2019-01-01 12:35:00 <jade135> "apk add virtualbox-additions-grsec" return error unsatisfiable constraints, added alpine testing repo but still get error. help?
2019-01-01 12:35:11 <jade135> (following https://wiki.alpinelinux.org/wiki/VirtualBox_guest_additions)
2019-01-01 12:37:33 <_ikke_> jade135: grsec has been renamed to hardened since 3.5 and abandoned since 3.8
2019-01-01 12:38:40 <jade135> thanks
2019-01-01 16:52:34 <jhumphrey> What are the most important differences (in your opinion) between slackware aand alpine linux, who should use which?
2019-01-01 17:02:55 <jhumphrey> What kind of people should use alpine linux?
2019-01-01 17:03:49 <dalias> well vs slackware, pretty much anyone who doesn't need glibc :) i <3 slackware -- it was the dist i first learned on decades ago -- but it has no package management and clean upgrade or system maintenance paths and very few packages
2019-01-01 17:05:03 <AinNero> i use alpine linux because im a terminal monkey
2019-01-01 17:05:11 <AinNero> musl seems somewhat more reasonable than glibc
2019-01-01 17:05:31 <AinNero> switched after the glibc 2.26 ucontext breakages
2019-01-01 17:07:11 <dalias> what was the ucontext breakage?
2019-01-01 17:07:45 <dalias> anyway if your interest in slackware is having a very light install without packages you don't need, alpine is really good for that
2019-01-01 17:08:43 <AinNero> i'd love to have some linux from scratch setup, but i usually only spend time designing the build system instead of doing it
2019-01-01 17:08:55 <AinNero> using musl and busybox, like alpine
2019-01-01 17:09:13 <jhumphrey> which is better in terms of speed?
2019-01-01 17:09:22 <AinNero> speed for what?
2019-01-01 17:11:10 <jhumphrey> I guess I mean which one _feels_ faster and lighter
2019-01-01 17:12:33 <AinNero> i dont know about slackware.
2019-01-01 17:16:26 <jhumphrey> The reason I ask is I am thinking about triple booting my computer (LinuxMint, slackware/alpine, openBSD)
2019-01-01 17:17:42 <jhumphrey> LinuxMint is for my guests.
2019-01-01 20:53:27 <aGaTHoS> how do i know which runlevels i have?
2019-01-01 20:58:03 <mps> aGaTHoS: rc-status -r
2019-01-01 21:22:38 <aGaTHoS> mps nop
2019-01-01 21:22:42 <aGaTHoS> just says default
2019-01-01 21:22:51 <aGaTHoS> but finally figured out howto
2019-01-01 21:22:58 <aGaTHoS> rc-status --list
2019-01-01 21:24:20 <aGaTHoS> does anyone what actually does klogd service? if i want /var/log/dmesg to be updated in real time have to put this in default or boot ?
2019-01-01 21:25:30 <aGaTHoS> knows*
2019-01-01 21:31:05 <mps> aGaTHoS: misunderstood you question, thought that you ask in which level you are. sorry
2019-01-01 21:31:43 <mps> of course, rc-status -l list available runlevels
2019-01-01 22:04:02 <Kettling> WRT "What kind of people should use alpine linux?" ... I got a free netbook (Atom N2600 + gma500 integrated VGA) and I was looking for something lean with 32bit support. Other distros seem determined to abandon support for intel Atom now.
2019-01-01 22:16:12 <dalias> hm? how would distros abandon support for atom?
2019-01-01 22:16:36 <dalias> it's not like it's some minimal isa subset
2019-01-01 22:16:57 <dalias> and fwiw atoms (all but the last generation or two) are the only semi-modern cpus immune to spectre
2019-01-01 22:29:55 <Kettling> As far as I know you can't run a 64bit distro on this hardware. I haven't really tried, memory is also tight.
2019-01-01 22:29:58 <kahiru> maybe Kettling meant distros dropping 32bit support?
2019-01-01 22:39:49 <dalias> some early atoms lack 64-bit support, but most have it
2019-01-01 22:40:08 <Kettling> I did a search, suposedly the N2600 can handle 64 bit instructions, but people have reported the BIOS won't boot it (haven't tried). Even then it's only got 1G of RAM. Might be interesting to try but I'm guessing something like CentOS-7 and Gnome desktop won't be pleasant.
2019-01-01 22:40:12 <dalias> i didn't remember this but the atom i always used in 32-bit mode (D510 iirc) actually had 64-bit support
2019-01-01 22:40:41 <dalias> "bios won't boot it" makes no sense. the bios never sees it go into 64-bit mode
2019-01-01 22:40:57 <dalias> bios would just hand off control to grub/lilo/syslinux/whatever
2019-01-01 22:41:11 <dalias> and they'd hand off control to the vmlinux entry point
2019-01-01 22:41:21 <dalias> which would then put the cpu in 64-bit mode
2019-01-01 22:41:41 <Kettling> I have another of these running 32 bit CentOS-6 with a gnome desktop but most of the gnome features ripped out. It's usable but I wouldn't recommend it, and I cannot upgrade because there is no 32 bit upgrade path. RedHat promise never again to support 32bit.
2019-01-01 22:41:52 <dalias> what "bios won't boot it" might mean is that the bios has some issues that prevent running 64-bit *windows* on it.
2019-01-01 22:42:19 <Kettling> Could be. I'll download a 64bit CDROM and try it myself. See what happens.
2019-01-01 22:43:04 <dalias> actually i remember reading the same about my D510 -- i think that me be why i never tried a 64-bit kernel on it before
2019-01-01 23:20:42 <Kettling> Our entire cloud services division runs exclusively on recycled netbooks. When the other virtual hosting providers say, "We got Spectre under control" can you really be sure? Maybe you can't afford to take that risk? Well, now you don't have to. Made from Atoms!
2019-01-01 23:23:45 <dalias> yeah N2600 should be safe
2019-01-01 23:23:59 <dalias> only avoton is out-of-order afaik
2019-01-01 23:26:06 <dalias> hm N2800 might even be a better choice than the S1260 i got
2019-01-01 23:26:33 <dalias> just because it's available on boards without all the datacenter backdoor crap
2019-01-01 23:27:13 <dalias> plus it has a gpu
2019-01-01 23:27:22 <dalias> downside is lack of ecc memory
2019-01-01 23:27:30 <mps> Kettling: how do you know that there is no yet uncovered bug in netbooks which could be security risk like spectre
2019-01-01 23:28:35 <dalias> mps, bugs are generally possible to work around. fundamental design flaws are usually not
2019-01-01 23:29:11 <dalias> spectre isn't an implementation bug. it's a fundamental design flaw in out-of-order cpu microarchitecture that affects _all_ such cpus
2019-01-01 23:29:27 <mps> of course, but we cannot be sure that atom doesn't have something similar
2019-01-01 23:30:16 <mps> you know how long time passed before spectre discovered
2019-01-01 23:31:07 <mps> and most of used these cpus thinking that they are secure
2019-01-01 23:31:50 <mps> ok, actually some of 'know' that the most cpus are unsecure
2019-01-01 23:31:53 <AinNero> maybe the mistake is running multi-tenacy on hardware not designed for multi-tenacy, but for performance
2019-01-01 23:32:14 <dalias> ainnero, *everyone* has multi-tenancy, even on desktop and mobile
2019-01-01 23:32:25 <dalias> you visit pages with somebody else's js
2019-01-01 23:32:39 <AinNero> IMHO also a bad idea
2019-01-01 23:32:43 <dalias> you run apps in isolated per-app privilege domains
2019-01-01 23:33:37 <dalias> ooh this looks nice, too bad it's discontinued: https://www.logicsupply.com/dn2800mt/
2019-01-01 23:34:37 <dalias> funny thing about my atom s1260 board...
2019-01-01 23:34:48 <dalias> it's got this awesome atom with like 15W TDP...
2019-01-01 23:37:54 <dalias> then
2019-01-01 23:37:56 <dalias> https://pbs.twimg.com/media/DUZjyzaX0AAoXKN.jpg:orig
2019-01-01 23:38:57 <dalias> shitty backdoor management chips putting out 3-10x as much heat as the cpu
2019-01-01 23:39:29 <AinNero> reference 122F = 50C, 100F = 37C
2019-01-01 23:39:30 <dalias> totally defeating the intended purpose of being economical for cloud services
2019-01-01 23:40:35 <Kettling> For the record ... the CentOS-7 CDROM does boot into 64bit rescue mode so the hardware does support 64 bit Linux. That means at least going for the 64 bit install might be an option. The 1G of RAM can be upgraded to 2G if I was skillful enough to pop the keyboard out and then get it back in again so it still feels like a keyboard afterwards.
2019-01-01 23:40:35 <dalias> you're literally paying several times as much to have a useless management interface as to run the actual workload
2019-01-01 23:41:13 <AinNero> is the NEEDED libc.* entry relevant for something? it seems that alpine uses nonstandard values there
2019-01-01 23:41:31 <AinNero> LDSO path is still the same
2019-01-01 23:42:18 <dalias> it's not used at all by musl as long as the part before the . is libc
2019-01-01 23:42:32 <dalias> however, alpine uses it for tracking package dependency info
2019-01-01 23:43:19 <AinNero> mh, makes sense
2019-01-01 23:43:33 <AinNero> but not obvious from the main/musl history
2019-01-01 23:44:47 <dalias> this doesn't look bad: https://www.amazon.com/Supermicro-X9SCAA-Motherboard-Mini-ITX-Mini-PCIe/dp/B00CGW3FZ2
2019-01-01 23:57:17 <psiconauta> hey dudes :)
2019-01-02 00:56:38 <Kettling> Re "useless management interface" ... the vendors feel the need to come up with something proprietary so they can deliver yet another incompatible protocol to a world that has too many of those already.
2019-01-02 04:51:50 <Kettling> What's the best way to dual boot both 32bit and 64bit Alpine on the same SATA drive? Can they share a boot partition or would that be a bad idea? Is LVM worth the trouble??
2019-01-02 05:10:13 <daurnimator> 
ACTION returns.

2019-01-02 05:10:25 <daurnimator> Anyone able to help out with ambient capabilities?
2019-01-02 05:20:18 <dalias> kettling, why?
2019-01-02 05:41:38 <Kettling> Why? I dunno, just can't make up my mind which to install and I guess I might want to flip back and forth between them.
2019-01-02 05:42:27 <Kettling> Hmmm, can I share a root filesystem? Does Alpine support dual libraries on everything?
2019-01-02 05:48:56 <dalias> no
2019-01-02 05:49:09 <dalias> you can of course share a /home or something if you want
2019-01-02 05:49:42 <dalias> but alpine doesn't support multiarch the way debian does
2019-01-02 05:51:27 <Kettling> That's ok, I have heaps of disk space, dual boot is fine for trying it out. See if there's any significant difference.
2019-01-02 05:51:53 <Kettling> Sharing swap partition should work I guess.
2019-01-02 05:53:37 <dalias> might have to be re-mkswap'd each time
2019-01-02 05:54:02 <dalias> i suspect swap structures depend on kernel's wordsize and endianness etc.
2019-01-02 07:10:50 <qman__> Kettling: I'd suggest just going 64-bit, 32-bit is missing a lot of packages that 64-bit has and there's not really any pressing reason to use it other than hardware limitations
2019-01-02 07:50:31 <Kettling> Is btrfs a good idea for installing? Does anyone actually use it? There's an option available ...
2019-01-02 08:16:28 <mepholic> probably not a good idea
2019-01-02 08:16:45 <mepholic> BTRFS has a history of losing data IME
2019-01-02 08:43:23 <Lord> not really
2019-01-02 09:02:01 <kahiru> on the other hand, zstd compression at the filesystem level is pretty neat
2019-01-02 09:23:44 <Kettling> I think the idea of writable snapshots is amazing. You could upgrade your system while it is running and then pivot everything smoothly into the upgrade ... but if it doesn't work, you can go right back again!
2019-01-02 09:25:34 <Kettling> I done work on remote routers three day's drive from anywhere and trying to upgrade over the network with minimal downtime. Every single time you reboot one of those things you sit there never expecting it to come back.
2019-01-02 09:40:53 <kahiru> I feel you
2019-01-02 09:58:10 <TBB> how does that differ from xfs freezes, lvm snapshots or, to pick a more exotic example, nilfs?
2019-01-02 10:27:57 <kahiru> btrfs snapshots should be cheap compared to lvm snapshots
2019-01-02 10:28:23 <AinNero> is btrfs still eating data?
2019-01-02 10:33:19 <Kettling> This netbook with Alpine is pretty darn snappy! Only problem is the battery seems stuck at 0% charge. The voltage has gone up a tiny bit. If I can't charge the battery that would spoil my fun a lot.
2019-01-02 11:09:30 <silverwillow> Good morning all, merry 2019!!
2019-01-02 11:10:50 <silverwillow> I'm seeing extremely slow MariaDB operation with ASCII. Wordpress CSV importer plug-in takes an enormous time to import a measly 50 records.
2019-01-02 11:11:23 <silverwillow> on Devuan it took a great deal less time. (no exactl numbers, but i didn;t need to go get coffee...)
2019-01-02 11:12:38 <silverwillow> is this a known issue? any fixes or pointers to info that will help me fix this?
2019-01-02 15:32:40 <_ikke_> Working on week 2 from cis194. I have a 'IO [LogMessages]' value. Is there an easy way in ghci to use that in something that just expects [LogMessages]?
2019-01-02 15:33:41 <_ikke_> (wrong chat)
2019-01-02 16:23:30 <danieli> iirc that's an introduction to haskell course @ _ikke_
2019-01-02 16:23:43 <danieli> yup, it is :) i had to check
2019-01-02 16:24:24 <_ikke_> it is indeed
2019-01-02 16:24:37 <_ikke_> dabbling in some haskell again
2019-01-02 16:25:05 <danieli> i haven't written a lot of haskell at all recently, it's a beautiful language, but fairly useless to me
2019-01-02 16:41:24 <ddevault> that's my review of haskell as well: an interesting language with which it is impossible to write useful programs
2019-01-02 16:43:04 <mps> ddevault: :D
2019-01-02 16:53:29 <danieli> ddevault: that isn't true, but it has a couple traits that need extra consideration
2019-01-02 16:53:43 <danieli> the purity can make it hard to be memory-efficient if you aren't used to haskell, for instance
2019-01-02 16:53:46 <danieli> but that's off topic :p
2019-01-03 17:30:30 <sy> can anyone recommend some IPS package for my alpine boxen?
2019-01-03 17:36:50 <Lord> IPS ?
2019-01-03 17:40:01 <sy> oops, IDS (intrusion detection systems)
2019-01-03 17:40:17 <sy> like systraq
2019-01-03 19:11:33 <Kanarok> hey. I'm on a Raspberry Pi with Alpine linux. So far I managed to get the otg working. But I cannot login after sshd setup getting some weird error that my password is expired and a PTY allocation request failed on channel 0. Any ideas?
2019-01-03 19:19:57 <danieli> did you run passwd on it before trying to ssh in?
2019-01-03 19:24:06 <Kanarok> yes but only for the root user and not the new user I created
2019-01-03 19:27:05 <Kanarok> I tried and failed. Might be more related to "PTY allocation request failed on channel 0"
2019-01-03 19:28:36 <danieli> root login is disabled by default (ssh), if that's what you're trying
2019-01-03 19:28:56 <danieli> what's the ssh command you're running?
2019-01-03 19:32:04 <Kanarok> I tried to enable it and also created a different user to try. I tried ssh sysop@ip with -t with -T and also with "/bin/ash" as last parameter
2019-01-03 19:33:16 <Kanarok> I also tried PermitTTY yes in sshd_config on the server side
2019-01-03 19:33:28 <danieli> what do the sshd logs say?
2019-01-03 19:37:23 <Kanarok> well I only found /var/log/messages telling me, that either root or sysop password has expired (root forced)
2019-01-03 22:39:40 <Kettling> Kanarok, there's a special mount on /dev/pts so make sure that looks good. Try "ls -l /dev/pts" and you should see some character devices in there.
2019-01-03 22:42:43 <Kettling> Not sure about the password expired problem. I have noticed that Alpine is sensitive to clock skew at boot if your hardware clock is set wrong but maybe that's unrelated. The standard Raspberry Pi has no hardware clock so I guess it must have a way to initialize that from the network??
2019-01-03 22:46:49 <AinNero> it has
2019-01-04 03:46:18 <bleb> does apk try to match the interface of freebsd's pkg?
2019-01-04 03:46:47 <bleb> i notice from this old old website that it was once "more or less a simplified, busybox-compatible clone of the FreeBSD pkgtools suite"
2019-01-04 03:46:51 <bleb> http://apk-tools.sourceforge.net/
2019-01-04 03:47:17 <bleb> and i notice that apk_add became apk add just as pkg_add became pkg add
2019-01-04 03:47:58 <bleb> have these tools diverged or is there still some effort to keep the interfaces similar?
2019-01-04 04:04:11 <Kettling> I'm really happy the way this Atom netbook is running on Alpine. It's dual booting 32 bit and 64 bit, and the wifi works and the GUI is pretty fast and smooth (not 3D accel) and gets about 4-5 hours of battery. Only strange thing is that if you suspend the wifi lamp goes off, but waking up from suspend the wifi lamp stays off (although wifi keeps working) ... I gues no one cares enough to stress over these things.
