings you put depend on `swaymsg -t get_outputs` for the naming/res/frequency 2022-06-01 17:48:14 usually sway picks the correct defaults 2022-06-01 17:57:24 When trying to use fbset, I get the error: 2022-06-01 17:57:24 fbset: ioctl 0x4601 failed: Invalid argument 2022-06-01 18:02:19 what do you pass to it 2022-06-01 18:03:07 -fb /dev/fb0 -g 1920 1200 1920 1200 16 2022-06-01 18:03:37 Trying to manually edit the files doesn't work either. 2022-06-01 18:04:18 can you post just `fbset` output 2022-06-01 18:05:11 at least the config.txt editing of the resolution should 2022-06-01 18:05:48 what did you put in there 2022-06-01 18:06:18 mode "1280x720-0"# D: 0.000 MHz, H: 0.000 kHz, V: 0.000 Hz 2022-06-01 18:06:36 Sorry, about that, accidentally hit enter 2022-06-01 18:07:34 gemoetry 1280 720 1280 720 16 2022-06-01 18:07:34 Reba 5/11,6/5,5/0,0/0 2022-06-01 18:07:34 Timings 0 0 0 0 0 0 0 2022-06-01 18:07:34 Access true 2022-06-01 18:07:34 Endmode 2022-06-01 18:08:34 I didn't edit config.txt, I tried to edit /sys/class/graphics/fb0/mode and "modes" in the same fb0 directory 2022-06-01 18:10:32 How would I edit resolution in the config.txt? 2022-06-01 18:15:54 Could I edit resolution by adding video=HDMI-A-1:1920x1200M@60 ? 2022-06-01 18:16:12 To cmdline.txt 2022-06-01 18:20:48 lol no, that prevents output to the screen 2022-06-01 18:40:29 hdmi_cvt and hdmi_group 2022-06-01 18:40:32 https://www.raspberrypi.com/documentation/computers/config_txt.html 2022-06-01 18:40:54 er, _mode. but really just read all of it 2022-06-01 21:50:54 i am once again here to whine about my no audio problems 2022-06-01 21:51:03 this time with more information 2022-06-01 22:03:26 what kind of error are you getting? 2022-06-01 22:05:26 sorry 2022-06-01 22:05:32 i got distractd 2022-06-01 22:05:37 ill pulll it up 2022-06-01 22:06:51 yeah, so i have configured alsa according to the wiki, have tried sndio as well with similar results 2022-06-01 22:07:11 basically, when i try to use amixer or aucat or acat or similar, it says "host is down" or "no devices found" 2022-06-01 22:07:25 but the thing is i can open amixer on my specific sound card, it will open 2022-06-01 22:07:38 and i can see all the inputs and outputs and their volume and i can toggle their volume 2022-06-01 22:07:43 mute, change it 2022-06-01 22:08:01 but no audio, and actually just trying to open amixer returns - host is down, and any kind of other audio test errors out similarly 2022-06-01 22:08:09 so heres a sniplit of dmesg 2022-06-01 22:08:14 https://pastebin.com/5X3iMSac 2022-06-01 22:08:25 https://pastebin.com/ZuVtB64d lshw 2022-06-01 22:08:32 aaand "aplay -l" 2022-06-01 22:08:34 https://pastebin.com/pfEJkHgN 2022-06-01 22:09:07 starting the sndio daemon it sees and enumerates all my sound cards inputs and outputs, and aplay can see them all too 2022-06-01 22:09:20 i am not trying to use hdmi audio for the record 2022-06-01 22:09:37 i am in the audio group, so is root 2022-06-01 22:09:41 the alsa deamon is running 2022-06-01 22:10:36 as for hardware, its just a ryzen system. b450 tomahawk, ryzen 2600, audio seems to just be snd_hda_codex_realtek 2022-06-01 22:10:50 i did set my default sound device to my sound card, which is 1 2022-06-01 22:10:57 which is an integrated sound device 2022-06-01 22:13:50 i just blacklisted the module for hdmi audio, goinjg to reboot it and try it out. brb 2022-06-01 22:14:03 i do have pipewire installed but it had similar issiues so it is not in use. it is not running 2022-06-01 22:14:22 i have no experience with alsa on alpine. did you try pipewire? 2022-06-01 22:16:05 works like a charm for me. even the bluetooth mic and sbc-xq on my headphones 2022-06-01 22:17:16 yes 2022-06-01 22:17:33 how did that go? 2022-06-01 22:17:33 i did, although the wiki pages are a bit outdated it seems. pipewire relies on alsa right? 2022-06-01 22:17:40 it couldnt find any devices 2022-06-01 22:17:59 also tried pulseaudio and sndio 2022-06-01 22:18:48 i had hoped to abvoid pipewire anyways i didnt use a distro like alpine just to end up being reliant on more poterringware like dbus 2022-06-01 22:18:57 poeteringware* 2022-06-01 22:19:08 dont see OSS in the repos, guess i cant try that 2022-06-01 22:19:31 Dbus and pipewire are not created by Poettering 2022-06-01 22:19:37 i know 2022-06-01 22:19:47 i only call it that for being similar in nature 2022-06-01 22:20:55 never said pipewire was , have no issue with pipewire, its good software, my problem is with dbus, but i would be willing to deal with it if pipewire is really waht it takes to get working audio 2022-06-01 22:21:34 i can give it another go ill probably get more support here if i do 2022-06-01 22:21:36 Afaik many desktop software utilizes dbus, so choices here are limited, alas 2022-06-01 22:21:52 oh i know 2022-06-01 22:21:59 i dont use a lot of those anyways 2022-06-01 22:22:20 not a lot of gtk3 programs i use or need 2022-06-01 22:22:33 here's a funny thread about dbus on freebsd forums https://forums.freebsd.org/threads/why-do-we-need-d-bus.78868/ 2022-06-01 22:23:10 moose771: do you see your sound device in lspci? 2022-06-01 22:23:34 uh, i didnt see lspci in the alpien repos, because i dont have it installed. what is the pkg called? 2022-06-01 22:23:49 pciutils 2022-06-01 22:23:51 yeah like every freebsd desktop package pulls in pulse and dbus even though the people on it advocate avoiding it 2022-06-01 22:24:00 freebsd is a mess these days 2022-06-01 22:24:05 why though? 2022-06-01 22:24:23 why is freebsd a mess is a topic id rahter not spam about right now iam very easily distracted 2022-06-01 22:24:39 check out this guide https://wiki.gentoo.org/wiki/ALSA 2022-06-01 22:24:39 sec 2022-06-01 22:25:15 thanks 2022-06-01 22:25:26 i can look at it, though the guide is probably written with systemd in mind 2022-06-01 22:25:42 you would be wrong 2022-06-01 22:25:45 gentoo supports openrc 2022-06-01 22:25:46 ah, okay 2022-06-01 22:26:01 cool, need to refer to it for alpine troubleshooting more often then 2022-06-01 22:27:45 Btw you can use apk search cmd:lspci to find packages with this command 2022-06-01 22:28:52 https://dpaste.com/64BH2GP4B.txt 2022-06-01 22:28:53 thanks 2022-06-01 22:29:06 is there a similar one for finding the name of a binary of ain installed package? sometimes its not obvious 2022-06-01 22:29:16 lets see 2022-06-01 22:29:56 1e:00.3 Audio device: Advanced Micro Devices, Inc. [AMD] Family 17h (Models 00h-0fh) HD Audio Controller i hope thats not hdmi audio its referring to 2022-06-01 22:30:27 it's not, it's the same thing i have 2022-06-01 22:31:36 alright 2022-06-01 22:31:58 i figured it was at least being recognized by the kernel since of all the spew a bout snd_hda_codec_realtek in dmesg and similar 2022-06-01 22:32:34 apparently snd_hda_codec_hdmi isnt the right thing to block for a card as ancient as my hd 6450 2022-06-01 22:33:11 not sure if this is relevant https://www.alsa-project.org/wiki/Matrix:Module-interwave 2022-06-01 22:33:49 custom kernel 2022-06-01 22:33:53 testing guide https://www.alsa-project.org/wiki/SoundcardTesting 2022-06-01 22:34:05 what does aplay -vv file.wav say? 2022-06-01 22:34:13 still learning grub, not sure how to boot another kernel from it it, been using syslinux for the last decade. one sec 2022-06-01 22:34:41 you're running a custom kernel? 2022-06-01 22:34:49 no no 2022-06-01 22:35:06 i'm on lts, the default one, just that the alsa wiki page you sent for the interwave sound card states id have to 2022-06-01 22:35:10 aplay: main:831: audio open error: Host is down 2022-06-01 22:35:15 it doesn't say that 2022-06-01 22:35:25 but you don't have this card anyway pretty sure, so it's not very relevant 2022-06-01 22:35:31 ah, okay. my bad 2022-06-01 22:36:55 can you post uhh 2022-06-01 22:36:58 ls -la /dev/snd 2022-06-01 22:37:56 https://dpaste.com/HD2RRPTHC 2022-06-01 22:37:58 yeah 2022-06-01 22:38:06 one moment, door 2022-06-01 22:39:10 now post `groups`. also check if the modules looks something like https://img.ayaya.dev/ORfPUIXlBW8m 2022-06-01 22:41:09 back. 2022-06-01 22:41:32 sorry about that, sure, thanks for helping by the way 2022-06-01 22:41:47 groups as normal user: spinner wheel audio input video 2022-06-01 22:42:03 groups as root 2022-06-01 22:42:06 root bin daemon sys adm disk wheel floppy audio dialout tape video 2022-06-01 22:42:34 looks fine so far, hm 2022-06-01 22:42:45 yeah thats the wall i ran into as well 2022-06-01 22:43:07 i think i will have a read of that gentoo article then 2022-06-01 22:44:58 yeah, sadly i have no idea how to debug alsa stuff 2022-06-01 22:45:38 pipewire still relies on alsa at its core right 2022-06-01 22:45:41 if you want to try pipewire it's a bunch of stuff; namely, start a session in dbus-run-session -- anything, then start pipewire; pipewire -c pipewire-pulse.conf; wireplumber 2022-06-01 22:45:47 cus it doesnt seem to be a alsa issue 2022-06-01 22:45:48 everything on linux relies on alsa in the kernel 2022-06-01 22:46:00 as for userspace alsa, pipewire-alsa relies on alsa-lib, sure 2022-06-01 22:46:17 but that's only to support 'alsa clients' 2022-06-01 22:46:22 ah 2022-06-01 22:46:43 well, i can give pipewire another roll 2022-06-01 22:47:08 you should run those 3 components in separate shells instead of autostarting them somewhere, so you can see anything in the logs if they output something 2022-06-01 22:47:26 alright 2022-06-01 22:47:50 psykose: there's a launcher script in /usr/libexec/pipewire launcher. i think it's preferrable to use that now 2022-06-01 22:48:07 then once they're started you can look at `wpctl status` for an overview of stuff, and manage the devices with any pulse frontend (funnily, nobody has really made a native pipewire frontend yet that ive seen), i use pulsemixer 2022-06-01 22:48:23 bubblz: yeah, and the point is to run it yourself so you see the output better, not just start them all in this case 2022-06-01 22:49:07 it would still work, but it literally just does that anyway, so 2022-06-01 22:49:09 starting some more do i need to be running a gui for this? 2022-06-01 22:49:19 i don't actually know, but i don't think so 2022-06-01 22:49:23 just being inside a dbus session 2022-06-01 22:49:25 also i didn't have to edit any /etc/pipewire/pipewire.conf in the edge at the very least 2022-06-01 22:49:28 alright 2022-06-01 22:49:36 i am on edge 2022-06-01 22:49:41 you don't have to edit pipewire.conf if you use the launcher, no 2022-06-01 22:49:47 or start the stuff yourself 2022-06-01 22:49:50 brb giving this a try 2022-06-01 22:52:21 i guess ill have to grab an interative tui wav/mp3 player to test this with then 2022-06-01 22:52:29 cant very well dbus-launch-session acat 2022-06-01 22:53:38 dbus-run-session -- ash 2022-06-01 22:53:46 it's in dbus-x11 2022-06-01 22:54:27 and i start my wm like that, just `exec dbus-run-session -- sway` 2022-06-01 22:55:41 ah 2022-06-01 22:56:05 sorry, im an average person which means i happen to be retarded 2022-06-01 22:58:10 where is pulse.conf? 2022-06-01 22:59:58 you shouldn't need to touch it at all 2022-06-01 23:01:11 also afaik there is usually client.conf and daemon.conf in /etc/pulse, but in this case only client would matter, but there's nothing to really change 2022-06-01 23:01:24 i just needed to change to the directory for pipewire -c conf to work 2022-06-01 23:01:47 that doesn't sound right 2022-06-01 23:02:06 -c pipewire-pulse.conf should work from anywhere 2022-06-01 23:02:10 it checks the config dirs first 2022-06-01 23:02:13 perhaps i misinterepted, oh 2022-06-01 23:02:27 (perhaps you didn't use the full name) 2022-06-01 23:02:46 https://dpaste.com/7EHKM8SDD.txt 2022-06-01 23:03:09 you probably forgot to install pipewire-pulse 2022-06-01 23:03:19 and perhaps wireplumber too, and perhaps pipewire-alsa 2022-06-01 23:04:47 Install all an uninstall unneeded parts 2022-06-01 23:05:12 definitely want all of those 2022-06-01 23:05:21 i i have those installed 2022-06-01 23:05:29 ah, actually didnt have pipewire-pulse 2022-06-01 23:05:31 just grabbed it 2022-06-01 23:06:09 https://dpaste.com/4HFGJEQLK.txt starting to lose track of things with my tiny brain 2022-06-01 23:06:56 you also need to `service start dbus` 2022-06-01 23:07:06 and you also need to run `pipewire` with no arguments 2022-06-01 23:07:45 got to see what th edamn landlord wants. oh i see, that alway s errored out 2022-06-01 23:07:48 one sec 2022-06-01 23:19:24 sorry about that 2022-06-01 23:19:26 back to work 2022-06-01 23:19:52 ~ 2022-06-01 23:20:04 i'd recommend just doing it in order 2022-06-01 23:20:13 start pipewire, ctrl-z, bg 2022-06-01 23:20:15 you know the usual 2022-06-01 23:20:25 then pulse, then wireplumber 2022-06-01 23:20:33 it was my understanding that i was to do it in seperate shells 2022-06-01 23:20:51 sure, because that's how i do it, as i'm in a gui 2022-06-01 23:20:56 it matters not :p 2022-06-01 23:21:10 ill use tmux then 2022-06-01 23:21:26 yeah anyways, i havent been able to just start pipewire alone yet 2022-06-01 23:21:38 more logs for me 2022-06-01 23:22:00 https://dpaste.com/7V6AD2U4H.txt 2022-06-01 23:22:13 you have no XDG_RUNTIME_DIR 2022-06-01 23:22:20 well ,i dont have a gui installed 2022-06-01 23:22:22 or xorg 2022-06-01 23:22:31 easy fix for now is just export XDG_RUNTIME_DIR=/tmp 2022-06-01 23:22:38 the gui part is not related :) 2022-06-01 23:22:54 thought xdg was apart of the freedesktop stuff 2022-06-01 23:22:55 alright 2022-06-01 23:23:04 sure, but that stuff is part of a login session 2022-06-01 23:23:09 i have one merely by logging in 2022-06-01 23:23:27 if you begin to lose your patience with my low iq do let me know 2022-06-01 23:23:46 well, pipewire starteda nd didnt crash right away, at least 2022-06-01 23:23:51 sikk 2022-06-01 23:24:05 did get some spew though 2022-06-01 23:24:12 ill post it while i spin up some shells and run the other cmds 2022-06-01 23:24:20 if it's the rtkit shit, doesn't matter 2022-06-01 23:24:24 onto pulse 2022-06-01 23:25:36 ah okay 2022-06-01 23:26:05 i really should just haul it in here next to me.. 2022-06-01 23:26:43 don't tell me you get up and go back and forth between each message 2022-06-01 23:27:02 oh no. 2022-06-01 23:27:06 im ssh'd in 2022-06-01 23:27:11 i get up to do actual audio testing 2022-06-01 23:27:19 its a dual motherboard case, its very heavy 2022-06-01 23:27:23 i dont want to move it 2022-06-01 23:27:42 makes sense 2022-06-01 23:27:55 but see my cat recently molested and broke my second monitor which was my main pc testing moniotor 2022-06-01 23:28:09 also note that if you open other tmuxes, they will have a separate shell 2022-06-01 23:28:13 which would mean they are not in the dbus session 2022-06-01 23:28:22 ah. okay, 2022-06-01 23:28:23 so you have have to open tmux in the place you have the dbus session, then go to the others 2022-06-01 23:28:27 i will pg pulse and do it there 2022-06-01 23:28:31 oh, i did 2022-06-01 23:28:33 sure, works 2022-06-01 23:30:39 https://dpaste.com/GZAG35RGZ have them all started, running. going to walk to the machine and play some audio 2022-06-01 23:31:57 you might have to also add pulseaudio-utils for pulse to work and restart the pulse part (forgot about pactl), but playing just alsa should work anyway 2022-06-01 23:32:38 so, to be clear, im gonna need pulse audio installed to rpoperly use pipewire? 2022-06-01 23:33:13 no pipewire handles pulseaudio clients itself 2022-06-01 23:34:11 pulseaudio-utils is just the pa* client utilities that configure/use a pulseaudio server 2022-06-01 23:34:31 pipewire handles pulseaudio clients in the sense of it knows all of those 2022-06-01 23:34:51 in the same way it handles alsa clients by being an 'alsa device' for them to connect to, and handles jack clients by being a jack server, etc 2022-06-01 23:35:33 ah 2022-06-01 23:36:15 well, aplay still says the host is down, cmus, moc and mgp123 fail similarly stating no sound device/driver 2022-06-01 23:36:46 actually. 2022-06-01 23:36:48 hm 2022-06-01 23:36:50 im not playing it in this bg'd session 2022-06-01 23:37:00 ill break into my tmux on the other pc and try it there 2022-06-01 23:37:06 good idea 2022-06-01 23:37:11 aplay foo.wav Playing WAVE 'foo.wav' : Signed 16 bit Little Endian, Rate 44100 Hz, Stereo 2022-06-01 23:37:14 ah, interesting 2022-06-01 23:37:15 thats new 2022-06-01 23:37:18 does it play 2022-06-01 23:37:23 good question. 2022-06-01 23:37:28 ill find some speakers 2022-06-01 23:37:36 you were testing audio without any damn speakers 2022-06-01 23:37:45 not really 2022-06-01 23:37:51 i had speakers for most of it 2022-06-01 23:37:59 i just wasnt getting any of it to actually run or recognzie sound devices 2022-06-01 23:38:03 so i relocated them temporarily 2022-06-01 23:38:27 for a moment there i was about to wonder if you had some new wireless-to-brain technology 2022-06-01 23:38:40 lol 2022-06-01 23:39:11 only have one pair as my old oners suffered heat death if you can believe that 2022-06-01 23:39:20 happens to the best of them 2022-06-01 23:39:23 pour one out 2022-06-01 23:40:03 i think i maybe just used the wrong power supply for it and it juiced too hard 2022-06-01 23:40:38 be back in a few got to traverse the wire jungle without knocking out power 2022-06-01 23:50:45 things play now, no actual audio though, might not be a problem, might be muted. so how do i get this to just work across all login sessions? what a messy ass sound server. 2022-06-01 23:52:58 just start it, with all the other parts. it works, so just use /usr/libexec/pipewire-launcher or some shit inside the session 2022-06-01 23:53:13 start your wm with dbus-run-session, add dbus to default in openrc, set an actual runtime dir 2022-06-01 23:53:28 you can run pulsemixer or some shit right now to see if it's muted or the default device is wrong or some shit 2022-06-02 00:02:04 back. sorry about that. 2022-06-02 00:02:14 sndio seems like a good fit for alpine, i wonder why pipewire is used over it 2022-06-02 00:02:15 alright 2022-06-02 00:03:05 still, i wonder why alsa wasnt working on its own... i wont be satisfied until i do know, butat least it might be working 2022-06-02 00:03:10 this pulse stuff pulls in so many dependencies 2022-06-02 00:04:06 https://dpaste.com/63Z4QUHCF will just re-xport it for the moment 2022-06-02 00:06:21 ah, host is down agian. wonder what happened to my ash dbus session, well i am going to go set htis up the right way real quick 2022-06-02 00:09:25 maybe pipewire being installed and not used was interfering with alsa 2022-06-02 00:11:19 pipewire-alsa existing places some alsa conf files, and probably made the clients try and connect to the nonexistent pipewire device 2022-06-02 00:11:49 ill set a , xdg_runtime_dir in my ash .profile i guess. /tmp/ is fine for logn term directory? 2022-06-02 00:11:57 its owned by root, which i thought might be aproblem 2022-06-02 00:12:25 no, you want it to be a bit more advanced than that 2022-06-02 00:12:39 technically you have to delete the directory when you log out, and it has to be user owned 2022-06-02 00:12:45 and chmod 700 2022-06-02 00:14:34 does. ash have a logout conf i can delete it in, then? 2022-06-02 00:15:18 normally my DE sets it to something in var/run.. 2022-06-02 00:15:23 but i am running windowmaker 2022-06-02 00:15:50 the de doesn't do it, usually it's pam_rundir afaik 2022-06-02 00:16:42 ah. 2022-06-02 00:16:46 well 2022-06-02 00:16:57 you have helped me so much more than i asked for, and i really appreciate that 2022-06-02 00:17:26 either way audio seems to be working now 2022-06-02 00:17:30 thanks for everything 2022-06-02 00:18:33 :3 2022-06-02 00:18:38 hope you enjoy your audio adventure 2022-06-02 00:22:00 lol 2022-06-02 00:22:22 well if pipewire ends up working im just going to remove it and use alsa or sndio anyways but il.l stick with it if i must 2022-06-02 00:22:51 avoiding linixusms in a linux desktop is probably a mistake 2022-06-02 00:23:18 i was just gonna try to get ol gnustep or tk stuff from cdde working anywayss and dont really need dbus 2022-06-02 00:27:02 but i guess that stuff will barely compile on a nmodern distro so it'll never work on musl 2022-06-02 00:27:32 shrug, probably just a patch or two 2022-06-02 00:30:45 probably, but im not a coder, i know some erlang and lua for funsies, sadly none of the fun gnustep even works in modern linux, windowmaker keeps patching otu support for it, no idea why 2022-06-02 00:46:44 gcompat probably isnt gonna work out for gui programs right 2022-06-02 00:47:20 worth a try 2022-06-02 00:47:26 works for spotify! if you can believe it 2022-06-02 00:49:54 hey, do you know if there's an 'easy' way to get zfs working with uefi and encryption? i tried it on bare metal but failed miserably. 2022-06-02 01:33:30 Wish I knew how to actually use gcompat. If I understood correctly I need to use elftools to patch what I want to run 2022-06-02 01:37:00 I think that's only necessary if the interpreter (ld-linux) path is different 2022-06-02 01:44:51 Oooh, okay. Then I just have not had success I guess 2022-06-02 01:45:09 My main intent was with AppImage 2022-06-02 01:50:31 What have you tried so far? 2022-06-02 01:53:07 😬 I don't recall in specific anymore 2022-06-02 01:53:44 I flatpak what I could and gave up on the rest. I think one was p3xonenote 2022-06-02 02:31:04 do you guys use alpine to run your alpine containers 2022-06-02 03:38:31 how can i replace grub with syslinux on alpine? 2022-06-02 07:33:52 Is it possible to use runit instead of s6? (that's what Alpine uses by default, right?) 2022-06-02 07:34:26 nothing against s6, just familiar with runit and didn't feel like Void was a good choice for a server. 2022-06-02 07:36:44 Alpine uses OpenRC. It is possible to use whatever you want, with differing amounts of work. 2022-06-02 07:37:41 busybox init is used for init 2022-06-02 07:37:59 openrc is used for service management 2022-06-02 07:45:11 humm: A bit of work is fine. I saw Google results which mentioned runit for "service management" (which I imagine is all I'm looking for; not sure if there are other differences to look out for), but wonder if there's a canonical reference. 2022-06-02 07:45:22 The wiki? docs.alpinelinux.org 404s 2022-06-02 07:48:03 Alpine gives you a working (busybox init +) OpenRC system. You can change the inittab or the init or whatever else you want however you want. 2022-06-02 07:49:00 If you're familiar with runit, I assume you know how to use runit. 2022-06-02 07:49:36 talismanick: works for me 2022-06-02 07:50:31 panekj: Really? Did you read the steps from somewhere or figure it out by trial and error? 2022-06-02 07:51:13 talismanick: Have you taken a look at the runit website? 2022-06-02 07:51:22 humm: yeah 2022-06-02 07:51:36 talismanick: meant docs.alpinelinux.org 2022-06-02 07:51:37 the smarden one, right 2022-06-02 07:52:00 panekj: oh, you're right. It works for me now. 2022-06-02 07:52:09 not sure what happened 2022-06-02 07:52:56 humm: oh, you mean follow the sysvinit steps? Is OpenRC based on it? 2022-06-02 07:54:18 You can do whatever you want with your system. Indeed, busybox init uses an /etc/inittab, which is similar to the sysvinit one. 2022-06-02 07:54:34 The Gentoo wiki implies it's a layer (my words) launched by sysvinit... hm 2022-06-02 07:55:49 Ah, I see. BusyBox has its own version of the init binary that sysvinit uses by default to launch the OpenRC overlay on Gentoo 2022-06-02 07:56:31 So it is the sysvinit instructions on the runit docs I should follow 2022-06-02 07:56:50 https://0x0.st/oBkS.jpg 2022-06-02 07:57:44 (being in my 20s, I've never used sysvinit) 2022-06-02 08:00:18 as shrimple as that 2022-06-02 10:39:57 Hi all 2022-06-02 10:40:00 Quick question, I'm building a small music device and want to use Alpine because of it's low overhead. I want to use a realtime kernel to minimise audio latency. Would that be a recompile, or is that available as a switch somewhere? 2022-06-02 10:41:01 Building on a raspberry pi 2022-06-02 10:45:41 jingles: I don't think we have any alternatives for the rpi kernel ourselves 2022-06-02 10:47:36 Yeah, I had a look. So probably a recompile on my part then 2022-06-02 10:48:11 There is a linux-edge kernel that is also compiled for arm*, which has some preempting enabled, but not sure if that's what you're looking for 2022-06-02 10:48:46 thank you 2022-06-02 10:49:40 To be clear, that's not the rpi kernel, so not sure if it will work 2022-06-02 10:50:18 Yeah, that might need a recompile too :D 2022-06-02 10:51:01 What I heard is that the upstream kernel should be able to boot rpi's nowadays, but the device trees are still missing, or something like that 2022-06-02 10:52:39 There are probably a number of distros with realtime kernels for pi actually. I haven't really looked too hard. I was just attracted to Alpine because of its minimalism 2022-06-02 10:53:35 right, we don't offer one ourselves 2022-06-02 10:53:42 There is also a patch-set involved 2022-06-02 10:53:46 https://wiki.archlinux.org/title/Realtime_kernel_patchset 2022-06-02 10:53:54 so you would need to patch + build it yourself 2022-06-02 10:58:02 It's been a very long time since I rebuilt a kernel, but I'm not averse to that at all 2022-06-02 10:58:22 I'll need to read up some more, but I can do it 2022-06-02 11:02:54 Thanks for your help! 2022-06-02 11:24:40 doing a realtime os is a lot more than just the kernel tho 2022-06-02 11:25:01 you probably know this but, for a raspberry pi, that just seems like it'd be really painful 2022-06-02 11:27:48 i dunno if the chips on the rpi4 support that better, though 2022-06-02 11:28:05 on the audio, that is 2022-06-02 12:11:44 invoked: True. I will be looking at various audio codecs to achieve what I want, however, the first step in this journey is a rt kernel. If I know that my audio threads will get interrupt and scheduling priority, then that's one big weight off my shoulders 2022-06-02 12:12:17 roger 2022-06-02 12:13:05 I'll be honest, it's been a while since I went digging deep in linux, so that's why I'm focussing on the kernel side first 2022-06-02 12:13:22 De-risk 2022-06-02 12:58:30 what i've seen/heard from the ppl who work on audio a lot, your audio chipsets need to have support for realtime or it can get glitchy 2022-06-02 12:58:52 you can make it work, but it may not work reliably 2022-06-02 12:59:11 s/realtime/low latency/ 2022-06-02 12:59:11 invoked meant to say: what i've seen/heard from the ppl who work on audio a lot, your audio chipsets need to have support for low latency or it can get glitchy 2022-06-02 22:04:41 Hello guys. Whats a current status of discussion about building flags as 02, 03, 0s? Are all builds still use -0s? 2022-06-02 22:30:55 the default cflags are -Os -fomit-frame-pointer and that's it 2022-06-02 22:32:14 psykose: For each program? 2022-06-02 22:32:32 unless the build system throws something else in, yes 2022-06-02 22:32:48 or something else is specified in the apkbuild of a package 2022-06-02 22:32:56 https://lists.alpinelinux.org/~alpine/devel/%3C1593625212.dirkptm3b0.none%40localhost%3E 2022-06-02 22:32:56 Here is what I've found briefly 2022-06-02 22:33:55 Some performance increasing is dramatically 2022-06-02 22:34:49 With -O2 2022-06-02 22:37:29 indeed, O2 is generally a better default 2022-06-02 22:42:30 Why not to change it? 2022-06-02 22:57:50 Size optimization is a big thing behind Alpine, no? 2022-06-03 02:46:53 i feel like i have seen this thread before 2022-06-03 03:08:01 Hmm 🤔 2022-06-03 03:08:01 You may well have written it, even 2022-06-03 08:41:54 isn't this basically https://gitlab.alpinelinux.org/alpine/abuild/-/issues/10057 2022-06-03 12:59:10 I'm a little confused. How do Alpine packages comply with licenses requiring the license text to be included in all copies? 2022-06-03 13:00:14 poorly 2022-06-03 13:02:12 in practice, SPDX license plus website is pretty reasonable attribution (assuming the license and website are correct), but it doesn't really fulfill the legal requirements 2022-06-03 13:03:04 I see 2022-06-03 17:56:26 hello alpiners. i need a little help. i am out of options on google. 2022-06-03 17:56:26 error msg ```mounting uuid on sysroot failed``` then initramfs emergency shell 2022-06-03 17:56:26 one of my vm with alpine crashed and wont boot up 2022-06-03 17:56:26 it seems to me sda is not recognized for some reason 2022-06-03 17:56:26 dmesg shows ```mounting root failed``` 2022-06-03 17:56:27 i booted up with alpine live cd, mounted sda and fsck /dev/sda1 and /dev/sda3 all clean 2022-06-03 18:03:11 i have no backup :( any guru could help me to boot it up again, much appreciate 2022-06-03 18:04:36 What kind of disks do you have? 2022-06-03 18:06:17 What disks do you see from emergency shell? 2022-06-03 18:07:28 8GB virtual, it is a proxmox vm 2022-06-03 18:08:19 how can i check disks from emergency shell? no fdisk there 2022-06-03 18:10:36 ls /dev/ 2022-06-03 18:12:43 yes, i checked /dev by ls. cant see any disk there. i will try grab some outputs. 