2019-01-04 04:05:22 <Kettling> Only remaining thing, I would like to get some heavyweight browser running, like palemoon perhaps. Might have a go at the unofficial alpine-pkg-glibc
2019-01-04 04:07:10 <dalias> palemoon maintainer is really hostile
2019-01-04 04:08:17 <danieli> bleb: that's a very, very old site
2019-01-04 06:25:24 <daurnimator> When is 3.9 due?
2019-01-04 06:25:52 <daurnimator> ncopa: could you bump the libcap package?
2019-01-04 07:16:14 <ncopa> daurnimator: pushed libcap. thanks
2019-01-04 07:16:33 <ncopa> i think we will have 3.9 rc1 within a few days
2019-01-04 09:05:26 <daurnimator> ncopa: great, thanks :)
2019-01-04 09:08:28 <kahiru> ooh, exciting. looking forward to it
2019-01-04 11:13:49 <Kettling> "palemoon maintainer is really hostile" => I searched on what has been said about him, oh! people are so cruel. I tried it with Sasha Gerrand's glibc-compat and what I found was that although the palemoon download has heaps of libraries included when you unpack it, does not have libstdc++ so glibc is insufficient.
2019-01-04 11:14:22 <Kettling> To be honest, I don't quite understand... why not a static link for something like that?
2019-01-04 11:15:18 <Kettling> It's a behamoth anyway, and the user expects this so just so brutally monolithic and deliver a static link... *sigh*
2019-01-04 11:15:57 <Kettling> I'll never understand software designers.
2019-01-04 11:39:20 <clandmeter> kristoferus75, https://git.alpinelinux.org/aports/tree/main/linux-vanilla
2019-01-04 11:39:25 <clandmeter> these are our build scripts
2019-01-04 11:39:38 <clandmeter> you can find more info on our wiki how to setup build env
2019-01-04 11:40:39 <kristoferus75> ok thanks !
2019-01-04 12:00:46 <kristoferus75> clandmeter Please send me the link where i can found the build env in the wiki !
2019-01-04 13:28:10 <asyd> morning folks. I have a question about timezone. I follow https://wiki.alpinelinux.org/wiki/Setting_the_timezone which works as far as I don't remove tzdata package
2019-01-04 13:29:02 <AinNero> yes. thats the expected behavior.
2019-01-04 13:30:21 <asyd> https://dpaste.de/UOm0
2019-01-04 13:30:44 <asyd> ah, because the wiki say "You can now remove the other timezones"
2019-01-04 13:31:33 <_ikke_> did you actually copy the specific timezone file?
2019-01-04 13:31:52 <asyd> yes
2019-01-04 21:11:31 <rfleming> Greetings!  How do I tell a service to wait for something else?
2019-01-04 21:11:58 <rfleming> My problem is that Docker uses cgroups, but when Docker starts, the cgroups have yet to mount
2019-01-04 21:41:31 <rfleming> OK, looks like it's the cgroups service itself that throws these errors when docker starts.
2019-01-04 22:40:22 <rfleming> problem solved! https://wiki.alpinelinux.org/wiki/Docker says to create cgroup in fstab, but there's a cgroups service that mounts cgroup as well.  Removed the fstab entry.
2019-01-05 01:01:06 <Kettling> Is there a trick to activating Javascript on the Midori browser? I can get pages to load but all JS activity is dead, and also the "Inspect Element" debug menu option is broken.
2019-01-05 01:01:32 <Kettling> Maybe some extra package is needed? Are other people using it?
2019-01-05 01:05:24 <Kettling> Another minor problem is that the cooling fan on the netbook does not restart after a suspend (fan stays off and CPU heats up).
2019-01-05 01:05:38 <Kettling> Is there a way to kickstart the cooling fan from software?
2019-01-05 01:23:07 <dalias> midori has js?
2019-01-05 01:23:18 <dalias> last i checked it was completely useless
2019-01-05 03:48:32 <Kettling> Midori claims to have JS haven't tried it on any other distro to confirm that. The general layout and rendering is nice.
2019-01-05 04:38:43 <tomato> I'm around 100% sure that midori js
2019-01-05 04:38:48 <tomato> *has js
2019-01-05 04:51:50 <Processus42> Hello there. What is the licensing for the aports repository ? I'd like to modify some of the APKBUILD files for personalisation purpose.
2019-01-05 05:17:31 <qman__> midori absolutely has js, but it does have limits and doesn't work properly on all websites
2019-01-05 05:43:36 <dalias> what is the alpine openssl/libressl situation right now?
2019-01-05 05:43:46 <dalias> hitting weird conflicts trying to upgrade some stuff and i wonder if it changed
2019-01-05 05:48:31 <qman__> migrating back to openssl from libressl due to compatibility issues and removed features with libressl
2019-01-05 05:49:14 <dalias> ah
2019-01-05 05:49:23 <dalias> so how to get back in a non-conflicting state?
2019-01-05 05:49:42 <dalias> is it just the -dev packages that conflict?
2019-01-05 05:49:57 <dalias> maybe i need to remove libressl-dev
2019-01-05 05:50:02 <qman__> I haven't run into any conflicts, so I suspect it is the dev packages
2019-01-05 05:51:02 <dalias> i got lots of packages stuck because i had qt5-qtbase-dev held back, in turn holding qt5-qtbase back
2019-01-05 05:51:12 <dalias> because new qtbase-dev depends on openssl-dev and i have libressl-dev
2019-01-05 05:52:14 <danieli> dalias: do `apk fix` if the issue is about libsomething
2019-01-05 05:52:23 <danieli> we moved from libressl to openssl
2019-01-05 06:53:27 <Processus42> I see the "expression" ${parameter/pattern} being used in the mkimage.sh script. It looks like it's doing the same thing than a # substitution, but I don't know about this syntax... Am I missing something ?
2019-01-05 08:25:36 <Processus42> Wait a minute, is this thing POSIX ?
2019-01-05 08:36:19 <mikroskeem> use shellcheck
2019-01-05 08:36:27 <mikroskeem> it'll tell you
2019-01-05 08:37:48 <Processus42> shellcheck.net ?
2019-01-05 08:40:30 <Processus42> It says it's not POSIX... Why is it in a script ran by /bin/sh ? I though sh was strictly POSIX.
2019-01-05 10:34:58 <wej> Hi! The 3proxy package on alpine is a little outdated and that version has a nasty bug preventing the IPv6 support to work properly. this is fixed in the latest 0.8.12 release. it would be awesome if the alpine package would be updated :)
2019-01-05 10:37:32 <wej> it looks like increasing the version number in the APKBUILD is enough, the package builds fine then. although abuild complains with ">>> WARNING: 3proxy: 3proxy.initd is not an openrc #!/sbin/openrc-run" that script uses #!/sbin/runscript instead. not sure what's up with that
2019-01-05 15:23:30 <milek7> hi
2019-01-05 15:23:33 <milek7> how to install apk from file in diskless mode?
2019-01-05 15:23:42 <milek7> normally it isn't preserved after reboot
2019-01-06 01:58:20 <electricityZZZZ> so i am doing something a little weird--i am installing alpine linux in a virtual box machine to use as a desktop environment. so far everything has gone brilliantly with the exception of sound... can i get a bix of help figuring out why alsa is telling me why i don't have a soundcard? i've tried setting virtualbox both to the intel HD audio and the ICH 97 sound driver
2019-01-06 02:02:58 <electricityZZZZ> SpaceToast: bat signal?
2019-01-06 02:03:09 <SpaceToast> \o/ alright, let's see
2019-01-06 02:03:22 <SpaceToast> are you using the -virt kernel?
2019-01-06 02:03:44 <electricityZZZZ> so i just ran amixer, and i get the error Mixer attach default error: no such file or directory
2019-01-06 02:04:04 <electricityZZZZ> dmesg reports the Intel HD audio device being present (that's the one i am using at the moment)
2019-01-06 02:04:11 <electricityZZZZ> i think that this uses the -virt kernel, i'll double check that
2019-01-06 02:04:48 <SpaceToast> it's possible -virt doesn't have sound drivers (like... at all), it's been a while since I checked, but -virt is generally *not* intended for "full desktop" use, iirc
2019-01-06 02:05:04 <electricityZZZZ> yeah, it's the -virt kernel. most recent alpine linux virt install (i think 3.8.2?)
2019-01-06 02:05:18 <SpaceToast> aha, can you try migrating to the vanilla kernel and rebooting?
2019-01-06 02:05:30 <electricityZZZZ> so i have one hesitation with that
2019-01-06 02:05:33 <SpaceToast> (just the kernel, it'll make things easier to debug and might be a solution right off)
2019-01-06 02:05:54 <electricityZZZZ> which is that generally i have been having trouble with running linux in a virtualbox machine insofar as it tends to be power hungry
2019-01-06 02:06:10 <electricityZZZZ> somehow stock freebsd and alpine linux-virt run cool as a cucumber
2019-01-06 02:06:23 <SpaceToast> virtualization in general is power hungry, but perhaps vbox manages to do some paravirt
2019-01-06 02:06:24 <electricityZZZZ> it is possible that the non-virt also runs cool, i haven't checked that one
2019-01-06 02:06:35 <electricityZZZZ> and tlp powertop etc doesn't make much of a difference for the power consumption
2019-01-06 02:06:50 <SpaceToast> still, this *is* for debugging :) if that immediately solves the issue, you could see if it runs cool or not, and if not you can just find the option in question and build your own kernel
2019-01-06 02:07:03 <SpaceToast> before we do that though...
2019-01-06 02:07:14 <SpaceToast> is /proc/asound/cards a file that exists?
2019-01-06 02:07:18 <SpaceToast> and if yes, what's in it?
2019-01-06 02:07:43 <electricityZZZZ> there is no asound folder in proc
2019-01-06 02:08:18 <SpaceToast> uh
2019-01-06 02:08:31 <electricityZZZZ> ubuntu 18 runs like crap in virtualbox (laggy etc) and alpine runs beautifully, except for this sound issue.
2019-01-06 02:08:48 <SpaceToast> what's `lsmod | grep snd` print out?
2019-01-06 02:08:51 <SpaceToast> (as root)
2019-01-06 02:08:54 <SpaceToast> (just the first one)
2019-01-06 02:09:16 <electricityZZZZ> lsmod as root reports nothing as snd
2019-01-06 02:09:33 <SpaceToast> sudo modprobe snd
2019-01-06 02:09:35 <SpaceToast> :)
2019-01-06 02:09:36 <electricityZZZZ> i have followed the instructions from here: https://wiki.alpinelinux.org/wiki/Sound_Setup
2019-01-06 02:09:59 <SpaceToast> (and then repeat)
2019-01-06 02:10:03 <SpaceToast> (the asound bit)
2019-01-06 02:10:04 <electricityZZZZ> modprobe FATAL not found
2019-01-06 02:10:19 <electricityZZZZ> i guess that meaks the kernel has no sound?
2019-01-06 02:10:26 <electricityZZZZ> means, not meaks.
2019-01-06 02:10:27 <SpaceToast> yup, you're gonna need -vanilla
2019-01-06 02:10:35 <SpaceToast> do try it out and see if it runs smoother though
2019-01-06 02:10:46 <SpaceToast> musl is very good at things like caching that glibc fails at, so there's a good chance it's pretty cool :)
2019-01-06 02:11:08 <electricityZZZZ> can you link me to instructions for how to install the vanilla kernel?
2019-01-06 02:11:33 <SpaceToast> apk add linux-vanilla
2019-01-06 02:11:40 <SpaceToast> that *should* be it (it'll remove -virt)
2019-01-06 02:11:42 <SpaceToast> and then a reboot
2019-01-06 02:11:51 <electricityZZZZ> wow and then i pick at the menu at start?
2019-01-06 02:12:22 <electricityZZZZ> ugh it is installing yamaha and sun and rockchip firmware...? bloatttt
2019-01-06 02:12:39 <SpaceToast> the two conflict, but installing a new kernel should proc your bootloader to re-detect stuff
2019-01-06 02:13:16 <electricityZZZZ> this is installing 80 ... things... for different architectures etc,... that's a bummer
2019-01-06 02:13:25 <SpaceToast> I suspect there's an install_if at play here
2019-01-06 02:13:35 <SpaceToast> anyway, just checked, -virt does indeed not enable CONFIG_SOUND
2019-01-06 02:13:48 <electricityZZZZ> so... recompile -virt is the "real" solution?
2019-01-06 02:14:05 <SpaceToast> no, modifying the configuration for virt and maintaining your own kernel port
2019-01-06 02:14:09 <SpaceToast> but just running -vanilla is fine :)
2019-01-06 02:14:14 <SpaceToast> anyway, let me see what's up with firmware
2019-01-06 02:14:33 <SpaceToast> what's going on is -vanilla depends on linux-firmware-any, and the default provider that is.. all of them
2019-01-06 02:14:34 <electricityZZZZ> wow i havent thought about 3com in a long time
2019-01-06 02:14:55 <electricityZZZZ> all these defunct companies
2019-01-06 02:15:30 <SpaceToast> electricityZZZZ: you can install linux-firmware-none
2019-01-06 02:15:38 <SpaceToast> to explain that you don't actually want any firmware
2019-01-06 02:15:47 <electricityZZZZ> heh, well it's a little late for that now
2019-01-06 02:15:57 <SpaceToast> (it's just full of nothing that "totally provides linux-firmware-any")
2019-01-06 02:16:03 <SpaceToast> well you can do that afterwards to clean up ;)
2019-01-06 02:16:52 <electricityZZZZ> ok i am rebooting,...
2019-01-06 02:17:32 <electricityZZZZ> ok lsmod now reports sound
2019-01-06 02:17:39 <SpaceToast> how about alsamixer?
2019-01-06 02:18:29 <electricityZZZZ> it ran
2019-01-06 02:18:35 <electricityZZZZ> but... startx broke now... ???
2019-01-06 02:18:51 <electricityZZZZ> oh, no,... it just took a strangely large amount of time to start
2019-01-06 02:19:00 <SpaceToast> :) enjoy
2019-01-06 02:19:19 <electricityZZZZ> works brilliantly
2019-01-06 02:19:20 <SpaceToast> if you really really want the -virt kernel, you'll need to build CONFIG_SOUND out-of-tree (not recommended), or run your own config file
2019-01-06 02:19:36 <SpaceToast> (but it's generally recommended to only use the -virt kernel for things like KVM servers and such)
2019-01-06 02:20:14 <electricityZZZZ> so if i want things to be clean maybe i should just install stock alpine?
2019-01-06 02:20:33 <SpaceToast> well, you're still bypassing quite a few things I'd imagine (e.g openrc)
2019-01-06 02:20:45 <SpaceToast> but generally, yeah, if you want a full like-desktop experience, you typically install the full-desktop version
2019-01-06 02:20:52 <SpaceToast> but if you're happy with what you've got that's what matters :)
2019-01-06 02:21:30 <electricityZZZZ> so... i was also having weird problems with sway
2019-01-06 02:21:39 <electricityZZZZ> trying to get it to work,... it was complaining about DRM
2019-01-06 02:22:40 <SpaceToast> gpu acceleration is a hard thing to virtualize - *especially* in virtualbox
2019-01-06 02:22:48 <SpaceToast> anyway, I'm gonna go get into bed \o
2019-01-06 02:23:18 <electricityZZZZ> ok you think the DRM complains were related to the fact that it is -virt, or running in a virtualbox?
2019-01-06 02:24:27 <electricityZZZZ> ok well thanks bat-SpaceToast
2019-01-06 04:16:09 <dalias> uhg why does updating linux-kernel force update of linux-firmware-* (slow and huge)
2019-01-06 14:15:57 <Processus42> Can I set any value to DEBUG envvar before running abuild ? Or to I have to specify a specific value ? (I'm not aware of how it could be interpreted by a software started by abuild)
2019-01-06 15:48:43 <Processus42> I'd want to build apk (via abuild) with the debug symbols but... It seems setting DEBUG to an arbitrary value is not working. (gdb says there is no debugging symbols)
2019-01-06 18:32:31 <daggs1> Greetings, is there any fix for the 1 minute delay for chronyd at edge?
2019-01-06 19:06:55 <mps> daggs1: maybe set 'offline' mode
2019-01-06 19:07:08 <TBB> Processus42: you could just specifg a -dbg subpackage for the abuild; of course if the software in question needs specific options to build with debug symbols you might have to make the necessary changes to the build() hook
2019-01-06 19:07:48 <daggs1> mps: offline e.g. no ntp at all?
2019-01-06 19:09:22 <mps> no, start in offline, i.e. doesn't try to connect to ntp server. and after some time set 'online' via chronyc (scripted) when network is available
2019-01-06 19:10:15 <mps> man page says all
2019-01-06 19:11:18 <daggs1> from the logs, I see that the system gets ip before running chronyd, am I wrong
2019-01-06 19:11:38 <daggs1> ?
2019-01-06 19:11:52 <mps> then you shouldn't have delay
2019-01-06 19:13:00 <daggs1> is there any rc log?
2019-01-06 19:13:09 <daggs1> where I can verify?
2019-01-06 19:13:17 <mps> it works in my cases as expected, delays when there is no net, and when the net is connected it have really small delay (it must wait for answer)
2019-01-06 19:15:57 <daggs1> assuming the boot logs are serial, I get ip before calling chronyd
2019-01-06 19:18:08 <mps> daggs1: look at /etc/init.d/chronyd, depend list. you will see that it needs these started before it starts
2019-01-06 19:19:11 <daggs1> sec
2019-01-06 19:19:52 <daggs1> need net, after firewall
2019-01-06 19:21:34 <mps> yes
2019-01-06 19:21:57 <daggs1> so net is up before
2019-01-06 19:22:41 <mps> right
2019-01-06 19:23:30 <daggs1> so your suggested fix won't work probably? maybe offline it and start it only when ping to that address works?