2022-06-03 18:49:20 How do you enable ~/.pam_environment ? Tried adding user_readenv=1 in /etc/pam.d/system-login. Doesn't seem to work. 2022-06-03 18:50:35 user_readenv=1 where exactly in system-login 2022-06-03 18:51:09 Well where it needs to go obviously, "after session required pam_env.so" 2022-06-03 18:52:09 i mean, how would i know that? for all i know you added randomly in the middle of the file 2022-06-03 18:52:14 do you have shadow-login or util-linux-login installed 2022-06-03 18:52:35 Not sure about that 2022-06-03 18:53:52 Nope don't have these installed. So I'm guessing PAM doesn't work with busybox login 2022-06-03 18:54:03 indeed it does not 2022-06-03 18:54:04 Which would explain why it doesn't work 2022-06-03 18:54:21 Well, thanks. I think I can fix it now. 2022-06-03 18:54:47 the default is also =1 afaik, so you just need to install one of those and it should work 2022-06-03 18:54:48 in theory 2022-06-03 18:55:13 $ apk add util-linux-login 2022-06-03 18:55:13 Magic, it works 2022-06-03 18:55:33 Yes, it seems so. 2022-06-03 18:55:36 Thanks again 2022-06-03 18:55:59 Oh and I get the XDG vars as a bonus, awesome 2022-06-03 18:56:20 yep 2022-06-03 18:56:56 there is pam_rundir.so in pam-rundir that i don't remember if it's in the files or not for runtime dir 2022-06-03 18:57:28 ah, it is there, just need to install it for that one 2022-06-03 21:12:35 i was comparing dmesg of the broken alpine vm with a working one and found two lines which are missing on healthy vm and may help to identify my problem 2022-06-03 21:12:35 `Freeing initrd memory: 6372K` 2022-06-03 21:12:35 `Initramfs unpacking failed: junk within compressed archive` 2022-06-03 21:12:35 seems like corrupted bootloader? could be this the casue my vm doesnt boot up? is this repairable? what do you guys think? 2022-06-03 21:13:58 gabes: booting it with a live image would help 2022-06-03 21:15:21 sounds like a corrupted initramfs 2022-06-03 21:22:59 is it possible to recover initramfs in case the root partiition is not corrupted? 2022-06-03 21:23:20 fullchroot into it and regenerate it 2022-06-03 21:24:01 via `apk fix mkinitfs`, easiest way 2022-06-03 21:27:34 gonna try this out, thank you psykose 2022-06-04 04:28:21 Okay, so I mistakenly thought Alpine (already) used s6 because I read https://ariadne.space/2021/03/25/lets-build-a-new-service-manager-for-alpine/ when it came out 2022-06-04 08:54:26 it's working again. my broken alpine vm is fixed and runs just like before the crash. 2022-06-04 08:54:26 as per suggestions, i booted up with a live image, mounted the broken image, chrooted into it and ran `apk fix mkinitfs` 2022-06-04 08:54:26 thank you ikke, psykose and a special thanks to minimal, with whom i had a long chat, debugging and looking for the solution. 2022-06-04 08:54:43 :) 2022-06-04 08:54:46 Nice to hear 2022-06-04 08:58:34 i also learned a few new linux things along the way. 2022-06-04 08:58:34 and most importantly i am going to create a vm snapshot and backup right now to avoid future headaches. 2022-06-04 09:08:55 that's good practice with vms, yes :) 2022-06-04 09:09:33 an advice from someone using vmware on a big scale: don't keep to many snapshots per vm, they hurt the performance real bad 2022-06-04 09:17:28 Ideally you should not keep them at all 2022-06-04 09:18:23 we do keep them around when we run regular patches for a few days 2022-06-04 15:30:34 Anyone know a good place where I can get quick regex help? Nothing mission-critical and should be a quick fix as all the elements I want are there, just one ends up in the wrong place. StackExchange? 2022-06-04 15:31:09 You could just ask here 2022-06-04 15:31:30 https://regex101.com/ is nice for that as well 2022-06-04 15:31:37 But it is only very adjacetly Alpine related, in that it is for an XFCE Custom Action I'll use on two Alpine machines 2022-06-04 15:32:08 ikke Thank you. I'll try that again 2022-06-04 15:32:46 you could ask in any linux channel really, as there should be someone with sed/grep/etc. knowledge 2022-06-04 15:33:26 Would the generic/top-level Linux channel be most appropriate then? 2022-06-04 15:33:32 Thanks, by the way 2022-06-04 15:34:46 Saijin_Naib: it depends on the customs of the channel 2022-06-04 15:37:33 Hmmm, good point. I have no community membership in those channels so it'll be harder. Ah well, I'll keep trying with the site you linked for a while 2022-06-04 15:37:54 Directly Alpine-Related, can someone check if this XFCE Custom Action makes sense to them and looks right? 2022-06-04 15:38:01 https://github.com/Saijin-Naib/Thunar-Custom-Actions/blob/trunk/.config/Thunar/ABUILD/uca.xml 2022-06-04 15:38:45 Just want it to be able to checksum & build when I have an APKBUILD in a directory or selected, but I get abuild permission errors which makes no sense to me 2022-06-04 15:39:54 ~~I could help you with regex but if you want to figure out it on your own~~ 2022-06-04 15:52:36 exiftool '-filename<${DateTimeOriginal} [${model;}]' -d "%Y-%m-%d %H.%M.%S%%-c.%%e" FILENAME.RAF 2022-06-04 15:52:52 what I get: 2021-11-25 19.56.50.RAF [X-T2] 2022-06-04 15:52:59 What super obvious thing am I over-looking here? 2022-06-04 15:53:20 That [X-T2] needs to go just before the .RAF 2022-06-04 15:53:40 And thanks for any pointers, panekj 2022-06-04 21:25:14 Hello! Trying to get alpine going on a Raspberry PI 4 and my usb keyboard is not working 2022-06-04 21:25:28 Latest stable alpine 2022-06-04 21:25:37 3.16 2022-06-04 21:25:58 On Raspberry PI OS, keyboard works 2022-06-04 21:27:54 I've noticed in the past there have been USB troubles, but all those issues seem to be resolved. I've also noticed that on 4GB version usb was somehow broken because of the memory size, however I have a 2GB version which shouldn't be affected by that 2022-06-04 21:31:35 So I'm fairly certain it's not a hardware issue since stock raspberry pi OS seems to be able to use usb just fine 2022-06-04 21:55:34 Ok 2022-06-04 21:58:01 I was able to get keyboard working using `otg_mode=1`. This seems to be specific to the Compute Module 4, not standard Pi 4. I noticed some errors as it booted so I'll need to dig a bit deeper, but at least I can run setup now 2022-06-04 22:00:05 it seems to affect things on both pi4 and cm4 2022-06-04 22:00:09 it's just default =1 on cm4 2022-06-04 22:00:14 https://www.raspberrypi.com/documentation/computers/config_txt.html#otg_mode-raspberry-pi-4-only 2022-06-04 22:30:24 Saijin_Naib: %%e already means extension 2022-06-04 23:15:18 Hello71: Yeah, which is great, but if I try to reorder anything so the camera model goes before the extension, I break the whole expression and ash throws an error about invalid substitution 2022-06-04 23:15:26 I have no idea what I'm doing wrong haha 2022-06-04 23:54:43 Is it possible to build the alpine OCI container image without using an existing OCI image as input? 2022-06-04 23:58:54 https://github.com/alpinelinux/docker-alpine/blob/c9515859c035acbb899c6738506954e36f57e7dc/x86_64/Dockerfile 2022-06-04 23:59:09 the base image is just the minirootfs and a cmd 2022-06-04 23:59:17 the ones from here https://alpinelinux.org/downloads/ 2022-06-05 00:02:02 now if you're asking 'is it possible to build it without even using the prebuild minirootfs? that's cheating!', well, it's just whatever generation of things you want to do yourself; the minirootfs img is generated with scripts/mkimg.minirootfs.sh in aports, using mkimg 2022-06-05 00:02:15 Saijin_Naib: well the narrow answer is that single quotes and double quotes are different in shel 2022-06-05 00:02:17 l 2022-06-05 00:03:45 i have a vague idea how to fix it but probably you should ask on the exiftool forms 2022-06-05 00:03:47 forums 2022-06-05 00:03:51 the creator is very active 2022-06-05 00:21:04 psykose: Thanks! 2022-06-05 00:32:33 Hello71: alright, I'll try that next then. Thanks 2022-06-05 12:26:02 After a couple of months of not rebooting, my laptop running alpine no longer boots. During early boot I get garbled text then the system freezes. The last lines I see mostly correctly are "initialized simpledrm for simple-framebuffer", "switching to colour frame buffer device", "simple-framebuffer: drm+plane_enable_fb_damage_clips() not called". The hardware is intel pentium n3520 and intel hd graphics bay 2022-06-05 12:26:05 I've had this issue before when I first installed alpine and with the help of this chat I was able to get it to work. Unfortunately I do not clearly remember what we did then. I believe I touched the mkinitfs configuration which currently contains: features="ata base cdrom ext4 keymap kms mmc nvme raid scsi usb virtio i915 cryptsetup". i915.modules contains: kernel/drivers/gpu/drm/i915/i915*. The linux-firm 2022-06-05 14:02:18 I have found my previous discussion of this problem in irc archives of this channel: https://irclogs.alpinelinux.org/%23alpine-linux-2021-12.log with my previous nick Carrot4123 . Back then we suspected the change to simpledrm caused this. 2022-06-05 14:03:20 Matsushita: if I recall correctly, to fix that it was enough to regenerate the initramfs 2022-06-05 14:03:40 I have regenerated it several times now during testing. That's not it. 2022-06-05 14:04:56 As in I run `mkinitfs` and I see that `/boot/initramfs-lts` has a recent timestamp. 2022-06-05 14:24:47 i've had a similar issue too when booting with efi; wasn't patient enough and just switched to legacy 2022-06-05 14:28:32 I'd really like to keep booting with efistub. Also this used to work. Originally it just worked on 3.14, then I ran into the problem for the first time on 3.15 fixed it through the previous irc discussion and now it is broken again. I am on 3.16 atm. 2022-06-05 14:33:05 I'm frustrated partially because I don't know how to debug this. Booting from the alpine live cd (usb stick) works so what is the difference in my installed system that makes it fail? 2022-06-05 14:48:38 Is there a bug tracker I should report this issue in btw? 2022-06-05 14:57:17 Matsushita: https://gitlab.alpinelinux.org/alpine/aports/-/issues 2022-06-05 16:48:04 minimal helped me solve this problem. (thank you so much.) The solution was to add `modules=i915` to my kernel parameters. 2022-06-05 16:48:39 Matsushita: thanks for feedback 2022-06-05 22:56:44 If I may ask, what sorts of patches does it typically take to make a program busybox-compatible? 2022-06-05 22:57:25 (I've seen would-be minimalists in other channels be told to check out Alpine if they want to try busybox, because the team has put effort into patching programs for it) 2022-06-05 22:59:21 What does this mean? If you write scripts that use busybox tools, you need to use only flags supported by these tools and make sure those flags are doing the thing you want 2022-06-05 23:03:49 if some shell script invokes utilities not supported by busybox then it's not compatible; so it's.. making it compatible, i guess 2022-06-05 23:03:56 although i have not exactly seen many such patches 2022-06-05 23:04:14 usually it's just sed/awk stuff 2022-06-05 23:35:18 Ermine: I'm not sure what it means, that's why I'm asking lol 2022-06-05 23:35:48 psykose: so, rewriting scripts to use POSIX sh instead of bash, etc? 2022-06-05 23:36:42 Do wish I remember who said that so I could ask them to clarify 2022-06-05 23:36:53 dunno, you tell me; i've certainly not seen much of rewriting entire things to 'not use bash', only where it's like 2 lines 2022-06-05 23:39:01 I was imagining there'd be more problems linking compiled programs, but I guess that's moreso a musl issue? 2022-06-05 23:39:20 i think you are confusing a lot of different terms 2022-06-05 23:39:45 probably 2022-06-05 23:40:04 Is it worth using the crypt filesystem option on a VPS? 2022-06-06 01:25:08 not to protect against whomever is running the vps, no 2022-06-06 01:25:29 because the hoster can just grab your keys from memory anytime 2022-06-06 01:26:51 so ultimately you have to trust them, and if you trust them, then luks isn't necessary 2022-06-06 01:27:39 unless you're running some service and you want to say in your privacy policy to your users that their data is encrypted "at rest", which doesn't really mean much in the grand scheme of things 2022-06-06 01:28:49 it could possibly protect against someone barging into the data center and stealing the server 2022-06-06 01:29:48 or the hoster backups leaking 2022-06-06 01:32:57 the latter is a good point. 2022-06-06 01:58:25 Hello71: That's actually a bit closer to what I had in mind - insurance against something happening on their end and losing control of my VPS's data. (although I do intend to keep everything possible ephemeral) 2022-06-06 06:50:05 Does Alpine use gnupg by default? (For example for signing packages?) 2022-06-06 07:35:45 Noclip[m]: it utilizes black magic, err, colored magic and pure luck ;) 2022-06-06 07:36:00 some info of signing: https://wiki.alpinelinux.org/wiki/Package_management and https://wiki.alpinelinux.org/wiki/Apkindex_format 2022-06-06 07:36:35 it signs the index using openssl 2022-06-06 07:36:49 but if that docs are true then its about time to replace sha1 with sha256 2022-06-06 07:36:58 or while you are at it sha512 2022-06-06 07:38:37 Apachez: APKv2 will not change anymore 2022-06-06 07:38:51 APKv3 should improve this 2022-06-06 07:39:59 what will APKv3 be using? 2022-06-06 07:40:10 and those examples should be updated anyhow 2022-06-06 07:40:21 its just very bad practice to rely on MD5 and SHA1 nowadays 2022-06-06 08:28:04 "it signs the index using openssl" <- Why not libressl? 2022-06-06 08:29:23 "but if that docs are true then..." <- Is it using some additional encryption/signing method like TLS/https besides that? 2022-06-06 08:29:40 "encryption/signing" 2022-06-06 08:29:46 Or is sha1 all it takes to compromise package delivery? 2022-06-06 08:29:59 what purpose do you want it for? privacy or integrity? 2022-06-06 08:30:20 grawity: If possible both but at least integrity. 2022-06-06 08:31:07 I think integrity is the least that should be expected of every package delivery system. 2022-06-06 08:31:34 And sha1 doesn't sound like it can provide that. 2022-06-06 08:37:59 "https://dl-cdn.alpinelinux.org/alpine/v3.16/main 2022-06-06 08:37:59 https://dl-cdn.alpinelinux.org/alpine/v3.16/community" 2022-06-06 08:37:59 Looks like it's using https. 2022-06-06 08:39:52 Could a downgrade-attack downgrade it to http? 2022-06-06 08:40:43 Noclip[m]: sha1 is not fully compromised yet. Though it's feasible to get collisions when you control both files, generating a collision when you need to match an existing file has not been shown yet 2022-06-06 08:41:24 That does not mean you should not migrate away from it, but it's not that it's an immediate issue 2022-06-06 08:41:34 that kind of downgrade attacks only works if the client starts off by doing http and expects to be upgraded 2022-06-06 08:42:01 if the client wants to speak TLS from the beginning, and the server tries to respond with plain HTTP, it's not going to work 2022-06-06 08:42:16 grawity: So with https in my repo file (as it's by default) downgrading shouldn't be possible? 2022-06-06 08:42:51 can't see it happening with any HTTP client 2022-06-06 08:43:20 one way to test – what happens if you try to request https://foo:80 (which will give a plaintext HTTP response)? 2022-06-06 08:44:11 Noclip[m]: the command is openssl even if you got libressl or tigerssl or bearssl or boringssl or polarssl etc installed 2022-06-06 08:46:13 Apachez: And which of those is used by default? 2022-06-06 08:46:27 Alpine uses openssl 2022-06-06 08:53:34 Ok 2022-06-06 08:55:46 ikke: maybe apkv3 is suitable time to upgrade just in case? 2022-06-06 08:57:09 Ermine: afaik it doesn't use sha1 2022-06-06 08:57:26 When is apkv3 going to be released? 2022-06-06 08:57:26 ah ok 2022-06-06 08:58:06 whenitsdone.jpg 2022-06-06 08:58:21 👍 2022-06-06 09:02:16 FYI, APKv3 uses a different index / package structure. Data wise, there is no backwards / forwards compatibility 2022-06-06 09:49:22 is flatpak not defaulting as --user for anyone? 2022-06-06 09:50:29 this been happening since the latest apkbuilds I think; doing "flatpak install ..." is defaulting as it was used by --system rather than user 2022-06-06 09:51:09 same with commands like "flatpak update" 2022-06-06 10:08:49 it's never defaulted to --user for me 2022-06-06 10:09:17 oh, i just used it without --user and it did 2022-06-06 10:09:17 hah 2022-06-06 10:10:55 flatpak remotes --system should be empty 2022-06-06 10:10:59 then it.. should default to user 2022-06-06 10:11:11 https://github.com/flatpak/flatpak/issues/3552 2022-06-06 10:11:43 so, make sure the remotes are only in --user and it should work 2022-06-06 10:11:54 aside from that you can probably also clean out the system repository if anything was there 2022-06-06 10:12:17 psykose: yes that's the problem 2022-06-06 10:12:41 i'll see if i can make the remote only for user 2022-06-06 10:12:44 no idea how to do that 2022-06-06 10:13:22 you just add it with --user 2022-06-06 10:14:02 there's remote-add/remote-remove commands 2022-06-06 10:14:18 ok thank you! will try and come back with results 2022-06-06 10:55:59 psykose: no chance 2022-06-06 10:56:25 i used flatpak remote-add --user flathub etcc etcc... 2022-06-06 10:56:33 can you post remotes --system and remotes --user 2022-06-06 10:57:43 Name Options 2022-06-06 10:57:45 flathub 2022-06-06 10:57:55 that is $ flatpak remotes --user 2022-06-06 10:58:23 and remotes --system gives me the typical cannot open repo because no root permissions 2022-06-06 11:00:49 so.. run it with root 2022-06-06 11:02:32 ok 2022-06-06 11:02:43 gives nothing 2022-06-06 11:02:50 nothing at all 2022-06-06 11:03:17 then i guess you have some other issue like others that posted there that it still doesn't work, i guess 2022-06-06 11:03:27 if you want you can try nuke everything and start from scratch 2022-06-06 11:04:07 will do that :) 2022-06-06 11:04:08 thank you 2022-06-06 11:04:52 to do that you need to basically rm -rf /var/lib/flatpak ~/.local/share/flatpak 2022-06-06 11:05:00 then flatpak repair as root 2022-06-06 11:05:04 then do everything with --user 2022-06-06 11:05:05 i guess 2022-06-06 11:05:19 you can probably trim ~/.var/app too if you didn't have anything saved 2022-06-06 11:05:54 and if it still doesn't work, well, just your usual flatpak issue :) it didn't work for me either for a long time, then somehow just did 2022-06-06 11:07:31 psykose: now it is listing user under Options 2022-06-06 11:07:54 after removing everything in .cache/flatpak .local/share/flatpak 2022-06-06 11:07:57 one of those did it 2022-06-06 11:08:44 strange things happen sometimes :P 2022-06-06 11:09:04 i thought it was apkbuild that wasn't compiled properly because I never changed anything 2022-06-06 11:09:50 haha, nope :) just flatpak things 2022-06-06 11:17:08 Hello, I'm looking to extend the life of my SD card on a sys install. The most significant thing I can think of is disabling logging. How would I go about doing this on a default install? 2022-06-06 11:17:37 Also, is there anything else that I can do to prolong my SD card's life? Thanks. 2022-06-06 11:32:11 judedon: assuming you are using ext4 then you could use a mount option (typically by adding to /etc/fstab) like "commit=300" to reduce the frequency of journal writes. Unfortunately for the rootfs such a fstab entry is not curently handled for the rootfs - I have had a MR open for more than 1 month to address this: https://gitlab.alpinelinux.org/alpine/mkinitfs/-/merge_requests/103 2022-06-06 11:34:12 if you want to just 'disable all the logs', you could mount /var/log as tmpfs, so they are in memory only, i suppose 2022-06-06 11:34:41 i think the early-boot few lines of logs will still be written (and then the mount will shadow them after each time), but after that, it shouldn't hit the disk 2022-06-06 11:34:49 i'm not sure if you run into other issues doing that though :) 2022-06-06 11:40:07 Thanks for the reply. I'm not so experienced with linux, so I may be wrong, but if I did that wouldn't it slowly eat up RAM on machine running for a long time? I also wouldnt actually know how to do that, lol. 2022-06-06 11:41:29 minimal: could you put it in laymans terms? Or is it as simple as adding a 'commit=300' line to /etc/fstab 2022-06-06 11:42:19 What ive tried so far is removing dmesg and syslog from their respective runlevels, though that didnt seem to do anything. 2022-06-06 11:47:21 sure, it would. though my /var/log is.. 4MB, and i don't have it mounted as that 2022-06-06 11:47:35 (well, up to the size cap of the tmpfs, past that it would probably fail to write) 2022-06-06 11:49:03 judedon: ext4 has a journal where it commits data, by default, every 5 seconds. Increasing the commit time will reduce the frequency of writes to storage (at an increased risk of unwritten data being lost during a crash). The 4th field of entries in /etc/fstab is a list of comma-separated mount options (e.g. rw,relatime). For any ext4 filesystems you could add ",commit=300" to their options. However I mentioned that such an option is 2022-06-06 11:49:03 currently ignored for the rootfs "/" 2022-06-06 11:50:07 judedon: if you are worried about SDcard writes then why not use Alpine's run-from-ram configuration rather than sys mode? 2022-06-06 11:52:11 i may do that yet, i suppose i just preferred a classic install since thats what im more familiar with 2022-06-06 11:54:21 thanks for the explanation minimal, hopefully your merge request is accepted soon then 2022-06-06 11:55:17 ill do more research into mounting logs in tmpfs, thank you 2022-06-06 11:58:58 judedon: are you using the default syslog daemon from Busybox? or rsyslog/syslog-ng - both of those, from memory, write a "MARK" entry are frequent intervals that you can disable to reduce disk writes 2022-06-06 12:00:24 default busybox syslog. Is this explained in the man pages? 2022-06-06 12:14:45 minimal: your MR makes sense. i'm just trying to think of the argument against it. (because in that case, why isn't it already done that way) 2022-06-06 12:16:45 no idea, but i have a gut feeling this causes a regression with somebody in some particular way 2022-06-06 12:17:14 the only regression it can cause is someone added a bunch of garbage into the mount options 2022-06-06 12:17:20 then forgot to remove them, and they were never used 2022-06-06 12:17:22 and now they are 2022-06-06 12:18:08 besides that. :) 2022-06-06 12:18:09 invoked: how would that be a regression though? if you add options into fstab you'd expect them to be used 2022-06-06 12:18:43 i agree 2022-06-06 12:18:57 if you can't come up with an example then you are speaking from fear 2022-06-06 12:19:07 and there is always 'the unknown' 2022-06-06 12:21:13 invoked: I guess the only "risk" might be if you add options that Busybox's mount can't handle - but again why add such options for rootfs if they can't be used? 2022-06-06 12:21:30 not fear. i'm not saying it shouldn't be merged. just has that look of something that should have been done a long time ago, for reasons, but nobody wrote down what the reasons were. 2022-06-06 17:16:58 Why does "git add -p" not work? 2022-06-06 17:17:06 Am I mssing some package? 2022-06-06 17:17:20 I have the regular "git" package installed. 2022-06-06 17:18:06 "$ git add -p 2022-06-06 17:18:06 git: 'add--interactive' is not a git command. See 'git --help'." 2022-06-06 17:18:20 Is that a bug? 2022-06-06 17:23:57 Noclip[m]: apk add git-perl 2022-06-06 17:24:18 Thanks 2022-06-06 17:24:45 Installing alpine packages is kinda like a puzzle game, lol. xD 2022-06-06 17:25:48 That's the flipside of having a small distro 2022-06-06 17:26:00 Is there a limit on how much I'm allowed to download from the alpine repo? 2022-06-06 17:26:05 no 2022-06-06 17:26:18 Great 2022-06-06 17:26:43 Then I don't need to fear rebuilding my container image too often. 2022-06-06 17:38:44 perhaps setup your own repo instead? 2022-06-06 17:46:43 Apachez: Nahh, I don't think I will rebuild my images that often. On avarage probably just every view days. Since the stable branch gets not a ton of updates most rebuilds will probably happen when I changed something in my Containerfile. 2022-06-06 17:47:52 I'm not rebuilding every hour or something like that. xD 2022-06-06 18:17:04 you can always lazycache 2022-06-06 18:17:20 keep a local apk cache, mount apk cache into container and enable it there too 2022-06-06 18:17:21 :) 2022-06-06 18:19:55 psykose: Sounds good but I have no idea how to do that. xD 2022-06-06 18:20:09 mounting folders into `docker build` is not really supported afaik, there's a feature for cache folder in the buildkit backend though: https://vsupalov.com/buildkit-cache-mount-dockerfile/ 2022-06-06 18:20:26 Ohh, you mean I just mount the cache from one persistent container into another container? 2022-06-06 18:22:20 no, you'd write something like `RUN --mount=type=cache,target=/var/cache/apk apk add ...` to make this folder persistent across multiple `docker build`s :) 2022-06-06 18:22:21 Or rather just create a volume for the cache? 2022-06-06 18:23:16 you don't need to configure a source, it works magically (somehow, never looked into it) 2022-06-06 18:23:44 I think the magic behind it is called container volumes. 2022-06-06 18:24:24 And those are basically persistent directories managed by the container engine. 2022-06-06 18:24:35 (I'm using Podman, not docker, btw.) 2022-06-06 18:24:55 ah, you'd need to check if podman has this feature then :) 2022-06-06 18:25:43 ah, it's possible to just use podman with buildkit apparently https://pythonspeed.com/articles/podman-buildkit/ 2022-06-06 18:25:55 According to the manpage it has that feature. 2022-06-06 18:36:46 Would the cache also reduce the amount of data written to disk? 2022-06-06 20:26:48 Noclip[m]: I would still consider setting up a local mirror :) 2022-06-06 21:01:13 Good afternoon. Do you have any material that teaches you how to set up plymouth? Thank you in advance. 2022-06-06 21:15:05 wolf: ubuntu plymouth? 2022-06-06 21:16:45 @panekj alpinelinux 3.16 2022-06-06 21:19:07 meant as package origin, but I just learned that it's apparently a freedesktop project 2022-06-06 21:31:44 I don't get it 2022-06-07 01:30:54 I think panekj was just trying to be sure you two were talking about the same package. The boot splash/animation program, right? 2022-06-07 01:31:08 I've tried to set it up a while back and didn't get anywhere, unfortunately 2022-06-07 01:31:45 yeah, I thought it was ubuntu-specific but it originated from fedora into freedesktop project 2022-06-07 01:46:16 Yeah, Ubuntu is where I first and only saw it used. It would be fun to have haha 2022-06-07 15:23:30 So I happen to know that apk lets you use cmd to select packages, as in `apk search cmd:zgrep` or even `apk add cmd:zgrep`, which is *super* cool. But looking around, I can't see where that's documented; I tried grepping all the manpages and none of them seem to contain "cmd". Am I missing something, or is it actually undocumented? And if so, are 2022-06-07 15:23:30 there other cool features that never got written down? 2022-06-07 15:24:18 I've found that feature "documented" here: https://docs.alpinelinux.org/user-handbook/0.1a/Working/apk.html 2022-06-07 15:25:10 Hm. Maybe documented shouldn't be in quotes, I guess this is documentation. 2022-06-07 15:25:30 technically this is part of abuild, not apk 2022-06-07 15:27:03 What apk offers is allowing packages to provide anything 2022-06-07 15:30:44 And this is why I ask: `apk add so:libmpack.so.0`and `apk add pc:msgpack` are the kind of feature that I wish every package manage had:) 2022-06-07 15:31:11 (I do wonder if that page shouldn't be merged into a manpage, though) 2022-06-07 15:32:16 apk manpages are very vague and uninformative unfortunately 2022-06-07 15:33:32 Huh. They're also weirdly short. I mean, there's a part of me that loves manpages that are actually a single page, but ... 2022-06-07 15:33:46 Is this intentional, or just that nobody's cared? 2022-06-07 15:34:02 Because I'm willing to open MRs if that would be well-received 2022-06-07 15:34:34 perhaps a question for fabled and/or Ariadne 2022-06-07 15:35:15 those features are unrelated to apk 2022-06-07 15:40:25 How are they unrelated to apk? Or perhaps, what *are* they a part of then 2022-06-07 15:40:35 specifically, semantic providers are something specific to alpine, not apk-tools (though other apk distributions have implemented them in part) 2022-06-07 15:40:59 yjftsjthsd: apk-tools does not know specifically about cmd: so: pc: etc 2022-06-07 15:41:25 Oh. Then how is it implemented? 2022-06-07 15:41:32 apk-tools knows about provides, and abuild adds these specific provides to packages 2022-06-07 15:41:44 also, some other distributions do have these features 2022-06-07 15:41:53 Ah. That's... actually kind of a cool way to implement it 2022-06-07 15:42:05 makes it a pain to document, though 2022-06-07 15:43:00 because as an end-user, I would totally expect that to be in `man apk-add` or at least `man apk-search`, but yeah that's not really an apk feature 2022-06-07 15:43:07 for example, with RPM distributions, you can frequently do `dnf install SONAME(libmpack.so.0)` or `dnf install pkgconfig(msgpack)` 2022-06-07 15:43:54 yjftsjthsd: except for cmd:, most of the time you would not even directly use them as a user 2022-06-07 15:44:25 hmm, not sure about that 2022-06-07 15:44:43 depends on the kind of user; I would call myself a "user" but compile things from source 2022-06-07 15:45:03 (and would definitely benefit from being able to install libraries by their pkgconfig name) 2022-06-07 15:45:23 but I agree that cmd is the *most* useful for wide audiences 2022-06-07 15:45:26 ikke: I love cmd: when someone tells me about some command that I'm missing or they use, I can just find it easily 2022-06-07 15:45:37 panekj: yes, I agree 2022-06-07 15:45:56 oh, misunderstood the statement 2022-06-07 15:46:00 oops 2022-06-07 15:47:16 anyway, I find those "features" very useful even if they are a byproduct 2022-06-07 15:47:31 not a byproduct, intentionally there 2022-06-07 15:47:43 of apk 2022-06-07 15:47:47 just not a feature of apk per-se 2022-06-07 15:48:01 it's a feature of the metadata that alpine bakes into packages 2022-06-07 15:48:29 open-closed principle in action 2022-06-07 15:48:30 anyway there is alpine handbook under construction 2022-06-07 15:48:33 maybe put it there 2022-06-07 15:48:42 see #alpine-docs 2022-06-07 15:50:27 though it seems discussed here already: https://docs.alpinelinux.org/user-handbook/0.1a/Working/apk.html 2022-06-07 15:52:50 yes, which was already linked to 2022-06-07 18:52:56 hello. I am seeing an issue with the update for the alpine baselayout. The pre upgrade script trying to convert /var/run from a directory to a symlink that points to /run. The issue is that the script doing the check to see if /var/run is a directory, and getting a false positive. Here is a paste bin showing how the test command is returning true when testing if /var/run is a symlink and a directory: 2022-06-07 18:52:56 https://pastebin.com/ccfn25wt Here is a link to the upgrade script that i believe is causing the issue: https://git.alpinelinux.org/aports/tree/main/alpine-baselayout/alpine-baselayout.pre-upgrade#n18 2022-06-07 18:53:13 im not sure if this is the right place to report this kind of issue, or if there is a github issues page or something... 2022-06-07 18:53:42 you can find link to issue tracker in /etc/os-release 2022-06-07 18:54:48 ty 2022-06-07 19:03:46 filed an issue: https://gitlab.alpinelinux.org/alpine/aports/-/issues/13917 2022-06-07 19:07:32 Good afternoon. Do you have any material that teaches how to configure plymouth in aplinelinux 3.16? Thank you in advance. 2022-06-08 00:33:38 hello, I'm trying to fix a broken lxc template for alpine images, and had a question about a specific apk signing key 2022-06-08 00:34:38 this is the upstream lxc template I'm trying to fix: https://github.com/lxc/lxc-templates/blob/master/templates/lxc-alpine.in 2022-06-08 00:36:07 so what I think is happening is that fetches the apk keys from https://alpinelinux.org/keys/ for verification, however, there's a key that comes bundled with apk.static, which is not at that URL (alpine-devel@lists.alpinelinux.org-6165ee59.rsa.pub) 2022-06-08 00:36:45 it's in alpine-keys 2022-06-08 00:37:01 we rolled some new keys for.. 3.15? or around that time 2022-06-08 00:37:24 ahh OK, that makes sense, I never had an issue with building 3.14 containers 2022-06-08 00:37:29 yep 2022-06-08 00:37:45 you can apk info -L alpine-keys on 3.15+ and see what the new ones are, it's the /etc ones 2022-06-08 00:37:48 (pretty sure) 2022-06-08 00:38:11 cool, so what I think I might do for this lxc template is to grab and extract that package so I have all the keys I need to verify 2022-06-08 00:38:18 I was just curious why it wasn't listed with the other keys 2022-06-08 00:38:40 in the template? 2022-06-08 00:38:46 because it hasn't been touched since 2018 2022-06-08 00:39:07 I meant at https://alpinelinux.org/keys 2022-06-08 00:39:22 that's what the template looks at to be able to verify packages 2022-06-08 00:39:47 that i have no idea about 2022-06-08 00:40:20 it should look at https://git.alpinelinux.org/aports/plain/main/alpine-keys/ 2022-06-08 00:40:54 awesome, I came across that but wasn't sure, I can update the template to look from there 2022-06-08 00:40:55 thanks! 2022-06-08 01:39:23 https://alpinelinux.org/keys/ should probably be updated 2022-06-08 01:40:02 or, hm, it was already? 2022-06-08 01:41:43 oh, it missed 4096 bits 2022-06-08 01:41:59 and also the new riscv64 2022-06-08 01:42:22 so theoretically it should be up-to-date, but in practice nobody checks 2022-06-08 02:13:09 TY psykose for reviewing and assigning my merge request. 2022-06-08 02:21:11 :) looks fine to me, was probably just an oversight (and so it actually does that every upgrade..) 2022-06-08 02:41:42 Yeah seems to be breaking some cicd pipelines. Found an issue from 2020 in the kaniko GitHub that I linked in my issue 2022-06-08 04:21:14 hello, i just migrated to wayland. Is there any panel with taskbar buttons such as the xfce-panel but for wayland? 2022-06-08 04:24:01 you mean other than the 'default' recommended one in the wm you picked? there's probably at least 10, though it depends what you want 2022-06-08 04:24:44 psykose: yes, I used swaybar and waybar but absolutely need a classic taskbar for desktop usage 2022-06-08 04:25:31 if you are on sway, then the click-icon-to-be-taken-to-that-workspace kind of panel is a bit backwards from what sway is supposed to be 2022-06-08 04:25:36 i /did/ see one though, let me check 2022-06-08 04:26:11 i saw nwg-panel but it is not built in the alpine repos 2022-06-08 04:29:37 ah, it was sfwbar 2022-06-08 04:31:02 nice! i didnt know about sfwbar 2022-06-08 04:31:18 looks good 2022-06-08 04:31:50 gonna enable the testing repos just for that :) 2022-06-08 04:32:47 voids your warranty a little :p 2022-06-08 04:32:55 thank you psykose 2022-06-08 04:33:15 i wouldn't have found it without your help 2022-06-08 04:33:27 i just happened to see it the other day by chance, lucky you :) 2022-06-08 04:33:31 not sure if it's any good 2022-06-08 04:33:47 it just crashed LMAO 2022-06-08 04:33:52 gonna test it a lot i guess 2022-06-08 04:33:52 see? fantastic 2022-06-08 04:34:03 if there's any logs those are useful 2022-06-08 04:34:20 maybe it's a segfault specific more to us than the project 2022-06-08 04:34:49 not sure 2022-06-08 05:48:34 hi, retarded question. how do i image an sd card with the "alpine-rpi-3.16.0-armhf.tar.gz" 2022-06-08 05:48:46 its not an .img like i have worked with in the past 2022-06-08 05:49:03 just uncompress it onto a fat32 sd card or what 2022-06-08 05:51:59 walrus77123: yeah exactly that. format the card yrself as fat32 and then extract the tarball into it 2022-06-08 05:52:21 or if you want you can do a fat32/ext4 split if you want an ext4 partition for persistent data 2022-06-08 05:52:53 by default the rpi alpine does not have any persistence aside from using lbu_commit to save installed packages and the /etc folder if i recall correctly 2022-06-08 05:53:06 (but that stuff goes on the fat32 partition) 2022-06-08 05:54:13 check the wiki for more info on the pi stuff cause it works a bit different from alpine on desktop outof the box 2022-06-08 05:57:59 oh i see. i was hoping to install and use it as a server normally, thanks i will 2022-06-08 05:58:06 been ahving trouble accessing the wiki lately 2022-06-08 05:58:10 maybe its my isp 2022-06-08 06:02:15 running the distro in a ram disk on arm socs that typically have less tha n4gb of ram seems counter productive 2022-06-08 06:04:45 but sd cards tend have their firmware lock up permanently randomly so i guess there is some wisdom in it 2022-06-08 06:07:56 to execute/read anything you need to read it into ram anyway; so if you have 2GB of stuff to load into ram but you weren't going to use it, that means it shouldn't be there 2022-06-08 06:08:35 i have installed alpine to an sd card and run it on a rpi4 before 2022-06-08 06:08:48 it was a persistent installation 2022-06-08 06:10:26 it is also nice on devices you expect to rudely power off instead of doing proper shut down 2022-06-08 06:10:46 haha 2022-06-08 06:11:04 i use a pi zero w as a wifi serial adapter for my other single board computers. i plug it in and then ssh in + screen to use the serial from my desktop 2022-06-08 06:11:05 yeah the lack of proper shut down procedure unless you add & script one is kind of annoying 2022-06-08 06:11:23 but because it is entirely from ram i dont have to worry that i just yank the power off 2022-06-08 06:11:25 (on arm socs) 2022-06-08 06:11:26 when im done 2022-06-08 06:12:15 what better shutdown procedure would they have? most of them have no 'on' button, so merely being powered is on 2022-06-08 06:12:35 for the rpi4 that i have at least it doesn't boot after a poweroff and just stays in the weird idle state or something (i think?) 2022-06-08 06:13:51 some arm boards can actually do a full power down procedure 2022-06-08 06:14:03 like the C.H.I.P. can for example (RIP) 2022-06-08 06:14:45 i guess, but i don't really see a point to it as long as the system actually synced stuff to disk and the os exited 2022-06-08 06:14:57 you're going to have to physically touch it to turn it back on anyway, so.. 2022-06-08 06:15:00 well for the CHIP it was important because that was intended to be battery powered 2022-06-08 06:15:10 so you do not want to be in a state that is consuming power at all 2022-06-08 06:15:30 ah 2022-06-08 06:16:51 but i also dont mind boards just syncing everything to disk and then halting the cpu when they are wall powered i think. the idle current is not that large in the grand scheme of things, just only relevant when its a battery 2022-06-08 06:22:13 haha 2022-06-08 06:22:38 my rpi4 always rudely power off 2022-06-08 06:22:57 it's either me yanking the cord for emergency situtations or a power cut happened 2022-06-08 06:24:37 psykose, i noticed, i think you need to short two pins on the gpio header to power it on again 2022-06-08 06:24:48 or just replug 2022-06-08 06:24:59 which acts like a power button on atx systems 2022-06-08 06:25:09 hah, or replug yup :) 2022-06-08 06:25:54 but i think most embedded systems just power on as soon power is delivered 2022-06-08 06:27:04 the isp tv set top box at home powers on as soon power came back, and it also doesn't go to sleep mode which is a weird decision for a software 2022-06-08 06:27:17 maybe the pmic doesn't support PON states 2022-06-08 09:40:49 hello. disclaimer: im tired, have migraine and im not a chat-person, so sorry if i type nonsense. problem: i use sway on alpine, seatd (not seatd-launch), did everything whats written inthe wiki, can login via "dbus-launch sway" in the login shell on tty1, everything is working,but if i poweroff, i get a lot of red error messages that do not appear in the openrc-log. its too fast to read it. something like libseat could not ... 2022-06-08 09:41:35 if i exit per swaymsg exit before i poweroff, no warnings 2022-06-08 09:58:57 they are probably there because when you poweroff shit gets killed randomly and then something 'errors' in the stack of stuff running in the wm etc 2022-06-08 09:59:00 i see much of the same 2022-06-08 09:59:02 all harmless 2022-06-08 09:59:27 specifically they would be outputted by sway itself, as in you ran `sway` there, so it's logging to stdout/err 2022-06-08 09:59:40 and so its logs are still there when it closes right before it powers off 2022-06-08 09:59:50 if you want to read them, add a `&> log` to the end of the sway start 2022-06-08 09:59:59 then do poweroff, and on next boot read the file 2022-06-08 10:05:47 thx psykose, if its hamless, ergo cosmetic, i do no care that much. an "dbus-launch exec sway" whould be bullshit , right - if im using seatd...? or whould that change something? 2022-06-08 10:13:45 that's what you do already? 2022-06-08 10:14:09 if you mean changing `dbus-launch ..` with `exec dbus-launch ..` then it would remove the extra shell in the middle 2022-06-08 10:14:25 generally you should always exec like that at the end :) 2022-06-08 10:20:45 hmm,thx again alice - to comet o an end: i willmake a fish function like "poff", "swaymsg exit && doas poweroff" (poweroff = permit nopass :wheel cmd poweroff -- doas config). this way, i have a clean poweroff i guess. your last point, my last question: so,you mean "exec dbus-launch sway"? 2022-06-08 10:22:10 sry, im in pain and cannot concentrate enough at the moment 2022-06-08 10:30:42 that's okay 2022-06-08 10:30:42 yeah 2022-06-08 10:36:03 whoo, thank you very much and for your work - so much work/compiling stuff like hell... . and thx to the whole ttam and i hope drew is getting better 2022-06-08 10:36:53 hope your migraine gets better :) 2022-06-08 10:37:53 ah, in 2 days im getting better - its always 3 days :( . but thx! 2022-06-08 11:41:11 Hi guys 2022-06-08 11:41:38 Does alpine have an unattended installer? 2022-06-08 11:42:09 There are some articles on the wiki, but they seem to target RPI 2022-06-08 11:42:23 https://docs.alpinelinux.org/user-handbook/0.1a/Installing/setup_alpine.html#_answer_files 2022-06-08 11:43:00 there are a few ways to make it execute some stuff at boot and set things up but i don't remember how good they worked 2022-06-08 11:43:06 and there's a bunch of them i guess 2022-06-08 11:43:13 i don't think any of them were integrated very well 2022-06-08 11:43:17 another option is to use apkovl file 2022-06-08 11:45:10 Is it possible to pass this file via kernel cmdline? 2022-06-08 11:45:17 In order to use e.g. cobbler 2022-06-08 11:45:21 or packer 2022-06-08 12:21:08 Anyone else not able to access files using any flatpak apps since about a week or two? Directories can be listed, files are visible to the picker, but actually picking a file(s) will return an empty list to the program and usually hang them 2022-06-08 13:10:21 hmm, no, works for me 2022-06-08 13:20:28 Hello, is there a lighter alternative for curl in Alpine? 2022-06-08 13:21:56 wget I guess 2022-06-08 13:22:25 something like uclient-fetch is for wget 2022-06-08 13:39:06 yeah you can write a main.c and open a socket and write stuff into it to make a tcp request /s 2022-06-08 13:39:20 realistically, busybox is always there already, so if that wget interface is fine you can just use that 2022-06-08 14:11:56 hey folks, is there plans to implement zstd support for modprobe? 2022-06-08 14:14:54 orange: ask the busybox devs? it has been discussed there in the past but they were not happy with the zstd patches provided to them 2022-06-08 14:16:35 they weren't? iirc it was just ignored 2022-06-08 14:16:47 also `kmod` modprobe supports zstd 2022-06-08 14:16:55 works fine, i've used zstd modules for a while 2022-06-08 14:17:08 psykose: I seem to remember some comments on the list about the size of the zstd code and changes they wanted made to it 2022-06-08 14:17:14 hmm 2022-06-08 14:18:01 or am I getting mixed up with some other form of busybox zstd support? 2022-06-08 14:18:42 orange: specifically, mkinitfs moves /sbin/modprobe to the initrd, and also depends on the `kmod` apk, so /sbin/modprobe is the full kmod one that supports zstd, so as far as alpine initrd goes they should work fine 2022-06-08 14:18:54 if you are doing fancy stuff with busybox-only that is another matter :) 2022-06-08 14:19:00 minimal: no idea, my memory suckks 2022-06-08 14:19:06 also tired 2022-06-08 14:19:50 psykose: doesn't the initramfs use "full fat" kmod/modprobe specifically because at the time around 3.15.0 that modules were compressed using zstd (before the switch to gzip) that busybox mobprobe didn't support zstd? 2022-06-08 14:20:04 yeah, and then it was never removed 2022-06-08 14:20:20 oor 2022-06-08 14:20:50 no, it's depended since aaa4611f40331a6ec7f0096fbbaabdc259820f91 just for modinfo 2022-06-08 14:20:55 i think the actual change was in the move 2022-06-08 14:20:56 psykose: yeah, I tested it locally with the intention of submitting a MR to switch to BB but at the time saw another issue which I seem to remember has since been address. Must test again with BB kmod 2022-06-08 14:21:25 yeah 2022-06-08 14:21:33 it's from mkinitfs itself, since 4f33e938240697175880406fb6e2ec840b434cfe 2022-06-08 14:21:50 'add support for compressed (zstd) .. ' 2022-06-08 14:21:53 thanks minimal psykose, installing kmod solved my issue 2022-06-08 14:22:09 before, it was not moved; but busybox --install filled it in with the bb one 2022-06-08 14:22:18 orange: it should already be installed by mkinitfs, what environment are you in? 2022-06-08 14:22:36 i'm using the docker images 2022-06-08 14:22:44 ah 2022-06-08 14:22:55 orange: why are you using kernel modules in a docker image? 2022-06-08 14:22:56 okay, makes sense :) indeed that is what you want 2022-06-08 14:23:00 though yes 2022-06-08 14:23:43 it's a wireguard container, utlizing the kernel modules from the host 2022-06-08 14:24:02 yeah but you should just have them loaded before you start the container 2022-06-08 14:24:18 why does the container modprobe them for you; they still get put into the host, unless i misremember 2022-06-08 14:24:25 psykose: that modinfo dep on "full fat" version may have been due to the use of "modinfo -k" - which BB modinfo didn't support until recently 2022-06-08 14:25:47 yeah, that was ancient; but it was non-mkinitfs anyway 2022-06-08 14:25:48 makes sense 2022-06-08 14:27:35 psykose: the kernel modules aren't included in the image, loading them from the host seemed like the best option 2022-06-08 14:28:11 you wouldn't be able to load them from the image even if they were there; but that's what i mean 2022-06-08 14:28:16 why does the container load them 2022-06-08 14:29:10 just add `wireguard` into /etc/modules, or some modprobe as a dependency of starting the container, etc 2022-06-08 14:39:56 psykose: what seems to be the last discussion of zstd in Busybox: http://lists.busybox.net/pipermail/busybox/2021-December/089377.html 2022-06-08 14:53:52 psykose: wireguard was explicitly stating that it could not find /lib/modules when trying to enable the interface with 'wg-quick up $interface' 2022-06-08 14:58:25 you load modules from host to host, not into container, containers don't have any knowledge about modules since they are talking to host kernel 2022-06-08 15:00:00 oh 2022-06-08 15:00:03 hang on 2022-06-08 15:02:45 orange: if you look at many of the online articles about running wireguard inside a container they indicate bind-mounting the host's /lib/module to inside the container precisely for this reason 2022-06-08 15:03:13 they also indicate giving the wireguard container other setting required to make it work 2022-06-08 15:06:00 minimal so binding /lib/modules is the correct way of going about this? 2022-06-08 15:07:30 orange: reading and understanding one or more of the various articles about how to use wireguard inside a docker container would be the "correct" way to do proceed 2022-06-08 15:10:06 thanks for clarifying it for me 2022-06-08 15:11:27 for reference, without /var/lib mounted https://0x0.st/oMJC.txt 2022-06-08 15:11:44 with /var/lib mounted but without kmod https://0x0.st/oMyz.txt 2022-06-08 15:12:06 with /var/lib mounted and kmod installed everything works 2022-06-08 15:12:14 orange: explained (to a certain extent) in some of those articles 2022-06-08 15:13:04 plus some of the security implications of getting it working 2022-06-08 15:14:22 thanks minimal, i've got some reading to do 2022-06-08 15:49:18 what does /var/lib have to do with anything 2022-06-08 15:50:19 my bad, i meant /lib/modules 2022-06-08 15:50:23 also why do i smell --privileged 2022-06-08 15:50:39 it is not used 2022-06-08 16:13:46 removed /lib/modules mount and enabled ip6_tables kernel module, that's where my issue was, ip6_tables isn't loaded by default in my kernel 2022-06-08 16:14:11 yeah, you need to do that on the host (kernel), 2022-06-08 16:15:40 got a little better understanding of how it works at least 2022-06-08 16:41:46 Firefox has stopped playing videos. I'm using plasma and x11. When I run from terminal I get this message: 2022-06-08 16:41:46 alpine-lap:/$ firefox 2022-06-08 16:41:46 Missing chrome or resource URL: resource://gre/modules/UpdateListener.jsm 2022-06-08 16:41:46 ATTENTION: default value of option mesa_glthread overridden by environment. 2022-06-08 16:43:49 On youtube, for instance, it nags me to try restarting if video doesn't start. 2022-06-08 17:23:04 psykose: Thanks for testing. Turns out I didn't have the exact combination of XDG Desktop Portals needed (xdg-desktop-portal + xdg-desktop-portal-gtk) 2022-06-08 17:23:20 And this is for XFCE as the DE 2022-06-09 02:21:09 pixelkeen: the attention is unrelated, it says the same for me but plays video 2022-06-09 09:49:55 How can I make single-user mode ask for a password before dropping you in a root shell? 2022-06-09 09:52:15 I tried adding :S:wait:/sbin/sulogin (having previously installed sulogin) and it doesn't work 2022-06-09 09:52:29 adding it in /etc/inittab 2022-06-09 10:16:56 marderbot: you can always set init=/bin/sh and bypass all that 2022-06-09 10:18:33 ikke: I know but if you add a grub password then it can't be bypassed 2022-06-09 10:19:07 the idea is having a lot of layers of security 2022-06-09 10:21:15 marderbot: https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/initramfs-init.in#L507 2022-06-09 10:21:41 It just drops you into a shell 2022-06-09 10:25:25 The inittab only comes into play when openrc starts 2022-06-09 10:26:40 nice, that's exactly the line. The bad thing is that it's insecure and can't be changed without recompiling. Do you think that a pull request may be heard? 2022-06-09 10:27:45 It would certainly be considered. Including tests certainly helps 2022-06-09 10:30:09 what types of tests? (never contributed anything) 2022-06-09 10:31:40 Oh, it was alpine-conf where ncopa added a test suite for, not yet for mkinitfs 2022-06-09 11:01:09 i think it's kinda bad that the initramfs hijacks single-user mode 2022-06-09 11:01:16 it probably shouldn't be involved there at all 2022-06-09 11:08:18 busybox can have sulogin 2022-06-09 11:08:42 I have compiled busybox with that option and it asks for root password 2022-06-09 11:09:03 the problem is that I don't know where it expects the password to be set (rootfs has not been mounted yet) 2022-06-09 12:05:45 okay, I have it working now. The problem is you need to add the shadow file to the initramfs so sulogin works 2022-06-09 12:20:46 I have opened an issue on gitlab, thanks ikke! 2022-06-09 13:58:37 hi 2022-06-09 13:59:11 Is alpine secure enough to run it on server ? 2022-06-09 14:06:17 That's a weird question, it really depends on what security you need. I'm sure there are tons of people running it on servers 2022-06-09 14:08:53 libredev: all our infrastructure runs on Alpine Linux 2022-06-09 14:13:20 sometimes for some reasons GDM does not autologin ... this is an issue for a kiosk machine that really needs to autologin 2022-06-09 14:13:58 can i do whole system updates directly without any break because it is very important for me. 2022-06-09 14:16:20 ok i found the steps in wiki. but is there any chance to break the system. 2022-06-09 14:20:08 libredev: speaking from experience, if you stay away from the testing repo it's typically rock solid. I have all of my servers automatically update and reboot every friday morning 2022-06-09 14:20:31 ok great. 2022-06-09 14:20:33 I've been doing that for 3-4 years now, have never had any issues from updates. 2022-06-09 14:21:25 That said, I also keep my systems light, so there's less to break if all you're updating is the dependencies for one thing. Like my Zabbix server is stand alone, so I only need to worry about Zabbix + its dependencies 2022-06-09 14:21:26 markand: have you tried tinydm? it's designed for autologin and I feel that should start all desktops fine 2022-06-09 14:22:16 let me check, for now I think it only happens when the kiosk didn't power off correctly (maybe a security issue?) 2022-06-09 14:22:46 what are the licenses alpine accept 2022-06-09 14:23:02 because openbsd only accepts ISC 2022-06-09 14:24:12 OpenBSD a 2022-06-09 14:24:20 OpenBSD accepts more than ISC as long as it's enough free 2022-06-09 14:25:58 can i install system without non free blob. 2022-06-09 14:33:38 depends on your hardware libredev 2022-06-09 14:34:06 an intel or broadcom wireless card will always require a firmware 2022-06-09 14:34:06 that means no 2022-06-09 14:34:28 if i don't want then 2022-06-09 14:37:58 then you don't get wireless 2022-06-09 14:38:41 donoban, wow tinydm documentation is pretty anaemic 2022-06-09 14:39:11 ouch, I complained the same months ago and didn't anything for help :( 2022-06-09 14:39:28 it's pretty simple, you enable it as rc-service 2022-06-09 14:39:32 do you know how to run GNOME with it? 2022-06-09 14:39:44 you have to set up with.. 2022-06-09 14:40:22 tinydm-set-session /usr/share/wayland-sessions/gnome.desktop? 2022-06-09 14:40:27 let me check 2022-06-09 14:40:27 do you use wayland or x11? 2022-06-09 14:41:49 you can change default UID (1000) on /etc/conf.d/tinydm 2022-06-09 14:43:10 I've used wayland, seems to work :) I just get a notification that screen locking is disabled, will check how to get rid of this 2022-06-09 14:43:29 nice 2022-06-09 14:43:59 thanks donoban! 2022-06-09 14:44:12 :P 2022-06-09 22:08:38 what it the correct usage case(s) for "apk -p " Docs are scarce... 2022-06-09 22:12:51 maybe only useful for apk operations inside nested container or chroot from an existing alpine install? 2022-06-09 22:30:58 openrc-0.45 has a bug if you use openrc-init. this isn't used by alpine yet but careful 2022-06-09 22:43:50 jammers: apk add -p root alpine-base, that's about it 2022-06-09 22:44:39 how best to install packages to alternate destinations, say.. /opt/ or similar then? 2022-06-09 22:47:34 chroot root apk add 2022-06-09 22:48:55 only then you have to have full chroot system contents per directory :( 2022-06-09 22:50:10 or make a mount bind mess ;) 2022-06-09 22:50:31 there is no way to make apk 'just install things to /opt instead of the usual /usr/* paths' 2022-06-09 22:50:34 a new root is well, a new root 2022-06-09 22:50:52 the package itself specifies its layout 2022-06-09 23:41:27 well, if you use apk -p then it will move the files there as you request 2022-06-09 23:41:33 but that doesn't mean you can run them there 2022-06-09 23:59:14 cant use apk -p with an empty destination, just error out, as it cant find the db or related apk deps there... 2022-06-10 00:04:20 https://wiki.alpinelinux.org/wiki/Alpine_Linux_in_a_chroot 2022-06-10 00:05:44 yes, can't run directly from /opt 2022-06-10 00:05:46 can symlink disk contents mounted to /opt into the real ramfs root where needed tho. 2022-06-10 00:09:08 jammers: `apk add -p /opt --initdb`, `cp /etc/apk/repositories /etc/apk/keys /opt/etc/apk/`, `apk -p /opt -U add ` 2022-06-10 00:09:32 Try those commands, and see if it does what you want 2022-06-10 00:13:27 @rbq thanks, will try that. 2022-06-10 00:13:28 This is not the standard container / docker / chroot setup. 2022-06-10 00:13:30 very small base alpine root (ramfs) with larger apps symlinked from exterior mounted /opt 2022-06-10 00:17:08 external mounted /opt/ to be precise... 2022-06-10 00:18:11 Ok, then change /opt to /opt/ in what I wrote 2022-06-10 00:19:12 The thing is, you'll have to manage upgrades separately for each of the /opt/ directories 2022-06-10 00:50:07 jammers: this is just an poorly implemented overlayfs? 2022-06-10 00:50:49 @rbq - good call.. will add that in the setup scripts as well for upgrading 2022-06-10 00:52:42 @Hello71 - Kernel is locked down on device, no loop, no fuse, no overlayfs 2022-06-10 00:54:10 but yes, a poor mans overlay | union | whatever fancy new fs you like :) 2022-06-10 00:58:08 i'm not sure this is superior to just using a chroot 2022-06-10 00:58:14 it sounds much more fragile 2022-06-10 01:37:14 very minimal hardware. not upgrade able, locked bootloader, yadda yadda 2022-06-10 01:37:15 seeing if Alpine can make it useful again 2022-06-10 08:28:35 good morning folks, just installed alpine in a vm 2022-06-10 09:15:05 one thing i just learned today, if you change /etc/localtime, changes apply immediately 2022-06-10 09:20:46 Time is an illusion. Local time, doubly so. 2022-06-10 09:21:12 by the time you have read /etc/localtime, it is too late 2022-06-10 09:24:33 does that mean we can invent a time machine? :) 2022-06-10 09:25:12 try it. if you disappear, we'll know you were either successful or very, very wrong 2022-06-10 09:25:39 ok 2022-06-10 09:26:04 will do it later 2022-06-10 09:28:00 Is it safe to include `[[ $- != *i* ]] && return` at the top of .ashrc? 2022-06-10 09:28:01 well i don't remember you tell us last week that you were successful, so… 2022-06-10 09:28:23 (the default header for .bashrc so it exits on non-interactive shells) 2022-06-10 09:28:35 #bash will tell you if it's posix, and should work 2022-06-10 09:28:43 on libera 2022-06-10 09:28:44 doesn't return an error, so I guess it's fine 2022-06-10 09:28:54 i'll leave a scrot: https://transfer.sh/NjktKy/2022-06-10-162653_1360x768_scrot.png 2022-06-10 09:29:16 lopid: as a matter of fact, I'm not sure why I'm able to execute bashisms fine on Alpine 2022-06-10 09:29:27 I expected it to be like the time I used dash 2022-06-10 09:29:59 ash supports a lot of bashisms 2022-06-10 09:30:52 Ok, so it's intentional. That's neat, actually - having a still-minimal shell with sensible QoL improvements. 2022-06-10 10:05:09 i'm back 2022-06-10 10:05:27 i forgot who replied to me about the TM thing 2022-06-10 10:05:37 but i just did and the whole thing disappeared 2022-06-10 10:06:29 as in the computer itself 2022-06-10 10:07:04 it was a school computer anyway, so that's why i don't authenticate using nickserv on that 2022-06-10 15:26:10 I have a problem with my crontab. For some reason what I put in /etc/periodic/*/whatever does not run 2022-06-10 15:26:56 does it end with a .extension 2022-06-10 15:27:00 and do you have crond running 2022-06-10 15:27:17 crond is running 2022-06-10 15:27:23 and what do you mean .extension? 