2019-01-06 19:26:54 <mps> hard to say because I'm not at your premise and can't see and 'feel' what actually happens. You have to debug it and post results and state you find, best to bugs.a.o
2019-01-06 19:37:34 <daggs1> well, not a chronyd issue, removed it and still stuck
2019-01-06 19:46:12 <daggs1> found it, sshd
2019-01-06 19:46:30 <daggs1> chronyd has a small delay, few seconds, sshd is the issue
2019-01-06 19:52:54 <daggs1> replaced openssh with dropbear and all works great
2019-01-06 19:57:12 <mps> nice
2019-01-06 20:00:50 <daggs1> no idea what was the issue but now it works, all I'm missing is the host hotplug and I'm all ready to go.
2019-01-06 20:01:09 <daggs1> no idea why it happeend
2019-01-06 21:28:22 <pheoxy> hey guys, just trying alpine with docker on x86_64 and after starting docker service it just crashes
2019-01-06 21:28:35 <pheoxy> how can i debug this?
2019-01-06 21:32:55 <_ikke_> pheoxy: try to start docker manually
2019-01-06 21:38:52 <pheoxy> did that and tryed zapping it to stop it and start again and even rebooted still crashes
2019-01-06 21:41:08 <pheoxy> "$ docker info" just asks if daemon is running
2019-01-07 00:17:06 <Processus42> @TBB It's for apk-tools. And I couldn't find a -dbg sub-package.
2019-01-08 07:53:47 <sphrak> hey, im trying to locate postgis and postgis-dev pkgs.. but I cannot find it. was it dropped? it used to be in testing but now its gone?
2019-01-08 07:56:51 <sphrak> nvm found it nuked here: https://github.com/alpinelinux/aports/commit/b8121970ccaa432fd01cd566e5dbe67273b80fd8
2019-01-08 09:16:36 <muhaha> ola. I am looking for smtp client which can send email with specifying smtp host + credentials from inline command.
2019-01-08 11:37:56 <ncopa> muhaha: maybe ssmtp?
2019-01-08 11:40:09 <muhaha> @ncopa looking here https://www.systutorials.com/docs/linux/man/8-ssmtp/, but there is no argument for smpt host
2019-01-08 11:40:30 <muhaha> seems that heirloom-mailx handle this
2019-01-08 13:53:21 <mistnim> hello, what is firefox situation? Isn't the firefox esr version on alpine 3.8 not mantained?
2019-01-08 19:59:48 <sincere_fox> Guise,
2019-01-08 20:00:05 <sincere_fox> I havent been able to use xorg on virtualbox
2019-01-08 20:00:33 <sincere_fox> I am following this tutorial https://wiki.alpinelinux.org/wiki/XFCE_Setup
2019-01-09 09:46:47 <muhaha> Any idea? https://pastebin.com/9FWB4KPq mailx and moreutils dont go togheter... seems there is a problem openssl 1.0 vs 1.1
2019-01-09 09:50:19 <clandmeter> muhaha, let me check
2019-01-09 09:52:11 <clandmeter> muhaha, works for me
2019-01-09 09:52:29 <clandmeter> did you apk -U upgrade -a?
2019-01-09 09:54:04 <muhaha> no... 1. add repo, 2. update, 3. install mailx and moreutils
2019-01-09 09:54:23 <_ikke_> what repos do you have?
2019-01-09 09:57:31 <muhaha> Its docker image, so default
2019-01-09 09:58:42 <muhaha> https://pastebin.com/2JEMsExK
2019-01-09 10:07:01 <muhaha> It does not work with community repo
2019-01-09 10:09:13 <muhaha> It is so wierd, I am able to install heirloom-maix from main repo....
2019-01-09 10:10:13 <muhaha> But according to https://pkgs.alpinelinux.org/package/edge/community/x86_64/heirloom-mailx its in community repo.  If I add comunnity repo, error shows up....
2019-01-09 10:11:21 <muhaha> This is working apk add --update heirloom-mailx, this is not working apk add --update heirloom-mailx moreutils, of course moreutils is also from community repo
2019-01-09 10:24:20 <clandmeter> its in community
2019-01-09 10:28:07 <clandmeter> muhaha, you need both main and community repos
2019-01-09 10:32:56 <muhaha> jesus
2019-01-09 10:33:02 <muhaha> Do you have a docker?
2019-01-09 10:33:49 <muhaha> Then run this: FROM alpine:3.8 RUN apk add --update heirloom-mailx and this FROM alpine:3.8 RUN apk add --update moreutils,
2019-01-09 10:34:11 <muhaha> both packages are in community repo, but.... moreutils can not be installed..., but mailx can be.. So tell me
2019-01-09 11:03:24 <unmy> muhaha, you trying use main 3.8 together with community edge?
2019-01-09 11:12:10 <unmy> clandmeter, he already had community repository 3.8 but dunno why added another community but from edge
2019-01-09 11:14:49 <muhaha> oh I see
2019-01-09 11:14:54 <muhaha> my mistake, sorry
2019-01-09 11:15:05 <muhaha> moreutils is only in edge
2019-01-09 12:37:05 <franks2> Hi, how can i run and install steam on alpine?
2019-01-09 12:37:59 <_ikke_> not sure if that's going to be trivial, steam is built against glibc, while alpine runs on musl
2019-01-09 12:51:49 <danieli> correct
2019-01-09 13:41:38 <TBB> you probably could run it as a snap or a flatpak
2019-01-09 14:47:38 <fjear> I seem to be unable to find which utlities package telnet is in?
2019-01-09 14:48:08 <mps> fjear: busybox-extra
2019-01-09 14:48:33 <fjear> missing
2019-01-09 14:48:53 <fjear> ah s
2019-01-09 14:51:10 <mps> extras, ofc
2019-01-09 15:15:00 <TechArch> Hello
2019-01-09 15:15:30 <TechArch> Does anyone happen to know any workarounds for installing K8s on Alpine Linux ARM?
2019-01-09 15:15:38 <TechArch> Thanks a lot :)
2019-01-09 16:05:52 <aGaTHoS> can any developer considere make a package for picprog and gputils? http://hyvatti.fi/~jaakko/pic/picprog.html https://gputils.sourceforge.io/
2019-01-09 16:07:09 <aGaTHoS> also maybe gpsim gpsim.sourceforge.net
2019-01-09 16:11:45 <aGaTHoS> and sdcc http://sdcc.sourceforge.net/
2019-01-09 16:12:03 <aGaTHoS> all are tools for developing with microchip microcontrollers
2019-01-09 16:12:17 <aGaTHoS> which i need
2019-01-09 16:27:34 <fcolista> aGaTHoS, fill a request on bugs.alpinelinux.org
2019-01-09 17:48:16 <aGaTHoS> done
2019-01-09 20:25:52 <jvoisin> https://wiki.alpinelinux.org/wiki/Install_Alpine_in_Qemu I guess that {{{1}}} isn't the actual command to run ;)
2019-01-10 08:07:59 <ncopa> jvoisin: its a wiki ;)
2019-01-10 08:08:07 <ncopa> i have fixed it. thanks!
2019-01-10 09:03:11 <jvoisin> (I wasn't sure about what to put in there, hence why I didn't fixed it myself ♥)
2019-01-10 09:06:40 <ncopa> i didnt know either, so i had to google it :)
2019-01-10 11:25:16 <jvoisin> any pointers on how to investigate https://github.com/nbs-system/snuffleupagus/issues/278 now that I have a working alpine linux running?
2019-01-10 12:00:10 <ncopa> similar problem in other module https://github.com/igbinary/igbinary/issues/77
2019-01-10 12:00:39 <ncopa> this is how they fixed it in the other module: https://github.com/igbinary/igbinary/commit/491c5468ecb06bd5d5ee99e1583ed8ae32d8fc09
2019-01-10 12:01:12 <ncopa> i think it happens when php is built with --with-session=shared
2019-01-10 12:05:12 <jvoisin> wonderful ♥
2019-01-10 12:05:58 <jvoisin> ho, it does make sense, there is a php7-session package in Alpine
2019-01-10 12:06:23 <jvoisin> thank you ncopa ♥
2019-01-10 12:09:11 <ncopa> jvoisin: i think you owe me a beer now ;)
2019-01-10 12:20:43 <jvoisin> I do for sure, I really like Alpine :)
2019-01-10 12:28:29 <ncopa> 3.9.0 release candidate 1 is out
2019-01-10 20:26:12 <Sytis> Good evening, I'm having trouble getting Alpine to boot on an EFI Installation on a NVMe drive. The initrd has no nvme support and therefore can't find anything nvme related under /dev. Does someone else had this problem and knows the culprit?
2019-01-10 20:27:07 <jn__> the culprit is likely that the initrd isn't built with nvme support by default
2019-01-10 20:29:42 <jn__> Sytis: you probably have to enable the nvme feature in /etc/mkinitfs/mkinitfs.conf and run mkinitfs
2019-01-10 20:30:55 <Sytis> So just adding nvme_core would suffice? Btw. I haven't been able to make the initrd build verbose so I can actually see what's being built, is there an option I haven't seen?
2019-01-10 20:31:35 <jn__> the "feature" is called nvme, not nvme_core
2019-01-10 20:31:43 <jn__> see /etc/mkinitfs/features.d
2019-01-10 20:32:26 <Sytis> Ok, I'll add that then and see if it works.
2019-01-10 20:39:25 <Sytis> I don't know what happened but after booting the live system again the nvme drive doesn't show up either. But it does show up in Bios/UEFI as a boot option and I can also boot from it (but only to initrd rescue shell).
2019-01-10 20:41:08 <Sytis> Loading nvme_core in the live system doesn't help apparently :/
2019-01-10 20:44:57 <jn__> does modprobing nvme help?
2019-01-10 20:45:24 <jn__> both module names exist, but i don't know which does what
2019-01-10 20:47:52 <Sytis> Tried, both get loaded successfully but the drive remains invisible. Really strange it is...
2019-01-10 20:48:51 <jn__> hmm, no idea :(
2019-01-10 20:50:51 <Sytis> Well, thanks anyway. I'll try debugging further (@_@)
2019-01-10 21:00:56 <Sytis> Found the error in case someone is interested: Switching from RST to AHCI or whatever it was called solved it. Now the nvme shows up and I can try fixing the initrd :)
2019-01-10 21:01:57 <jn__> ah, is RST one of those vendor-specific RAID modes?
2019-01-10 21:02:50 <Sytis> It is, yes. Got enabled by accident and caused this trouble for the live image.
2019-01-10 21:13:13 <Sytis> THANKS! Adding nvme to the feature list worked. :D
2019-01-11 04:05:16 <sodface> New Alpine Linux user here.  Installed on raspberry pi tonight. Had to compile squeezelite from source but that was it, all other packages were in the repo.  About an hour and half from download to having Alpine/squeezelite connected to my LMS music server and playing music.  Really pleased and impressed!  Thanks!!
2019-01-11 05:48:58 <danieli> :))
2019-01-11 07:41:06 <nidheeshdas> Is there a way to specify a script as kernel parameter during pxe boot? I want run a bunch of computers as workers. I want them to use PXE to boot alpine and then run a bash script that will load my app and join my cluster.
2019-01-11 08:56:40 <ncopa> sodface: thanks for sharing!
2019-01-11 08:58:02 <_ikke_> sodface already left
2019-01-11 09:31:46 <jbwdevries1234> libcurl and apk-tools in 3.8 depend on libressl, but in edge they depend on libssl. Does that mean alpine is switching back to OpenSSL?
2019-01-11 09:32:25 <mps> jbwdevries1234: yes
2019-01-11 09:32:55 <jbwdevries1234> hm, interesting, I'm having trouble with libressl so that's good news for me
2019-01-11 09:33:20 <jbwdevries1234> Is this documented some where? Is there a planned date or something?
2019-01-11 09:33:21 <mps> you are not alone :)
2019-01-11 09:33:33 <jbwdevries1234> Apparently :) Always nice to know
2019-01-11 09:34:31 <mps> it was discussed on mailing list. date? with 3.9 openssl is default
2019-01-11 09:35:15 <jbwdevries1234> Ooh, there's talk of a RC
2019-01-11 09:35:16 <jbwdevries1234> nice
2019-01-11 09:36:26 <jbwdevries1234> Alright, that saves me a bunch of work. Thanks for the info, and for the good product :)
2019-01-11 09:36:28 <mps> 3.9 rc1 is announced yesterday
2019-01-11 09:37:06 <jbwdevries1234> Yeah, just found that on the mailing list, thanks :)
2019-01-11 11:55:17 <floWenoL> hi
2019-01-11 11:57:26 <floWenoL> I was trying to create an AWS AMI with alpine but there seems to be missing a package called aws-ena-driver, from google searches it seems to have been available but I can't find it anymore
2019-01-11 15:00:47 <geosmin> in regards to the $hostname.apkvol.tar.gz file. does $hostname need to match the host in order for it to get loaded?
2019-01-11 15:58:45 <clandmeter> geosmin, what defines the "host"
2019-01-11 18:13:17 <maldridge> on 3.8 php7-openssl identifies its library version as LibreSSL 2.7.4, is this intentional?
2019-01-11 18:21:47 <TBB> 3.8 is still libressl-based, so I guess so
2019-01-11 18:22:37 <maldridge> ah, I was mistaken then, I thought 3.8 was the first to be back on openssl
2019-01-11 18:24:19 <TBB> stable is with libressl but some packages in edge are already on openssl
2019-01-11 18:24:35 <TBB> that's the impression I have
2019-01-11 18:37:00 <wej> what's the reason for going back to openssl?
2019-01-11 18:41:04 <_ikke_> wej: https://lists.alpinelinux.org/alpine-devel/6308.html
2019-01-11 18:44:27 <wej> _ikke_: thanks!
2019-01-11 19:20:04 <essence> does anyone here know how to use Plymouth on Alpine Linux? I see that there's no "splash" implementation on initramfs' init using Plymouth, only fbsplash, and couldn't find any reference on the internet about that
2019-01-11 19:20:40 <essence> creating a custom init for that seems like an overkill :/
2019-01-11 19:23:08 <dalias> what is plymouth?
2019-01-11 19:27:07 <mps> dalias: some graphic boot add-on, unnecessary imo
2019-01-11 21:22:50 <TBB> the package adds a feature named plymouth if I remember correctly, adding that to your feature set in mkinitfs.conf should be enough to get that stuff into the initramfs
2019-01-11 21:23:09 <TBB> but also IIRC there's no code for it in the early init
2019-01-11 21:48:39 <mps> after six hours of compiling firefox-esr on aarch64 build process is killed, dmesg: Out of memory: Kill process 10677 (cc1plus) score 324 or sacrifice child :(
2019-01-11 21:58:13 <unmy> ouch! my condolence mps :\
2019-01-11 22:11:09 <mps> no, will continue will less parallel processes
2019-01-11 22:11:40 <geosmin> clandmeter: /etc/conf.d/hostname sets $hostname based on external config
2019-01-11 22:12:08 <geosmin> but tl;dr: the hostname won't always matc hthe apkvol's
2019-01-11 22:12:14 <geosmin> match the*
2019-01-11 22:12:34 <mps> just mumbling where software goes today, can't be compiled on 4GB of RAM XD
2019-01-12 18:02:49 <hatramatra> Evening, running in LBU mode (from USB stick), is it possible to upgrade the kernel? I can't fathom how this would possibly worked.
2019-01-12 18:10:46 <hatramatra> OK, found it in the wiki.
2019-01-12 18:11:05 <_ikke_> hatramatra: yes, but you have to update the boot media
2019-01-12 18:14:26 <hatramatra> I need to try kernel 4.19 for 'ip vrf', so I guess I will try with 3.9_rc1.
2019-01-12 18:15:08 <hatramatra> _ikke_: Can I create new USB media and just copy apkovl from the old one?
2019-01-12 18:16:01 <mps> hatramatra: I installed 4.19 from edge on 3.8, works fine
2019-01-12 18:16:50 <hatramatra> mps: is that run from ram installation?
2019-01-12 18:17:34 <mps> no, sys install
2019-01-12 18:19:21 <hatramatra> i'm reading the wiki, there seems to be all what i need to know
2019-01-12 18:22:31 <_ikke_> hatramatra: yes, should be possible
2019-01-12 18:41:56 <hatramatra> something else, if apkovl is readable by all, does it raise any concerns in run-from-ram installations?
2019-01-12 18:44:50 <_ikke_> good question
2019-01-12 19:35:58 <hatramatra> back to 3.9_rc1 with 4.19 kernel, even there CONFIG_BPF_SYSCALL is not set
2019-01-12 19:36:23 <hatramatra> because of that, 'ip vrf exec' doesn't work
2019-01-12 19:37:12 <hatramatra> is there any chance to get that CONFIG_BPF_SYSCALL and friends depending on it like CONFIG_CGROUP_BPF set before 3.9 is released?
2019-01-12 19:48:58 <hatramatra> i need to disconnect, i sent email about it to alpine-devel
2019-01-13 11:53:50 <gecku> hey, I've got a problem with fsck where `rc-service fsck start` won't work ("Filesystems couldn't be fixed"), but I can mount the drive later just fine and fsck does not report any errors on the drive when I try it from other machines. Help? :/
2019-01-13 12:13:27 <gecku> i can work around the problem by specifying to not check the fs in /etc/fstab, but that doesn't feel like a solution
2019-01-13 13:24:01 <debuggerboy> Is it possible to upgrade from alpine linux 3.8 to edge. I found url : https://wiki.alpinelinux.org/wiki/Edge all I have to do it to update "/etc/apk/repositories" file. And do a upgrade ?