2022-06-10 15:27:27 script.sh 2022-06-10 15:27:42 a dot in the name 2022-06-10 15:27:47 my scripts don't have an extension 2022-06-10 15:27:48 Why? 2022-06-10 15:27:55 just one of the reasons it wouldn't work 2022-06-10 15:28:00 aside from that, are they executable 2022-06-10 15:28:14 oh my god 2022-06-10 15:29:20 yep, one of them wasn't executable 2022-06-10 15:30:05 FYI, it's the run-parts command that handles this and has these requirements 2022-06-10 15:30:15 I saw 2022-06-10 15:30:34 https://www.systutorials.com/docs/linux/man/8-run-parts/ 2022-06-10 15:30:54 do I need to restart crond after changing scripts in /etc/periodic? 2022-06-10 15:31:03 (though, we have the busybox version) 2022-06-10 15:31:04 no 2022-06-10 18:27:10 Okay even with the executable bit set, my scripts in /etc/periodic won't run 2022-06-10 18:31:25 I also made a test script that writes to /tmp/test123.log and set the executable bit on that in /etc/periodic/15min but it just never runs 2022-06-10 18:35:09 Soa nyone know what's going on? 2022-06-10 18:35:16 latex: which cron are you using? busybox? cronie? dcron? 2022-06-10 18:35:23 is the cron process actually running? 2022-06-10 18:43:22 is DROPBEAR_OPTS="-w -s" what I need to connect with SSH keys - no password auth - when the public key for my client is in /etc/dropbear/authorized_keys? 2022-06-10 18:43:44 because -s disables password login - I'm not going to lock myself out when I restart the machine, right? 2022-06-10 18:45:05 minimal: busybox and rc-service reports it is running 2022-06-10 18:45:47 latex: what is the name of the file you added to /etc/periodic/15min/ ? 2022-06-10 18:45:56 test 2022-06-10 18:46:33 wait I am stupid again. It was in daily. Let's wait 15 minutes again 2022-06-10 18:47:19 You can just try to execute command from crontab 2022-06-10 18:48:17 latex: does /var/log/cron.log show anything? 2022-06-10 18:51:42 /var/log/cron.log does not exist 2022-06-10 18:53:14 latex: have you a syslog daemon running? 2022-06-10 18:53:55 Btw my alpine system does not have crond.log too 2022-06-10 18:54:41 have you guys tried groff? kinda interesting alternative to latex (just happened to think of this for some reason) 2022-06-10 18:55:03 frag: more of a competitor with texinfo, no? 2022-06-10 18:55:22 can it typeset math and all? 2022-06-10 18:56:06 minimal: well I have logrotate 2022-06-10 18:56:20 yes if you use eqn 2022-06-10 18:56:28 ah, it has a preprocessor addon. interesting 2022-06-10 18:57:09 Ermine: cron.log, not crond.log 2022-06-10 18:57:09 huh, eqn is only 4 years older than TeX 2022-06-10 18:57:34 minimal: this one missing too. Apparently it gets created on first write 2022-06-10 18:58:48 talismanick: yes i think it can typeset math yes, but i dont personaly do it 2022-06-10 19:00:06 tex and groff serve different auditories to my mind 2022-06-10 19:00:42 groff - man pages, tex - scientific papers/books/etc 2022-06-10 19:00:57 Almost everyody on my faculty use tex 2022-06-10 19:01:17 I use plaintext 2022-06-10 19:01:23 despite my name 2022-06-10 19:01:30 latex: I see an entry written to /var/log/cron.log each time a periodic run happens (NOT the individual scripts, just that the 15min, hourly, etc is happening) 2022-06-10 19:01:52 I see it in /var/log/messages 2022-06-10 19:01:56 Ermine: well, thats why i mentioned it, seem few ppl know you can use groff for papers also 2022-06-10 19:02:00 And those who do not use tex create documents with shiity formulas 2022-06-10 19:02:01 or just letters etc 2022-06-10 19:02:02 latex: ah, forgot I'm running rsyslog here, not busybox syslogd 2022-06-10 19:02:26 anyway, frag, try both and make your choice 2022-06-10 19:02:34 Okay so /etc/periodic/15min just ran, but I don't see my test file, so my test script did not run 2022-06-10 19:02:50 latex: so what's in your script? 2022-06-10 19:03:02 latex: try doas /etc/periodic/15min 2022-06-10 19:03:05 Ermine: i have :] 2022-06-10 19:03:56 echo test >> /tmp/test123.log 2022-06-10 19:04:14 did you add shebang line? 2022-06-10 19:04:14 latex: literally just that? no 1st line with "#!/bin/sh" ? 2022-06-10 19:04:28 my system doesn't have doas or any privilege escalation utility 2022-06-10 19:04:38 I dsabled su too 2022-06-10 19:04:58 login as root then 2022-06-10 19:04:59 oh my god the shebang line 2022-06-10 19:05:11 thanks for reminding me 2022-06-10 19:05:43 also do you need to specify the full path to executables inside cron scripts? 2022-06-10 19:05:50 Because I see other scripts do it too 2022-06-10 19:06:18 crond may or may not set PATH 2022-06-10 19:06:27 you can and probably should set it yourself 2022-06-10 19:15:14 welp, locked myself out 2022-06-10 19:15:35 thanks, dropbear(1) 2022-06-10 19:16:20 copied .pub to both /etc/dropbear/authorized_keys and /home//.ssh/authorized_keys 2022-06-10 19:16:28 so not sure why it denied me 2022-06-10 19:17:00 ah, nvm 2022-06-10 19:17:25 the key wasn't in my keychain somehow - interesting... 2022-06-10 19:38:15 Okay thanks everyone, my cron tasks work now 2022-06-10 19:38:22 Shebang line was the issue 2022-06-10 19:49:39 Also does anyone know of a way to measure cpu and other statistics per process over time? 2022-06-10 21:31:23 latex: Prometheus. If you do `apk add prometheus prometheus-node-exproter` you will get something like this: 2022-06-10 21:31:26 https://metrics.sr.ht/graph?g0.expr=node_load1&g0.tab=0&g0.stacked=0&g0.show_exemplars=0&g0.range_input=1h 2022-06-11 05:17:43 hello all! alpine 3.15.4 user. installed docker, having some trouble with the containers networking though. if i use networkmanager to connect to my wifi, the containers cannot access any network (local, internet, etc). if i disable networkmanager and use wpa_supplicant and udhcpc, docker container networking works fine. i've written a little network triage script to debug all 2022-06-11 05:17:45 system networking info, and the only difference i can see between the two is /proc/sys/net/ipv6/conf/wlan0/addr_gen_mode and /proc/sys/net/ipv6/conf/wlan0/use_tempaddr are set differently 2022-06-11 05:18:04 was hoping someone has seen this issue with networkmanager and docker and figured out a fix 2022-06-11 06:16:26 perlfu: are the docker iptables rules still in place? 2022-06-11 06:17:06 I don't use network manager, so no experience with that 2022-06-11 09:19:06 I'm having experience with networkmanager, docker, firewalld - albeit not on alpine. If there's firewalld support, you might be able to put the docker stuff into a zone that is allowed to connect 2022-06-11 13:31:24 ikke: yes, iptables rules look fine. i've diffed routing tables, iptables, bridges, ip addr show, and all of /proc/net, /proc/sys/net, and /sys/devices/virtual/net, and the only difference between docker networking working and not working are those two entries i mentioned 2022-06-11 13:31:58 i know something else must be going on but i don't know what other networking info to look for 2022-06-11 13:32:38 i guess next step is to do actual debugging of networkmanager and udhcpc and see what's going on under the hood 2022-06-11 16:56:14 Hey guys. I faced up with bug using lynx browser in alpine. Can you please help me to fix it? 2022-06-11 17:19:07 what bug? 2022-06-11 17:44:34 "what bug?" <- @Shifulor @mckaygerhard... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/HSikTgTzfaZxbrSDFWNAPMbb) 2022-06-11 17:56:55 ? 2022-06-11 18:08:43 "?" <- What? 2022-06-11 18:27:56 boombim[m]: can you reproduce bug on non-alpine system? 2022-06-11 18:44:49 boombim: can you reproduce bug on non-alpine system? 2022-06-11 18:44:57 Yes... 2022-06-11 18:45:01 On void as well 2022-06-11 18:45:09 So it's not alpine issue 2022-06-11 19:53:49 hello 2022-06-11 19:54:40 can somebody help me with a pi problem? 2022-06-11 19:54:54 randomjohn: works better if you just ask an actual question 2022-06-11 19:55:02 sorry, first time 2022-06-11 19:55:11 i'm trying to change the boot order 2022-06-11 19:55:16 it keeps trying to boot to usb 2022-06-11 19:55:28 but I can't find the utilites like rpi-eeprom 2022-06-11 19:56:18 and even if I use the ones in github it still doesn't work 2022-06-11 20:09:23 Hello. How to change the boot order in the raspberry pi? 2022-06-11 20:40:17 Is Varnish still faster than Squid? 2022-06-11 20:44:11 Has https://queue.acm.org/detail.cfm?id=1814327 held up to the test of time? It's certainly still entertaining. 2022-06-11 22:51:31 An Awall rule for squid apparently looks like https://0x0.st/oM7V.txt 2022-06-11 22:51:39 Do I need to write it differently for varnish? 2022-06-12 00:00:19 psykose: hi, I found that waybar has a taskbar module called wlr/taskbar which does the same as sfwbar :P 2022-06-12 00:00:45 that's great :) 2022-06-12 00:01:34 thank you for helping me the other day 2022-06-12 00:23:10 ^~^ 2022-06-12 00:23:22 presently i was having fun building the latest electron version on alpine 2022-06-12 01:14:47 "fun" 2022-06-12 01:33:27 it was 2022-06-12 02:16:38 ash: /usr/lib/ssh/sftp-server: not found 2022-06-12 02:17:43 I already added vsftpd to the default runlevel and rebooted (thought maybe there was some user without permissions before logging back in, but that's evidently not the issue) 2022-06-12 02:18:20 Symlink the vsftpd binary? 2022-06-12 02:21:02 that is probably from openssh-sftp-server 2022-06-12 02:21:28 or, hm 2022-06-12 02:21:47 psykose: that was my guess, but the wiki even calls vsftpd the "default" 2022-06-12 02:21:53 vsftpd is a completely different thing, it's a standalone ftp server 2022-06-12 02:21:59 could you link the wiki page? 2022-06-12 02:22:01 not sure what uses what, if sshd uses vsftpd to provide sftp, then you want to set the sftp server in sshd_config to that 2022-06-12 02:22:02 so I figure I must be something wrong 2022-06-12 02:22:17 https://wiki.alpinelinux.org/wiki/FTP#vsftpd 2022-06-12 02:22:37 oh, yeah, that's completely unrelated to SFTP (file transfer over SSH) 2022-06-12 02:22:56 Ah, so I likely need to add something to dropbear 2022-06-12 02:23:38 you just need to install the package psykose mentioned earlier, openssh-sftp-server 2022-06-12 02:23:42 it has the file you're missing 2022-06-12 02:24:08 for sshd specifically you can use `internal`, for dropbear you probably need to add that 2022-06-12 02:24:39 ah, nice 2022-06-12 02:24:42 thanks 2022-06-12 02:25:14 I guess it's separately packaged /so/ I don't have to install the whole OpenSSH suite? 2022-06-12 02:25:53 or if you just want the openssh server without sftp, yeah 2022-06-12 05:15:58 is there something like ifplugd that I could use to automatically do something when a network interface goes up or down due to gaining or losing its physical connection? 2022-06-12 10:56:27 Is there any way I can add a speak up to the Alpine image to get speech for an install? 2022-06-12 11:21:28 sethmhurst[m]: https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/mkimg.standard.sh defines what's inside the image 2022-06-12 11:21:38 so you could add packages to there 2022-06-12 11:21:46 but not sure how to integrate it in the installation further 2022-06-12 16:14:47 What I'm way what I'm looking for is a way to add it while booted into a live CD. 2022-06-12 16:43:05 do there exist some packages for alpine who can take a current install of alpine and make that into an iso? 2022-06-12 16:43:13 similar to pinguybuilder for debian/ubuntu? 2022-06-12 16:43:21 livecd that is 2022-06-12 16:45:32 I've only ever used it on linode so I'm not really sure if there is. 2022-06-12 16:46:24 I know you can run it you can run the setup script but I don't think you speak up is listed in the APK. 2022-06-12 17:24:36 \q 2022-06-12 17:24:40 quit 2022-06-12 17:58:21 That's what we call, a dramatic exit :P 2022-06-12 22:51:58 Computer: End program. 2022-06-13 02:55:27 when i fullscreen mpv, the video gets distorted unless I move my mouse around 2022-06-13 02:55:32 this issue started with 3.16 2022-06-13 02:55:36 anyone else have this issue? 2022-06-13 03:53:18 sethmhurst[m]: Maybe brltty is something that could work for you? 2022-06-13 09:59:07 In my alpine container the characters ü,ö,ä,€,° show as "_". How can I fix that? 2022-06-13 10:02:07 docker? 2022-06-13 10:18:30 lopid: No, podman. 2022-06-13 10:18:51 (But I'm using the official image from dockerhub.) 2022-06-13 10:19:04 is it a ui problem? 2022-06-13 10:20:42 lopid: Mhh, I just noticed that the issue appears only in tmux. 2022-06-13 10:20:54 Out of tmux the characters are displayed correctly. 2022-06-13 10:21:55 Wait, there is more to it. 2022-06-13 10:22:27 If I enter the container and then start tmux it works fine, too. 2022-06-13 10:23:28 Looks like the issue only happens if I parse tmux as initial parameter to the "podman run" command.` 2022-06-13 10:27:33 With `./run /bin/sh -c "tmux"` the issue appears, too. 2022-06-13 10:27:58 As well as with /bin/bash instead of /bin/sh. 2022-06-13 10:30:18 perhaps the environment is different 2022-06-13 10:30:36 It works with `./run /bin/bash -l -c "tmux"` 2022-06-13 10:31:14 I guess if I directly invoke tmux then the login-shell functionality is missing?! Does that make sense? 2022-06-13 10:31:32 env | sort 2022-06-13 10:38:35 lopid: `env | sort` (invoked from within tmux) shows the same output in every case. 2022-06-13 10:40:29 Yea, `env | sort | sha256sum` shows that it's exactly the same. 2022-06-13 13:46:08 LANG needs to be set correctly when running tmux 2022-06-13 13:46:58 or, as a hack, invoke tmux with -u 2022-06-13 13:47:08 but then it will still use the wrong language 2022-06-13 17:39:05 I'm trying to load x86_64 alpine in a VM on macos (with virtualization.framework). the framework needs a kernel image to boot, so I extracted vmlinuz-virt from the .iso, but for some reason it gets stuck without any console output for the standard image, while it seems to work with the netboot image vor some reason. does anybody know what's different about the 2 kernel images? the header etc. looks v 2022-06-13 17:50:50 oh, I meant the virt image, but it's the same for the standard one, that they don't load. but netboot does 2022-06-14 15:18:35 Anyone tried to get Alpine headless running on RPi 4 lately? There are a lot of roadbumps 2022-06-14 15:19:19 Currently cannot ssh into a fresh install, though I can ping it 2022-06-14 15:27:10 not yet 2022-06-14 16:24:59 jingles: do you have other means to connect to your rpi 4 and check sshd status? 2022-06-14 16:32:54 Hey, minimal helped me out in a DM, thank you. 2022-06-14 16:33:30 Although to answer your question, this pi is purely headless in this immediate moment. I don't have the right cables yet 2022-06-14 16:35:35 Figured it'd be easy to just ssh in and set it up that way, but it seems sshd isn't running or is refusing connections. I'm going to switch from alpine headless to rasp os headless just to get this particular project up and running for now 2022-06-14 16:35:56 I would like to convert it to alpine some time in the future but I only have 2 brain cells left 2022-06-14 16:43:48 wondering if any of you guys has the same issue with lxc on 3.16, after upgrading lxc is not found anymore but installed from apk point of view, I guess a ln is missing. I am using lxd and lxc is supposed to be a dependancy so no more lxc command are working.. 2022-06-14 16:44:45 We have not upgraded our LXC hosts to 3.16 yet 2022-06-14 16:45:16 blt[m]: is lxc-client installed? 2022-06-14 16:45:30 sorry, lxd-client 2022-06-14 16:46:03 lxc itself only has lxc-* commands 2022-06-14 16:46:26 😲 2022-06-14 16:46:59 ikke: I am discovering that! it was the case in the past? I don't recall installing lxd-client (it is indeed working... thanks for that) 2022-06-14 16:47:35 We just use plain lxc, so not sure 2022-06-14 16:47:44 lxd is anyway flagged out as 5.0 is LTS for already a couple months ago but I guess some dependancies are tricky to make the move on Alpine? 2022-06-14 16:47:46 but I see that 3.15 does not have a lxc command 2022-06-14 16:48:05 fcolista maintains it, so he might know better 2022-06-14 16:48:37 blt[m]: I just checked https://pkgs.alpinelinux.org/contents?file=lxc&path=&name=&branch=edge&arch=x86_64 btw 2022-06-14 16:50:10 https://pkgs.alpinelinux.org/packages?branch=edge&repo=community&name=lxd this is flagged and indeed 5.0 is LTS (https://linuxcontainers.org/lxd/news/) 2022-06-14 16:56:02 blt[m]: If you are in the opportunity, you could try to make an MR to upgrade it 2022-06-14 17:35:31 ikke: I wish I could do it but I believe it is more complicated than just changing the version variable in the APKBUILD as the dependencies may need update as well 2022-06-14 17:36:04 yes, that's possible 2022-06-14 19:13:17 Hello!! I'm attempting to dual boot Alpine on my system 2022-06-14 19:13:57 The installation works without error, except for an MBR fix warning - however I use UEFI 2022-06-14 19:14:21 Do you have CSM off? 2022-06-14 19:14:45 On bootup, I get a mount error: UUID=XXX on /sysroot failed: no such file or directory 2022-06-14 19:15:04 do you get emergency shell? 2022-06-14 19:15:29 Yeah! 2022-06-14 19:15:43 I can check if CSM is on or off, I don't know 2022-06-14 19:16:19 When I run blkid, I get no output - none of my partitions are registered on /dev 2022-06-14 19:16:31 ls /dev ? 2022-06-14 19:16:42 ^ anything relevant? 2022-06-14 19:17:01 you can also check /dev/disk/by-uuid 2022-06-14 19:17:40 I'm looking for nvmen0p*, but none of them are available :P 2022-06-14 19:18:03 And /dev/disk has only 'by-id' and it's an empty directory 2022-06-14 19:18:04 So you have nvme? 2022-06-14 19:18:07 Yeah 2022-06-14 19:18:33 And which Alpine version do you try to install? 2022-06-14 19:19:30 Standard x86_64 2022-06-14 19:19:38 Latest release 2022-06-14 19:20:01 so v3.16? 2022-06-14 19:20:30 Yeah! 2022-06-14 19:20:58 ikke: looks like that nvme story is not finished? 2022-06-14 19:21:04 Maybe 2022-06-14 19:21:16 AndroUser: you are in an emergency shell, right? 2022-06-14 19:21:48 Yeah 2022-06-14 19:21:51 Emergency shell 2022-06-14 19:22:00 What does cat /proc/cmdline return? 2022-06-14 19:24:18 BOOT_IMAGE=/boot/vmlinuz-lts root=UUID=XXX ro modules=sd-mod,usb-storage,ext4,nvme quiet rootfstype=ext4 2022-06-14 19:24:56 XXX is the wrong UUID, however when I edit the grub boot menu I get the correct UUID shown there 2022-06-14 19:25:13 and lsmod | grep nvme? 2022-06-14 19:26:14 nvme 49152 0 2022-06-14 19:26:21 So it did properly detect nvme 2022-06-14 19:26:28 so seems to be a different issue 2022-06-14 19:27:31 AndroUser: and there is no /dev/sd*? 2022-06-14 19:27:58 AndroUser: "except for an MBR fix warning" - what warning exactly? 2022-06-14 19:28:11 No, however I don't have any /dev/sd* device in the first place so not an issue 2022-06-14 19:28:40 'You might need fix the MBR to be able to boot.' 2022-06-14 19:29:20 https://serverfault.com/questions/1048119/alpine-setup-on-virtual-machine-uefi-does-not-boot this post said that shouldn't be relevant 2022-06-14 19:29:57 (in my case, there's no virtual machine) 2022-06-14 19:31:57 AndroUser: if you don't have /dev/sd* normally, what _do_ you have normally? 2022-06-14 19:33:39 I have /dev/nvme* 2022-06-14 19:36:10 My current artix linux and windows installation boots off nvme partitions 2022-06-14 19:37:01 AndroUser: so when it fails to boot, do you get an error message? 2022-06-14 19:37:07 except for that MBR warning? 2022-06-14 19:37:56 I get the MBR warning after I run setup-disk to the partitions 2022-06-14 19:38:01 On bootup, I get a mount error: UUID=XXX on /sysroot failed: no such file or directory 2022-06-14 19:39:40 AndroUser: did you partition the disk yourself or let setup-alpine/setup-disk do it? 2022-06-14 19:40:21 Myself, following instructions on https://wiki.alpinelinux.org/wiki/Dualbooting 2022-06-14 19:42:26 AndroUser: so how many partitions (and of what size/type) did you create? 2022-06-14 19:43:16 and what type of Alpine install did you do? Sys-mode? Data-mode? Diskless-mode? 2022-06-14 19:44:41 Sys mode 2022-06-14 19:44:55 I currently have 9 partitions, it's a gigantic mess 😅 2022-06-14 19:45:30 that's fine :D 2022-06-14 19:46:27 AndroUser: all 9 partitions are on the same disk? 2022-06-14 19:47:15 Yeah! They range between 250MiB and 750GiB in size, ext4/NTFS/vfat 2022-06-14 19:50:29 AndroUser: so did the Alpine install then write a EFI files to the efi/alpine/ and efi/boot/ directories of the ESP partition? 2022-06-14 19:53:38 AndroUser: you said "XXX is the wrong UUID, however when I edit the grub boot menu I get the correct UUID shown there", that makes no sense unless you booted via a different menu item than the one you edited 2022-06-14 19:58:51 There's just one menu entry, however I'll confirm just to be sure 😃 2022-06-14 19:59:14 Alpine live environment did write to the efi directories during install yeah! 2022-06-14 20:00:35 AndroUser: so is this just basic rootfs on ext4 or did you specify LVM or encryption or a different filesystem type (btrfs, xfs, zfs, etc)? 2022-06-14 20:02:20 Basic rootfs on ext4 2022-06-14 20:02:31 No encryption or LVM 2022-06-14 20:02:54 nothing in "dmesg" output to show why no NVME device was discovered? 2022-06-14 21:27:43 Anyone able to offer some guidance to an attempt to get 3.16.0 sys install up via UEFI? I'm apparently losing something which is leading to me losing my mind. :) 2022-06-14 21:30:30 PedanticAvenger: you are the 2nd person in the past couple of hours to have mentioned the saem/similar problem 2022-06-14 21:30:41 what is the problem? 2022-06-14 21:30:54 Heh, I was certain I had it working but fell apart later. 2022-06-14 21:31:54 I have a build that "seems" to go well but on reboot no boot options found. I've got a gdoc I've been documenting my steps in but even to gett he this point I am confused by some apparently conflicting wiki info. 2022-06-14 21:32:40 https://docs.google.com/document/d/1PEL4LeWBfgaRV9yFYhmkVgAacP_-_UZWLsYtGXFLMow 2022-06-14 21:33:08 what does "no boot options found" mean? Does Grub load? If so does it not have any menu items? 2022-06-14 21:34:00 the motherboard can't see that there is an appropriate partition/boot entry. 2022-06-14 21:34:58 so in other words Grub is not being loaded? 2022-06-14 21:35:41 booting from a NVME device or a SATA HDD/SSD? 2022-06-14 21:35:53 SATA HDD 2022-06-14 21:36:15 have a dell R220 server I'm trying to use for the build. 2022-06-14 21:36:49 did you let setup-alpine/setup-disk partition the disk and format filesystems or did you do it manually? 2022-06-14 21:37:17 manual partition setup and mount on /mnt followed by setup-disk 2022-06-14 21:37:28 and the ESP partition? 2022-06-14 21:37:47 100Meg vfat, mounted as /mnt/boot/efi 2022-06-14 21:38:32 512 byte or 4K sector drive? 2022-06-14 21:39:01 Good question. Dunno 2022-06-14 21:39:22 as the size of the ESP partition may be a factor if a 4K logical drive 2022-06-14 21:40:44 Using... 2022-06-14 21:40:58 parted /dev/sda 2022-06-14 21:40:58 mkpart ESP ext2 1 100 2022-06-14 21:40:58 mkpart primary ext4 100 81000 2022-06-14 21:40:58 mkpart primary linux-swap 81000 97000 2022-06-14 21:40:58 I get no erors. 2022-06-14 21:41:11 and which tool did you use for format the vfat partition? (just to be sure it is FAT32 rather than FAT16) 2022-06-14 21:41:23 mkfs.vfat 2022-06-14 21:41:24 you just created a ext4 ESP part then, not a vfat one... 2022-06-14 21:42:43 Does the partition label actually care for ESP or the format? 2022-06-14 21:43:04 but I'll re-do that setup, I've got it mostly scripted now so I'll kick it. 2022-06-14 21:44:05 yes it is important to create the ESP properly 2022-06-14 21:44:20 Right, on it. 2022-06-14 21:45:11 I use parted myself so don't know the fdisk/cfdisk commands for it 2022-06-14 21:46:02 Does that partition need to be 100 Meg? I'd rather have an esp part mounted as /boot but the wiki docs are confusing around this. 2022-06-14 21:46:57 well that goes back to my note about FAT32 vs FAT16 - they are limits on min and max sizes for FAT12, FAT16, and FAT32 2022-06-14 21:47:39 only support for FAT32 is mandatory in the UEFI specs AFAIK, so some/many machines may not boot from a FAT12 or FAT16 ESP partition 2022-06-14 21:48:55 and also the min/max size varies depending on whether disk has 512 byte or 4K logical sectors 2022-06-14 21:49:17 Ok, so if I make a 500Meg FAT32, and mount it as boot... 2022-06-14 21:49:31 I'm on the server, there quick way to check sector size 2022-06-14 21:49:33 ? 2022-06-14 21:50:10 Oh 512B 2022-06-14 21:51:41 setup-disk seems to assume ESP partition is mounted as /boot/efi 2022-06-14 21:51:41 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-disk.in#L298 2022-06-14 21:52:08 also assumes that partitioning for UEFI is GPT, not MBR 2022-06-14 21:52:13 it is 2022-06-14 21:52:46 Argh, it assumes /boot/efi but doesn't automatically copy files from /boot into /boot/efi. Sigh 2022-06-14 21:53:11 copy which files from /boot into /boot/efi ? 2022-06-14 21:53:43 initramfs-lts and vmlinuz-lts 2022-06-14 21:54:00 why would it copy those into /boot/efi? grub loads them from /boot 2022-06-14 21:54:51 Ok, then this is back to wiki confusion. Namely the Bootloaders section at the very bottom 2022-06-14 21:54:56 when using Grub then the only files typically in the ESP partition is/are the Grub EFI file(s), i.e. EFI/alpine/grub???.efi and EFI/boot/boot???.efi 2022-06-14 21:55:15 I ignore most wiki articles are they are inaccurate or out-of-date 2022-06-14 21:55:23 I'm getting that..... 2022-06-14 21:55:46 Is the boot stub stuff still appropriate/relevant? 2022-06-14 21:55:59 Or just ignore and discover via motherboard utilities? 2022-06-14 21:56:10 the Grub EFI file needs to be in the ESP partition for the UEFI ti load it, it then knows where to find the grub.cfg file which tells it where the kernel and initramfs are 2022-06-14 21:56:55 boot stub? you mean EFI stub? 2022-06-14 21:57:52 yes, using efibootmgr to update the boot list/order 2022-06-14 21:58:44 well I guess that depends on whether you are trying to multiboot 2022-06-14 21:59:46 Ok, so in this scenario efi partition is a mirrored partition. identical on both HDs. 2022-06-14 22:00:02 I don't think that setup-alpine/setup-disk is designed with multiboot in mind 2022-06-14 22:00:06 So really I should be ok ignoring it as I dont care which drive it grabs. 2022-06-14 22:00:22 mirrored how? by manual copying? 2022-06-14 22:00:41 raid1 2022-06-14 22:00:45 typically the UEFI should look at the ESP partition on each drive to see what EFI files are present in them 2022-06-14 22:01:30 raid1 where? the motherboard's drive controller is doing hardware RAID? or pseudo "hardware" RAID? 2022-06-14 22:01:38 softraid 2022-06-14 22:02:13 If it were hard raid the system would only ever see one disk 2022-06-14 22:02:22 don't think Alpine has drivers for that in the initramfs so it might see them as 2 separate drives 2022-06-14 22:02:22 (and I don't mean pseudo) 2022-06-14 22:02:49 And at boot that should be fine, it should start the raid for that part after the system comes up and allow it to be updated/managed/etc without hassle. 2022-06-14 22:03:10 if I use the right metadata format it should still look normally useable as a FAT32 partition 2022-06-14 22:03:18 you mean MD? when I used the term "pseudo" RAID I was thinking of the controllers that some motherboars used that also required software support to work 2022-06-14 22:03:30 I do mean MD 2022-06-14 22:03:55 so the UEFI then will see 2 ESP partitions, not one 2022-06-14 22:04:01 yes 2022-06-14 22:04:32 and I was trying to use efibootmgr to setup both of them with sequential boot priority so if one disk died it would just go on to next. 2022-06-14 22:05:28 so when you booted to do the install did Alpine see 2 sets of partitions or single /dev/md* devices? 2022-06-14 22:06:07 to do install I mounted the /dev/md devices in /mnt then ran setup-disk -m sys -v /mnt/ 2022-06-14 22:06:17 and for the ESP? 2022-06-14 22:06:29 also with bootloader, disklabel, use_efi environment variables. 2022-06-14 22:06:48 esp mounted as /mnt/boot/efi originally via /dev/md0 2022-06-14 22:07:56 originally via /dev/md0 - that might be a problem, as the underlying partition would include metadata AFAIK so it would not be a "real" ESP partition but rather be a softraid partition 2022-06-14 22:08:16 which them the UEFI likely wouldn't like 2022-06-14 22:08:53 I think you can't use softraid for ESP 2022-06-14 22:10:02 it "should" be ok if I use version 1 metadata which puts the superblock at the end of the device. But if it is at the front I would certainly agree. 2022-06-14 22:10:27 1.2 might be ok too as it's 4k from start. But that is theoretical certainly. 