2019-01-13 13:25:35 <_ikke_> ...
2019-01-13 13:26:22 <_ikke_> debuggerboy: yes, change the repository urls from the versioned ones to edge
2019-01-13 13:26:29 <_ikke_> debuggerboy: and then apk upgrade -U --available
2019-01-13 13:30:16 <debuggerboy> _ikke_: thanks, 👍 will try it today!
2019-01-13 13:32:26 <mps> debuggerboy: maybe first upgrade apk-tools-static and then the rest with with it, apk.static upgrade -U --availabe
2019-01-13 13:33:42 <mps> I always do that and never had serious problem with upgrades
2019-01-13 13:34:31 <debuggerboy> mps: is the "apk-tools-static" static linked binaries for "apk tools" ?
2019-01-13 13:34:37 <_ikke_> yes
2019-01-13 13:35:10 <_ikke_> So it does not depend on a working installation
2019-01-13 13:35:50 <debuggerboy> Thanks for that info, sure will keep it my upgrade checklist. 👍
2019-01-13 13:36:35 <mps> although never needed it, i usually also first upgrade busybox-static
2019-01-13 13:45:21 <debuggerboy> @mps : I think it's a good step as well.
2019-01-13 21:26:11 <sauercrowd> Hey Folks. I'm currently trying to use alpine with qemu using the -kernel and -initrd options. When trying to use the vanilla initramfs from the netboot download it tries to mount a "Root VFS" (which I don't have), same story when I try to build my own initrd from the "Mini Root Filesystem". Is it somehow possible to prevent mounting that root Filesystem and just go with the initrd?
2019-01-13 21:31:28 <sauercrowd> My goal is to be able to quickly run alpine with different kernels and without all the formatting of "real" FS, happy for any suggestions or tips!
2019-01-13 21:42:48 <clandmeter> It should work  out of the box
2019-01-13 21:43:10 <clandmeter> But it needs modloop
2019-01-13 21:45:30 <sauercrowd> Modloop is a squashfs right? Is it somehow possible to use that with qemu?
2019-01-13 21:48:03 <clandmeter> Yes you can use ipxe
2019-01-13 21:48:11 <clandmeter> Or just use the iso
2019-01-13 21:48:59 <clandmeter> Or specify the modloop as boot arg via http
2019-01-13 23:15:56 <sauercrowd> I will try that out, thanks!
2019-01-14 09:59:14 <mps> what are usage of consolekit and polkit
2019-01-14 10:00:19 <mps> i remember time when they didn't existed and everything was fine, even better imo
2019-01-14 11:13:22 <kahiru> mps: iirc consolekit is some kind of session tracking and polkit is a privilege elevation thing
2019-01-14 11:53:32 <rain1> hi
2019-01-14 11:59:06 <mps> kahiru: I know what they are 'by definition', but what practical usage of them is/are? Btw, i rebuilt slim (X DM) without consolekit and use it for months without any problem and don't see any difference
2019-01-14 12:01:04 <mps> new versions of polkit depends on mozjs, so for simple system we will have to install mozjs
2019-01-14 12:05:16 <kahiru> ah, right. the only use ive seen for polkit is that when i start up virt-manager I get a window to enter my pass and I am then allowed to connect to qemu:///system
2019-01-14 12:05:29 <kahiru> and the rules for polkit are written in js
2019-01-14 12:06:01 <mps> eh, system software in js :|
2019-01-14 12:23:05 <TBB> networkmanager uses polkit, udisks* uses polkit... plenty of software depends on it, although I imagine systemd etc etc idontneedtocontinuethatthought
2019-01-14 12:27:47 <mps> TBB: I'm happy 'nouser' of these :)
2019-01-14 15:41:56 <geosmin> any good way to get a package from edge on stable?
2019-01-14 15:42:13 <_ikke_> Not really
2019-01-14 15:42:24 <_ikke_> It might work in some cases, but in other cases it won't work
2019-01-14 15:42:32 <geosmin> well maybe there's a less stupid solution than syncthing for what i'm trying to do
2019-01-14 15:43:06 <geosmin> i have 3 identical alpine aarch64 boards deployed and i'm trying to do some rapid prototyping
2019-01-14 15:43:14 <geosmin> manually mirroring changes is a bit of a chore
2019-01-14 15:43:43 <CcxWrk> geosmin: dsync, unison, or plain old version control system
2019-01-14 15:43:57 <_ikke_> I use a simple script with inotifywait + rsync
2019-01-14 15:44:00 <geosmin> currently dumping all relevant bits of code to a single dir, files are then symlinked to the right places
2019-01-14 15:44:22 <CcxWrk> Or just have an authoritative version on NFS or something.
2019-01-14 15:44:35 <geosmin> yeah, was also considering NFS
2019-01-14 15:46:48 <CcxWrk> I usually do that with my workstation being the server and the various development boards boot over PXE+NFS either into live system or into system that will rsync the NFS source to their storage.
2019-01-14 15:56:15 <clandmeter> geosmin, syncthing is in stable?
2019-01-14 15:56:44 <clandmeter> looks like it doesnt build for arm
2019-01-14 15:59:11 <geosmin> clandmeter: yeah, but not on aarch64
2019-01-14 15:59:36 <geosmin> though frankly the other suggestions are probably better
2019-01-14 15:59:50 <clandmeter> they probably are
2019-01-14 16:00:11 <clandmeter> but i also use syncthing for setups behind fw's out of my control
2019-01-14 16:00:42 <geosmin> i think another option would be to self-host a local repo and do my own apkbuilds
2019-01-14 16:02:45 <clandmeter> i wonder why syncthing is not  in repo
2019-01-14 16:05:03 <mps> clandmeter: it is: apk search syncthing says: syncthing-0.14.54-r0
2019-01-14 16:05:29 <mps> on aarch64 v3.9_rc1
2019-01-14 16:05:33 <clandmeter> yes you are right
2019-01-14 16:05:39 <clandmeter> i didnt look correctly
2019-01-14 16:05:47 <clandmeter> its only arm32 thats missing
2019-01-14 16:06:25 <clandmeter> geosmin, https://pkgs.alpinelinux.org/package/v3.8/community/aarch64/syncthing
2019-01-14 16:06:33 <mps> on armhf/armv7 it is not, because of go-lang issue. I will work on go-bindata which is probmlematic on armv7
2019-01-14 16:07:11 <mps> I started but switched to rust problem on ARM
2019-01-14 16:08:11 <mps> think, that is more important for release. (and I prefer mozilla instead of go-ogle)
2019-01-14 16:42:47 <FreeBDSM> hello, how to add a dir to PATH systemwide?
2019-01-14 16:43:23 <FreeBDSM> I've tried to add ~/.profile with `echo 'PATH=${PATH}:/app/node_modules/.bin;export PATH' > ~/.profile`, but root user still doesn't have that dir in the PATH envvar
2019-01-14 16:43:33 <FreeBDSM> I had to manually source it
2019-01-14 16:43:39 <FreeBDSM> I am using ash
2019-01-14 16:44:32 <FreeBDSM> well, I'm using busybox
2019-01-14 16:47:32 <FreeBDSM> how to permanently alter PATH envvar using busybox?
2019-01-14 19:23:54 <sy>  /w 6
2019-01-14 19:23:56 <sy> ugh
2019-01-14 22:35:13 <CcxWrk> Hmm, installing s6-linux-init on edge bugs out with ERROR: unsatisfiable constraints:   so:libskarnet.so.2.6 (missing):     required by: s6-linux-init-0.4.0.0-r0[so:libskarnet.so.2.6]
2019-01-14 22:35:20 <CcxWrk> Is there a rebuild in progress?
2019-01-14 23:54:34 <maldridge> once a release ships is that done, or is it still possible to get packages into 3.8?
2019-01-15 00:25:26 <maldridge> I should revise my question.  The package I want is in edge, how bad of an idea is it to mix edge and v3.8 in prod
2019-01-15 00:25:38 <maldridge> I didn't realize the package I wanted existed already
2019-01-15 00:26:43 <mps> maldridge: you can try, if the package doesn't install a lot of dependent packages
2019-01-15 00:26:53 <mps> I do that sometimes
2019-01-15 00:27:09 <mps> for small packages, mostly
2019-01-15 00:27:14 <maldridge> I think I'll just roll wiht edge and see what happens here
2019-01-15 00:27:24 <maldridge> we have good QA procedures and reasonably high test coverage
2019-01-15 00:27:30 <mps> which package
2019-01-15 00:27:42 <maldridge> php7-pecl-msgpack and php7-pecl-igbinary
2019-01-15 00:28:34 <mps> huh, what arch? I worked to fix some bugs on php7-pecl-msgpack
2019-01-15 00:28:40 <maldridge> x86_64
2019-01-15 00:29:00 <maldridge> since this is going into a container, I'm less worried about rolling along a few steps behind HEAD
2019-01-15 00:29:23 <mps> what version of the php is on the 3.8
2019-01-15 00:29:32 <maldridge> 7.2 iirc
2019-01-15 00:30:01 <mps> you can rebuild it on 3.8, more safe option
2019-01-15 00:30:27 <mps> s/more safe/safer/
2019-01-15 00:30:29 <maldridge> hmm, harder to justify
2019-01-15 00:30:37 <maldridge> then I need build infrastructure
2019-01-15 00:30:44 <maldridge> and an internal repo
2019-01-15 00:31:11 <mps> well, yes, but not complicated at the end
2019-01-15 00:31:22 <mps> just follow alpine dev wiki
2019-01-15 00:32:02 <maldridge> sure, I run void's build infrastructure so I'm not that bothered by it, but right now I'm still trying to validate that these php packages are needed
2019-01-15 00:32:27 <maldridge> I would personally have rather fixed the problem that prevents php7-pecl from being able to install it that way, but neither here nor #php knows what's wrong
2019-01-15 00:33:42 <mps> I don't use php, just worked to fix some bugs for next Alpine release on armv7/aarch64
2019-01-15 00:34:05 <mps> there were no problems on x86_64, iirc
2019-01-15 00:36:44 <maldridge> ah, good to know
2019-01-15 00:39:59 <mps> maldridge: I'm going to bed now, it's late here, if you try to build and have problem I'll be here tomorrow, probably
2019-01-15 00:40:13 <maldridge> sure, thanks for the replies so far
2019-01-15 00:40:22 <maldridge> I just built the container and I'm pushing it to a test cluster
2019-01-15 00:40:30 <maldridge> so I should know in a bit if its okay or not
2019-01-15 00:40:59 <mps> during our chat I rebuilt it on 3.8 without any issue, so it should work
2019-01-15 00:41:21 <maldridge> good to know
2019-01-15 00:41:47 <mps> ok, good night now
2019-01-15 00:41:52 <maldridge> 'night
2019-01-15 00:53:03 <kxzaon> Hello guys. Can you tell me why init.d is not included in lbu /etc by default ?
2019-01-15 00:57:02 <johnnyfive> Howdy! I'm having issues with consuming signed files. I receive an "untrusted signature" when I try to `apk update`. Is there some way to get more information about what is wrong? I have performed lots of checking against the .SIGN. file inside the APKINDEX.tar.gz, and lots of other steps, but I just can't figure out why it is untrusted.
2019-01-15 01:15:43 <johnnyfive> My hope is for some other tool I can use that will be very explicit about why the signature fails.
2019-01-15 09:45:55 <strta> Hello everyone, is the "mysql_embedded" binary packaged with mariadb broken or something? I am trying to set up a bootstrap init-script to set up a database and user before mariadb starts, but running "mysql_embedded" just instantly exits with the message "Aborted" regardless of which parameters I pass.
2019-01-15 09:46:08 <strta> I'm tearing out my golden locks in frustration.
2019-01-15 09:47:15 <strta> It appears to work in a Debian container when testing.
2019-01-15 09:50:45 <strta> Running Alpine 3.8.
2019-01-15 10:02:44 <thresh> hello.  I'm building commercial software to run on top of alpine linux, and was successful in building an .apk.
2019-01-15 10:03:12 <thresh> however, I wish to sign the package;  is that something that can be done not on the build host, but on a separate machine?   I don't trust buildhosts enough to store my private gpg key on them.
2019-01-15 10:06:17 <terra> thresh: you must find a secure host to store your keys then.
2019-01-15 10:07:06 <terra> otherwise signing a file without a private key seems meaningless
2019-01-15 10:08:19 <thresh> I do have a secure host.
2019-01-15 10:08:56 <thresh> Basically that's what I was doing for every other linux distro and *bsd: ssh to a buildhost, do a build, fetch the .deb/.rpm/.txz and sign them locally, and create the indexes
2019-01-15 10:09:05 <thresh> I'd like to have the same way to do it for Alpine, is that possible?
2019-01-15 10:11:15 <terra> when "abuild -r" finishes, it uses a private key that assigned to builder's personal info which acquires from /etc/abuild.conf
2019-01-15 10:13:05 <terra> https://wiki.alpinelinux.org/wiki/Abuild_and_Helpers
2019-01-15 10:13:50 <thresh> is there a way to skip this invocation?
2019-01-15 10:14:12 <thresh> or, maybe, use a fake key, and then re-sign both package contents and indexes with a proper key?
2019-01-15 10:14:44 <thresh> s/fake/unimportant/
2019-01-15 10:18:49 <terra> then extrack a .apk file and check the contents, especially metadata
2019-01-15 10:19:36 <terra> if there's no hashed data (or similar) you may able to restore signer info
2019-01-15 10:25:52 <clandmeter> thresh, check the abuild src and see how signing is done. you can easily extract the apk and resign it.
2019-01-15 10:31:58 <thresh> thank you terra, clandmeter
2019-01-15 10:34:36 <mps> clandmeter: built and running firefox-esr-60.4.0 on armv7 v3.9_r1. thanks for help :)
2019-01-15 14:24:36 <unrznbl[m]> new to alpine linux, I was trying to use setup-interfaces to get dhcp working on either a wlan0 or usb0 interface but I never seem to get an IP addr. Wasn't sure "where" dhclient or equiv was located. And I just figured it out: /etc/init.d/networking restart and udhcpc is the client. :) I suppose I might add some udev hooks to restart networking when I insert a usb tether or wifi device?
2019-01-15 16:22:26 <unrznbl[m]> there is one package: "pass" that I would like... but it is only in "edge". I imagine a sane choice is to build manually and not switch to "edge" on a computer I'd like to "work" most of the time... right?
2019-01-15 16:22:37 <unrznbl[m]> Or is there a way to pick and choose if the dependencies aren't too complicated?
2019-01-15 16:23:25 <sy> you can mask edge
2019-01-15 16:24:02 <sy> just add @edgemain to the start of the repository line
2019-01-15 16:25:17 <unrznbl[m]> mm... this example is probably bad since it is so easy to install manually... so add @edgemain where?
2019-01-15 16:25:37 <unrznbl[m]> "apk add pass" becomes something else? or add something to /etc/<something>?
2019-01-15 16:28:22 <unrznbl[m]> sy: thank you for the clue. I see how it works now: https://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management
2019-01-15 16:29:12 <sy> no problem :)
2019-01-15 16:37:57 <unrznbl[m]> I managed to finally get alpine installed on a thinkpad 240 by using these instructions: https://wiki.alpinelinux.org/wiki/Replacing_non-Alpine_Linux_with_Alpine_remotely
2019-01-15 16:38:23 <unrznbl[m]> but... I had to do setup-alpine and THEN setup-disk, otherwise setup-disk would complain about missing packages.
2019-01-15 16:38:49 <unrznbl[m]> I also used a usb stick to hold the sda things and made symlinks instead of copying around.
2019-01-15 16:40:16 <unrznbl[m]> had to pull the harddrive out and work with it on a another computer to get the install iso in, then used those instructions. Worked great! :)
2019-01-15 16:41:55 <pgauret> hi, quick question, how does one install a cross compiler on Alpine (i.e. build armhf binaries on x86_64 platform)?
2019-01-15 16:42:07 <pgauret> couldn't find the relevant apks
2019-01-15 16:42:31 <_ikke_> I don't think they are packaged
2019-01-15 16:44:43 <pgauret> ah :( and that can't be an easy one to package
2019-01-15 16:46:28 <_ikke_> alpine itself mostly compiles on native hardware
2019-01-15 16:47:33 <pgauret> kernel builds on my old cubieboard won't be fun :)
2019-01-15 16:47:41 <_ikke_> heh
2019-01-15 17:03:10 <_ikke_> pgauret: the gcc package does have some kind of cross-compile support, but I don't know if that works for you
2019-01-15 17:03:33 <_ikke_> But that's probably to cross-compile gcc itself
2019-01-15 17:04:54 <pgauret> I've found toolchains on https://toolchains.bootlin.com
2019-01-15 17:05:08 <pgauret> but running into issues with some missing libs to execute
2019-01-15 17:05:26 <TBB> I guess you have two options, either use a toolset like scratchbox or such, or doing your compilation inside qemu
2019-01-15 17:06:35 <pgauret> /tmp/armv7-eabihf--musl--stable-2018.11-1/bin # ./tar
2019-01-15 17:06:35 <pgauret> Error relocating ./tar: wmempcpy: symbol not found
2019-01-15 17:07:30 <TBB> I don't have much scratchbox experience but it seems to not support musl libc
2019-01-15 17:08:02 <TBB> (at first glance that's my impression, could be really wrong in other words)
2019-01-15 17:12:24 <squarebracket> busybox wget should work with https if i have certain libraries installed, right?