2022-06-14 22:11:13 BUT at this point I'm willing to stop putting RAID there and try a regular sync job to keep them in alignment instead 2022-06-14 22:11:56 minimal, fwiw most UEFI implementations share a common core 2022-06-14 22:12:02 sane behavior *usually* applies 2022-06-14 22:12:27 upstream tianocore supports FAT12,16 and 32 for instance 2022-06-14 22:13:22 (also ext4 support written by me if you know how to set that up/are compiling your own fw image :)) 2022-06-14 22:14:00 heat_: yes I use FAT12 for ESP on QEMU. It is more likely to be a issue on physical machines (how many UEFI vendors are left these days?) 2022-06-14 22:14:33 certainly a good bunch. but afaik they all use tianocore's stuff 2022-06-14 22:15:10 you also have u-boot which implements UEFI... and that's about all the UEFI implementations I know of 2022-06-14 22:16:09 heat_: is some of that stuff in tianocore/EDKII configurable? as then there's nothing to stop a vendor simply not enabling some stuff when they create their UEFI 2022-06-14 22:16:29 everything is super configurable, up to an unreasonable extent :P 2022-06-14 22:16:40 but as for FAT? I don't think it is 2022-06-14 22:17:52 yup, FAT12/16 doesn't seem to be configurable 2022-06-14 22:18:59 PedanticAvenger: I think you need to set USE_RAID=1 USE_EFI=1 when calling setup-disk 2022-06-14 22:21:01 I'll add USE_RAID=1 already using USE_EFI=1, DISKLABEL=gpt and BOOTLOADER=grub 2022-06-14 22:22:02 USE_EFI=1 means use grub as Syslinux UEFI isn't supported by setup-disk 2022-06-14 22:22:54 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-disk.in#L1422 2022-06-14 22:23:55 I'm beginning to loath the wiki 2022-06-14 22:24:33 But those two weren't in conflict so that is fine. 2022-06-15 00:21:06 :wave: 2022-06-15 00:22:44 !nick abc 2022-06-15 06:47:59 blt[m], lxd is in community and the maintainer is rnalrd - I maintain lxd-feature (which is the cutting-edge of lxd branch), which is in testing 2022-06-15 07:48:05 fcolista: thanks 2022-06-15 07:49:14 rnalrd: do you know if something is blocking lxd to be on 5.2 (LTS), maybe some dependancies ? 2022-06-15 07:51:19 blt[m], where do you see that LXD 5.2 is LTS? 2022-06-15 07:52:22 https://discuss.linuxcontainers.org/t/lxd-5-0-lts-has-been-released/13723 2022-06-15 07:52:31 It seems to me that 5.0 is LTS, not 5.2 2022-06-15 07:53:09 I've just bumped lxd-feature to 5.2 btw 2022-06-15 07:53:18 alpine/aports:master | Francesco Colista | testing/lxd-feature: upgrade to 5.2 | http://dup.pw/alpine/aports/60167de60511 2022-06-15 07:56:36 blt[m]: 5.2 is not LTS, is "feature" branch 2022-06-15 07:56:36 fcolista: yes 5.0 branch is new LTS and .2 is fixing a lo 2022-06-15 07:56:41 s/lo/lot of bugs/ 2022-06-15 07:56:41 blt[m] meant to say: fcolista: yes 5.0 branch is new LTS and .2 is fixing a lot of bugs 2022-06-15 07:56:58 5.0 is LTS and I'm working on it 2022-06-15 07:57:17 ha very nice! 2022-06-15 08:00:20 when I spoke to the main lxc/lxd dev he advised to be on 5.2 because of a lot of bug fixed 2022-06-15 08:01:14 then use lxd-feature package, not lxd. 2022-06-15 08:02:53 I expect though, that being LTS, any serious bug would be fixed in the 5.0.x releases 2022-06-15 08:03:10 otherwise what's the point of being "LTS"... 2022-06-15 08:03:49 that's indeed what an LTS supposed to be, perhaps they have some patch release for the bug I havent checked 2022-06-15 08:04:15 not yet https://linuxcontainers.org/lxd/downloads/ 2022-06-15 08:04:33 for 4.0.x branch we have up to 4.0.9 2022-06-15 08:06:48 in fact they haven't released the patch so far, I am reading this on 5.1 :... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/BzTmdvLjyYQtoPADcyzhkXSx) 2022-06-15 08:07:06 so it supposes to have 5.0.1 and 5.0.2 indeed 2022-06-15 08:07:27 "all of which will also be in the upcoming LXD 5.0.1 LTS bugfix 2022-06-15 08:07:28 release" 2022-06-15 08:07:47 so wait for 5.0.1 or use lxd-feature-5.2 2022-06-15 08:08:29 I guess even 5.0.2 as this article was for 5.1 2022-06-15 08:09:49 apart from features (.1 / .2) would you made 5.0 LTS available to apk ? 2022-06-15 08:10:32 yes 2022-06-15 08:12:17 cool many thanks 2022-06-15 08:12:26 yw 2022-06-15 08:12:34 oh hi 2022-06-15 08:12:39 alpine p. good 2022-06-15 08:13:41 https://open.spotify.com/episode/5hUDciwoe7uhV3goszkbSa?si=_m7B0UikQba1rmI51TjdaA 2022-06-15 08:14:28 oh shit 2022-06-15 08:14:33 mike comments on jan 6 2022-06-15 08:14:43 :D 2022-06-15 08:14:59 eh 2022-06-15 08:21:16 eh 2022-06-15 14:36:59 hi, can anybody explain or provide further resources on how the apk verify command works? i have been trying to verify some package signatures with it but it didn't work 2022-06-15 14:38:48 it verifies a .apk against the internal checksums and the keys in /etc/apk/keys for signature 2022-06-15 14:57:23 ok but i have to fetch the package first, right? and is every package in the official repositories signed? 2022-06-15 14:58:26 yes 2022-06-15 15:00:51 iirc apk fetch itself verifies the things you are fetching 2022-06-15 15:03:51 thank you! 2022-06-15 18:38:10 The description of the "extended" installer image says "Most common used packages included." Where can I find a list of those packages? 2022-06-15 18:39:16 I dug around in the wiki and other documentation, and so far I haven't found anything. 2022-06-15 18:40:16 aports/scripts/mkimg.standard.sh / profile_extended / apks 2022-06-15 18:41:13 https://git.alpinelinux.org/aports/tree/scripts/mkimg.standard.sh#n41 2022-06-15 18:42:34 Thanks. 2022-06-15 20:51:21 hi 2022-06-15 20:51:46 keep running into an issue with firmware (?) updates, where usb mouse starts double clicking on single clicks 2022-06-15 20:51:56 not sure what that's about 2022-06-15 20:52:52 have you tried it on another system? 2022-06-15 20:52:55 seems like every other set of firmware updates gets this bug 2022-06-15 20:54:26 it's the updates 2022-06-15 22:30:19 not sure if it's the kernel updates or the firmware tho 2022-06-15 22:30:25 but this keeps happening 2022-06-16 00:20:38 Hello. 2022-06-16 00:21:00 Is anyone here familiar with the init system or initrd/initramfs? 2022-06-16 00:22:08 I'm trying an experiment and seem to have hit a snag with the kernel not reading input from the keyboard. 2022-06-16 03:44:45 Hello World 2022-06-16 14:49:02 If I made an APKBUILD with the wrong version naming convention and I want to adjust it, is that an Update, Rename, or other, and how should it look if other? 2022-06-16 14:49:18 Note that the tagged version and commit are not changing, so it isn't really an update 2022-06-16 14:53:32 seems like you should just amend 2022-06-16 14:53:56 if i understand correctly, `git commit --amend` with new commit message 2022-06-16 14:54:17 assuming it's not merged already, of course 2022-06-16 14:54:25 ^ 2022-06-16 14:55:03 Habbie: I believe it wouldn't be if the naming convention is wrong? 2022-06-16 14:56:42 i hope so, yes! 2022-06-16 15:03:51 Saijin_Naib: also, dunno if you know, but `git push` with `--force` after you amend 2022-06-16 15:04:35 No, it is merged and in testing for a while now 2022-06-16 15:04:46 ok 2022-06-16 15:04:47 I went with the "Other" part of the COMMITSTYLE.MD and hope it is acceptable 2022-06-16 15:04:54 alternative answer: send the MR and see what feedback you get 2022-06-16 15:14:16 Given I messed it up already... haha. Gotta redo, squash, and force again 2022-06-16 15:32:29 Saijin_Naib: testing/aportname: correct port version 2022-06-16 15:32:46 Ah, whoops 2022-06-16 15:32:47 Not what I have 2022-06-16 21:21:22 hi there, anyone having issues with qemu high cpu usage in Alpine 3.16.0 even when guest is idle? 2022-06-16 21:22:44 i just upgraded a KVM host that was running on 3.15.4 and without changing any VM configuration, qemu process is taking 100% (a full core) while if I check CPU usage in the guest it's totally idle 2022-06-16 21:23:50 this is happening for Debian 11 and Windows 2012 R2 guests, yet I have one Alpine 3.16.0 guest that is not showing this problem, I'm really lost and considering rolling back to 3.15.4 :( 2022-06-16 21:28:38 was doing one last test and found the cause of the issue: iothreads configuration. I will need to investigate more but I just commented out the iothread configuration parameters from the VM XML and the issue is gone 2022-06-16 21:35:21 qemu went from 6.1.1 to 7.0.0, maybe they broke something 2022-06-16 21:35:34 regarding iothreads changing 2022-06-16 21:35:36 shrug 2022-06-16 21:42:35 yes. I will have to look further into it but I can confirm that removing the iothreads settings from the VMs configuration CPU usage is back to normal 2022-06-16 21:43:36 thanks! 2022-06-16 22:27:51 psykose: is irqbalance still around? 2022-06-16 22:28:08 not sure what you mean, where would it be 2022-06-16 22:28:17 ps aux 2022-06-16 22:28:27 what? 2022-06-16 22:28:34 running process 2022-06-16 22:28:46 how would i know if you are running irqbalance on your computer or not? 2022-06-16 22:29:05 well I asked you 2022-06-16 22:29:09 how would I know what you run? 2022-06-16 22:29:21 you asked me if it's still 'around', not what i am running 2022-06-16 22:29:29 i don't run it, no 2022-06-16 22:29:33 it's still around as a package 2022-06-16 22:29:40 or it might be a bit square 2022-06-16 22:29:42 and both debian and ubuntu use it by default 2022-06-16 22:29:49 unless that changed recently 2022-06-16 22:29:54 alpine has never used it as a default 2022-06-16 22:30:01 but it turns out to be a bad default to have it running 2022-06-16 22:30:19 psykose: fc7 was running alpine virtual 2022-06-16 22:30:38 i don't know what an alpine virtual or what fc7 is 2022-06-16 22:30:39 so the question is not what alpine is running but what the host is running since thats where the changes occured 2022-06-16 22:30:52 i just upgraded a KVM host that was running on 3.15.4 and without changing any VM configuration, qemu process is taking 100% (a full core) while if I check CPU usage in the guest it's totally idle 2022-06-16 22:30:58 perhaps time for you to read before answering? 2022-06-16 22:31:13 lmao 2022-06-16 22:31:59 ah, it was related to that 2022-06-16 22:32:17 panekj: roflcopter plays airwolf themesong 2022-06-16 22:32:39 i have no idea what irqbalance has to do with any of that, but it's not enabled by default and never has been 2022-06-16 22:32:48 not sure what in the heaven that means 2022-06-16 22:33:32 and 3.15/3.16 have the same irqbalance version 2022-06-16 22:33:45 again its about the host not the quest 2022-06-16 22:33:58 yeah, what about it 2022-06-16 22:34:10 they upgraded an alpine host from 3.15 to 3.16 2022-06-16 22:34:12 > was running 3.15.4 2022-06-16 22:34:21 maybe it's time for you to read before answering? 2022-06-16 22:34:32 anyway argue with someone that isn't me, i have no idea how any of these interact, and nothing is default 2022-06-16 22:36:35 so please tell me which part of this sentence is it that you dont comprehend? i just upgraded a KVM host that was running on 3.15.4 and without changing any VM configuration, qemu process is taking 100% (a full core) while if I check CPU usage in the guest it's totally idle 2022-06-16 22:37:05 i understand all of it, i have no idea what this has to do with irqbalance 'being around' 2022-06-16 22:41:31 psykose: i would imagine they're thinking that irqbalance would affect this issue somehow, but i have no idea why *your* computer would be relevant in any way though 2022-06-16 22:42:57 maybe, but only the host changed, and even if they were running irqbalance it's the same version between the two, for a vm host the only real big change is qemu to 7.0 2022-06-17 01:14:34 psykose, if I'm thinking more seriously about the TSC decision to kill anything stuck in testing for more than 6mo or whatever, what metric is there for a package in testing that I maintain being "tested" enough to promote to community? 2022-06-17 01:14:51 I'm not sure anyone but myself uses anything I maintain 2022-06-17 01:14:59 no idea :) 2022-06-17 01:17:34 Haha, nice. I assume anyone with stuff in testing will be notified before any action, right? I have a bit of time to learn more about backporting and security fix analysis and such, right? 2022-06-17 08:49:58 After updating my laptop just stopped booting... On edge. Just hangs when GRUB loads the kernel and initramfs 2022-06-17 08:51:07 i suppose it was the .47 kernel then 2022-06-17 08:51:11 not sure what was there before 2022-06-17 09:06:33 Probably. Not sure how to downgrade... 2022-06-17 09:08:08 unless you want to pull in 5.10 from 3.14 or build one yourself.. blech 2022-06-17 09:08:18 Actually maybe the initramfs generation? it got replaced recently 2022-06-17 09:08:32 I just installed linux-edge and it doesn't work either 2022-06-17 09:08:53 it didn't get replaced at all 2022-06-17 09:09:03 you should still have and use mkinitfs 2022-06-17 09:09:22 https://git.alpinelinux.org/aports/commit/main/linux-lts?id=4fac61c826cac81af72e095b645098900d3094a1 2022-06-17 09:09:43 yeah, and that pulls in mkinitfs 2022-06-17 09:10:15 unless it doesn't, in which case it's a Serious Bug, but.. 2022-06-17 09:10:18 and the rest is the same 2022-06-17 09:12:18 well it was gone, so that's a serious bug 2022-06-17 09:12:37 did you have booster installed 2022-06-17 09:12:57 oh derp yes, but that doesn't fully work yet 2022-06-17 09:13:08 but makes sense then 2022-06-17 09:13:26 ..working exactly as intended then :p 2022-06-17 09:15:13 yup, fun lol 2022-06-17 13:58:50 hi all, not sure if it has been reported already, but since yesterday (or so) if I do apk upgrade on my alpine 3.16 instances, I fail to be able to install and run docker due to iptables issues. If I boot with alpine 3.16, but do not apk upgrade, then everthing goes fine (as before) 2022-06-17 14:01:28 is this the right channel to report these kind of issues to alpine? 2022-06-17 14:07:16 fsxpinha: did you run "apk -aU upgrade"? 2022-06-17 14:07:32 no, just "apk upgrade" 2022-06-17 14:07:49 what iptables issues 2022-06-17 14:08:42 the docker won't start saying it couldn't find iptables in the modules or something like that. give some minutes to replicate and get you the exact error 2022-06-17 14:09:11 that sounds like you upgraded, it upgraded the kernel, and then docker cannot load the kernel modules for the currently running kernel, because you didn't reboot after upgrading 2022-06-17 14:09:56 the old kernel modules are gone on upgrade, so you have to reboot (with the changes) after, and it sounds like you instead reboot into some previous state and also undo that 2022-06-17 14:10:31 ok, that makes sense 2022-06-17 14:12:01 similarly, if you load the actual needed modules first, then upgrade, everything will continue to work anyway as you loaded what you needed (i think? unless it reloads them), but rebooting is good practice for new kernels 2022-06-17 14:13:48 yeah, you're right. the procedure is wrong on my end as i just booted the latest AMI, apk update, apk upgrade and continued with whatever I had to do. Thanks for helping me on this 2022-06-17 14:45:03 psykose, PureTryOut: tbf i think it's a bug that this results in a "hang" 2022-06-17 14:45:51 in this specific case i assume the initrd is there, and it was some booster-generated one that didn't boot 2022-06-17 14:46:00 if it was missing at all afaik the kernel outputs something else, but i forget 2022-06-17 14:46:00 all sorts of initramfs issues on current alpine manifest as "hang" or "black screen" when they could full well print a useful error 2022-06-17 14:46:07 ah, that specifically, yeah 2022-06-17 14:46:31 i think this is related to simpledrm, but not sure exactly how, and i haven't had time to rummage it 2022-06-17 14:47:27 "CONFIG_DRM_SIMPLEDRM=m" that's probably not a good configuration. i vaguely recall there was a reason why this was selected, but i think it's worth revisiting and seeing what other distros do 2022-06-17 14:48:35 because many users have reported drm/kms related issues, and if we can't figure out the root cause then we should at least try to harmonize our configuration with other distros to reduce the chances 2022-06-17 14:50:20 fedora makes _DRM=y, which i assume forces =y for _DRM_SIMPLEDRM too 2022-06-17 14:50:23 not sure why we changed it 2022-06-17 14:50:39 it was half the reasoning in the first place, to just copy that specific page https://fedoraproject.org/wiki/Changes/ReplaceFbdevDrivers 2022-06-17 14:52:02 it's incredibly annoying to actually test this shit though because for most people it just works, and the very few people that end up with a black screen can't usually meaningfully 'just go rebuild the kernel with =y instead to see if something useful happens' for a confirmation 2022-06-17 14:52:08 mmhmm 2022-06-17 14:52:20 we should make it =y anyway, but this doesn't give conclusive easy answers 2022-06-17 14:53:07 things like this should really have been on some -edge variant first instead of this ''''edge'''' we have now which is just lts with half the options turned off 2022-06-17 15:07:41 I'm running alpine on a RPi0 and i'm wondering if otg serial is supported? and if so, could i get some pointers about how te get started with it? I tried using documentation about otg for debian and other distros, but i am failing to apply them to alpine. 2022-06-17 15:08:04 I was not able to find much information about it that applies to alpine. 2022-06-17 15:10:48 otg serial as in using the rpi0 as a serial adapter? 2022-06-17 15:11:55 Yes, a serial adapter that serves a tty i can screen into from my host 2022-06-17 15:13:35 I've got it running on raspbian, but loading the same modules with alpine doesnt do much. as far as my host is concerned it just times out. 2022-06-17 15:13:52 sounds like you need `dtoverlay=dwc2` and `modules=dwc2,g_serial` (though maybe modules-load is different) 2022-06-17 15:13:58 aside from that none of these mention much else 2022-06-17 15:14:22 well, you'll need to manually start a getty on the serial port 2022-06-17 15:14:38 psykose: my proposal was to split -slim out from -edge 2022-06-17 15:15:13 how do i do that with a diskless alpine os? All i know is systemd lol 2022-06-17 15:15:43 /etc/inittab 2022-06-17 15:15:47 i would assume you would add an inittab entry for ttyGS0 2022-06-17 15:16:05 there's a comment for a 'serial port', you can use that but name it.. GS0 instead of S0 2022-06-17 15:16:35 and then you need to persist the inittab with however that is done on diskless normally that i forget 2022-06-17 15:17:33 lbu 2022-06-17 15:17:37 lbu ci 2022-06-17 15:17:55 :) 2022-06-17 15:18:27 ah right, i need access to it first tho. The only way i can interact with it at the moment is by using overlay files 2022-06-17 15:19:32 How can i view /etc/inittab without booting the diskless images? 2022-06-17 15:20:49 you could unpack the .iso or copy it from a container or something if you never changed it, it's part of alpine-baselayout-data 2022-06-17 15:21:08 https://git.alpinelinux.org/aports/tree/main/alpine-baselayout/inittab 2022-06-17 15:21:09 or here 2022-06-17 15:22:37 Oh awesome, thanks. i could append that line using the overlay image 2022-06-17 15:23:38 sllide: if "console=ttyAMA0" is passed to initramfs' init it will start a getty on the device 2022-06-17 15:25:09 That sounds even better, do you know if i should use 'modules=dwc2,g_serial' or 'modules-load=dwc2,g_serial' per chance? 2022-06-17 15:29:08 sllide: for things passed by bootloader to init it would be "modules=" followed by a comma-separated list 2022-06-17 15:32:15 anyone here has any pointers to getting an alpine linux installation with FDE and secure boot using TPM2.0? Any pointers to this setup would be greatly appreciated 2022-06-17 15:35:09 I think the arch wiki has some useful information regarding that 2022-06-17 15:35:33 https://wiki.archlinux.org/title/Trusted_Platform_Module 2022-06-17 15:38:57 I don't think it will help a lot since arch uses systemd. But gentoo's wiki has a page on it (I haven't read it but I'm sure it'll be useful for you) 2022-06-17 15:39:07 https://wiki.gentoo.org/wiki/Trusted_Platform_Module 2022-06-17 15:40:11 good luck 2022-06-17 15:40:35 thank you I'll have a look at the gentoo one. Arch didn't really help as you say they have a systemd service for that 2022-06-17 15:42:10 fsxpinha: using UEFI with "hardware" drive encryption? 2022-06-17 15:43:47 you can try with https://github.com/jirutka/efi-mkuki 2022-06-17 15:44:33 fsxpinha: so you want AFDE? 2022-06-17 15:45:48 minimal: yes, that's the idea. uefi with "hardware" drive encryption. running on bear metal 2022-06-17 15:46:30 dmesg still gives me the message that a new usb device is found. followed by a timeout. this is my cmdline.txt: `modules=loop,squashfs,sd-mod,usb-storage,dwc2,g_serial noquiet console=tty1 console=ttyAMA0` And i've added `dtoverlay=dwc2` to my usercfg.txt file 2022-06-17 15:47:05 panekj: thanks a lot! this is super interesting 2022-06-17 15:47:56 minimal: not sure what A from FDE stands for? 2022-06-17 15:48:13 fsxpinha: ah, you didn't originally specify the hardware (SED) aspect so I assumed you were referring to GRUB FDE which, for UEFI, is actually AFDE (*almost* FDE) as the ESP partition is unencrypted ;-) 2022-06-17 15:48:32 sllide: i don't think there's such a thing as "noquiet" 2022-06-17 15:48:41 but it should be ignored 2022-06-17 15:49:26 oh damn, i suppose it does the same as ommiting it then. I've red somewhere that that prints extra debug info. 2022-06-17 15:53:30 I'm trying to build mpv from source. I added clang using `apk` but then when I followed the build instructions on their github page, I got the following error. http://dpaste.com/C8A9KW4NB 2022-06-17 15:55:08 and using `mason` (their other method of building from source also profided in the README on github) I got this 2022-06-17 15:55:09 cc is provided by gcc only, i assume CC=clang would also work 2022-06-17 15:55:10 http://dpaste.com/9JSAGQNWY 2022-06-17 15:55:23 for the latter you need to post that actual build failure log 2022-06-17 15:55:34 sllide: i don't think modules= will work. iirc that applies to initramfs modules 2022-06-17 15:55:38 one second (I was going to do so) 2022-06-17 15:55:45 but i think g_serial is not in initramfs 2022-06-17 15:55:53 so either you have to add it to initramfs, or load it later 2022-06-17 15:56:06 i think last time i did this i put it in /etc/modules in lbu 2022-06-17 15:56:12 psykose: http://dpaste.com/9XDXT6W64 2022-06-17 15:56:50 which was confusing to me beacause ld should come with clang (afaik) 2022-06-17 15:56:58 yeah, just adding CC= works fine and gets to the next step 2022-06-17 15:57:06 Oh okay, thats good to know. i can load them both via a overlay file in that case 2022-06-17 15:57:17 as for ld, you need.. ld, which is in binutils 2022-06-17 15:57:21 just add build-base and forget about it 2022-06-17 15:57:58 psykose: okay .. it worked , thank you 2022-06-17 15:58:38 as for the rest of the deps, you can look at https://git.alpinelinux.org/aports/tree/community/mpv/APKBUILD 2022-06-17 15:58:51 we already have it packaged, i assume you just want to build the latest commit or some patches :) 2022-06-17 15:59:09 I would recommend modifying apkbuild to use commit/patches 2022-06-17 15:59:29 and then you can just `abuild rootbld` 2022-06-17 16:00:50 So if i go for /etc/modules, is using the kernel parameter `console=ttyAMA0` still applicable? 2022-06-17 16:01:26 I know about ttyUSB0 too. But I have no idea if that is a thing in alpine 2022-06-17 16:07:02 can't remember 2022-06-17 16:08:12 according to https://www.kernel.org/doc/Documentation/usb/gadget_serial.txt it should be ttyGS0 2022-06-17 16:08:44 but i think console= won't work with modules 2022-06-17 16:09:19 okay, i just wrote a script that adds dwc2 and g_serial to /etc/modules with a local.d script 2022-06-17 16:09:44 if this works and it doesnt timeout i'm one step futher atleast 2022-06-17 16:11:18 i think local.d is too late 2022-06-17 16:11:53 ah, i suppose modprobe would work in that stage? 2022-06-17 16:12:15 probably 2022-06-17 16:12:28 man, openrc is something else 2022-06-17 16:12:43 fun tho. bummer about the debug capabilities i have access to atm 2022-06-17 16:13:50 local.d is the very last service, so inittab/kernelstuff/whatever has all already ran (for whatever guarantee openrc gives you that they actual ran and exited, not really reached a state, but generally this means it's too late) 2022-06-17 16:14:12 (local.d declares 'after *') 2022-06-17 16:14:16 yep 2022-06-17 16:14:53 good to know too 2022-06-17 16:15:01 it's useful for random shit to run once at boot away from any other openrc management and that's about it, if you know more about where to put stuff you can create a whole service to shove into `boot` instead or whatever 2022-06-17 16:15:19 using local.d to modprobe dwc2 and g_serial isnt helping anyway 2022-06-17 16:15:59 panekj: never done that, can't seem to find a wiki page that describes the process. is there a guide online that you recommend ? 2022-06-17 16:16:56 i'm going to try to load a different otg module. Might be missing something serial related. Would be good to know 2022-06-17 16:17:48 I copied the APKBUILD to a file. `apk add`ed what I need (abuild and abuild-rootbld). and I was faced with a message that told me that I need to `abuild-keygen` which I did. Now I'm stuck with... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/WeFCbTosRmwbxOjbRwtcPaUU) 2022-06-17 16:18:03 sllide: sorry for interrupting you 2022-06-17 16:18:16 no problem bro 2022-06-17 16:19:09 ehh g_ether isnt responding to anything either. i'm going to the pub lol. thanks for the help. I might return 2022-06-17 16:19:21 cheers 2022-06-17 16:21:46 just use regular `abuild -r` instead of rootbld` 2022-06-17 16:22:05 or put your /etc/apk/repositories into `/home/jbara/mpv/jbara/.rootbld-repositories`, i guess 2022-06-17 16:22:09 not sure if there were more gotchas 2022-06-17 16:22:28 but there's nothing really wrong with manually building mpv either like you already were 2022-06-17 16:22:58 personally i use the aport as a place to run `abuild deps` in (so it's easy to have them in one virtual to clean up after, and so i don't miss any) and then i just manually run build commands somewhere 2022-06-17 16:23:16 the abuild integration is usually not that useful if you're not actually using it to feed an apk repository 2022-06-17 16:23:30 or at least doing some level of automation beyond 'i want to build this thing to use' 2022-06-17 16:23:38 You can do that also yourself with apk add -t foo dep1 dep2 2022-06-17 16:23:55 yeah, though then in this case you would be copying 35 lines :) 2022-06-17 16:23:59 :) 2022-06-17 16:47:11 psykose: yeah pretty sure it's the booster one, booting basically went the same as I got so far with trying booster 2022-06-17 17:53:50 Hi! I'm getting a compilation error when using "aio" with C on Alpine and I can't figure out where it comes from or how to fix it: "field 'aio_sigevent' has incomplete type". Not sure if I can post links, but can provide a CI log if postiing links is allowed! 2022-06-17 18:12:47 witcher01: no worries, post anything useful 2022-06-17 18:28:06 sounds like you forgot an include for 2022-06-17 19:25:21 good afternoon 2022-06-17 19:25:33 o/ 2022-06-17 19:26:14 Is it possible to run torbrowser on alpine 3.16? 2022-06-17 19:29:35 as easily as just running it? no, the shipped binaries are linked against glibc, and it doesn't build against musl 2022-06-17 19:29:40 with flatpak? sure, it's a few minutes 2022-06-17 19:30:05 think it was com.github.micahflee.torbrowser-launcher 2022-06-17 19:40:19 Note: graphical installation of Flatpak apps may not be possible with Alpine. 2022-06-17 19:40:56 https://flatpak.org/setup/Alpine 2022-06-17 19:42:05 Nao installs graphics software on alpine 2022-06-17 19:43:22 i have no idea what that means but i can run steam fine and i'm sure it doesn't get more graphical 2022-06-17 19:47:31 Are you using flatpack in alpine 3.6 and using steam? 2022-06-17 19:48:07 edge, but currently it would be identical 2022-06-17 19:50:23 Tomorrow I will update and test. Thanks. 2022-06-17 19:59:09 I have a problem installing samba with lxc in alpine 3.16, when I will provision this message appears. https://pastebin.com/CwPsVqD3 And does not run. It says it doesn't access the secrets.tdb file and stops. 2022-06-17 19:59:59 I sent a message to the samba list yesterday, but so far no one has responded. 2022-06-17 20:03:59 where does it say that? all that logs says is the underflow 2022-06-17 20:07:04 what it means? underflow! 2022-06-17 20:09:50 https://pastebin.com/cP6U8y0F 2022-06-17 20:15:03 that accesses the secrets.tdb fine (from what i'm reading) and then fails to find a specific entry 2022-06-17 20:15:13 unless they like writing confusing things in the log, wouldn't surprise me 2022-06-17 21:12:43 seems like s390x runner wants to take a long nap tonight 2022-06-17 21:16:06 He is very tired apparently 2022-06-17 21:20:35 Late reply, but this is the CI log for my problem with C compilation with "aio" on Alpine: https://builds.sr.ht/~witcher/job/782440#task-build. Source can be found here: https://git.sr.ht/~witcher/echo-algo/ 2022-06-17 21:29:44 https://img.ayaya.dev/7Ui5CEKgMzph 2022-06-17 21:32:14 source: on musl basically everything in signal.h which defines that is under https://img.ayaya.dev/o1IeICnxbe12.png 2022-06-17 22:37:08 where can I find the source for alpinelinux/alpine-gitlab-ci, the container image used for the CI jobs on aports? 2022-06-17 22:37:28 https://gitlab.alpinelinux.org/alpine/infra/docker/alpine-gitlab-ci 2022-06-17 22:38:08 thanks 2022-06-17 22:39:04 I'm making a custom package repo, in which there are dependencies between some of the packages, so when rebuilding everything I need to take this order into account 2022-06-17 22:39:16 when it comes time to rebuild all of alpine, how is this order determined? 