2019-01-15 17:12:43 <squarebracket> or am i stuck installing the wget package?
2019-01-15 17:13:13 <dalias> i think it supports https via execution of s_client
2019-01-15 17:13:18 <dalias> openssl s_client
2019-01-15 17:13:36 <dalias> so you need the openssl (or libressl) utils package or whatever installed
2019-01-15 17:13:45 <dalias> but real wget is probably better
2019-01-15 17:14:38 <dalias> ah ssl_client
2019-01-15 17:14:39 <squarebracket> and the library that calls s_client should tie in with ca-certificates, yes?
2019-01-15 17:15:13 <dalias> it's the ssl_client package
2019-01-15 17:16:06 <dalias> it depends on libtls-standalone which depends on ca-certificates
2019-01-15 17:16:11 <dalias> no idea what libtls-standalone is...
2019-01-15 17:16:20 <squarebracket> 
ACTION thinks

2019-01-15 17:16:22 <pgauret> dalias: maybe related to https://bugs.alpinelinux.org/issues/9830 ?
2019-01-15 17:16:23 <dalias> it's from libressl
2019-01-15 17:16:46 <squarebracket> then my custom CA should be working as i call update-ca-certificates... hmm...
2019-01-15 17:17:31 <dalias> that bug might be related
2019-01-15 17:19:16 <squarebracket> in the packages for 3.8, i don't see ssl_client depend on ca-certificates, and conversely i don't see ca-certificates required by anything installed in the base docker image
2019-01-15 17:20:02 <squarebracket> if it's a bug (like some linking issue) then good cuz it means it will be fixed... i'm just trying to figure out what to do since it looks like my "install company CA" script breaks since 3.5
2019-01-15 17:20:22 <dalias> i'm using edge
2019-01-15 17:20:38 <dalias> i suspect something broke deps in mess around libressl/openssl switch
2019-01-15 17:21:24 <squarebracket> yeah, that could be. has the switch been completed?
2019-01-15 17:23:43 <squarebracket> cool, no libressl in the edge docker image. guess that means yes
2019-01-15 17:26:03 <pgauret> on arm, how do you load the initramfs-vanilla at boot? is there a way without manually converting to uinitrd for uboot?
2019-01-15 17:52:28 <mps> pgauret: cross-tools are on my (not so short) todo list for Alpine
2019-01-15 19:17:56 <pgauret> mps: thanks for mentioning this, I'll keep a ubuntu vm in the meantime, looking forward to testing your package :)
2019-01-15 19:21:57 <mps> pgauret: I also have debian in lxc for cross compiling kernels for some ARM machines
2019-01-15 19:22:33 <mps> that is the reason I plan to make it on Alpine
2019-01-15 19:23:18 <mps> but, that is on the not so short todo list
2019-01-15 19:43:13 <pgauret> mps: side question then, did you ever manage to use the stock alpine kernel on arm?
2019-01-15 19:43:42 <pgauret> (found it didn't boot on Allwinner/sunxi, not sure uboot can load the initramfs as such)
2019-01-15 20:04:45 <mps> pgauret: didn't even tried that, so don't know to be honest
2019-01-15 20:06:07 <mps> but tried Alpine kernels in qemu for armhf and aarch64. there it worked both
2019-01-15 20:08:01 <TBB> I've done some stuff with lxc, but does it allow running other architectures?
2019-01-15 20:09:23 <_ikke_> lxc is not virtualization
2019-01-15 20:09:46 <_ikke_> so no
2019-01-15 20:11:38 <TBB> that was my assumption as well; other binfmts can probably be run, that was the thought behind the question
2019-01-15 20:14:13 <_ikke_> but other binfmts still need to be the same arch, right?
2019-01-15 20:14:37 <TBB> surely
2019-01-15 20:15:20 <mps> _ikke_: yes, i run some time ago armhf and aarch64 in chroot on x86_64
2019-01-15 20:15:33 <TBB> don't exactly even know what drove me to ask really, I'm well experienced in virtualization and containerization and know the difference
2019-01-15 20:16:18 <_ikke_> how did you manage that?
2019-01-15 20:16:48 <mps> that was under debian, yes it worked but was really slow
2019-01-15 20:16:58 <_ikke_> was it some kind of emulation?
2019-01-15 20:17:42 <mps> qemu_somehing have to be installed/copied to chroot-ed dir
2019-01-15 20:18:18 <TBB> so, basically, virtualization
2019-01-15 20:18:31 <mps> forgot exactly what, but worked. used it to debootstrap arm for some SBC's
2019-01-15 20:19:07 <TBB> I should probably be looking into that, I've got a potential next project that runs on arm something
2019-01-15 20:19:12 <mps> somewhere on debian site are guides about that
2019-01-15 20:20:08 <TBB> scratchbox is qemu based
2019-01-15 20:20:29 <mps> I remember that I bootstrapped debian on chromebooks
2019-01-15 20:21:01 <mps> bootstrapped it on x86_64, I mean
2019-01-15 20:21:54 <_ikke_> You even need emulation for arm / aarch64
2019-01-15 20:21:58 <mps> that worked, only have had to build kernels and flashed it on sd cards
2019-01-15 20:22:52 <mps> _ikke_: yes, there is something qemu_format-bin but forgot what is that exactly
2019-01-15 20:23:16 <mps> probably could be found on debian site
2019-01-15 20:28:50 <mps> TBB: search for qemudebootstrap or qemu-debootstrap or similar term
2019-01-15 20:47:33 <TBB> thanks for the pointers mps :)
2019-01-15 20:53:10 <mps> TBB: I bootstrapped Alpine for chromebooks and Allwinner A20 boards that way on Debian about year and half ago
2019-01-15 20:54:15 <mps> didn't tried that on Alpine because didn't had need for that
2019-01-15 23:53:09 <jwh> binfmt++
2019-01-16 01:19:56 <m4t> hey, has anyone noticed the new util-linux mount failing to mount if /etc/fstab has partitions specified by LABEL?
2019-01-16 01:20:29 <m4t> busybox mount works fine. i just pulled in a bunch of updates in edge and noticed after rebooting :/ util-linux-2.33-r0 on armhf
2019-01-16 01:21:06 <dalias> :o
2019-01-16 01:21:15 <m4t> e.g.: mount: /: can't find LABEL="pmOS_root"
2019-01-16 01:21:29 <m4t> blkid shows the correct label/uuid
2019-01-16 01:21:46 <m4t> doh, blkid is busybox ;)
2019-01-16 01:23:02 <m4t> yeah. so 'blkid' and 'liblkid' are busted i think. blkid-2.33-r0 only shows a vfat/dos partition. busybox blkid only shows f2fs and ext2.
2019-01-16 03:04:47 <m4t> :( https://gitlab.com/postmarketOS/pmaports/issues/156 so it's known
2019-01-16 03:11:13 <m4t> adding this script to sysinit seems to fix it. does some kpartx crap to make the sub-partitions visible to util-linux 2.33 i guess https://gitlab.com/postmarketOS/pmaports/raw/6d25c1bef690a9fe3b5f21977791c675e4c9843d/main/postmarketos-base/devmappings.initd
2019-01-16 14:51:19 <steven> sup guys. I am just double checking before I have to trash my alpine images, but could it be the case that there are no mpdecimal packages for alpine?
2019-01-16 15:09:46 <mps> steven: interesting, didn't know for that package. could be packaged probably. could you give me more info
2019-01-16 15:12:57 <steven> mps: not quite sure what information to provide, its a lib providing decimal/floating point arithmetic and is used by pythons decimal module and others
2019-01-16 15:13:18 <steven> so if you wanted to use pythons decimal module, this package is a requirement.
2019-01-16 15:13:41 <steven> i am compiling it manually right now and it does not have any other dependencies (as far as I can tell)
2019-01-16 15:13:55 <mps> ah, for python and C/C++ only. sorry I'm not versed to python
2019-01-16 15:14:27 <steven> php-decimal usees it too mps, but I obviously don't have a list of all decimal/floating point implementations using this lib :D
2019-01-16 15:14:48 <mps> doesn't look hard to make apk but that should do someone who knows python, not me
2019-01-16 15:16:02 <steven> I never made an apk package, but if its easy I might just do that instead of manually compiling it. can you link me to a resource/how to do that?
2019-01-16 15:16:07 <mps> but the library is interesting, really
2019-01-16 15:16:15 <steven> just so I can make it work locally :D
2019-01-16 15:16:55 <_ikke_> steven: https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2019-01-16 15:16:57 <mps> there are guides on Alpine wiki, it is not complicated to make apk
2019-01-16 15:17:09 <steven> cool, thanks guys!
2019-01-16 15:17:47 <steven> ah looks a bit like archs PKGBUILD
2019-01-16 15:17:51 <_ikke_> yes
2019-01-16 15:17:56 <_ikke_> it's very similar
2019-01-16 15:18:00 <mps> it takes me few hours to make first package after reading wiki on url _ikke_ just posted
2019-01-16 15:18:03 <steven> that I am familiar with, neat.
2019-01-16 15:18:29 <_ikke_> steven: there are some differences in how alpine packages things
2019-01-16 15:18:44 <_ikke_> that you should be aware of, but abuild will mostly let know those kinds of things
2019-01-16 16:04:17 <steven> ERROR: mpdecimal: Provide a valid RFC822 maintainer address
2019-01-16 16:04:38 <steven> _ikke_: ^ does this mean imma have to configure git? does abuild use git to get this info?
2019-01-16 16:07:34 <mps> steven: add '# Maintainer: email address of maintainer' in APKBUILD at top
2019-01-16 16:08:01 <steven> I did :D
2019-01-16 16:08:19 <steven> any special format? like My Name <e@mail.tld> or something?
2019-01-16 16:08:19 <mps> hmm, strange then
2019-01-16 16:08:48 <mps> no need for special format but usually just like you wrote
2019-01-16 16:08:50 <steven> I think there was a Contributor: thing which I've removed jfyi
2019-01-16 16:09:11 <mps> set it both
2019-01-16 16:09:23 <mps> same address is ok
2019-01-16 16:10:50 <steven> cool, that did it
2019-01-16 16:11:20 <steven> the error really should say provide a contributor address :D
2019-01-16 16:11:34 <mps> btw, it good idea to git clone aports, then create new branch in which you add new package, so you can send your 'patch' to Alpine easily
2019-01-16 16:14:53 <steven> abuild -F eventually errors out saying don't run it as root
2019-01-16 16:15:10 <steven> -F == forceroot :D buggy tooling
2019-01-16 16:18:44 <_ikke_> yes, someone already mentioned it
2019-01-16 16:18:45 <mps> no, it is better to add yourself in abuid group and add in /etc/sudoers
2019-01-16 16:20:27 <mps> so you don't have to force root
2019-01-16 16:25:06 <steven> mps: its in a docker container
2019-01-16 16:26:08 <_ikke_> steven: I use this in Docker: https://tpaste.us/vboM
2019-01-16 16:26:44 <steven> bruv, this might just fixes my permission issues
2019-01-16 16:26:52 <steven> thanks, i'll take the bits I am missing out of it
2019-01-16 16:29:59 <steven> close, still throwing errors
2019-01-16 16:30:39 <steven> >>> mpdecimal: Entering fakeroot...
2019-01-16 16:30:41 <steven> /usr/bin/install -c -d -m 755 /usr/local/include
2019-01-16 16:30:43 <steven> /usr/bin/install -c -m 644 libmpdec/mpdecimal.h /usr/local/include
2019-01-16 16:30:45 <steven> install: can't create '/usr/local/include/mpdecimal.h': Permission denied
2019-01-16 16:31:01 <_ikke_> You should add DESTDIR="$pkgdir"
2019-01-16 16:31:04 <_ikke_> to make install
2019-01-16 16:31:45 <steven> like literally? just copy / paste?
2019-01-16 16:32:59 <_ikke_> make install DESTDIR="$pkgdir"
2019-01-16 16:33:49 <steven> running
2019-01-16 16:34:59 <steven> >>> mpdecimal*: Running postcheck for mpdecimal
2019-01-16 16:35:01 <steven> >>> ERROR: mpdecimal*: Packages must not put anything under /srv, /usr/local or /opt
2019-01-16 16:35:06 <steven> looks like mpdecimal doesn't like it
2019-01-16 16:35:36 <_ikke_> no, this is abuild complainging :-)
2019-01-16 16:35:41 <_ikke_> do you have a configure step?
2019-01-16 16:35:48 <steven> in build, ues
2019-01-16 16:35:50 <steven> yes*
2019-01-16 16:35:59 <_ikke_> add PREFIX=/usr
2019-01-16 16:36:19 <_ikke_> or was it --prefix=/usr
2019-01-16 16:36:22 <_ikke_> I think that
2019-01-16 16:36:27 <steven> the latter makes more sense
2019-01-16 16:36:51 <steven> when I did it manually tho it didn't complain about it
2019-01-16 16:36:53 <steven> weird
2019-01-16 16:37:04 <_ikke_> Like I said, abuild complains about it
2019-01-16 16:37:37 <_ikke_> it does some checks afterwards to make sure it follows alpine practices
2019-01-16 16:38:09 <steven> nice, worked!
2019-01-16 16:38:24 <steven> thanks, appreciate it!
2019-01-16 16:41:52 <_ikke_> yw
2019-01-16 17:21:23 <pgauret> Is there any documentation on how to run Alpine as a LXD host? I've figured a few things on what's requried but still struggling to actually launch a container
2019-01-17 05:02:50 <jwh> well this is fun, stick keeps crashing :(
2019-01-17 05:17:47 <jwh> trollop:~# dd if=/dev/zero of=/swap bs=1M count=2048
2019-01-17 05:17:54 <jwh> instant crash, love it
2019-01-17 06:48:22 <minsec> I found the recent mysql version is 10.1.32 at alpine 3.7, but how can I install an old version like 5.6
2019-01-17 06:49:46 <minsec> I want use apk to install an old version software, how can i do
2019-01-17 07:42:53 <clandmeter> minsec, you cannot
2019-01-17 09:00:12 <danieli> yes, we don't keep archives of old versions
2019-01-17 09:12:23 <_ikke_> minsec: you'd have to build it yourself
2019-01-17 09:28:33 <minsec> ok, thank u
2019-01-17 09:50:02 <steven> _ikke_: totally forgot about it last night, but if I wanted (yes. I am that selfish, its about me) the pacakge to be moved to the official repos, what one one have to do to make this happen?
2019-01-17 09:50:38 <_ikke_> steven: Usually you first submit it to the testing repo
2019-01-17 09:50:49 <_ikke_> you can create a pull request on the github aports repository for example
2019-01-17 13:34:18 <jwh> to use an old version you should probably also stop confusing mariadb with mysql :D
2019-01-17 13:36:43 <steven> myriasql is best
2019-01-17 14:31:28 <nindustries> Hi, does setuid work under Docker?
2019-01-17 14:33:42 <_ikke_> Seems to work
2019-01-17 14:33:49 <_ikke_> I can run fping as a non-root user
2019-01-17 14:35:18 <nindustries> hm.
2019-01-17 14:36:30 <nindustries> chown root:"$APP_USER" /usr/local/bin/installer $APP_DIR/secscan.sh && chmod 4550 /usr/local/bin/installer $APP_DIR/secscan.sh
2019-01-17 14:36:45 <nindustries> And if I run /usr/local/bin/installer, i'm still $APP_USER and not root
2019-01-17 14:39:21 <tw> shell scripts always ignore suid.
2019-01-17 14:40:25 <nindustries> ohhh.
2019-01-17 14:40:29 <nindustries> Any way to get that behavior?
2019-01-17 14:42:23 <tw> my preferred method is sudo with a specific line for the full script path, enabled for all users.
2019-01-17 14:43:53 <tw> If you actually need linux to honor suid on interpreted scripts... I think you have to patch the kernel.
2019-01-17 14:44:44 <nindustries> Darn it, I stripped off sudo
2019-01-18 00:03:45 <CcxWrk> I have issue creating tun interfaces (EPERM) under regular user with CAP_NET_ADMIN, kernel 4.19.15-0-vanilla. Anyone has any clue?
2019-01-18 00:20:11 <dalias> ccxwrk, see what strace shows failing
2019-01-18 00:21:02 <CcxWrk> ioctl(…, TUNSETIFF
2019-01-18 00:42:02 <ccx^xmpp> dalias: seems my modem crapped out the moment I've sent my response
2019-01-18 00:43:29 <ccx^xmpp> The error would be expected under regular user, but I've set and verified with capsh to have the cap_net_admin (inheritable)
2019-01-18 01:05:18 <dalias> can you keep the program from exiting somehow (maybe make stderr something that blocks) and inspect the capability mask via /proc?
2019-01-18 01:05:24 <dalias> to make sure it really has it
2019-01-18 01:05:31 <dalias> it could be a kernel regression i guess
2019-01-18 02:11:25 <jwh> hmmm, wonder whats causing this minute delay between networking and sshd
2019-01-18 02:11:37 <jwh> same with static/dhcp
2019-01-18 02:24:00 <dalias> service start delay?
2019-01-18 02:24:27 <jwh> hm, is there?
2019-01-18 02:24:44 <dalias> i'm asking what you mean by "delay between networking and sshd"
2019-01-18 02:25:23 <dalias> like a delay between the service startup messages when booting? or..?
2019-01-18 02:25:39 <jwh> well, it sets up networking, then does nothing for like a minute until it starts sshd
2019-01-18 02:25:56 <jwh> no output so I don't think theres anything else starting up, fresh box so just networking and sshd really
2019-01-18 02:26:25 <dalias> any nondefault configuration on sshd?
2019-01-18 02:26:36 <jwh> nope
2019-01-18 02:26:46 <dalias> openssh or dropbear?
2019-01-18 02:26:58 <jwh> openssh, I haven't checked dropbear actually
2019-01-18 02:27:31 <jwh> let me see if its actually sshd
2019-01-18 02:28:23 <jwh> yeah, straight to console with sshd disabled
2019-01-18 02:28:30 <jwh> with or without dropbear
2019-01-18 02:28:37 <jwh> 
ACTION tinkers

2019-01-18 02:30:37 <jwh> none of the commands the init script runs actually has any kind of delay, hm
2019-01-18 02:31:16 <dalias> the ssh service looks like it has a nonsensical dependency on dns
2019-01-18 02:31:22 <dalias> not sure if that could be the cause
2019-01-18 02:31:23 <jwh> https://bugs.alpinelinux.org/issues/6635
2019-01-18 02:31:25 <jwh> probably that
2019-01-18 02:31:26 <jwh> yeah
2019-01-18 02:31:48 <dalias> i don't think that's it
2019-01-18 02:31:49 <jwh> actually hm, no it isn't that as the keys are already generated
2019-01-18 02:31:51 <dalias> what kinda box is this?
2019-01-18 02:32:19 <jwh> I've actually seen it on 2 now, I thought ti was just dhcp on a box here but its doing it on a new VM with static config
2019-01-18 02:32:26 <jwh> only thing they have in common is they're both edge
2019-01-18 02:32:46 <dalias> i mean is it something weird like a rpi?
2019-01-18 02:32:50 <jwh> oh, no
2019-01-18 02:32:51 <dalias> or normal x86 server-type box?
2019-01-18 02:32:56 <jwh> x86 physical, x86 vm
2019-01-18 02:33:14 <dalias> rng init should only take a matter of milliseconds on any remotely modern x86
2019-01-18 02:33:23 <jwh> yeah thats what I'm thinking
2019-01-18 02:33:40 <dalias> this sounds more like dns timeouts
2019-01-18 02:33:51 <jwh> whats curious is the delay is very similar despite being completely different hardware and environments, so I expect its not the rng
2019-01-18 02:33:53 <dalias> do you have anything weird in resolv.conf ?
2019-01-18 02:33:57 <jwh> yeah, thats what I initially thought
2019-01-18 02:34:19 <dalias> and can you tcpdump to see if there's network traffic during the delay?
2019-01-18 02:34:23 <jwh> but I can't see why that would be a problem, sshd doesn't do any lookups
2019-01-18 02:34:31 <jwh> and doesn't need to dns to startup or function
2019-01-18 02:34:44 <dalias> well the init script has a "depend" on dns
2019-01-18 02:34:44 <jwh> but in both cases dns is available during startup
2019-01-18 02:34:53 <dalias> so maybe there's something that probes if dns is up yet by trying a lookup
2019-01-18 02:34:54 <jwh> oh hm, does that do lookups?!
2019-01-18 02:34:59 <jwh> that seems... silly
2019-01-18 02:35:06 <dalias> yes indeed if it's doing that it's idiotic
2019-01-18 02:35:15 <dalias> there's no reason for sshd to depend on _anything_, not even network
2019-01-18 02:35:31 <dalias> normally you bind to 0.0.0.0/:: which works fine with no network up
2019-01-18 02:35:35 <jwh> but even so, both boxes have valid ptrs for their configured addresses <10msec response
2019-01-18 02:36:19 <jwh> is there some sort of debug i can enable for openrc that will show dependency resolving and such?
2019-01-18 02:36:26 <jwh> without flooding the console
2019-01-18 02:37:01 <dalias> i don't know. i'm just a user (and musl author:) not an expert on alpine init system
2019-01-18 02:37:36 <jwh> ah, worth a try :)
2019-01-18 02:42:20 <jwh> nuked the dns depend, isn't that
2019-01-18 02:47:13 <jwh> heh, was rng
2019-01-18 02:47:16 <jwh> guess both boxes suck
2019-01-18 02:47:28 <jwh> or something is broken
2019-01-18 02:48:15 <dalias> how did you test it?
2019-01-18 02:48:27 <jwh> enable haveged
2019-01-18 02:49:11 <jwh> neither probably has RDRAND and maybe theres not enough entropy from other sources before sshd whats to start
2019-01-18 02:52:36 <jwh> yeah, so theres a ~30 second delay before crnf init completes without haveged
2019-01-18 02:52:58 <jwh> at least on this VM
2019-01-18 02:54:51 <jwh> bit of googling suggests entropy sources are limited these days
2019-01-18 02:56:51 <jwh> I mean I should probably run haveged anyway so I don't need to wait all day for key generation and such, but 30 seconds or more seems excessive
2019-01-18 02:57:16 <dalias> haveged sounds bogus
2019-01-18 02:57:28 <dalias> you can't magically make entropy and the kernel already collects it from valid sources
2019-01-18 02:57:44 <jwh> its really just low quality entropy filling
2019-01-18 02:58:01 <dalias> you can save old entropy across reboots to solve this tho
2019-01-18 02:58:06 <jwh> specifically for systems with limited entropy sources (like embedded ones)
2019-01-18 02:58:34 <jwh> but it does make gpg take a reasonable amount of time to generate a key heh
2019-01-18 02:58:34 <dalias> but i wonder what the actual mechanism stalling it here is. sshd should *start* just fine without urandom ready, but take a while to respond to connections
2019-01-18 02:58:46 <jwh> yeah thats what I would expect
2019-01-18 02:59:08 <jwh> unless... its using /dev/random and because its blocking, it's stalling startup
2019-01-18 02:59:08 <dalias> so i wonder if there's some broken init-system rule making it wait for /dev/random
2019-01-18 02:59:13 <jwh> yeah
2019-01-18 02:59:22 <dalias> even though sshd doesn't need to do that
2019-01-18 02:59:38 <dalias> (maybe in case there are no keys yet and it needs to do keygen on first run?)
2019-01-18 03:00:15 <jwh> thats what I initially thought but since it seems to block key generation too, that works fine after a whle
2019-01-18 03:00:21 <jwh> while*
2019-01-18 03:00:36 <jwh> looks like it's in the init scripts, or openrc, rather than sshd
2019-01-18 03:01:50 <dalias> it looks like by default it's saved/restored as /var/lib/misc/random-seed
2019-01-18 03:02:04 <jwh> I wonder...
2019-01-18 03:02:15 <dalias> in the urandom service
2019-01-18 03:02:37 <dalias> does rc-update show "urandom" service in "boot" columnb?
2019-01-18 03:02:40 <jwh> both of these were "installed" from the alpine-make-vm image that the script creates
2019-01-18 03:03:09 <dalias> i wonder if that does something bogus different from what the normal installer does
2019-01-18 03:03:13 <dalias> (e.g. omitting this service)
2019-01-18 03:03:25 <jwh> I just checked and theres no output from urandom, so I'm guesing its not enabled when it should be
2019-01-18 03:04:16 <jwh> yeah, ok lets see
2019-01-18 03:06:38 <jwh> hm, urandom doesn't seem to make any diffrence - added to boot but still nothing during init
2019-01-18 03:13:37 <dalias> well it needs to have had a chance to save the old entropy before it can do anything
2019-01-18 03:15:27 <jwh> yeah
2019-01-18 03:17:57 <jwh> obviously it also needs some entropy to save...
2019-01-18 03:18:19 <jwh> if the VM isn't generating any that the kernel is using (no i/o devices etc)
2019-01-18 03:20:21 <jwh> entirely possible that there just isn't enough to seed it on boot
2019-01-18 03:21:42 <jwh> arch box suffers the same problem but systemd doesn't wait for it
2019-01-18 03:21:48 <jwh> (same hardware)
2019-01-18 03:22:40 <jwh> question is why sshd (or init) is waiting, but dropbear doesn't
2019-01-18 03:24:00 <dalias> well it's needed to make safe transient/session keys
2019-01-18 03:24:16 <dalias> but i'm not sure why it doesn't do the wait in the background
2019-01-18 03:25:27 <jwh> yeah, it could start to let init finish but just not exchange keys on new sessions until its ready
2019-01-18 03:26:04 <jwh> but if the entropy availability is so terrible I'm gonna need something to refill it anyway
2019-01-18 03:26:36 <jwh> if its only entropy source is disk i/o it isn't gonna have much of that either
2019-01-18 03:28:31 <jwh> hm, physical box has rdrand - can the kernel not use that on its own?
2019-01-18 03:28:56 <dalias> not if the hypervisor doesn't expose it; i'm not sure if it's even allowed to
2019-01-18 03:29:19 <dalias> there's a virtio driver by which the hypervisor can expose its own kernel entropy source to the guest
2019-01-18 03:29:21 <jwh> yeah the VM is a lost cause for useful entropy
2019-01-18 03:29:31 <jwh> but this physical box that exhibits an identical problem has rdrand
2019-01-18 03:29:34 <dalias> but that requires trusting that it's not intentionally nerfed
2019-01-18 03:29:41 <dalias> (which would be trivial for a host to do)
2019-01-18 03:29:45 <jwh> yah
2019-01-18 03:30:59 <jwh> it would appear the kernel will not use rdrand on its own, patches exist but bleh
2019-01-18 03:31:29 <jwh> so that may explain why the same problem exists since although its physical it probably has marginally more entropy available than the VM (no i/o either)
2019-01-18 03:32:14 <jwh> although crappy entropy isn't ideal, I'm not sure blocking is either
2019-01-18 03:33:36 <jwh> especially as it doesn't inform the user so others may end up down the same rabbithole figuring out why its broken
2019-01-18 03:33:51 <jwh> not an ideal user experience
2019-01-18 03:34:45 <m4t> rng-tools can use rdrand
2019-01-18 03:34:54 <m4t> on debian anyhow... ;)
2019-01-18 03:35:22 <jwh> yeah I was just considering that, but still sucky
2019-01-18 03:35:36 <m4t> nah, works fine
2019-01-18 03:35:53 <m4t> tbh haveged works fine too, though the nsa hasn't tried to haxx me yet (to my knowledge)
2019-01-18 03:36:05 <jwh> really the kernel should either default to, or at least have the option to just use rdrand (at least during boot) if no other entropy is available and userspace is wanting randomness
2019-01-18 03:36:19 <jwh> shouldn't need to start userland to... start userland
2019-01-18 03:36:19 <jwh> :D
2019-01-18 03:37:05 <jwh> theres likely enough entropy at boot time to at least partially seed it with that, but fill enough to service boot time requirements from rdrand
2019-01-18 03:37:15 <jwh> it can always be rekeyed as soon as more is available
2019-01-18 03:38:16 <m4t> it is kinda silly to trust intel cpus to run the OS but not seed entropy lol
2019-01-18 03:38:24 <jwh> just a little
2019-01-18 03:38:45 <jwh> and really, most users to not need ultra safe randomness sources
2019-01-18 03:38:59 <jwh> those that do (for crypto ops etc) can configure appropriately
2019-01-18 03:39:23 <jwh> I don't think I need an ultra safe rng to watch TV :D
2019-01-18 03:40:21 <jwh> but on my production servers I obviously wouldn't want pseudo randomness or to use just one source of entropy
2019-01-18 03:40:42 <jwh> but then they have rdrand, plus other sources, and those that really need it I can add another hardware rng
2019-01-18 03:42:34 <jwh> I actually thought the kernel considered rdrand these days, it appears it does but only when other sources are also available (or maybe that patch is still pending)
2019-01-18 03:48:41 <jwh> at least virtio has the option for an rng (disregarding the host potentially manipulating it), vmware has *nothing*
2019-01-18 03:48:58 <jwh> no devices to gather entropy from, no rng, not even emulated rdrand or anything useful
2019-01-18 03:49:56 <jwh> it at least appears to be an option in newer vmware at least, guess the cpu doesn't have it anyway
2019-01-18 04:52:43 <dalias> m4t, its not. difference jn detectability and deniability is huge
2019-01-18 04:53:33 <m4t> why trust any hw rng then?
2019-01-18 04:54:09 <m4t> there's a semirecent commit that makes tpm-rng built-in and directly feeds the kernel from any number of tpm chips
2019-01-18 04:54:33 <m4t> including infineon, who had the fiasco with weak key generation
2019-01-18 04:55:02 <m4t> not saying it should be the default. but it should be available e.g. with cmdline param
2019-01-18 04:55:54 <m4t> we can forcefully turn off mitigations for meltdown/l1tf, but not feed entropy from rdrand
2019-01-18 04:55:56 <m4t> ¯\_(ツ)_/¯
2019-01-18 05:03:34 <dalias> m4t, indeed dont. instead collect entropy and feed it to deterministic algorithms
2019-01-18 05:56:11 <terra> After last clang-5.0.2 upgrade, it seems compiler is not usable on armhf.
2019-01-18 05:57:08 <terra> https://build.alpinelinux.org/buildlogs/build-edge-armhf/testing/mame/mame-0.205-r0.log
2019-01-18 05:57:48 <terra> https://build.alpinelinux.org/buildlogs/build-edge-armhf/testing/advancemame/advancemame-3.8-r1.log
2019-01-18 05:58:34 <terra> clang-5.0: error: unable to execute command: Segmentation fault
2019-01-18 05:59:53 <terra> Oh, this is not #alpine-devel
2019-01-18 14:50:23 <fiftysix> hi!
2019-01-18 14:50:55 <fiftysix> i just did apk upgrade without doing my homework first, and now there's a version mismatch between docker and shorewall that i must fix quickly.
2019-01-18 14:51:04 <fiftysix> is there a way to downgrade docker to what it was before?
2019-01-18 14:52:10 <fiftysix> (or even find out what that was? i don't have apk caching enabled and there don't seem to be logs :( )
2019-01-18 14:52:18 <mps> fiftysix: something like 'apk add docker=version', forgot exactly
2019-01-18 14:52:50 <mps> there is wiki page about that on alpinelinux.org
2019-01-18 14:54:46 <clandmeter> Did u use upgrade -a
2019-01-18 14:55:23 <clandmeter> Repo only supports one version of a package
2019-01-18 14:55:36 <fiftysix> no, "upgrade"; didn't know about "-a"
2019-01-18 14:56:36 <fiftysix> also, just inadvertently clicked on the "Flag" button of https://pkgs.alpinelinux.org/package/v3.6/main/x86_64/shorewall and hit a 404. probably not how you intended it.
2019-01-18 14:57:11 <clandmeter> You cannot flag stable
2019-01-18 14:58:40 <fiftysix> i guess then you might want a message to appear saying "you cannot flag stable"
2019-01-18 14:58:48 <fiftysix> or for the button to be not there, or something
2019-01-18 14:59:55 <fiftysix> anyways, http://dl-cdn.alpinelinux.org/alpine/latest-stable/community/x86_64/ only shows the *latest* packages ... is there a list of older, available ones, too?
2019-01-18 15:56:51 <geosmin> so i'm split 50/50 on whether i'm encountering some strange bug or whether i'm being an idiot
2019-01-18 15:57:40 <geosmin> serial console is disabled everywhere i know of yet i'm still getting a tty on /dev/ttyAMA0 on rpi3
2019-01-18 15:58:34 <geosmin> trying to use the serial connection to interface with a piece of equipment but getting a login prompt on that line which is muddying the comms
2019-01-18 15:59:14 <geosmin> uart is disabled, no tty setup, etc. across cmdline.txt and config.txt
2019-01-18 16:01:14 <geosmin> FWIW have bluetooth disabled with dtoverlay=pi3-disable-bt
2019-01-18 16:03:38 <geosmin> dmesg also seems to constantly be flooding with 'ttyS ttyS0: tty_port_close_start: tty->count = 1 port count = 2 ' and i'm getting an I/O error whenever i try to do anything with ttyS0
2019-01-18 17:10:04 <divx118> Hi, just a question will the mpd package https://pkgs.alpinelinux.org/packages?name=mpd&branch=edge be updated to 0.21.4 ?
2019-01-18 19:12:17 <kxzaon> Hello guys. When I try to make a sys install on HDD, I can't boot after installation. In the emergency iniramfs shell, I have to mount /dev/sda2 on /sysroot to boot properly. I can't find many references about sysroot on boot. Can you help me?
2019-01-18 19:33:46 <fiftysix> did alpine at any point in the last 3 months remove physdev matching from kernel or iptables?
2019-01-18 19:34:39 <_ikke_> 0
2019-01-18 19:36:26 <fiftysix> _ikke_, 0? anyways, the combination of shorewall+docker depends on physdev matching. it's not working anymore after today's apk upgrade.
2019-01-18 19:37:05 <fiftysix> and "alpine physdev" are not helpful google keywords :(
2019-01-18 19:37:21 <_ikke_> fiftysix: accident
2019-01-18 19:37:37 <fiftysix> oh
2019-01-18 19:37:44 <_ikke_> I know nothing about physdev, so can't help you
2019-01-18 19:37:47 <fiftysix> :-(
2019-01-18 19:37:56 <fiftysix> and you're an admin in here...
2019-01-18 19:37:59 <fiftysix> _ikke_, who would?
2019-01-18 19:38:26 <_ikke_> Some people here, but it's weekend for most of them
2019-01-18 19:38:35 <fiftysix> :(
2019-01-18 19:38:42 <fiftysix> not for me now, because of this
2019-01-18 19:39:19 <_ikke_> sad to hear
2019-01-18 19:39:44 <_ikke_> is it a kernel module?