2022-06-17 22:39:26 buildrepo from lua-aports builds the repo 2022-06-17 22:40:31 which iirc is just some ap buildorder command from the same thing 2022-06-19 09:34:59 It's been 2 days so I'll ask my question again since I haven't found a fix yet: I'm getting a compilation error when using "aio" with C on Alpine and I can't figure out where it comes from or how to fix it: "field 'aio_sigevent' has incomplete type". CI: https://builds.sr.ht/~witcher/job/783206#task-build src: https://git.sr.ht/~witcher/echo-algo/ 2022-06-19 09:35:57 As far as I can tell aio.h is included since that is where the error originates from. It builds just fine on Arch Linux 2022-06-19 09:36:15 i answered and even sent a patch to fix it 2022-06-19 09:36:40 Oh, sorry, where to? I didn't notice 2022-06-19 09:36:55 https://img.ayaya.dev/7Ui5CEKgMzph 2022-06-19 09:37:53 I see it now, completely missed it. I'm so sorry, thank you! 2022-06-19 13:43:51 i think this is *technically* a bug in musl 2022-06-19 13:48:42 indeed, everything in aio.h is essentially unusable without.. any of the _SOURCE defines 2022-06-19 13:49:04 technically correct or not i would rather just have an #error or some shit 2022-06-19 13:50:55 ah, no, i see what you mean; merely including a header should not error so unclearly without either not exposing the definition at all or throwing an explicit error 2022-06-19 14:28:54 something like that 2022-06-19 14:40:52 hi, short question - after the last mesa update, chrome://gpu log messages are full of "Skia shader compilation error Errors: link failed but did not provide an info log" -- can someone confirm this behavior? thx 2022-06-19 16:23:12 Hi there. I have a question about the re2c package: https://pkgs.alpinelinux.org/package/edge/main/x86_64/re2c It looks like the v3.0 package is missing the --case-ranges option. Is this intentional? The binary compiled from scratch has it. 2022-06-19 16:34:36 Also it won't accept the re2c:eof option. 2022-06-19 17:01:14 gm: it's not explicitly disabled or something like that 2022-06-19 17:01:24 Not sure why it would be missing 2022-06-19 17:02:06 gm: is that a rust specific thing? 2022-06-19 17:02:20 No, it's C specific. 2022-06-19 17:02:24 ok 2022-06-19 17:03:20 I haven't tried much harder, but the lack of re2c:eof suggests other things may be missing. eof is not C-specific. 2022-06-19 17:04:14 checking config.log if something is in there 2022-06-19 17:11:16 cannot find any reason why it would not be present myself 2022-06-19 17:57:23 Why does `/etc/mavenrc` have `MAVEN_OPTS="$MAVEN_OPTS -Xmx512m"` set? 2022-06-19 18:00:15 Hi, I'm running into an issue with the alpine-xen installation media. I'm booting with uefi, I get the grub menu, I get the Xen kernel output. I'm expecting the output for openrc but all there is a blank screen. Is anyone able to help me? 2022-06-19 18:26:31 Ah it is definitely an issue with efi. Boots fine with crm enabled 2022-06-19 18:28:41 PureTryOut: https://gitlab.alpinelinux.org/alpine/aports/-/issues/8762 2022-06-19 18:39:17 Thanks. So no need for it anymore then, it was definitely causing me problems just now... 2022-06-20 05:43:28 gm: it has both options though? not sure what error you are getting 2022-06-20 13:29:34 jello. quick question regarding alpine. I'm currenlty working on a not-yet-mainline armv7 soc, so I do frequentish kernel patches. Is there a mechanism in alpine I can use to rebuild local packages with patches? 2022-06-20 13:30:07 I would use gentoo but its a 'weak' soc, so a primarily binary distro would be best suited to my purposes. 2022-06-20 13:35:14 hanetzer: you can clone the aports repo, adjust the APK BUILDS 2022-06-20 13:35:31 you want to patch and build them with abuild 2022-06-20 13:35:44 ahhh. ok, that should be fine. 2022-06-20 13:36:07 is there a cross-build mechanism? 2022-06-20 13:37:04 Not really, we do very little cross-compiling 2022-06-20 13:37:35 So we don't have a general purpose cross-compile toolchain 2022-06-20 13:38:30 fair. I'll look into it more :) 2022-06-20 13:38:42 hanetzer: my "cross-build" mechanism is based on qemu 2022-06-20 13:39:57 I saw it on alpine's wiki 2022-06-20 13:40:11 works like a charm 2022-06-20 13:40:27 Yes, that's a possibility if there is qemu support 2022-06-20 13:42:01 heh. yeah I could spin up a qemu vm. I wonder, if qemu building is faster than native building, given I have a fairly decent host machine. 2022-06-20 13:42:29 you don't actually need a complete vm, just userspace emulation 2022-06-20 13:42:31 There is qemu-user 2022-06-20 13:42:41 true that. 2022-06-20 13:42:52 but even then, I wonder the metrics. 2022-06-20 13:43:13 We build riscv64 that way 2022-06-20 13:43:26 that being said. how are kernel packages setup in alpine? one big package? a number of sub packages like linux-{image,modules,headers} ? 2022-06-20 13:44:06 Modules and image are combined 2022-06-20 13:44:30 adventage of qemu-user over qemu-system: you less likely end up in OOM 2022-06-20 13:44:45 and if memory serves, apkbuilds are somewhere between portage and pacman? 2022-06-20 13:45:06 Very similar to pacman 2022-06-20 13:46:38 I notice only linux-lts exists under main; is there *somewhere* I can get something closer to mainline? 2022-06-20 13:47:33 linux-edge in edge? 2022-06-20 13:47:42 ^ 2022-06-20 13:48:45 danke. that's under community, according to find? 2022-06-20 13:49:18 Yes 2022-06-20 13:56:09 5.18.5, not bad. 2022-06-20 13:56:27 pretty sure I can work with that, since my patches are *mostly* whole new files :) 2022-06-20 13:59:41 pretty sure a cross-toolchain is faster than either qemu-* 2022-06-20 14:00:08 yeh. 2022-06-20 14:00:19 but if there's no tooling for it :) 2022-06-20 14:00:30 if you know how to build one that works, though admittedly you'll probably have to adjust the apkbuild to use it :) 2022-06-20 14:00:32 plus, my host is gentoo, and lacking an ebuild for abuild 2022-06-20 14:27:41 docker-abuild is an option if you don't mind docker 2022-06-20 14:28:30 proper mechanism of running alpine netboot on arm? its failing to mount boot media (which should be initrd?) 2022-06-20 14:53:12 dunno, iirc the netboot stuff (ipxe?) was either outdated or missing stuff for specific systems or whatever 2022-06-20 14:53:37 if you're just using a vm then.. pick anything else and have an easier time :p 2022-06-20 14:54:28 well, installing on real hw atm :) 2022-06-20 14:54:57 untarred the minirootfs, seems like it should work for now. that said. is there a listing of the default world set I can use to 'convert' it to full? 2022-06-20 14:57:26 afaik it's +openrc +linux-lts (a kernel), and.. i think that's it? 2022-06-20 14:57:34 but you need to set up the openrc things and i forgot what you need 2022-06-20 14:59:02 you can probably run `setup-alpine` and go through it and it will put it into the correct state 2022-06-20 14:59:26 making the install bootable is another matter, past just unpacking it onto a disk 2022-06-20 15:00:21 currently the u-boot on the device (2010.06, yeah I know) doesn't support loading images from sata so I'll prolly still be tftp'ing lulz 2022-06-20 15:00:29 hehe 2022-06-20 15:00:40 if it works then it works 2022-06-20 16:49:03 could someone provide a recursive listing of /etc/runlevels ? 2022-06-20 16:55:41 hanetzer: I can sure play a mean pinball 2022-06-20 17:00:07 hanetzer: any of them? mine is https://img.ayaya.dev/CHcxHk43GAwc 2022-06-20 17:17:32 psykose: danke. :) 2022-06-21 04:10:29 oh, that 'async' runlevel is a good idea. 2022-06-21 11:56:01 I am trying to make my Fibocom L860-GL M.2 WWAN 4G adapter to work with Alpine without success. It seems that this adapter is based on Intel XMM7560 chipset and I read that some [patches](https://www.phoronix.com/scan.php?page=news_item&px=Intel-M.2-Modem-Linux-5.14) had been applied 1y ago to the linux kernel 5.4 to support that but the... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/HeqArIimOBeuIpiFIayntBnP) 2022-06-21 12:43:29 is there a way to make abuild fetch distfiles through HTTP server with authentication? (for a private repo) or my only bet is to fetch manually in /var/cache/distfiles 2022-06-21 12:43:35 because abuild checksum is quite handy though 2022-06-21 12:44:05 http://user:password@hostname/ ? 2022-06-21 12:44:49 ah yes if that's as simple it's great :) 2022-06-21 12:44:53 let me check 2022-06-21 12:45:05 I assume so 2022-06-21 12:45:56 I'm not even sure if I sent it correctly because irccloud formatting 2022-06-21 14:52:52 markand: did you check :)? 2022-06-21 14:53:20 not yet, I need to wait that my teammate setup the HTTP auth first ;p 2022-06-21 15:16:07 Trying to use borg mount, and it complains about the fuse kernel module not being loaded. What's the 'alpine' way to have it automatically load? 2022-06-21 15:16:27 I've loaded it manually and it works, so just need to get it to load automatically. 2022-06-21 15:16:42 add it to /etc/modules 2022-06-21 15:17:10 Thought that might be it, just wanted to double check there wasn't a different way to do it that was more 'alpiney' :) 2022-06-21 15:17:54 eventually /etc/modules-load.d iirc 2022-06-21 15:18:34 That directory does exist on my install, but is empty. Any docs regarding it? That'd be cleaner. 2022-06-21 15:19:23 modules-load.d(5) 2022-06-21 15:19:39 tl;dr: *.conf files with same syntax as /etc/modules 2022-06-21 15:21:02 Perfect, thanks panekj, much appreciated. 2022-06-22 01:55:13 Anyone have familiarity with gdk-pixbuf providers and thoughts on adding a few more to Alpine's gdk-pixbuf packge (or as new subpackages)? 2022-06-22 01:56:07 Reason being, I'd appreciate more image format support in programs that use gdk-pixbuf providers like Ristretto, Thunar, tumbler, etc etc etc. AVIF, HEIC, RAW formats, JPEGXL, etc 2022-06-22 01:56:34 I am not knowledgeable in this problem space, so looking for advice/feasibility assessment or whatever 2022-06-22 08:47:14 Hello! ;) Is there on alpine some tools helping me to define the last reboot reason? Last reboot occured 73 days ago and my /var/log/message doesn't log this day anymore... 2022-06-22 09:15:43 not if you've deleted the logs 2022-06-22 09:40:09 I didn't manually deleted logs, but there is a rotate log on /var/log/messages and /var/log/messages.0 only, so I can't look far in the past... Is there a way to configure this rotate log for the next time I have an unexpected reboot? 2022-06-22 09:40:59 /etc/logrotate* 2022-06-22 09:44:31 I have acpid chrony openrc in my /etc/logrotate.d/, but none of there refer to messages log. Is that normal? 2022-06-22 09:46:23 you may have it in /etc/logrotate.conf 2022-06-22 09:48:28 nope, I haven't this file... 2022-06-22 09:50:08 is logrotate installed? 2022-06-22 09:55:18 no indeed, sorry my mistake, I didn't explain myself very well. In fact mile file /var/log/messages is renamed in /var/log/messages.0 when it reach 200ko. But I don't know who is in charge of that (log rotate or other). 2022-06-22 09:55:37 my file* 2022-06-22 09:57:47 i think that's syslog 2022-06-22 10:05:35 yes, you're right, I will check that! thanks! 2022-06-22 10:07:00 looks like adding "-s 0" to /etc/conf.d/syslog should disabled that 2022-06-22 10:20:47 yes that's it, thanks lopid. 2022-06-22 11:37:05 xfce puzzle: I have an eee-pc, alpine 3.16, lightdm+xfce4. Mostly works, however, on resume from RAM the xfce display is frozen. 2022-06-22 11:38:42 Weirder even, after booting into cli, pm-suspend, and resume, lightdm starts and works, but after login gives a blank screen with (working) mouse cursor. 2022-06-22 11:38:50 Where to start looking? 2022-06-22 12:29:35 mercenar-: Do you have all the right firmware loaded? Same for xf86-video-* package? 2022-06-22 12:30:51 sounds like you need to make pm-suspend switch to a cli before doing suspend 2022-06-22 12:30:58 all kinds of hardware have weirdnesses 2022-06-22 12:32:22 I don't remember which one of the tools pm-suspend was (there's swsusp, s2ram, ...), but it may have a config in /etc for this 2022-06-22 12:32:30 does anyone remember TuxOnIce? :> 2022-06-22 12:37:05 Saijin_Naib: yeah. it seems like an xfce thing, running something like xinit xterm survives suspend, so X itself handles it 2022-06-22 12:38:17 satmd: even if the machine hasn't had X running at all, boot into command line, suspend/resume, and then service lightdm start, xfce fails 2022-06-22 12:38:22 Huh, that's super odd then 2022-06-22 12:38:42 maybe... 2022-06-22 12:38:44 What do your runlevels look like? 2022-06-22 12:39:06 that could be a compositor's doing things with the GPU/driver 2022-06-22 12:39:36 it's worth trying with compositing disableds 2022-06-22 12:40:16 hmm. does default now have compositors? Will have a look 2022-06-22 12:41:34 I have no idea, just years-long experiences with s2ram since even a IBM T32 laptop (ans I also wrote a keyboard driver for MS-DOS to put a PC into APM S3) 2022-06-22 12:42:12 aka "I'm old" 2022-06-22 12:42:32 o/ 2022-06-22 12:42:41 do you know if there is a special package to install in GNOME to enable screen sharing? 2022-06-22 12:43:00 gnome-control-center complain about rdp schema not found and in sharing tab there is nothing I can tune 2022-06-22 12:56:52 weird. looks like after suspend xfce points to a non-existent display. running DISPLAY=:0 xeyes pretends to run, just nothing appears. 2022-06-22 13:46:09 "hmm. does default now have..." <- XFWM4 is a compositor, not sure if it's hardware composting is enabled by default or not though 2022-06-22 13:51:00 don't want my hardware composted 2022-06-22 13:53:03 lol 2022-06-22 13:55:38 That is an oddly appropriate typo for compositor. 2022-06-22 14:16:01 satmd: you were right. Options "Composite" "Disabled" in xorg.conf fixes, thank you 2022-06-22 15:39:31 mercenar-: heh. I had a fun situation in the past. I found that when I'd kexec into a new kernel, a ghost cursor would be 'stuck' where it was on the 'previous' boot 2022-06-22 15:41:47 haha. isn't accelerated hardware fun? State spread out over multiple places that can go out of sync 2022-06-22 20:22:16 has anyone encountered an issue with apk being unable to select packages on an add/del/upgrade command? I'm seeing ERROR: unable to select packages: Huh? Error reporter did not find the broken constraints. On one of my systems and am not entirely positive how to approach the issue 2022-06-22 20:23:31 Probably this https://gitlab.alpinelinux.org/alpine/apk-tools/-/issues/10847 2022-06-22 20:24:58 yep that looks like the exact problem, glad it's a known issue 2022-06-22 20:27:07 ncopa had same issue but I do not remember how it was fixed exactly, I think it was trial by error by removing packages from world one by one 2022-06-22 20:30:20 we downgraded apk back to previous version, so it would be some other bug 2022-06-22 20:30:28 or rather the same bug but for another reason 2022-06-22 20:38:46 that's what I'm trying now, just removing the packages from the world file and hoping it'll pick up 2022-06-22 20:47:35 durrendal: thats what I did too 2022-06-22 20:48:05 the nice thing is that it is super easy to make backup: cp /etc/apk/world /etc/apk/world.backup 2022-06-22 20:53:36 oh absolutely, even though something's broken this is much easier than mucking around with something like dpkg 2022-06-22 21:01:55 mercenar-: you're welcome :) 2022-06-22 23:23:18 After installing chroot with --no-scripts, how do I make sure the scripts get run? 2022-06-22 23:23:26 Namely I'm missing busybox symlinks 2022-06-22 23:27:42 `apk fix busybox` would rerun them probably 2022-06-23 08:56:20 to just create the busybox links (in case apk fix fails due to missing symlinks): busybox --install -s 2022-06-23 09:14:08 Hello 2022-06-23 09:14:10 hi 2022-06-23 09:14:45 Is alpine 3.15 considered an lts release. It mentioned lts kernel but not exactly if the release itself is lts 2022-06-23 09:15:07 Forza: 2 years 2022-06-23 09:15:08 all releases are supported for 2 years 2022-06-23 09:15:12 for main/ 2022-06-23 09:15:20 https://www.alpinelinux.org/releases/ 2022-06-23 09:15:21 6 months for community/ 2022-06-23 09:17:44 my ISP finally gives me ipv6 and I'd like to get my alpine router to use ipv6 and provide ipv6 to my home network. I use awall for firewall, and dnsmasq for dhcp/DNS. First step seems to be to get a routable ipv6 address from ISP on the WAN interface. But apparently I dont get any ip address 2022-06-23 09:18:11 the idea here is to get ipv6 to my home net so I can test ipv6 only setups in a VM 2022-06-23 09:19:01 eth0 is the WAN interface. it only gets an ipv6 scope link addr 2022-06-23 09:19:52 I am not sure that ISP give ipv6 addrs via SLAAC or dhcp6. Most likely it is the latter, but I'd like to learn how to test the former to be 100% sure that they dont have SLAAC enabled 2022-06-23 09:20:42 First step was to make sure ipv6-icmp is enabled. I did: ip6tables -A INPUT -i eth0 -p ipv6-icmp -j ACCEPT 2022-06-23 09:22:06 how can i (re)trigger a request for ipv6 and monitor the status? 2022-06-23 09:22:15 via SLAAC 2022-06-23 09:23:35 I do: echo 2>/proc/sys/net/ipv6/conf/eth0/accept_ra 2022-06-23 09:25:06 and then? 2022-06-23 09:25:12 i would guess the dhcp client that connects does it, so whatever probes the isp router for dhcp, so udhcpc or something, no? 2022-06-23 09:25:17 i don't actually know though 2022-06-23 09:25:48 so, I'd like to try SLAAC first, which is stateless, without dhcp as I understand 2022-06-23 09:25:54 dhcpv6 is stateful 2022-06-23 09:25:59 ncopa, panekj thanks. I see. I need to plan for this. Is upgrading supported, for example from 3.15 to 3.17? 2022-06-23 09:26:05 yes 2022-06-23 09:26:23 ah, right, slaac is not dhcp 2022-06-23 09:26:26 ACTION is asleep 2022-06-23 09:26:30 you replace version in /etc/apk/repositories then apk upgrade -Ua 2022-06-23 09:27:02 well, i think we officially only support 3.15 -> 3.16, but upgrades 3.15 -> 3.17 should work 2022-06-23 09:27:22 in my experience the direct upgrade has always worked, more or less 2022-06-23 09:27:39 but yeah, it's silly to aim for infinite jumps, so it's only monitored for one version at a time 2022-06-23 09:27:56 it depends a bit on what application you use. if you for example use postgresql and there is a new major version, you may need to do a dump/restore of the db 2022-06-23 09:27:58 I never had issues with jumping multiple versions but I'll also say it's not impossible for something to break :) 2022-06-23 09:29:25 i did an upgrade from alpine 3.10 to 3.16 today, on my router. i ended up need to find a screwdriver to find the SD card and edit syslinux.cfg by hand, because i forgot one change of initramfs-vanilla -> initramfs-lts 2022-06-23 09:29:29 Thanks. This is a file server so mostly nfs, iscsi and samba stuff 2022-06-23 09:29:39 because we changed name of the kernel at some point 2022-06-23 09:29:51 other than that, everything seems to just work 2022-06-23 09:30:29 Great. I think I'll stay one version back instead. So when 3.17 comes, I'll go to 3.16 instead 2022-06-23 09:30:54 that doesn't really make any sense but i won't tell you what to do with your computer 2022-06-23 09:31:36 doing an upgrade from 3.15 -> 3.17 will most likely work well 2022-06-23 09:32:07 a better idea would be to not upgrade from 3.15.x to 3.16.0 but wait for 3.16.1 or 3.16.2 2022-06-23 09:32:23 if you want "lts" feel 2022-06-23 09:33:36 otherwise why not just use debian at this point 2022-06-23 09:46:19 It's not lts feel, but just not jump on the latest and greatest :) but I don't want lts like in ubuntu where they still use chisel and stone to write ;) 2022-06-23 09:56:47 To be fair, I haven't had any issues with Alpine so far 2022-06-23 12:44:01 How can I increase kernel dmesg buffer? The default is too little so my boot dmesg gets truncated even in /var/log/dmesg 2022-06-23 13:33:41 log_buf_len or somesuch 2022-06-23 15:55:24 hello. is there a way to change colors in tty on alpine? back on arch linux it was doable with escape sequences but it doesn't seem to work on alpine 2022-06-23 15:55:53 what shell are you using? 2022-06-23 15:56:01 zsh 2022-06-23 15:56:39 tput setaf 5; echo cyan 2022-06-23 15:57:17 oh, cyan is 6, 5 is magenta 2022-06-23 15:57:47 iow, "yes" 2022-06-23 15:57:49 lopid: i mean change colors of the linux tty, not output of single command 2022-06-23 15:58:09 that did 2022-06-23 16:05:40 no it didn't, what i need is to set some hexadecimal value for green for example 2022-06-23 16:06:12 please ask the question in full 2022-06-23 16:23:53 oh, escape sequences do work but only as long as xorg is not running 2022-06-23 16:32:48 why would it be related to xorg 2022-06-23 16:35:43 it sounds like you're running a different shell 2022-06-23 16:45:59 colors cannot be changed while xorg is running i swear) i set my colors with https://gist.github.com/txgk/59112b684f9a60bad37b18d6c4ff2655 2022-06-23 16:46:25 what do you mean by "linux tty"? 2022-06-23 16:47:36 thing you see before getting inside your desktop environment/window manager 2022-06-23 16:50:07 you can output colour using that, before starting x. you start x. you ctrl-alt-fn to a console. using that again, shows no colour? 2022-06-23 16:56:28 previous colors remain, but new ones cannot be set while xorg is running. everything is okay now, otherwise why should i change the palette of the console if i run X 2022-06-23 16:56:42 I think they mean is there a way to theme a Linux framebuffer pseudoconsole 2022-06-23 16:58:02 there is a way, script i mentioned above does the trick. you just have to shutdown your x session 2022-06-23 17:02:09 that's unrelated to xorg 2022-06-23 17:02:33 it depends on whether your terminal or terminal emulator supports those escape codes 2022-06-23 17:03:13 maybe they're seeing some redraw quirk of switching back and forth 2022-06-23 17:05:02 Hello71: i see with my own eyes that the colors change only after X is turned off)) 2022-06-23 17:05:34 maybe there is a little misunderstanding here? here's a tool that does what i mean https://bbs.archlinux.org/viewtopic.php?id=168392 2022-06-23 17:06:05 yes, but that's unrelated to xorg. it's like saying that it only works when the text is fullscreen 2022-06-23 17:06:21 or when the font is 8x16 2022-06-23 17:06:39 these are all *correlated* but not causally related 2022-06-23 17:06:58 the underlying cause, as i have explained already, is whether your terminal or terminal emulator supports OSC P 2022-06-23 17:07:06 which i *think* is linux-specific 2022-06-23 17:09:20 the xterm extension is OSC 4 set, which is widely supported according to https://github.com/lemnos/theme.sh 2022-06-23 17:09:38 why would i run terminal emulator? i'm inside linux tty. when i send these escape codes while x session is on - it does nothing, when i shutdown xorg - it changes my colors 2022-06-23 17:09:52 what does it mean "while x session is on" 2022-06-23 17:10:10 if x is occupying the vt then you cannot type into linux terminal 2022-06-23 17:10:36 i3 is running on first tty, i switch to second tty and run script that sends these escape codes 2022-06-23 17:11:13 it does nothing in this case 2022-06-23 17:12:19 but when i run killall i3 and send escape codes again, i get nice looking terminal :> 2022-06-23 17:16:35 ok, i understand now 2022-06-23 17:16:51 it works fine with sway open on my computer 2022-06-23 17:17:06 what is your (kernel) video driver 2022-06-23 17:17:56 5.15.48 Intel(R) HD Graphics 500 2022-06-23 17:18:17 Mesa 21.3.9 2022-06-23 17:18:19 you are not using nomodeset or such nonsense? 2022-06-23 17:18:24 none of those are the kernel video driver 2022-06-23 17:18:35 what does it say in /proc/fb 2022-06-23 17:18:58 i915drmfb 2022-06-23 17:19:34 really i would recommend just running alacritty in cage. fbcon is effectively deprecated at this point 2022-06-23 17:20:07 in cage? what do you mean? 2022-06-23 17:21:08 cage is a Wayland compositor for running single programs ala kiosk style, effectively running a gui program (like alacritty) 2022-06-23 17:21:43 so the suggestion is to stop using fbcon and use x/wayland 2022-06-23 17:21:49 what xf86-video-* have you installed 2022-06-23 17:22:20 and is your linux text vt at native resolution 2022-06-23 17:22:53 I can't imagine how a running xorg session can affect a different fbcon vt though 2022-06-23 17:23:12 i don't know guys it just does :D 2022-06-23 17:23:23 Hello71: i have xf86-video-intel 2022-06-23 17:23:34 might work if you uninstall that 2022-06-23 17:27:00 whoa, now it works, thank you 2022-06-23 17:29:56 handlerug: cage is an interesting thing, but i doubt it fulfills my setup requirements 2022-06-23 20:39:42 o/ - is there nethack-x11 in repo? 2022-06-23 20:41:00 https://pkgs.alpinelinux.org/packages?name=nethack*&branch=v3.16&arch=x86_64 2022-06-23 20:41:32 hello. i'd like to ask. does apk verify checksums of downloaded packages & extracted files? and if yes then is there a file with pkg checksums like packagesite om freebsd? 2022-06-23 20:46:48 yes of course, and i don't know what packagesite is. 2022-06-23 21:20:56 think APKINDEX, but in yaml 2022-06-23 22:55:40 How to convince package developer that they should use original package names and not debian idiosyncrasies? 2022-06-23 22:56:38 Oh, this one should go to -offtopic, sorry! 2022-06-23 23:03:44 what kind of idiosyncrasies 2022-06-23 23:10:23 Like prepending stuff like gir1.2- to some stuff or renaming packages in non-trivial manner 2022-06-24 07:02:14 I've resurected a Pi 3 b+ which I throw an alpine aarch64 (that I usually use with a Pi 4) and looks like the wifi is detected but wpa_cli reports SCANNING all day long 2022-06-24 07:02:18 on pi4 it works fine 2022-06-24 07:49:54 what does `iw wlan0 scan` do, out of curiosity 2022-06-24 07:49:59 or iwd, etc 2022-06-24 07:51:48 could also be some firmware-related thing, if you're not on edge where it's one version ahead, but i forget what the 3b+ needs 2022-06-24 07:53:39 post the dmesg if you can 2022-06-24 08:00:01 oh wait, we have various wireless issues in fact, checking this at the moment 2022-06-24 14:06:33 hello 2022-06-24 14:06:39 does alpine have a non-network install 2022-06-24 14:07:33 https://alpinelinux.org/downloads/ 2022-06-24 14:09:10 Even the extended image would require internet 2022-06-24 14:09:46 maybe ikke can mail you a CD 2022-06-24 14:10:25 lmao 2022-06-24 14:10:49 i just dont want to go through the hassle of doing ethernet internet sharing 2022-06-24 14:10:54 since my room doesnt have ethernet lol 2022-06-24 14:17:05 aryakk: I've had the same issue before, I keep a Mikrotik mAP lite in my work bag just in case I run into the environment you described 2022-06-24 14:17:22 super cheap and decently easier to setup as a wireless ethernet adapter essentially 2022-06-24 14:17:54 s/easier/easy/ 2022-06-24 14:17:54 durrendal meant to say: super cheap and decently easy to setup as a wireless ethernet adapter essentially 2022-06-24 14:17:55 wow the ethernet sharing thing was surprisingly easy to configure with networkmanager lol 2022-06-24 14:32:18 sucks that the laptop hinge is broken so the display is 180 degrees down 2022-06-24 14:32:30 my neck is paining after usingit for a few minutes like that :P 2022-06-24 14:37:10 then stop using your neck 2022-06-24 16:12:10 well why i asked about the apk checksum thing yesterday.. i've got a laptop with nvme drive running freebsd and when installing packages it sometimes gets checksum errors that follow with ZFS I/O failure error 97. i basically need to verify on windows os that it can write files properly. 2022-06-24 16:12:59 i've been thinking about using alpine wsl that does the checksums for this task 2022-06-24 22:58:43 Hi. Is the blktrace utility available for Alpine Linux? 2022-06-24 23:00:42 https://pkgs.alpinelinux.org/contents?file=blktrace 2022-06-24 23:00:51 Hello71: thanks 2022-06-24 23:01:17 I'll have to make a nice kernel with CONFIG_BLK_DEV_IO_TRACE and get the git sources then. Thank you. 2022-06-25 05:58:32 Praise fathers for running away if you praise women for having abortions. 2022-06-25 06:04:03 did you post this in the wrong channel 2022-06-25 08:01:31 Good morning 2022-06-25 08:03:08 A while back I was trying to do a custom kernel. It works sort of, but I never figured out the initrd bit. I am reading https://wiki.alpinelinux.org/wiki/Custom_Kernel 2022-06-25 08:33:58 I suppose I simply make mkinitfs -o initrd.new on my buildhost and trasnfer the result? 2022-06-25 08:59:44 Is there a place where I can submit feature requests to alpine kernel options? 2022-06-25 09:20:47 Forza: https://gitlab.alpinelinux.org/alpine/aports/-/issues 2022-06-25 09:22:31 ikke: thanks 2022-06-25 09:23:03 Hm. It seems that there is still a problem with grub.cfg. on 3.15 and btrfs/raid1 2022-06-25 09:25:45 https://dpaste.com/AE8DRCVHT 2022-06-25 09:26:33 Should look like this https://dpaste.com/2KQBRBT35 2022-06-25 11:46:30 When doing abuild -r to make the linux-lts kernel. Is it possible to skip the linux-lts-dev compilation? 2022-06-25 12:39:50 I notice that libeatmydata depends on dpkg-architecture 2022-06-25 12:40:00 But it isn't installed automatically 2022-06-25 12:41:05 also the script /usr/bin/eatmydata uses "export `dpkg-architecture|grep DEB_BUILD_MULTIARCH`" and then "shlib="/usr/lib/$DEB_BUILD_MULTIARCH/eatmydata.sh"" which is not correct. The eatmydata.sh is in /usr/libexec/eatmydata.sh 2022-06-25 12:41:11 Should I make a bugreport? 