2019-01-18 19:39:48 <_ikke_> or how does it work?
2019-01-18 19:40:47 <fiftysix> _ikke_, good question! so far, i know that "shorewall show capabilities | grep Physdev" tells me whether i have it or not.
2019-01-18 19:41:01 <fiftysix> apparently it must be compiled into both kernel and iptables.
2019-01-18 19:41:17 <fiftysix> also see http://shorewall.org/Docker.html
2019-01-18 19:41:20 <_ikke_> You can check the aports repo for changes in both
2019-01-18 19:42:48 <fiftysix> especially weird since i'm on 3.8 ... which shouldn't suffer such changes, right?
2019-01-18 19:43:03 <_ikke_> correct
2019-01-18 19:43:53 <fiftysix> https://git.alpinelinux.org/aports/log/?h=3.8-stable&qt=grep&q=physdev comes back empty ... that's what you meant?
2019-01-18 19:52:20 <fiftysix> omg
2019-01-18 19:52:24 <fiftysix> _ikke_, i apologize!
2019-01-18 19:52:29 <fiftysix> after a reboot all is fine!!!
2019-01-18 20:00:01 <_ikke_> heh
2019-01-18 20:00:13 <_ikke_> So probably kernel ugrade and the module wasn't loaded yet
2019-01-18 20:00:31 <_ikke_> so it cannot find the module when needed
2019-01-18 20:15:04 <fiftysix> both kernel upgrade and iptables upgrade at once i think
2019-01-18 20:17:20 <_ikke_> perhaps
2019-01-18 20:27:53 <fiftysix> anyways, sorry for causing a ruckus and thanks for still listening to me being somewhat of a madman. also, happy weekend! :)
2019-01-18 20:42:26 <_ikke_> no worry
2019-01-18 20:47:21 <mchasard> hello
2019-01-18 20:47:55 <_ikke_> hi
2019-01-18 20:48:27 <mchasard> i hear about alpine linux but never install it
2019-01-18 21:28:34 <mcr1> the udhcpc package does not include /usr/share/udhcpc/default.script to actually configure the address.  What package might that be in?
2019-01-18 21:29:00 <mcr1> it seems that udhcpc came with busybox, which I suspect might be a mistake.
2019-01-18 21:29:27 <_ikke_> https://pkgs.alpinelinux.org/contents?file=&path=%2Fusr%2Fshare%2Fudhcpc&name=&branch=edge&arch=x86_64
2019-01-18 21:32:36 <mcr1> thank you.  I am using the container mini-container, and adding things as I want a very minimal container.
2019-01-18 21:32:59 <mcr1> er, mini-root-filesystem.
2019-01-18 21:33:50 <mcr1> what is the official way with openrc to mark init.d scripts to be run?
2019-01-18 21:40:52 <_ikke_> rc-update add <service>
2019-01-18 22:05:48 <mcr1> thank you!
2019-01-19 00:39:00 <elxa> docker alpine edge seems unavailable for arm?
2019-01-19 00:41:22 <elxa> maybe my docker is buggy
2019-01-19 00:42:09 <elxa> it's falling back to 3.7 https://dpaste.de/Hean
2019-01-19 00:59:48 <danieli> sounds like a docker issue
2019-01-19 01:03:42 <danieli> ncopa: might know, he works there
2019-01-19 01:22:18 <ncopa> elxa: can you please report that on bugs.alpinelinux.org and i`ll have a look at it when i get a chance
2019-01-19 01:22:36 <elxa> ok
2019-01-19 01:27:02 <elxa> I am not getting my restistration confirmation mail :(
2019-01-19 01:30:06 <ncopa> clandmeter: can you please help us with ^^^ when convenient?
2019-01-19 01:30:17 <ncopa> he probably asleep now
2019-01-19 01:33:02 <elxa> it seems to be something about this raspberry pi 3 architecture
2019-01-19 01:33:34 <elxa> the same dockerfile that pulls 3.7 on the pi pulls edge on my desktop
2019-01-19 01:34:27 <elxa> FROM alpine:edge
2019-01-19 01:34:30 <elxa> should be good enough no?
2019-01-19 01:43:03 <elxa> with alpine:3.8 I get 3.8, with alpine:edge I get 3.7 :D
2019-01-19 09:56:17 <clandmeter> elxa, pm me your username and ill check
2019-01-19 11:31:49 <elxa> clandmeter: konstantin.blaesi
2019-01-19 11:38:13 <clandmeter> elxa, we are experiencing mail issues
2019-01-19 11:38:27 <elxa> I noticed :)
2019-01-19 11:38:30 <clandmeter> issue is resolved but could take some time to really clear up
2019-01-19 11:45:50 <clandmeter> elxa, acc activated
2019-01-19 11:45:56 <elxa> ty
2019-01-19 11:46:06 <clandmeter> yw
2019-01-19 13:10:48 <elxa> created the ticket https://bugs.alpinelinux.org/issues/9873
2019-01-19 13:54:38 <unrznbl[m]> would it make sense at all to make a speakup-linux-vanilla package? to enable speakup (https://git.alpinelinux.org/aports/tree/main/linux-vanilla/config-vanilla.x86#n6281) , and probably have a default dependency on an espeakup package? (https://github.com/williamh/espeakup) This is something I intend on working on for postmarketOS eventually but have an x86 laptop I'd like to play with now.
2019-01-19 13:55:28 <unrznbl[m]> or would I just make a change request to make that a module by default in linux-vanilla?
2019-01-19 15:53:20 <mps> unrznbl[m]: less complicated is to enable module in kernel and make espeakup package, imo
2019-01-19 15:57:41 <unrznbl[m]> Cool. Making speakup built in has only a slight advantage of earlier loading maybe. I can test. Can an espeakup package depend on loading a module?
2019-01-19 16:00:30 <mps> if the module is in kernel, then not, I think. It should depend on linux-vanilla
2019-01-19 16:01:13 <_ikke_> should it? If another kernel package is installed, I don't think it should pull in linux-vanilla
2019-01-19 16:03:31 <mps> hmm, if the speakup module enabled in kernel (linux-vanilla) then I think that is ok for package which need that module to depend on linux-vanilla
2019-01-19 16:04:34 <mps> if the speakup module is in separate package then of course not
2019-01-19 16:05:38 <_ikke_> but there may be more than one kernel that has this module enabled
2019-01-19 16:05:39 <mps> but if the separate linux-vanilla-speakup package is made, that package will depend on linux-vanilla
2019-01-19 16:07:38 <mps> _ikke_: right, I'm speaking in general. OP could look in some of modules with their packages (drbd, for example) how that is done
2019-01-19 16:08:04 <unrznbl[m]> mps: ah. Good. An example! Will check it out. Thanks.
2019-01-19 16:08:20 <mps> by OP I mean unrznbl[m], sorry for impolite write
2019-01-19 16:20:07 <unrznbl[m]> mps: No problem at all. Really appreciate the feedback. :) am trying to get a small no-screen laptop (thinkpad 240) setup for fosdem trip. (Screen is broken).
2019-01-19 16:21:45 <unrznbl[m]> And using a computer w/o a screen is one of my goals. Packaging is just interesting thing I need to learn for pmos. :)
2019-01-19 16:25:16 <mps> unrznbl[m]: jokes aside, but I have some number of computers without screen/display and keyboards/mices. but serial consoles :)
2019-01-19 16:38:28 <unrznbl[m]> Ah. Right. For me though I want to use them without any other devices so need some output mechanism. Either voice or morse code are my current ideas.
2019-01-19 16:39:14 <unrznbl[m]> But this tp240 has a serial port so that might be a good choice too as a backup instead of a monitor.
2019-01-19 16:42:17 <mps> I also thought to add morse code for output
2019-01-19 16:42:35 <mps> mostly for console messages
2019-01-19 16:45:00 <unrznbl[m]> Nice! Unixcw is a pretty good console option for text to alsa audio as far as i know. haven't used it a ton.
2019-01-19 16:47:48 <mps> yes, I'm packaging in private repo some of the ham software for personal use, and thought to make some of the morse code 'players'
2019-01-19 22:23:36 <unrznbl[m]> mps: "private repo" do you mean like your own aports somewhere local that you build packages to and then isntall from those packages?
2019-01-19 22:24:44 <unrznbl[m]> I would try to package-ize things like unixcw and espeakup if that makes sense. Wasn't sure how curated the main aports was.
2019-01-19 22:44:23 <mps> unrznbl[m]: no clone of aports, but just dir (git inited) with subdirs of packages which I think is not yet ready to be sent to Alpine aports or packages which I doubt will be useful for Alpine users
2019-01-19 22:45:56 <unrznbl[m]> ah right, becuase you can use abuild in any-old-dir with an APKBUILD file right? Sounds good. I'm following these instructions right now to build a new kernel: https://wiki.alpinelinux.org/wiki/Custom_Kernel. :)
2019-01-19 22:51:57 <mps> unrznbl[m]: keep in mind that is good idea to have that dir git init-ed
2019-01-19 22:58:13 <unrznbl[m]> yup.
2019-01-20 05:47:23 <aGaTHoS> why manuals are in a separated package?
2019-01-20 05:49:23 <m4t> to save space i believe
2019-01-20 05:49:42 <m4t> if you're using a minimal container doesn't make sense to have manpages :/
2019-01-20 05:59:09 <aGaTHoS> there should be a way to isntall automatically man pages
2019-01-20 05:59:45 <aGaTHoS> like a var or something
2019-01-20 06:15:17 <aGaTHoS> and i can never expect to nvidia propietary drivers being compiled against musl right?
2019-01-20 06:28:21 <m4t> those are compiled against the kernel, idk
2019-01-20 06:29:11 <m4t> most of the userland stuff like gl library, nvidia-settings, etc., would come linked against glibc
2019-01-20 06:30:36 <m4t> maybe something like this could help https://github.com/sgerrand/alpine-pkg-glibc
2019-01-20 06:31:15 <dalias> you should use the open src drivers bc theyre better
2019-01-20 06:31:47 <dalias> however supposedly the proprietary ones can work due to abi compat
2019-01-20 06:32:28 <dalias> glibc pkg will not help
2019-01-20 06:33:04 <dalias> bc the driver has to be loaded in programs that are using musl not glibc
2019-01-20 06:34:26 <dalias> that package also looks very very broken
2019-01-20 06:34:32 <dalias> eg https://github.com/sgerrand/alpine-pkg-glibc/blob/master/ld.so.conf
2019-01-20 06:36:24 <aGaTHoS> better?
2019-01-20 06:36:29 <aGaTHoS> dont think so
2019-01-20 06:37:05 <aGaTHoS> nvidia propietary drivers are always best and they make linux version...
2019-01-20 06:37:30 <aGaTHoS> in the gnu linux world nobody is gonna make free opensource better graphic drivers
2019-01-20 06:37:56 <aGaTHoS> in the past linux was greater than know because some people dont want to evolve
2019-01-20 06:44:01 <aGaTHoS> moreover i even cant use nouveau drivers
2019-01-20 06:44:09 <aGaTHoS> i found i think the problem
2019-01-20 06:44:34 <aGaTHoS> just the kernel had to be compiled with an option i filled the bug request time ago but nobody did
2019-01-20 06:45:15 <aGaTHoS> there was like a vesa  mode to be added to the kernel so it dont hangs when load nouveau dri or something like that i remember it was
2019-01-20 06:45:35 <aGaTHoS> im using nv
2019-01-20 06:45:40 <aGaTHoS> with x11
2019-01-20 06:45:55 <aGaTHoS> but it would be great if i could use propietary drivers
2019-01-20 06:46:43 <aGaTHoS> i dont know how to start with that alpine-pkg-glibc
2019-01-20 06:47:20 <aGaTHoS> its supposed i have to install manually nvidia propietary drivers then make opengl etc.. make to work with glibc apk?
2019-01-20 06:47:39 <aGaTHoS> if i install this package how is decided when a program uses glibc or musl?
2019-01-20 06:53:32 <dalias> the glibc apk is invalid and will not help you
2019-01-20 06:55:12 <dalias> what is it you want to switch to the proprietary driver for?
2019-01-20 06:55:49 <dalias> if its commercial games for linux theyll also be for a glibc ecosystem...
2019-01-20 06:56:09 <aGaTHoS> blender
2019-01-20 06:56:09 <dalias> and not work so well with alpine
2019-01-20 06:56:15 <dalias> ah
2019-01-20 06:56:42 <dalias> well i think ppl have made the driver work
2019-01-20 06:57:01 <dalias> it might require some binary patching
2019-01-20 06:57:10 <aGaTHoS> im googling and not finding anything
2019-01-20 06:57:15 <dalias> having glibc will not help you at all
2019-01-20 06:57:23 <aGaTHoS> what should i try?
2019-01-20 06:57:46 <dalias> see if someone who's done it is here
2019-01-20 06:58:20 <aGaTHoS> if someone have done maybe could write a howto in alpine wiki or something
2019-01-20 07:07:30 <dalias> kaniini maybe
2019-01-20 08:17:49 <aGaTHoS> im trying to update the megatools package and #abuild -r :>>> ERROR: megatools: builddeps failed
2019-01-20 08:18:04 <aGaTHoS> without any else indication
2019-01-20 09:05:43 <danieli> aGaTHoS: it will not be easy to make the proprietary nvidia driver work on alpine, and unless nvidia open sources it (or prebuild it on alpine), it will never make it into alpine
2019-01-20 09:05:55 <danieli> using glibc will indeed not help you much at all
2019-01-20 09:06:01 <danieli> there's more magic to it
2019-01-20 09:06:21 <danieli> it sounds like you're going down a very deep rabbit hole, so be careful
2019-01-20 09:06:54 <aGaTHoS> ?
2019-01-20 09:09:20 <danieli> ?
2019-01-20 09:24:04 <aGaTHoS> well but appart from nvidia-drivers
2019-01-20 09:24:25 <aGaTHoS> why i have just updated the package version in megatools and does not buld
2019-01-20 09:24:27 <aGaTHoS> build
2019-01-20 09:24:41 <aGaTHoS> >>> megatools: Analyzing dependencies...
2019-01-20 09:24:43 <aGaTHoS> >>> ERROR: megatools: builddeps failed
2019-01-20 09:24:46 <aGaTHoS> just says that
2019-01-20 13:08:48 <mi_schm> hi, is anyone using i3? if so, i get an error message status_commend not found or missing a library dependency.
2019-01-20 13:49:27 <elxa> ncopa: fyi https://bugs.alpinelinux.org/issues/9873
2019-01-20 18:26:40 <danieli> shame mi_schm left, they haven't installed i3status
2019-01-20 18:27:46 <danieli> elxa: it might be that the edge image is simply using the 3.7 repos, i think i heard of something along those lines not long ago
2019-01-20 19:07:14 <elxa> danieli: I switched to edge mostly to get a newer g++ supporting newer c++ standards as seen here https://pkgs.alpinelinux.org/packages?name=g%2B%2B&branch=edge
2019-01-20 19:08:14 <elxa> danieli: when I build the Dockerfile on my desktop with intel cpu I get edge. When I build the same Dockerfile on the raspberry pi I get 3.7 ...
2019-01-20 19:08:22 <danieli> that's stragne
2019-01-20 19:08:23 <danieli> strange*
2019-01-20 19:08:43 <danieli> and you're sure it's 3.7 and not just edge with a 3.7 repositories file?
2019-01-20 19:10:54 <elxa> what do you mean by that? I wrote my first Dockerfile, it's possible I did something wrong. I specified "FROM alpine:edge" and ran "docker build -t logger ." from the directory of the Dockerfile
2019-01-20 19:11:15 <elxa> on my desktop I see logs like
2019-01-20 19:11:18 <elxa> fetch http://dl-cdn.alpinelinux.org/alpine/edge/main/x86_64/APKINDEX.tar.gz
2019-01-20 19:11:27 <elxa> on the pi
2019-01-20 19:11:28 <elxa> fetch http://dl-cdn.alpinelinux.org/alpine/v3.7/main/armhf/APKINDEX.tar.gz
2019-01-20 19:12:11 <elxa> danieli: btw when I specifiy "FROM alpine:edge" the pi does pull 3.8 ...
2019-01-20 19:12:19 <elxa> erm sry I meant "FROM alpine:3.8"
2019-01-20 19:14:07 <elxa> packages being installed like "(33/36) Installing g++ (6.4.0-r5)" suggest it's not edge so ...
2019-01-20 19:15:04 <danieli> is the actual image 3.7, or is it an edge image that uses 3.7 repositories?
2019-01-20 19:15:09 <danieli> I'm not sure how else to explain it
2019-01-20 19:17:03 <clandmeter> https://github.com/gliderlabs/docker-alpine/blob/master/versions/library-edge/armhf/options
2019-01-20 19:17:53 <danieli> right, so it is an actual issue with the image, that's a shame
2019-01-20 19:17:55 <elxa> clackety: +1
2019-01-20 19:18:01 <elxa> whops
2019-01-20 19:18:11 <elxa> clandmeter: ^ iorc autocomplete fail
2019-01-20 19:18:13 <elxa> *irc
2019-01-20 19:20:24 <elxa> all except x86_64 are on 3.8, reasons?
2019-01-20 19:20:32 <elxa> erm 3.7 I meant
2019-01-20 19:21:10 <clandmeter> There is an related pr
2019-01-20 19:28:45 <elxa> if I want newer toolchains (g++) than available on 3.8 the only way to get it with apk is by using edge instead of 3.8 right?