2022-06-25 12:54:40 https://gitlab.alpinelinux.org/alpine/aports/-/issues/13957 2022-06-25 19:14:32 OK: 1408 MiB in 323 packages 2022-06-25 19:14:32 and then nothing happens 2022-06-25 19:14:32 i am in need of tech support lol, how do i install wine? when i run apk add wine it says: 2022-06-25 19:18:31 then it's already installed 2022-06-25 19:18:46 you already have wine 2022-06-25 19:18:56 you can check with apk info -e wine if it exits with non-zero code 2022-06-25 19:20:37 panekj: oh i see, thanks 2022-06-25 19:21:52 and now how do i open a file with wine? i never used wine before lol 2022-06-25 19:24:01 wine 2022-06-25 19:27:14 ikke: thank you 2022-06-25 19:28:37 it says /bin/ash: wine: not found 2022-06-25 19:29:41 from what I see, there is no cmd:wine 2022-06-25 19:29:51 wine64 2022-06-25 19:30:00 https://pkgs.alpinelinux.org/contents?file=&path=&name=wine&branch=edge 2022-06-25 19:32:11 it won't work for most things because it's 64-only 2022-06-25 19:32:27 indeed 2022-06-25 19:35:11 "https://pkgs.alpinelinux.org/..." <- do i need to get it from the git repo? 2022-06-25 19:35:50 You need to use wine64 if you are on a 64-bit system 2022-06-25 19:36:03 sadly alpine does not support combined 32/64 bits 2022-06-25 19:36:43 ikke: i see, ok thanks 2022-06-25 19:39:50 so how do i get the 64 bit version? 2022-06-25 19:40:14 (sorry for being stupid btw lol) 2022-06-25 19:40:15 did you try wine64 ? 2022-06-25 19:40:22 ikke: not yet 2022-06-25 19:41:12 oh i think it worked, thanks 2022-06-25 19:44:48 ShellExecuteEx failed: File not found. 2022-06-25 19:44:48 i'm sure i typed the exact name of the file though, i even tried with and without the .exe part 2022-06-25 19:46:53 might be that the exe is 32-bits 2022-06-25 19:49:00 ikke: i see, ok thanks 2022-06-25 19:49:26 https://unix.stackexchange.com/questions/643633/wine-not-running-this-exe-setup-file-in-void-linux 2022-06-25 19:49:42 Note that the solution there will probably not work on Alpine 2022-06-25 20:08:19 "Note that the solution there..." <- ok, thanks 2022-06-25 22:43:10 did anyone run steam on flatpak fine? it's not creating the virtual input for me 2022-06-25 22:43:33 i loaded the uinput module on boot 2022-06-25 22:44:17 user groups (input and uinput) are fine too, maybe i'm missing something 2022-06-25 22:44:26 what virtual input is it supposed to make? i've never had any issues 2022-06-25 22:45:53 psykose: i remember that steam creates a temporal virtual input, like an xbox 360 controller, temporarily 2022-06-25 22:48:04 no idea, never checked and also never had any problems 2022-06-25 22:48:34 do you use xorg 2022-06-25 22:48:38 no 2022-06-25 22:49:09 im on wayland too 2022-06-25 22:49:39 do you also have udev 2022-06-25 22:49:51 "did anyone run steam on flatpak..." <- works fine for me; i pretty much just set up xorg w libinput and setup dbus and ran with it on this install, but I remember it working fine with xwayland too 2022-06-25 22:50:19 psykose: the plain udev service? 2022-06-25 22:50:54 udev instead of mdev or mdevd, via setup-devd udev 2022-06-25 22:51:42 jcbsnclr[m]: does it create the virtual input and you can see it listed in evdev-joystick --l ? 2022-06-25 22:52:03 psykose: that might be! 2022-06-25 22:52:09 gonna give it a try and come back 2022-06-25 22:58:39 psykose: the udev service is now up 2022-06-25 23:18:35 psykose: is there anything else more than setting up udev as service? 2022-06-25 23:18:48 i don't know what you mean 2022-06-25 23:19:08 anything else to do 2022-06-25 23:19:21 steam can read the input in the client, but still not creating the virtual one 2022-06-25 23:19:45 read what input in what client 2022-06-25 23:20:18 my generic joystick in the controller configuration SDL UI where you can map everything 2022-06-25 23:20:35 (general controller settings) 2022-06-25 23:21:26 ah, i don't think i've ever used any 'controller configuration ui' except for what is in steam 2022-06-25 23:21:45 ahh 2022-06-25 23:54:21 psykose: ok found out the virtual input gets created by a setting per game 2022-06-25 23:54:28 so obscure 2022-06-26 11:05:18 Hi. I upgraded to Alpine 3.16 and am attempting a custom kernel. It seems it cannot handle a local version appended to the name. During creation of initramfs-lts it does not create the file as initramfs-lts-Forza (the appended string) 2022-06-26 11:05:29 Same issue in earlier alpine version too. 2022-06-26 11:07:05 yes, only one - 2022-06-26 11:07:11 name it -Forza instead 2022-06-26 11:07:19 not -lts-Forza 2022-06-26 11:07:32 I have CONFIG_LOCALVERSION="-forza" 2022-06-26 11:07:46 in what 2022-06-26 11:08:07 https://dpaste.com/3QVTMQ2UZ 2022-06-26 11:08:19 In lts.x86_64.config 2022-06-26 11:08:41 Also a change between 3.15 and 3.16 is the name of the config files 2022-06-26 11:09:00 [~/aports/main/linux-lts]$ nano lts.x86_64.config 2022-06-26 11:09:54 instead of that, copy linux-lts to linux-forza and don't set localversion 2022-06-26 11:10:56 also the paste you posted did make `Found initrd image: /boot/initramfs-lts-forza` 2022-06-26 11:11:25 Yes but that is the old initramfs file that I manually renamed 2022-06-26 11:11:40 ah 2022-06-26 11:11:53 If I remove it it fails and I get an unbootable system 2022-06-26 11:12:25 So I should copy lts.x86_64.config to forza.x86_64.config? 2022-06-26 11:14:12 probably 2022-06-26 11:15:01 Trying :) 2022-06-26 11:15:13 I am using a vm build host so no issues to test 2022-06-26 11:17:55 Thanks for fixing eatmydata update 2022-06-26 11:19:32 :) 2022-06-26 11:25:59 Alpine is really very nice. Both community and as a product. 2022-06-26 11:37:18 Changing config to forza.x86_64.config did not work 2022-06-26 11:37:59 /usr/bin/abuild: cd: line 195: can't cd to /home/forza/aports/main/linux-lts/src/build-lts.x86_64: No such file or directory 2022-06-26 11:38:11 Because the actual build root is build-forza.x86_64/ 2022-06-26 11:39:56 So I'll try without local version and see what happens 2022-06-26 11:40:34 anyone knows a simple&reliable way for sync maildir folders? 2022-06-26 11:40:41 the package is still -lts 2022-06-26 11:41:15 I have some cron errors on a server and would like to notice them on my laptop without too much complexity 2022-06-26 11:41:17 but also, in package() it calls _package() with `lts` 2022-06-26 11:41:28 psykose: yes. So the build scripts get confused. I suppose the point is to not change the name, and if so create a new package instead? 2022-06-26 11:41:52 well, that hurdle is changing the package() call to be `forza` instead, that always has to be changed 2022-06-26 11:42:02 but there will probably be another broken thing iirc, really just copy the folder 2022-06-26 11:42:42 cp -r linux-lts linux-forza, rename the configs, delete all the -virt stuff, rename any lts references to forza, aand i think that would be it 2022-06-26 11:42:49 You mean copy the individual vmlinuz and modules to the target machine? 2022-06-26 11:43:47 I suppose. Which way would be minimal effort to keep my custom kernel config, but still allow easy upgrade when Alpine upgrades new kernel sources 2022-06-26 11:44:17 no 2022-06-26 11:44:25 copy main/linux-lts to main/linux-forza 2022-06-26 11:44:30 uhM, Forza you could create a local branch with your config changes, when there is some update just merge it with master 2022-06-26 11:44:33 or make your own repo, etc 2022-06-26 11:44:39 ah I see 2022-06-26 11:44:44 it's your own kernel, you don't have to keep anything 2022-06-26 11:46:19 you could also just create a patch 2022-06-26 11:46:26 and add it to sources 2022-06-26 11:46:28 also an option 2022-06-26 11:47:03 anyway, are your changes related with hardware support? 2022-06-26 11:47:04 I'll try to do the ˜/aports/main/linux-forza/ option 2022-06-26 11:47:29 Yes I wanted to add some tracing options to debug a SAS HBA controller. Also enable BFQ support 2022-06-26 11:47:52 Nothing major otherwise 2022-06-26 11:48:00 ah, if they are debug options it's unlikely they be enabled official 2022-06-26 11:49:01 I have an annoying bug where the LSI/Broadcom HBA 9500-8i SAS/SATA controller has a hard coded timeout of 30 seconds. So when some SSDs take longer, the controller attempts to reset the bus 2022-06-26 11:49:28 Broadcom support wasn't helpful. They suggested to disable fstrim 2022-06-26 11:49:50 ouch 2022-06-26 11:50:03 Yup. Lovely :o 2022-06-26 11:52:39 What I don't get is that they don't use the variable /sys/block//device/timeout values 2022-06-26 12:15:10 Hm. The apk was still named linux-lts-5.15.48-r101.apk 2022-06-26 12:20:44 you forgot to change pkgname probably, or never built it, dunno 2022-06-26 12:21:00 (or are looking at the wrong apk) 2022-06-26 12:31:02 https://dpaste.com/DJPU3LEFB that's the APKBUILD 2022-06-26 12:32:50 the _flavors gets set to the config names, so the lts.x86_64 generates a -lts 2022-06-26 12:33:09 so you have to change that too 2022-06-26 12:33:14 Minimal is helping me :) 2022-06-26 12:34:23 Trying this now https://dpaste.com/5YQZDZUQN 2022-06-27 11:23:44 anyone tried aspell from repo with dictionary from https://ftp.gnu.org/gnu/aspell/dict/0index.html ? im trying 'nb', but it triggers on _every_ word 2022-06-27 16:29:36 I'm trying to get networkmanager to manage eth0 on a fresh 3.16 install, but `nmcli device status` shows eth0 as "unmanaged" even though the `networkmanager` service is running and the `networking` service is not. 2022-06-27 16:29:52 is this something to do with the ifupdown plugin not being installed/configured correctly? 2022-06-27 16:30:21 if you want to use nm via ifupdown plugin, you need to specify it in the config 2022-06-27 16:31:10 I don't really have a preference for which plugin to use, just that I need nm to manage eth0 2022-06-27 16:31:20 that being said, I tried all permutations of configs and plugins and nm is ****, sometimes it just doesn't want to manage devices at all (: 2022-06-27 16:32:58 if you don't want any plugin, just remove interfaces from /etc/network/interface and reload 2022-06-27 16:33:17 should eth0 get managed by default without any plugins? 2022-06-27 16:33:36 yes? 2022-06-27 16:34:04 Currently /etc/network/interfaces only mentions the lo device 2022-06-27 16:34:27 also I thought this file was to do with the `networking` service, might be mistaken 2022-06-27 16:34:46 it is 2022-06-27 16:35:23 do I also need this service running? I've disabled it because I thought it might be competing with nm for managing the connections 2022-06-27 16:35:51 eddsalkield_: what does you /etc/NetworkManager/NetworkManager.conf look like? 2022-06-27 16:36:15 if you don't have it in /etc/... and it's unmanaged you are probably hitting the situation where NM just doesn't want to manage no matter what 2022-06-27 16:36:41 anyone tried aspell from repo with dictionary from https://ftp.gnu.org/gnu/aspell/dict/0index.html ? im trying 'nb', but it triggers on _every_ word 2022-06-27 16:38:04 durrendal: /etc/NetworkManager/NetworkManager.conf didn't contain anything at first, then I've added dhcp=internal and plugins=ifupdown to [main] 2022-06-27 16:38:25 I also added [ifupdown] managed=true 2022-06-27 16:38:51 didn't seem to work either with or without the ifupdown plugin, but that might be because I was configuring it incorrectly 2022-06-27 16:39:04 `ip a` shows eth0 as DOWN 2022-06-27 16:39:37 http://ttm.sh/wCI.conf <- try something like this config 2022-06-27 16:40:05 though the wifi.backend stuff can be ignored since you're looking to control eth0 2022-06-27 16:40:25 and dhcp could be internal 2022-06-27 16:40:39 yes dhcp absolutely could be internal if you don't have/want dhcpd 2022-06-27 16:42:29 okay I've edited the config to be identical except the wifi stuff 2022-06-27 16:42:41 do I need any additional packages for dhcp or ifupdown? 2022-06-27 16:43:10 on reboot eth0 is still down and `nmcli device status` shows eth0 as unmanaged 2022-06-27 16:46:31 I know for sure you'd need to restart networkmanager for the config change to take effect, you may want to restart networking as well 2022-06-27 16:46:46 ah but if you rebooted and it's still down, then that's moot 2022-06-27 16:48:35 do you have networkmanager-elogind installed? 2022-06-27 16:49:08 also if you add auto eth0 iface eth0 inet dhcp to /etc/network/interfaces does the interface state change on a reboot? 2022-06-27 16:50:02 i'll try this, thanks 2022-06-27 16:51:44 like panekj said earlier, NM can be a real pain 2022-06-27 16:52:30 I didn't have networkmanager-elogind installed, and I've added those lines to /etc/network/interfaces 2022-06-27 16:54:40 eth0 is still down on reboot 2022-06-27 16:54:55 but I can bring it up with `ifconfig eth0 up` 2022-06-27 16:55:05 and then manually running udhcpc gets me an ip 2022-06-27 16:56:52 what's the program that's supposed to read /etc/network/interfaces and bring the interfaces up? 2022-06-27 16:59:15 I believe it's ifupdown-ng, but that's required by openrc, so it should be installed already 2022-06-27 17:01:06 I disabled nm and reenabled networking, now I'm getting eth0 up on boot which is nice 2022-06-27 17:01:28 I didn't get this before however - not sure why 2022-06-27 17:07:18 sounds like the /etc/network/interfaces change helped, I'd bet that if you checked ncmli the device still shows as unmanaged though 2022-06-27 18:07:19 wondering why I have to manualy apply 'sysctl -w net.ipv4.ip_forward=1' after a reboot even if it is in /etc/sysctl.conf (and sysctl net.ipv4.ip_forward is showing 1 after the reboot) ? 2022-06-27 18:14:01 did you rc-update add sysctl ? 2022-06-27 18:19:15 indeed I hadnt... I thought it was already here.. let me check 2022-06-27 18:20:23 durrendal: arf... thanks indeed that's better now 🤐 2022-06-27 18:22:30 :) glad to hear it worked 2022-06-27 21:58:18 Does Qt and/or OpenAL in Alpine support sndio? 2022-06-27 22:03:29 looks like a no 2022-06-27 22:03:51 :-( 2022-06-27 22:05:09 does sndio not support alsa clients or something 2022-06-27 22:05:48 because if it doesn't that sounds quite useless to use, as you can get audio from like.. 3 programs 2022-06-27 22:20:11 There is a bridge you can make ALSA spawn that talks sndio protocol. 2022-06-27 22:24:24 An ALSA plugin to be precise. But it has several downsides, like no per-app volume control because ALSA downmixes everything before it gets sent off. Nor per-app selection of sound device, etc. 2022-06-27 22:34:07 the pipewire-alsa bridge handles per-app everything fine, as does (iirc) the pulseaudio alsa interface 2022-06-27 22:34:16 you really can't use sndiod with any features with non-sndio clients? 2022-06-27 22:39:07 Aha, misunderstood what the plugin doc says. Applications won't get access to their per-app volume knob, only the global one. 2022-06-27 22:40:00 that makes more sense 2022-06-27 22:40:07 Though I specifically won't as I want to avoid alsa here completely (or at least on kernel level, if plugins work with no alsa in kernel then maybe that would be a way). 2022-06-27 22:41:35 But I don't think alsa-sndio is packaged anyway. 2022-06-27 22:41:41 not sure how any audio would work without kernel alsa 2022-06-27 22:42:10 By sending it elsewhere over socket/network. 2022-06-27 22:42:33 where would the elsewhere be 2022-06-27 22:44:22 Anything that exposes sndiod over network, like OpenBSD workstation. 2022-06-27 22:45:32 Though at the moment this was more of experiment in how much I can cut down container access to kernel interfaces. 2022-06-28 01:35:00 I don't understand the docs - is there an envvar I need to set, flag I need to pass, etc for awall to take YAML instead of JSON? Do I just feed it the file and it guesses which it is based on the extension? 2022-06-28 03:01:02 i don't know anything about awall but json is yaml 2022-06-28 03:01:17 so i assume it always takes yaml 2022-06-28 03:46:16 Hello71: I thought it was the other way around? And, I imagine they require different parsers in most cases - I would want a JSON library to "do JSON only and do it well", I'd think 2022-06-28 03:47:19 yaml is a superset of json 2022-06-28 03:47:31 my point is that if awall accepts yaml, then it automatically also accepts json 2022-06-28 03:48:02 ah, I see 2022-06-28 03:48:12 well, no harm in trying inside a VM, I suppose 2022-06-28 03:56:26 so it does not necessarily require any special heuristic to "detect" json files and pass them through a special parser. it could, and it might, but my guess is that it probably doesn't 2022-06-28 04:24:19 awall only accepts json 2022-06-28 04:25:54 talismanick: where did you get the impression awall had support for yaml? 2022-06-28 04:38:59 (though that would be a nice addition) 2022-06-28 04:54:07 ikke: My bad, I thought I read that somewhere 2022-06-28 08:30:08 wondering if adding http3/quic support to nginx would be possible ? https://git.alpinelinux.org/aports/tree/main/nginx/APKBUILD I don't see the http3 module in the build, anyone knows if Jakub the maintener is in this chan ? 2022-06-28 08:30:58 open an issue on gitlab 2022-06-28 08:31:03 jirutka is not on IRC 2022-06-28 08:32:22 s/gitlab/gitlab.a.o 2022-06-28 08:32:22 panekj meant to say: open an issue on gitlab.a.o 2022-06-28 08:34:39 nginx-quic is a separate build of nginx entirely is it not 2022-06-28 08:37:11 it seems it is a separate branch of nginx with quic 2022-06-28 08:37:22 so yeah, not a module 2022-06-28 08:38:30 or rather, it is a module, but only available for nginx builds in quic branch 2022-06-28 08:39:40 aye 2022-06-28 08:50:45 Hi all :-) 2022-06-28 08:50:56 o/ 2022-06-28 08:51:37 Is anyone using Yubikey/Nitrokey experiencing issues on alpine/edge ? 2022-06-28 08:53:19 Perhaps a better question would be: "I'm experiencing issue X with Yubi/Nitro on alpine edge. How can I fix that?" 2022-06-28 08:53:40 so people not using yubi/nitro/edge can also answer (: 2022-06-28 08:54:27 what broke 2022-06-28 08:54:47 The problem seems to be related to hardware devices as I can encrypt/decrypt using keys stored as file 2022-06-28 08:55:21 I'm just having a timeout, unable to decrypt anything 2022-06-28 08:56:13 was it broken before? i.e. on 3.16 world or something 2022-06-28 08:57:18 The hardware tocken is working on another computer (having another OS). I'm on edge since a long time and I haven't tried on 3.16 release 2022-06-28 08:58:52 no idea what the specific issue could be 2022-06-28 08:59:13 you can check all the logs / dmesg to see if anything is there, and any error you get 2022-06-28 09:01:08 Ahhh I get it 2022-06-28 09:01:40 I reverted gpg related packages to @3.16, and it's working again 2022-06-28 09:02:24 "or rather, it is a module, but..." <- in fact I saw that here https://quic.nginx.org/ , maybe it is more a nginx-quic needed instead of changing the current one, will open an "issue" on Gitlab 2022-06-28 09:02:29 which ones 2022-06-28 09:02:43 i don't think nginx-quic would get packaged 2022-06-28 09:03:05 I did "apk add -i gpg@3.16 gpg-agent@3.16 gnupg-scdaemon@3.16 pcsc-lite@3.16 pcsc-lite-openrc@3.16" 2022-06-28 09:03:34 I am more wondering if it is safe to build our current nginx to the quic branch and maybe better to have a separate nginx-quic 2022-06-28 09:04:03 well, pcsc-lite did get upgraded.. the others are identical 2022-06-28 09:04:23 I would say that switching standard nginx to quic branch is definite no 2022-06-28 09:05:02 yep that's what I meant 2022-06-28 09:05:20 better to have another version for the quic branch and not touching to our current nginx 2022-06-28 10:25:42 Hi 2022-06-28 10:26:15 I want to set up a syslog server on Alpine. It should receive logs from other computers. What syslog server do you recommend? 2022-06-28 10:26:59 On Gentoo I use syslog-ng but it was years ago since I set it up 2022-06-28 10:28:37 rsyslog is quite nice 2022-06-28 10:54:50 is it possible to use busybox/mount to mount drive with permission for user to write to it? (without editing fstab) 2022-06-28 11:03:10 that depends on the filesystem on the drive 2022-06-28 11:04:19 ah, and busybox mount help seems to be lacking uid= for -o anyway, but maybe it's supported and undocumented (doubt) 2022-06-28 11:04:49 but if that was, there's a uid=1000 or whatever for your user you can pass, only for.. non-posix-acl(?) filesystems like fat and the like 2022-06-28 11:07:33 psykose: i see 2022-06-28 11:24:21 after many upgrades without rebooting I did it yesterday and my swaybar fonts seems weird 2022-06-28 11:24:54 what font do you have set 2022-06-28 11:24:56 e.g. it misses mayus/minus, points, @... 2022-06-28 11:25:02 good question :P 2022-06-28 11:25:57 psykose: do you mean on swaybar? other things looks fine 2022-06-28 11:26:01 yes 2022-06-28 11:26:35 there is nothing special on .config/waybar/config 2022-06-28 11:26:43 that's not swaybar, that's waybar 2022-06-28 11:26:55 yeah 2022-06-28 11:26:57 in that case.. it defaults to uh 2022-06-28 11:26:59 monospace something 2022-06-28 11:27:14 can you post uh 2022-06-28 11:27:19 `fc-match -s monospace` 2022-06-28 11:27:41 http://ix.io/416B 2022-06-28 11:28:10 hmm 2022-06-28 11:28:56 info: fcft.c:839: /usr/share/fonts/ttf-dejavu/DejaVuSansMono.ttf: size=8.00pt/8px, dpi=75.00 2022-06-28 11:29:10 yeah i've no idea why it would be missing common symbols.. 2022-06-28 11:29:16 it's on tinydm.log 2022-06-28 11:29:55 there are any 2022-06-28 11:30:01 if you open an editor with a font you can set (or well, that would be default monospace too), and type with that same font.. does it also break 2022-06-28 11:30:42 let me check 2022-06-28 11:31:18 uhM 2022-06-28 11:31:28 psykose: waybar loads /etc/xdg/waybar/style.css 2022-06-28 11:31:42 yea 2022-06-28 11:31:49 there many fonts there 2022-06-28 11:31:52 http://ix.io/416F 2022-06-28 11:32:09 font-family: FontAwesome, Roboto, Helvetica, Arial, sans-serif; 2022-06-28 11:32:12 ah, then it would be using those 2022-06-28 11:32:45 the file changed recently 2022-06-28 11:33:11 does your waybar config have icons 2022-06-28 11:33:25 the defaults 2022-06-28 11:33:33 do you have font-awesome added 2022-06-28 11:33:34 I can see the temp icon 2022-06-28 11:33:42 and battery 2022-06-28 11:33:45 and then.. one of those fonts 2022-06-28 11:34:18 problem is that eth0 192.168.1.1/24 2022-06-28 11:34:20 looks like 2022-06-28 11:34:25 ETH192168110024 2022-06-28 11:34:44 or ETH0 192168110024 2022-06-28 11:34:59 so do you feel that I need to install some font package? 2022-06-28 11:35:16 uhM 2022-06-28 11:35:24 look 0345e2585a3e73ffe0e048c80c6ed8994aca633a 2022-06-28 11:38:13 looks that I have both installed 2022-06-28 11:38:51 yeah, not related 2022-06-28 11:48:37 Lord: thanks. Had a look. Seems the config files are a bit harder to maintain 2022-06-28 13:52:17 blt[m]: i think the whole point of it being a separate repo/branch is that the nginx maintainers consider the code unstable. otherwise they would put it in the main nginx repo and turn it off by default or something 2022-06-28 15:30:07 I have a misfeature in busybox 'df'. It fails on btrfs subvolumes that are not also mounted somewhere. 2022-06-28 15:30:33 I guess it's an upstream bug that I should report? 2022-06-28 15:31:12 open it in aports first 2022-06-28 15:32:57 Ok :) a bit later. I'm not at home now 2022-06-28 15:34:03 How is Alpine pronounced? Alpeeen or Pine as in pineapple? 2022-06-28 15:34:20 latter 2022-06-28 15:34:37 Cool. Thanks. 2022-06-28 15:34:57 I heard someone on YT say the first so I got wondering 2022-06-28 16:37:06 Hi, I can't get hwclock working on my rasbperry pi 2. I have hwclock enabled. it prints `rtc-ds1307 1-0068: setting system clock to 2022-06-28T15:48:56 UTC (1656431336)` during boot (dmesg). But somehow it resets to some months before. When I reset it manually using hwclock -s I get the right date again. I'm using chronyd with the rtcsync option. Syslog starts at the wrong date in the past (may 2nd). Did someone experience something simila 2022-06-28 17:16:24 you must write to it every now and then for it to store the time 2022-06-28 17:24:51 that's not the issue, setting it again with -s fixes it, there's just something that overrides it that i'm forgetting 2022-06-28 17:45:04 minimal helped me figuring it out. I had hwclock in the default runlevel, but swclock was also still in the boot runlevel. I didn't get it to use `rc-status -a` to show the boot runlevel and didn't find log entries of swclock... Removing swclock fixed it. 2022-06-28 17:49:36 ahh yes that "small" issue :) 2022-06-28 17:49:39 conflicting clocks... 2022-06-28 17:56:40 maybe it would make sense for swclock and hwclock to conflict 2022-06-28 19:55:15 Has anyone any experience with Busybox HTTPD and serving PHP-based content like PiwiGo? 2022-06-28 19:55:33 Along those lines, is there anything special about packaging PHP for PiwiGo? 2022-06-28 20:24:25 No experience 2022-06-28 20:39:40 Mmm, alright, thank you. Trying to do some pre-research on PiwiGo, PHP, Busybox HTTPD before I give myself a massive project and burn time I don't have setting up my Pi 1B+ as an Alpine+PiwiGo box 2022-06-28 20:48:06 theoretically it should work with cgi or interpreter mode 2022-06-28 20:48:18 but will likely be slower than fpm or module moe 2022-06-28 22:00:40 Security-wise, is busybox httpd a good choice or should I be using nginx or something? 2022-06-28 22:13:24 nginx receives much more testing and security audits than busybox httpd 2022-06-28 22:19:00 But larger surface, or does that sort of come out in the wash with the increased scrutiny/usage it has? 2022-06-28 22:19:29 I'd say that busybox is bad for anything (: 2022-06-28 22:20:02 And yet the only Linux distro I've ever used that doesn't make me want to rip my eyes out in frustration is built around it, so it must not be awful :P 2022-06-28 22:20:46 By default nginx does not do much 2022-06-28 22:20:53 https://gitlab.alpinelinux.org/alpine/tsc/-/issues/39 mhmmmm 2022-06-28 22:21:11 So basically unless you configure it to do something crazy you should be very well tested 2022-06-28 22:23:29 Hmm... That sounds good, consus. I'm not well-versed with network security so I'm looking for something that will make it hard to shoot myself in the foot with if at all possible 2022-06-28 22:23:42 All it has to do is serve PiwiGo, nothing more 2022-06-28 22:23:50 Hence my interest in busybox httpd 2022-06-28 22:30:26 You want to make a public site? 2022-06-28 23:54:16 Want... No. Need? Feels like it. I have need to share photos with family that will not use anything other than Facebook, and I'm not putting data there anymore, so I settled on a Public self-hosted PiwiGo I can share a link to them with 2022-06-28 23:55:14 Considered pixfed but I don't think I can get the type of control I want without having everyone make a pixfed login, which they will not do 2022-06-28 23:56:04 what about other options, not exactly php-based and more self-contained-web-server? 2022-06-29 01:06:49 Is the PHP thing a big problem to deal with? I don't know of anything else like PiwiGo, and I didn't know how it was built until I saw we don't package it and I wanted to give that a go 2022-06-29 01:06:57 Do you have any recommendations? 2022-06-29 02:00:52 Saijin_Naib: https://photoview.github.io/ 2022-06-29 02:01:01 downside: it's also not packaged in alpine 2022-06-29 02:01:09 and it's Go(lang) 2022-06-29 02:03:32 Oh no... Is Go hard to package or work with? 2022-06-29 02:03:37 nah 2022-06-29 02:04:36 but it's garbage collected language, so it uses a bit more memory 2022-06-29 02:08:13 Yo, this is amazing looking! Thank you so much 2022-06-29 02:08:28 (: 2022-06-29 02:08:44 I assume with PHP and whatnot and how beefy PiwiGo is, this might be a lighter choice for the Pi 2022-06-29 02:09:33 Add this to the massive packaging backlog too 🤣 I'll do them someday, I hope 2022-06-29 02:10:00 You might want to disable face recognition and video encoding if you are going to use rpi 2022-06-29 02:10:30 photoview during scan/thumbnail generation can be quite heavy 2022-06-29 02:11:21 but I also have very big libraries of pics/vids so can't tell if you will have same experience 2022-06-29 02:16:41 Ahhh, good points... Hmm. It won't be many photos, just selections for family members to be able to view. My internal library is like 300GB, but not what all I plan to share 2022-06-29 04:45:13 Is there current/correct documentation on how to install Alpine in system mode on a Raspberry Pi 1? 2022-06-29 04:55:03 can't find wireguard module in virt version. 2022-06-29 04:56:59 are you sure? 2022-06-29 04:57:31 did you reboot after updating kernel? 2022-06-29 04:58:30 linux-virt 5.15.50-r0 in alpine v3.16 does have wireguard 2022-06-29 05:02:15 no let me reboot. 2022-06-29 05:05:28 thanks, now works 2022-06-29 05:06:29 yeh wireguard-tools package 2022-06-29 05:09:19 does alpine linux have openbsd's pf. 2022-06-29 05:10:56 oh nervermind, I guess pf is BSD specific. 2022-06-29 06:07:24 Good morning 2022-06-29 06:08:09 good morning 2022-06-29 06:09:26 I've been testing syslog-ng and found a problem in syslog-ng.conf. The way it is set up it will never use included syslog-ng/conf.d/*.conf because it sends everything not matched to var/log/messages before the include *.conf 2022-06-29 06:11:15 Solution is to move the @include above the last line: https://dpaste.com/F329JUYHC 2022-06-29 06:49:49 Does someone have the freeipmi.plugin for netdata available? 2022-06-29 08:54:03 anyone aware of what changes are needed for NFS? my previous setup was working fine and now with updating to 3.16 it complains of "unsupported version" when trying to mount 2022-06-29 08:54:41 did nfs update? 2022-06-29 08:56:09 unsure to be honest, i updated and left it while it did its thing 2022-06-29 08:57:53 from 3.15 to 3.16 nfs-utils has different version 2022-06-29 08:58:18 are you using the latest now? 2022-06-29 08:59:58 yeah 2.6.1 shows installed 2022-06-29 09:00:14 previous was 2.5.4 2022-06-29 09:01:01 gawd need to find the magic fix between debian and alpine by the looks of it 2022-06-29 09:01:38 what is the exact error message? 2022-06-29 09:06:43 mount.nfs: mount options: "bg,hard,nolock,nointr,nfsvers=4,addr=xxx.xxx.xxx.xxx" 2022-06-29 09:06:43 repeat the two lines above, then 2022-06-29 09:06:43 * ERROR: nfs failed to start 2022-06-29 09:06:43 mount.nfs: backgrounding "nfs-server:/" 2022-06-29 09:06:43 2: Unsupported version 2022-06-29 09:07:03 hm 2022-06-29 09:07:27 and that same config worked with the previous version? 2022-06-29 09:07:36 yeah 2022-06-29 09:08:31 i even tried changing the /etc/exports from /mnt/nfs to /mnt/nfs4 2022-06-29 09:11:46 removing nfsvers=4 does not help 2022-06-29 09:12:39 hmm oh well ill try to find the solution to my problem 2022-06-29 09:12:53 sorry i couldn't help 2022-06-29 09:13:02 np 2022-06-29 11:34:09 https://odysee.com/@freq301:7/Beavis.and.Butthead.Do.the.Universe.2022:3 2022-06-29 14:22:16 hi. im trying to debug a performance problem that has appeared on my alpine router after I enabled ipv6. Im comparing the performance of a web page when using ISP provided (also ipv6 enabled) router with my alpine router 2022-06-29 14:22:45 Firefox says that the page (https://vg.no) loads in 300-400ms with ISP provided router 2022-06-29 14:22:56 and 7-10 seconds with the alpine router 2022-06-29 14:23:18 I suspect it was me enabling ipv6 that introduced it 2022-06-29 14:24:15 weird, now it seems it loads in 1.5 seconds 2022-06-29 14:24:29 Happy eyes should make clients to fall back to Ipv4 when ipv6 is slow / timing out 2022-06-29 14:24:54 happy eyes? 2022-06-29 14:25:28 I suspect its dns failures or similar, or that ipv6 fails and falls back to ip4 after a timeout 2022-06-29 14:25:38 but I have no idea how to troubleshoot it 2022-06-29 14:27:15 I wonder if it is my firewall, or unbound resolving dns 2022-06-29 14:27:34 ok now it took 6 seconds 2022-06-29 14:28:07 any ideas what I can do to figure out what is going on? 2022-06-29 14:29:12 Im not able to reproduce it with curl 2022-06-29 14:30:59 is vg.no supposed to be down? i'm getting an error 2022-06-29 14:34:25 bad gateway? 2022-06-29 14:34:31 I saw it a couple of times 2022-06-29 14:34:44 "error 559 down" with a message in norwegian saying to contact the webmaster 2022-06-29 14:34:58 its up for me 2022-06-29 14:35:08 it does say HTTP 559 2022-06-29 14:35:16 as the status code, not just on the webpage 2022-06-29 14:36:18 https://downforeveryoneorjustme.com/vg.no 2022-06-29 14:37:07 but internet in general has been horribly sluggish after I enabled ipv6 2022-06-29 14:37:23 but it appears to work oldish with the ISP provided router 2022-06-29 14:37:34 I guess you could exclude DNS issues using `curl -kv https://[]/' -H 'Host: www.vg.no'` (with the Host to bypass any redirect or IP blocks) 2022-06-29 14:38:00 or maybe curl -6 2022-06-29 14:38:48 yeah maybe you can compare between -4 and -6 on both routers 2022-06-29 14:41:45 ok, I can reproduce now. it takes 5 seconds 2022-06-29 14:41:47 with -6 2022-06-29 14:42:07 and now its snappy again 2022-06-29 14:42:32 I had a short period with 3 tries and it was 5 seconds 2022-06-29 14:42:43 does it lag intermittently like that with just a ping too? 2022-06-29 14:42:55 curl -v already shows if dns is slow 2022-06-29 14:43:14 if it is only slow on firefox, then it smells like doh 2022-06-29 14:43:38 vg.no shows 559 on my side and downforeveryoneorjustme.com says vg.no is down. may be it is just host? 2022-06-29 14:44:23 possibly 2022-06-29 14:44:45 ipv4 works mostly, but I just had 5 sconds with curl -4 too 2022-06-29 14:46:01 I had an impression that it was not only vg.no 2022-06-29 14:46:35 * connect to 2001:67c:21e0::16 port 443 failed: Network is unreachable 2022-06-29 14:46:35 * Failed to connect to vg.no port 443 after 4052 ms: Network is unreachable 2022-06-29 14:47:59 "Network is unreachable" doesn't that mean that there is no default route? 2022-06-29 14:48:04 locally 2022-06-29 14:50:53 hmmm yeah that's a local route issue, considering that it didn't crash in 0ms i'll say it tried to connect to your router but couldn't (there probably is a route, but to nowhere) 2022-06-29 14:51:01 Jun 29 14:50:15 rtr-copa daemon.info dhcpcd[26041]: eth0: fe80::ae4b:c8ff:fe9a:f29e is unreachable 2022-06-29 14:51:01 Jun 29 14:50:16 rtr-copa daemon.info dhcpcd[26041]: eth0: soliciting an IPv6 router 2022-06-29 14:51:13 I see that frequently on the router 2022-06-29 14:58:07 lol 2022-06-29 14:59:07 so I decided torrid what vg.no actually says (Norwegian newspaper) and there is a headline: "VG is affected by attacks" 2022-06-29 14:59:16 to read* 2022-06-29 14:59:38 that explains the issues with vg 2022-06-29 14:59:59 but it does not explain why my wife complains about slow internet last few days 2022-06-29 15:00:26 she says its been slow last week 2022-06-29 15:00:47 I guess the cyberattack does not help me troubleshoot my network setup 2022-06-29 15:01:02 better try investigate other day 2022-06-29 17:33:08 im beginning to dislike NFS... why did i update grr 2022-06-29 18:29:24 i have stumbled upon a problem: project i compile and run on Alpine without any leaks or segfaults (checked with valgrind) just spits out a bunch of invalid reads and leaks on my Debian machine. what do i do in this situation? is there are some compile options i need to explicitly set on glibc based distributions? https://gist.github.com/txgk/aec7e1bac25ed04ce8ef50711474516a shows that libraries are 2022-06-29 18:29:30 faulty, not my code (i hope) 2022-06-29 19:58:00 I don't see invalid reads in your paste 2022-06-29 19:58:10 I also wonder if you picked the worst possible channel for your question ;) 2022-06-29 20:10:29 oh i'm sorry about that, i thought it is related because some compiler settings make my program work flawlessly and i need to set these settings manually on other distributions 2022-06-29 20:11:06 i could be wrong - but "what do i need to do on glibc" seems weird to me in this musl-focused channel :) 2022-06-29 20:16:30 i see, okay 2022-06-29 20:17:22 also, a small collection of individual leaks is not uncommon with some libraries, including glibc, in my experience 2022-06-29 20:17:33 some things are allocated once, when used for the first time, but not when used again 2022-06-29 20:17:36 so they're not really leaks 2022-06-29 20:18:15 you should check if, say, calling wrefresh more also causes more leaked blocks 2022-06-29 20:22:00 grisha: it looks like a typical ncurses/terminfo leak, they allocate memory and never free it, because it's supposed to be used throughout the program's runtime 2022-06-29 20:23:01 you can compile ncurses with --disable-leaks in the configure to check for yourself 2022-06-29 20:31:06 ptrc: i know that still reachable blocks are not a problem especially for ncurses, but there are possibly lost blocks which i don't get when program is compiled on Alpine. also this paste i mentioned above refers only to the first seconds of the programs operation, then after some actions it segfaults (valgrind reports invalid reads in vsnprintf a lot). again, on alpine everything is okay, and i 2022-06-29 20:31:12 don't want to litter this channel with glibc issue, as Habbie pointed out 2022-06-29 22:53:43 well alpine doesn't use glibc 2022-06-30 02:03:44 Is xf86-video-fbdev the appropriate video package for the Rasbperry Pi platform, or is it just the compatibility fallback so we have some video for XOrg? 2022-06-30 02:22:19 Does the raspi-config CLI tool work on Alpine? I see it uses the /bin/sh shebang, so it should work in Busybox ASH, right? 2022-06-30 02:23:53 theoretically raspi-config could work, but i suspect nobody uses it 2022-06-30 02:24:01 also many of the functions would certainly not work 2022-06-30 02:24:29 i'm not terribly familiar with video drivers on rpi; my understanding is the answer is "it's complicated" 2022-06-30 02:26:54 probably the config.ini stuffs works 2022-06-30 02:28:59 but also you don't really need that either, raspi wiki documented all sort of config stuff 2022-06-30 02:36:04 Yeah, it was just nice to set a "safe" profile, but I'm poking through what I need to set it manually in the config.txt 2022-06-30 02:36:24 Alpine is way faster than Debian on this Pi 1B+, but wow, is this thing absolutely horrible 2022-06-30 02:36:48 Ah, then I guess if video is complicated, I'll just have to hope I never need to manage anything graphically on it 2022-06-30 02:37:03 If I'm lucky, I can just set & forget PhotoView once it is setup 2022-06-30 02:38:15 alpine is hellafast, that's why containers and embedded systems love them 2022-06-30 02:38:52 but i'm a fan of debian more, they seem to keep support for vast majority of architectures 2022-06-30 02:39:03 i think the m68k is even supported there, no? 2022-06-30 02:39:26 or mips, i forgot 2022-06-30 02:40:36 there's no safe profile :) 2022-06-30 02:40:39 also ^ 2022-06-30 02:48:04 Safe was quoted for a reason 😛 2022-06-30 02:57:28 .. 2022-06-30 02:57:39 i clearly didn't see that 2022-06-30 03:18:34 Nah, all good. Not being snarky, haha. My platform is old and finicky and has the USB inrush current bug and can't overclock well besides 2022-06-30 03:18:43 Might have to be Pi4 time finally 2022-06-30 03:22:59 yep! except.. 2022-06-30 03:23:23 *fears* 2022-06-30 03:27:15 i hate reauthenticate with the wifi network because it disconnects me from irc for a moment 2022-06-30 03:27:24 for pi4 afaik all you need is dtoverlay=vc4-kms-v3d 2022-06-30 03:40:24 Oooh, that could be nice... I'm thinking the Pi 2B is so much better than what I have that it might make sense to move to over the 3B+/4B for price and wattage... Hmm 2022-06-30 03:42:00 i'd move to the pi4, more powerful and still draw around the same amount as the old 3b 2022-06-30 03:42:38 and the latest model is "overclocked" out of the box 2022-06-30 03:43:01 except the price currently.. 2022-06-30 03:43:19 and i dont think you can buy them easily at retail anymore 2022-06-30 03:46:25 Ahhh, yeah, I'm just seeing that now... Hmm. I don't need a Pi exactly, just a cheap SBC that I can run Alpine on with Docker for PhotoView... Maybe now is a bad time for this project with supply chain pricing weirdness still 2022-06-30 03:47:27 a little dirty secret, you can get something that's not a pi but have support for alpine :) 2022-06-30 03:47:48 i think the odroid do, not sure 2022-06-30 03:48:41 or if you're like me then i can grab some amlogic board and port alpine to it, i'm planning to do it for my tv box but it's currently a wireguard device now 2022-06-30 03:51:29 I guess I'm just looking for low financial, physical, and electrical footprint, Alpine support, and enough oomph to serve photos 2022-06-30 03:51:45 Pi-class hardware is all I know, but I'll check the ODroid line, thank you 2022-06-30 04:08:03 Oh wow, SBC market is nuts right now. Everything is either out of stock or way, way over-priced 2022-06-30 04:12:38 yep 2022-06-30 04:12:46 everything is nuts these day s 2022-06-30 04:12:51 unfortunately.. 2022-06-30 04:19:31 Good night. There are no nft modules in 3.16. 2022-06-30 04:19:46 https://pastebin.com/C19X9g54 2022-06-30 04:22:32 that looks like you upgraded your kernel so they're not there anymore 2022-06-30 04:24:11 as in, you have to reboot 2022-06-30 04:40:16 psykose. Thanks. I rebooted and it worked. 30 days online. Hug. 2022-06-30 04:40:51 hug :3 2022-06-30 04:45:23 don't reboot it just patch :( 2022-06-30 04:47:53 I didn't understand 2022-06-30 04:55:02 it was a joke about system uptime 2022-06-30 04:55:45 it's like a game where who has the longest uptime wins 2022-06-30 04:56:52 and also was referring to the suse song parody "uptime funk" 2022-06-30 04:57:08 rsrs. It was uptime to 30 days. And I had already updated the kernel a few times. I didn't pay attention to that. 2022-06-30 05:04:30 now you will :) 2022-06-30 05:06:35 :) :) :) 2022-06-30 05:06:52 I really will 2022-06-30 05:25:57 test 2022-06-30 05:34:49 just as i said this, power at home completely went out 2022-06-30 05:35:02 rip me 2022-06-30 05:39:19 :) 2022-06-30 05:54:36 ROCK64 have decent Alpine support? 2022-06-30 05:55:21 I am just now finding out that the ARMV6 in my Pi 1B is not supported by many docker images, so PhotoView is right out the window... This thing is a brick, haha 2022-06-30 05:56:38 may or may not be buildable by hand, didn't try 2022-06-30 05:59:00 for rock64.. dunno? i assume it's at least bootable mainline based on https://wiki.pine64.org/wiki/ROCK64#Upstreaming_Status 2022-06-30 06:13:05 Hmmm.... Bootable is good, but I'm not sure how these devices without EFI/CSM/BIOS load stuff 2022-06-30 06:25:28 https://odysee.com/@freq301:7/Beavis.and.Butthead.Do.the.Universe.2022:3 2022-06-30 07:31:40 my colleague decided to use docker to run our app, my app though uses OpenRC (using supervise-daemon to start) 2022-06-30 07:33:13 I guess I have to change the entry point to start openrc if I want my service to run? 2022-06-30 07:33:28 ACTION never used docker and don't like that very much 2022-06-30 07:34:31 you don't need to supervise it at all if it's just one process/service/app/whatever, just make it the CMD entrypoint 2022-06-30 07:35:16 no it's more than one process actually 2022-06-30 07:35:26 I need cron, logrotate, his app and my app to be running 2022-06-30 07:36:36 the professional advice would be 'you dont need any of that in the same container, use more than one, define cross-container dependencies outside (k8s, docker-compose, etc), manage logs through there as well' 2022-06-30 07:36:55 the less professional advice would be.. dunno, i guess s6-overlay is an easy way to run more than one thing in a container 2022-06-30 07:48:32 Yeah, don't treat docker containers like VMs is the best advise 2022-06-30 07:48:47 Running openrc inside docker is not trivial 2022-06-30 07:49:32 as long as I can daemonize my app that will be fine 2022-06-30 07:53:11 okay looks like supervisord can help to start our two apps 2022-06-30 07:53:14 thanks guys 2022-06-30 08:02:48 oh looks like 42.fr mirror's broken 2022-06-30 08:04:11 removed by request 2022-06-30 08:04:30 oh :( 2022-06-30 08:04:37 was one of the fastest 2022-06-30 08:04:44 and easy to type by heart :P 2022-06-30 08:04:56 same-country is quite fast usually yeah 2022-06-30 14:12:55 do anybody use gitea in alpine? How much ram does it consume? 2022-06-30 14:14:06 seems to be like ~150mb for some personal stuff after a week and a bunch of random shit that scraped it 2022-06-30 14:31:20 psykose: thank you. By the way, does it need mail daemon? 2022-06-30 14:32:14 not required 2022-06-30 14:36:18 ok, thank you again 2022-06-30 14:48:17 Hello. I have two alpine machines and I wanted to use mc with sftp to copy some files. But it seems mc fails with 'failed to open file (-16)` 2022-06-30 14:49:03 Sftp server is installed and running as I can use sftp from command line 2022-06-30 14:49:54 -16 == -EBUSY I guess 2022-06-30 14:51:09 What does that mean? Too many sessions? I have the same on both machines 2022-06-30 14:56:39 got `RuntimeError: Unknown network config version: None` error in cloud-init using version 1 or version 2 cloud-init.network-config, any idea what can be wrong ? my file : https://paste.neviani.fr/?b7ec9cab0b877439#DseawYbpqH47wJL6fx6UFQpbxjjPuefAe6od2DxKhog2 2022-06-30 14:57:33 Ermine: indeed, EBUSY 16 Resource busy 2022-06-30 14:59:46 Forza: no, I guess some file was opened and opener prevented anybody else from opening the same file 2022-06-30 14:59:58 Just thoughs, because idk 2022-06-30 15:00:12 can you connect directly with sftp command? 2022-06-30 15:03:32 blt[m]: can you use a pastebin-type service that does not require Javascript enabled? 2022-06-30 15:11:31 Ermine: it's odd. I can't connect at all. It just throws the error 2022-06-30 15:12:01 While from sftp in a shell everything is OK. Maybe some lib or something is missing with mc? 2022-06-30 15:12:36 maybe an mc bug 2022-06-30 15:31:18 It's possible :o I seem to hit lots of annoyances like this 2022-06-30 16:15:27 minimal: sure do you have a service you recomend? 2022-06-30 16:17:24 blt[m]: I believe tpaste is the one Alpine recommend: https://tpaste.us/ 2022-06-30 16:17:28 ix, bpaste, dpaste, debian, ubuntu... 2022-06-30 16:18:24 minimal: https://tpaste.us/Oq9X here we go 😇 2022-06-30 16:19:25 blt[m]: ok, BTW cloud-related discussions are probably best handled in the alpine-cloud channel. 2022-06-30 16:20:07 minimal: ha! didn't know this channel exists thanks for that, let me connect and copy my question thanks 2022-06-30 16:21:23 it doesnt seem to be bridged with matrix 2022-06-30 16:22:16 ok. so this is a network-config file you are providing via NoCloud DataSource? 2022-06-30 16:25:02 minimal: it is a file I am providing through terraform 2022-06-30 16:25:30 which DataSource are you using with cloud-init? 2022-06-30 16:28:26 minimal: not sure I know what you are refering as I don't know how terraform is providing that, if you are refering on which between the 3 available I am using cloud-init.network-config for the network and cloud-init.user-data for the other parameters, no one is working BUT I see that /var/lib/cloud/seed/nocloud-net/network-config in my Alpine containter got the right things (the exact copy of what I send on pastbin) 2022-06-30 16:28:57 blt[m]: so you are using NoCloud DataSource then 2022-06-30 16:29:26 minimal: I think I saw that indeed in cloud-init.log 2022-06-30 16:29:58 the network-config file you provided does not look syntatically correct 2022-06-30 16:29:59 but I still don't know why I got this version error and why it doesn't apply the configuration but was able to copy it to the container 2022-06-30 16:30:17 https://cloudinit.readthedocs.io/en/latest/topics/network-config-format-v1.html 2022-06-30 16:30:20 minimal: that could be the reason, you think something is missing? 2022-06-30 16:32:09 so I could try to delete :... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/ByHwbGrvxQrLvdMACxXWTRYV) 2022-06-30 16:34:44 doing so would fit with the defined v1 syntax 2022-06-30 16:44:24 minimal: many thanks it is much better at least for network ! I have to figure out why user.data is not apllied.. any clue if network v2 syntax is supposed to work in Alpine? 2022-06-30 16:44:52 blt[m]: yes I use v2 format all the time as I find it more flexible to v1 2022-06-30 16:45:42 minimal: I agree when I see the options I see the reason why there is a v2, will try to switch to v2 now conteptualy it is working with v1, many thanks for your help 2022-06-30 16:46:39 blt[m]: re user-data, like network-config it sounds like it is Terraform use issues you are having (as to what/how it presents to cloud-init) rather than cloud-init issues 2022-06-30 16:48:42 blt[m]: you may perhaps also be interested in my script for creating cloud-init enabled Alpine servers: https://github.com/dermotbradley/create-alpine-disk-image 2022-06-30 16:54:54 minimal: oh very nice ! thanks for sharing 2022-06-30 18:22:54 How can I get the latest version of Alpine Linux on wsl-2 when I update my current version its still showing Alpine Linux v3.11 2022-06-30 18:28:44 mwoodpatrick: how did you upgrade it? 2022-06-30 18:33:23 apk add --upgrade apk-tools 2022-06-30 18:33:23 apk update 2022-06-30 18:33:23 apk upgrade 2022-06-30 18:33:41 did you change the repositories in /etc/apk/repositories? 2022-06-30 18:34:53 no, what should I set it to? 2022-06-30 18:36:05 change it from v3.11 to v3.16 2022-06-30 18:36:21 after doing that, you can run: apk upgrade --update --available 2022-06-30 18:36:25 or better, change it to latest-stable 2022-06-30 18:39:42 What should /etc/apk/repositories look like when setting it to latest-stable, what URL should I be looking to see where this isa documented 2022-06-30 18:41:37 Take the version section of the url and put latest-stable 2022-06-30 18:41:49 I found it by browsing the mirror via http 2022-06-30 18:42:39 https://dl-cdn.alpinelinux.org/alpine/latest-stable/main/ 2022-06-30 18:42:44 https://dl-cdn.alpinelinux.org/alpine/latest-stable/community 2022-06-30 18:43:47 minimal: how do you match macadress in v2 format? it seems to be mandatory, I can find a variable but I am wondering if it can be not use ? 2022-06-30 18:46:28 blt[m]: do you actually want to match on mac address? 2022-06-30 18:46:55 minimal: nope I am trying to match on name but I see "network_state.py[DEBUG]: NetworkState Version2: missing "macaddress" info in config entry" 2022-06-30 18:47:32 can you show the network-config you're using? 2022-06-30 18:50:47 minimal: https://tpaste.us/wYMJ just tried that which is failing as well, I tried with match and name just before 2022-06-30 18:52:10 I wonder if I have an issue on address, it seems that even for a single one it is using adresses instead of address 2022-06-30 18:52:33 it is "addresses", not "address" 2022-06-30 18:53:26 thats was the issue indeed 2022-06-30 18:53:41 is "eth13" the real interface name? 2022-06-30 18:53:48 yes 2022-06-30 18:54:08 so you don't normally need to use "match:" then 2022-06-30 18:54:15 all good, gatewway as well, only resolv.conf is blank trying to understand why 2022-06-30 18:54:43 I got ERROR: http://dl-cdn.alpinelinux.org/alpine/latest-stable/main: UNTRUSTED signature 2022-06-30 18:55:40 blt[m]: is manage_resolv_conf set to True in your user-data? 2022-06-30 18:55:59 hum it is not! let me test that 2022-06-30 18:56:09 blkt[m]: maybe try reading the docs? https://cloudinit.readthedocs.io/en/latest/topics/modules.html#resolv-conf 2022-06-30 18:56:59 thanks I haven's seen that there is a dedicated section for that I beter have a look to it before.. 2022-06-30 18:57:45 "manage_resolv_conf: (boolean) Whether to manage the resolv.conf file. resolv_conf block will be ignored unless this is set to true. Default: false" 2022-06-30 18:59:36 thanks for that, I am wondering if it needs to be on userdata for the network config to be apply or should I delete the section nameservers from network config and add the one to userdata? 2022-06-30 19:00:01 just set that in the user-data and leave the rest as it is 2022-06-30 19:01:36 mwoodpatrick: We changes keys in 3.15, and you don't have the latest keys yet 2022-06-30 19:01:46 btn[m]: well nameservers can be set in either place. Don't sure if user-data will override network-config for that 2022-06-30 19:01:56 mwoodpatrick: apk add -X http://dl-cdn.alpinelinux.org/alpine/v3.14/main -u alpine-keys 2022-06-30 19:03:56 minimal: interesting, I have now to understand why my user-data is not consumed globaly now network config is working, I am trying to find some hints in log 2022-06-30 19:05:48 not consumed globally? 2022-06-30 19:07:02 I am trying to create a user, adding pass, shell, ssh and some package, nothing is done, I guesss I have an issue on the format but so far no error found in cloud-init.log 2022-06-30 19:07:46 I am reading the doc, wondering if the -default under users: is necessary 2022-06-30 19:08:30 blt[m]: check /var/lib/cloud/instances//user-data.txt to see what cloud-init was passed 2022-06-30 19:09:20 MarkSpectre14:~# apk update 2022-06-30 19:09:20 fetch http://dl-cdn.alpinelinux.org/alpine/v3.16/main/x86_64/APKINDEX.tar.gz 2022-06-30 19:09:20 fetch http://dl-cdn.alpinelinux.org/alpine/v3.16/community/x86_64/APKINDEX.tar.gz 2022-06-30 19:09:20 v3.16.0-242-ge8a8292729 [http://dl-cdn.alpinelinux.org/alpine/v3.16/main] 2022-06-30 19:09:20 v3.16.0-240-g6e6a2bc8ec [http://dl-cdn.alpinelinux.org/alpine/v3.16/community] 2022-06-30 19:09:21 blt[m]: are you specifying a "users:" section in your user-data? 2022-06-30 19:09:22 OK: 17023 distinct packages available 2022-06-30 19:09:23 MarkSpectre14:~# apk upgrade 2022-06-30 19:09:27 OK: 8 MiB in 18 packages 2022-06-30 19:09:28 MarkSpectre14:~# cat /etc/os-release 2022-06-30 19:09:30 NAME="Alpine Linux" 2022-06-30 19:09:32 ID=alpine 2022-06-30 19:09:34 VERSION_ID=3.11.5 2022-06-30 19:09:36 PRETTY_NAME="Alpine Linux v3.11" 2022-06-30 19:09:38 HOME_URL="https://alpinelinux.org/" 2022-06-30 19:09:41 BUG_REPORT_URL="https://bugs.alpinelinux.org/" 2022-06-30 19:10:27 it looks like I was able to fetch the packages for 3.26 but my release file is still showing 3.11 what am I doing wrong 2022-06-30 19:10:28 minimal: I see it passed what I asked, let me sanitized and sending to you 2022-06-30 19:10:34 btn[m]: https://cloudinit.readthedocs.io/en/latest/topics/modules.html#users-and-groups 2022-06-30 19:10:51 blt[m]: "Omission of default as the first item in the users list skips creation the default user." 2022-06-30 19:11:47 minimal: arf thanks for that I was reading this one https://cloudinit.readthedocs.io/en/latest/topics/examples.html yours is much better 2022-06-30 19:12:37 blt[m]: the "example" page is full of..........examples, the modules page is full of defintive instructions for how each module behaves 2022-06-30 19:21:30 mwoodpatrick: apk upgrade --available 2022-06-30 19:25:37 I did that 2022-06-30 19:25:39 OK: 8 MiB in 18 packages 2022-06-30 19:26:06 cat /etc/apk/world 2022-06-30 19:26:20 Probably the package that updates that file is not installed 2022-06-30 19:26:27 but /etc/os-release still shows Alpine Linux v3.11 2022-06-30 19:28:05 cat /etc/apk/world 2022-06-30 19:28:05 alpine-baselayout 2022-06-30 19:28:05 alpine-keys 2022-06-30 19:28:05 apk-tools 2022-06-30 19:28:05 bash 2022-06-30 19:28:06 busybox 2022-06-30 19:28:08 libc-utils 2022-06-30 19:28:57 yeah, alpine-base is not installed 2022-06-30 19:29:00 which would update those 2022-06-30 19:29:18 you can either install it, but it comes with some extra dependencies, or you ignore it 2022-06-30 19:29:22 minimal: stupid me... I cant believe I forgot to put `#cloud-init` at the first line... it is way better now... 2022-06-30 19:31:39 So I added that package but release is still not updated 2022-06-30 19:31:41 cat /etc/apk/world 2022-06-30 19:31:41 apk-tools 2022-06-30 19:31:41 alpine-baselayout 2022-06-30 19:31:41 alpine-base 2022-06-30 19:31:41 alpine-keys 2022-06-30 19:31:42 bash 2022-06-30 19:31:43 busybox 2022-06-30 19:31:45 libc-utils 2022-06-30 19:33:41 is there /etc/os-release.apk-new ? 2022-06-30 19:33:44 minimal: btw and just for your info , if I just put the bolean to True but keeping the parametes into network I got 'cc_resolv_conf.py[WARNING]: manage_resolv_conf True but no parameters provided!' will try to put the parameters like the doc suggest and removing from network to test 2022-06-30 19:34:15 mwoodpatrick: mv /etc/alpine-release.apknew /etc/alpine-release 2022-06-30 19:34:24 mv /etc/os-release.apknew /etc/os-release 2022-06-30 19:34:57 s/apknew/apk-new 2022-06-30 19:35:05 oh, thanks :) 2022-06-30 19:35:21 missed that you already suggested that 2022-06-30 19:35:27 (: 2022-06-30 19:35:43 blt[m]: was forgetting, nameservers specified in network-config will appear in /etc/network/interfaces 2022-06-30 19:37:35 That worked but how come I had to do this manually shouldn't the package manager take care of this, where can I find more details on the steps necessary to do the update/ 2022-06-30 19:43:12 mwoodpatrick: in the future it will 2022-06-30 19:43:22 but apk will not overwrite files in /etc that it did not manage before or have changed 2022-06-30 19:43:31 so it will place .apk-new files there instead 2022-06-30 19:44:05 thought that --clean-protected already did that 2022-06-30 19:45:01 --clean-protected will just prevent it from creating those files iirc 2022-06-30 19:50:25 good to know and really appreciate your help, is this stuff documented somewhere 2022-06-30 19:54:39 minimal: wondering if there is the equivalent of sudo in the user declaration for us using Alpine with doas ? 2022-06-30 19:55:49 mwoodpatrick: I'm not sure where that specific detail is documented 2022-06-30 19:56:04 blt[m]: yes there is. Did you read README.Alpine in the cloud-init-doc package? As mentioned on screen when setup-cloud-init is run... 2022-06-30 19:56:21 mwoodpatrick: https://docs.alpinelinux.org/user-handbook/0.1a/Working/apk.html#_upgrading_packages 2022-06-30 19:56:41 minimal: I haven't but I will right now 😇 2022-06-30 19:57:02 apk manpages should have it 2022-06-30 19:57:16 blt[m]: I added doas support to the Alpine package of cloud-init. It is not in upstream cloud-init as it relies on a doas modification that only Alpine (so far) have made 2022-06-30 19:57:36 minimal: ha wonderful ! 2022-06-30 19:58:17 minimal: https://git.alpinelinux.org/aports/tree/community/cloud-init/README.Alpine am on it 🙂 2022-06-30 20:35:25 minimal: I realised that even if I ask for chrony to be installed it is in crashed status, I just added theses to cloud init :... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/AUYUuXoyRDTAtfFGlQxftJUj) 2022-06-30 20:36:25 s/crashed/stopped/, s///, s/// 2022-06-30 20:36:25 blt[m] meant to say: minimal: I realised that even if I ask for chrony to be installed it is in stopped status, I just added theses to cloud init :... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/AUYUuXoyRDTAtfFGlQxftJUj) 2022-06-30 20:51:57 blt[m]: "if I ask for chrony to be installed" - not sure what you mean 2022-06-30 20:55:22 minimal: I added a... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/cazPEbvHZNwCXcwfYsMYIBtA) 2022-06-30 21:01:45 blt[m]: Isn't it (full message at https://tpaste.us/rnJ1 ) 2022-06-30 21:04:17 ikke ;-) 2022-06-30 21:04:46 :) 2022-06-30 21:06:10 blt[m]: so cloud-init will try and setup chrony before it tries to install chrony...as that's what the cloud-init module ordering tells it to do 2022-06-30 21:06:39 I always have the NTP server pre-installed in any cloud-init Alpine image I create 2022-06-30 21:06:47 s/server/server software/ 2022-06-30 21:06:47 minimal meant to say: I always have the NTP server software pre-installed in any cloud-init Alpine image I create