2019-01-20 19:29:45 <clandmeter> Edge is the latest
2019-01-20 19:30:00 <clandmeter> Or you can try 3.9
2019-01-20 19:30:57 <clandmeter> It's the same as edge
2019-01-20 19:31:12 <clandmeter> For now
2019-01-20 19:31:27 <elxa> manifest for docker.io/alpine:3.9 not found
2019-01-20 19:31:32 <elxa> doesn't seem like that'll work
2019-01-20 19:31:45 <clandmeter> No it's not yet released
2019-01-20 19:32:05 <clandmeter> You can use 3.8 and upgrade it
2019-01-20 19:34:53 <elxa> I see, something like this https://wiki.alpinelinux.org/wiki/Upgrading_Alpine
2019-01-20 19:36:58 <clandmeter> change repo file to 3.9 and apk -U upgrade -a
2019-01-20 20:34:38 <mps> anyone know where in wiki CBUILD, CHOST, CTARGET etc are described/explained
2019-01-20 21:23:55 <m4t> jwh: actually i'm going through 4.20.3 menuconfig right now and see this CONFIG_RANDOM_TRUST_CPU and random.trust_cpu=on/off
2019-01-20 21:37:36 <m4t> i remember reading this at the time but forgot https://lwn.net/Articles/760121/
2019-01-20 21:38:33 <jwh> I dont think I'm the person you're looking for, unless I've forgot a conversation :D
2019-01-20 21:38:41 <jwh> oh
2019-01-20 21:39:07 <jwh> so it got merged then, thats cool
2019-01-20 21:39:58 <jwh> bit of a pain in the ass to have to rebuild kernels though, a cmdline option would be nicer
2019-01-20 21:40:12 <jwh> that is, compile it in but explicitly enable it with cmdline option
2019-01-20 21:44:21 <m4t> it doesn't appear to be wrapped in ifdef
2019-01-20 21:44:42 <m4t> er
2019-01-20 21:44:44 <jwh> ohhh, so its not the original patch?
2019-01-20 21:45:23 <jwh> I think its a bit irresponsible to say cpus are any more at risk from backdoors than any other hardware RNG
2019-01-20 21:45:34 <jwh> but people keep bringing politics into it
2019-01-20 21:45:53 <m4t> looking at drivers/char/random.c its different
2019-01-20 21:46:43 <jwh> hm
2019-01-20 21:47:06 <m4t> i think that might determine the default value, but not prevent it from being set via cmdline
2019-01-20 21:47:54 <neplevitan> Hey there, does alpine linux support efi boot?
2019-01-20 21:48:00 <m4t> https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/drivers/char/random.c?h=v4.20.3&id=9b25436662d5fb4c66eb527ead53cab15f596ee0
2019-01-20 21:48:37 <neplevitan> I am asking since i tried both the 3.8.2 standard and extended versions (copied to a usb drive, the .iso file) and both fail at the grub stage
2019-01-20 21:48:55 <neplevitan> (as in, i get a grub rescue prompt to play around with)
2019-01-20 21:50:04 <jwh> m4t: oh, I guess they had exactly the same objection to a compile time option as I did
2019-01-20 21:50:23 <jwh> thats much better
2019-01-20 21:50:43 <m4t> need 4.19+ it seems
2019-01-20 21:51:55 <jwh> yeah
2019-01-20 21:53:21 <jwh> thanks
2019-01-20 21:53:55 <jwh> doesn't help for machins without cpu rng but nothing to be done there
2019-01-20 21:54:05 <m4t> anything reasonably modern has it
2019-01-20 21:54:14 <m4t> sandy bridge+ i think
2019-01-20 21:54:41 <m4t> nope ivy bridge
2019-01-20 21:55:02 <jwh> yeah, they were left out of the earlier low end socs though
2019-01-20 21:55:13 <jwh> as they do... features tend to trickle down after  a couple of generations
2019-01-20 21:56:10 <jwh> I'm busy figuring out what ghetto PoE kit to buy atm, to supply 5V to these atom sticks and stuff heh
2019-01-20 22:00:40 <PureTryOut[m]> neplevitan: I noticed the same thing today, I guess efi is not supported
2019-01-20 22:01:26 <jwh> I thought there was EFI support?
2019-01-20 22:01:54 <jwh> I did just this the other day, cant remember if the iso booted or not with efi now
2019-01-20 22:02:49 <neplevitan> Well, in efi mode the iso fails to boot, with bios mode i dont manage to get efi vars :g
2019-01-20 22:02:50 <jwh> the other option is to just create an image, bit of losetup -P, bootstrap alpine
2019-01-20 22:03:05 <jwh> then dd it somewhow
2019-01-20 22:03:06 <jwh> somehow*
2019-01-20 22:03:29 <neplevitan> Eh, my one machine is haiku, the other one is FreeBSD, i doubt i can just create an image easily ;)
2019-01-20 22:03:42 <neplevitan> I suppose i can try an older release and update once it is installed
2019-01-20 22:03:45 <jwh> actually you should be able to do it on either
2019-01-20 22:03:56 <jwh> well, maybe not haiku but freebsd
2019-01-20 22:04:15 <neplevitan> i've always had huge trouble trying to compile linux on FreeBSD
2019-01-20 22:04:30 <jwh> mdconfig for freebsd instead of losetup, linux emu for static apk-tools and chroot after
2019-01-20 22:04:43 <jwh> just use packages...
2019-01-20 22:05:12 <neplevitan> in that case i don't know what you mean by just creating an image
2019-01-20 22:07:19 <neplevitan> regardless, efi support is broken on 3.8.2, i'll try to use an older release then
2019-01-20 22:07:35 <neplevitan> (and hopefully be able to use a efistub kernel afterwards to bypass grub :D)
2019-01-20 22:15:11 <neplevitan> Well, tried 3.8.0 same problem
2019-01-20 22:15:26 <neplevitan> tried 3.6.0 and it didnt even boot grub
2019-01-20 22:22:54 <mps> neplevitan: try v3.9_rc2, it should have efi support
2019-01-20 22:23:15 <neplevitan> mps: I'd rather not run a rc :/
2019-01-20 22:23:52 <mps> it is next stable in few days :)
2019-01-20 22:24:30 <neplevitan> is it the final rc then?
2019-01-20 22:25:09 <mps> well, not sure 100% but is nearing that state
2019-01-20 22:25:27 <neplevitan> fine, i will try it
2019-01-20 22:26:20 <mps> I run it for days on some boxes without serious issue, although I must say I didn't tried efi, it is added and tested by someone in last few days
2019-01-20 22:27:06 <mps> but, beside that efi, it is stable in my usage on different boxes
2019-01-20 22:27:58 <mps> I run armv7 for 60 days, and it is new arch
2019-01-20 22:28:48 <neplevitan> huh, actually boots :)
2019-01-20 22:41:38 <neplevitan> hmm, it sais Installation finished, and now has a progress bar below that, what is the purpose of that?
2019-01-20 22:41:46 <neplevitan> it's moving verry slowly
2019-01-20 22:53:02 <neplevitan> anyway, finished installing
2019-01-20 22:53:09 <neplevitan> have a good day :)
2019-01-20 23:33:46 <elxa> I am trying to install flatbuffers using alpine v3.9 https://pkgs.alpinelinux.org/packages?name=flatbuffers&branch=edge
2019-01-20 23:33:51 <elxa> it failed with ERROR: unsatisfiable constraints:
2019-01-20 23:34:05 <elxa> flatbuffers (missing): required by: world[flatbuffers]
2019-01-20 23:34:10 <elxa> from my apk add command
2019-01-20 23:34:28 <elxa> do I need to enable that testing repo ?
2019-01-20 23:37:17 <elxa> found it https://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management#Packages_and_Repositories
2019-01-21 01:09:08 <jwh> hm, wonder why reboots are broken on this intel stick... only with alpine though, arch on the same reboots ok - weird graphics corruption but it doesn't reboot with text/nomodeset either
2019-01-21 01:18:07 <jwh> hm, works if I don't use efi to reboot, guess theres some difference in config/build
2019-01-21 02:14:30 <aGaTHoS> can someone help me with the compile of a package?
2019-01-21 02:15:18 <dimon222[m]> what are you trying to compile?
2019-01-21 02:18:23 <aGaTHoS> some time ago i did a apk for megatools , now that i have tried to solve more problems like try to get nvidia propietary drivers work etc.. i have read again how to make packages for alpine and end wanted to update also megatools package, i just changued version updated the checksum and tried to compile and get an error not even very verbose
2019-01-21 02:18:35 <aGaTHoS> https://privatebin.net/?7f5bde84fb536c12#8puZLMA/2Awll0PPyISLBnNzJmT/G52cLRqZrQ9oHOM=
2019-01-21 02:18:38 <aGaTHoS> here it si
2019-01-21 02:18:40 <aGaTHoS> is
2019-01-21 02:18:47 <aGaTHoS> the error and below my APKBuILD
2019-01-21 02:18:53 <aGaTHoS> any clue
2019-01-21 02:19:03 <aGaTHoS> I don't know why it fails
2019-01-21 02:19:53 <aGaTHoS> i have just updated version...
2019-01-21 02:19:55 <aGaTHoS> https://git.alpinelinux.org/aports/tree/testing/megatools/APKBUILD
2019-01-21 02:20:12 <aGaTHoS> this is the current apkbuild which seems someone merged
2019-01-21 02:20:19 <aGaTHoS> in the testing oficial git repo
2019-01-21 02:21:53 <aGaTHoS> i dont even remember the workflow, i know someone tell me to clone the repo not only locally in my pc but also in my git acoount and do a pull request? but in my local repositori i only have as origin the alpine aports repo , i suppose i have to push my local repo to my github repo then do a pull request? anyway the updated package does not compiles...
2019-01-21 02:22:06 <aGaTHoS> and dont know why
2019-01-21 02:28:44 <aGaTHoS> is it also because i now need to make packages for gpsim gputils sdcc, ive already enter a request but it seems nobody takes care of some requests for new packages so i decided to try to do myself even if i always have unknown problems as this, i always get stucked
2019-01-21 02:33:21 <dimon222[m]> >>> WARNING: megatools: No maintainer - try add maintainer
2019-01-21 02:33:29 <aGaTHoS> its just  warning
2019-01-21 02:33:37 <dimon222[m]> still do that
2019-01-21 02:33:50 <aGaTHoS> but the error its supposed to be related to dependencies
2019-01-21 02:34:03 <aGaTHoS> ok
2019-01-21 02:34:41 <dimon222[m]> also, workflow typically on github is next - you click Fork on github aports repo, it makes repo for you. You then clone yours repo, branch of out of it, and make one commit with your necessary change. On aports main repo it will automatically start showing icon for creating PR after you push your branch to github
2019-01-21 02:35:14 <aGaTHoS> so i should clone my own repo not alpine?
2019-01-21 02:35:20 <aGaTHoS> locally i mean
2019-01-21 02:35:32 <aGaTHoS> branch?
2019-01-21 02:35:44 <dimon222[m]> if u want to keep your forked main branch up to date, you can add additional source (add remote upstream) https://help.github.com/articles/configuring-a-remote-for-a-fork/
2019-01-21 02:35:44 <dimon222[m]> once in a while git pull upstream master - will give u update of tree in your fork
2019-01-21 02:35:58 <dimon222[m]> yep clone your fork that was created out of alpine/aports
2019-01-21 02:36:51 <aGaTHoS> when you say branch is locally not in github right?
2019-01-21 02:37:23 <aGaTHoS> no you mean in github
2019-01-21 02:39:14 <aGaTHoS> and any clue on the compile of the megatools apkbuild?
2019-01-21 02:39:30 <aGaTHoS> ive put the mantainer but obviously th error is still there
2019-01-21 02:39:59 <dimon222[m]> 
ACTION sent a long message:  < https://matrix.org/_matrix/media/v1/download/matrix.org/viVmCmKqEIMXlpkSrEnsEmmO >

2019-01-21 02:43:31 <dimon222[m]> ps i tried to run your build, it gives me proper errors and much more logs. Seems your patch is outdated
2019-01-21 02:43:32 <dimon222[m]> https://i.imgur.com/F2fRAKJ.png
2019-01-21 02:45:18 <dimon222[m]> considering it gives u builddeps errors instead... did u make sure your alpine is upgraded to latest within "edge" tree? default for 3.8 is 3.8 (aka stable tree), you might want latest what is in edge
2019-01-21 02:45:42 <aGaTHoS> thats another proiblem i had in the past, dont know how to compile "by hand"
2019-01-21 02:45:48 <aGaTHoS> so i can solvee the problems
2019-01-21 02:45:56 <aGaTHoS> as is i just download
2019-01-21 02:46:00 <aGaTHoS> the sources
2019-01-21 02:46:08 <aGaTHoS> i have also manually download dependencies
2019-01-21 02:46:24 <aGaTHoS> maybe abuild should be more "modular"
2019-01-21 02:46:31 <aGaTHoS> maybe it is but i dont know how to use
2019-01-21 02:46:46 <aGaTHoS> if it not compiles it ripes dependencies
2019-01-21 02:46:58 <dimon222[m]> abuild -r is supposed to download all the dependencies (as long as they present in alpine repos?)
2019-01-21 02:47:18 <aGaTHoS> it would be fine if i could do every step download dependencies... build ... clean build dependencies .. etc
2019-01-21 02:47:26 <aGaTHoS> but if it does not compile
2019-01-21 02:47:30 <aGaTHoS> it deletes
2019-01-21 02:47:32 <aGaTHoS> them
2019-01-21 02:47:41 <aGaTHoS> i want to make by hand so i can see the exact error
2019-01-21 02:48:12 <aGaTHoS> they are all
2019-01-21 02:48:15 <aGaTHoS> i checked
2019-01-21 02:48:27 <aGaTHoS> and why you get error when patching
2019-01-21 02:48:33 <aGaTHoS> and i have early i suppose
2019-01-21 02:48:40 <aGaTHoS> when resolving dependencies
2019-01-21 02:48:42 <aGaTHoS> ?
2019-01-21 02:49:39 <aGaTHoS> and the patch
2019-01-21 02:49:56 <aGaTHoS> i just copied the arch linux build
2019-01-21 02:50:09 <aGaTHoS> was there
2019-01-21 02:50:34 <dimon222[m]> honestly i just took your build file (in privatebin) + patch from https://git.alpinelinux.org/aports/tree/testing/megatools/libressl.patch
2019-01-21 02:50:35 <dimon222[m]> ... and ran "abuild -r" on alpine docker
2019-01-21 02:50:39 <aGaTHoS> or well if i could inspect the code maybe can make a new one
2019-01-21 02:50:56 <aGaTHoS> but my error is this
2019-01-21 02:51:08 <aGaTHoS> >>> megatools: Analyzing dependencies...
2019-01-21 02:51:10 <aGaTHoS> >>> ERROR: megatools: builddeps failed
2019-01-21 02:51:17 <aGaTHoS> and haven't any clue why it fails
2019-01-21 02:52:19 <dimon222[m]> check that you use "edge" in /etc/apk/repositories
2019-01-21 02:52:19 <dimon222[m]> You need "edge/testing"
2019-01-21 02:54:46 <aGaTHoS> of course
2019-01-21 02:54:52 <aGaTHoS> and have all repos
2019-01-21 02:55:33 <aGaTHoS> main community and testing
2019-01-21 02:55:40 <aGaTHoS> edge ones
2019-01-21 02:59:20 <aGaTHoS> well im trying to do step by step
2019-01-21 02:59:23 <aGaTHoS> i have done abuild deps
2019-01-21 02:59:28 <aGaTHoS> and this is the error
2019-01-21 02:59:59 <aGaTHoS> https://privatebin.net/?403e21623823677c#EU6ZwsjOodNTtBbwBv2bTaeSsVbz0KSs1GyvDpf+NP8=
2019-01-21 03:02:48 <dimon222[m]> for some reason i feel core is not to date.
2019-01-21 03:02:49 <dimon222[m]> try to do "apk update && apk upgrade"
2019-01-21 03:03:27 <aGaTHoS> ive made recently
2019-01-21 03:03:51 <aGaTHoS> # apk upgrade
2019-01-21 03:03:53 <aGaTHoS> 1 error; 2244 MiB in 566 packages
2019-01-21 03:03:57 <aGaTHoS> ...
2019-01-21 03:04:13 <aGaTHoS> keeps saying one error
2019-01-21 03:04:46 <aGaTHoS> https://privatebin.net/?e292b50cb4e90d30#RNxWdl3tCaPsWg3ELFt8kUM/DR97sOcOnhGKbrxXpWw=
2019-01-21 03:04:56 <aGaTHoS> i think my system is f**ked
2019-01-21 03:04:59 <dimon222[m]> so even upgrade doesn't work... just run latest docker of alpine for builds
2019-01-21 03:05:08 <aGaTHoS> docker?
2019-01-21 03:06:04 <dimon222[m]> ye if u run alpine in Docker/VM its isolated so if u break it, easy to reset
2019-01-21 03:07:41 <aGaTHoS> and what do i do with my broken system? well i had to reinstall anyway cause i have to pass the os to the internal ssd so anyway im doing this first then see if all is ok or still have problems
2019-01-21 03:07:50 <aGaTHoS> gonna reinstall alpine first
2019-01-21 03:09:26 <aGaTHoS> is cyclic...
2019-01-21 03:09:39 <aGaTHoS> another package i wanted to do was multiboot
2019-01-21 03:09:45 <aGaTHoS> ive filled a request
2019-01-21 03:10:03 <aGaTHoS> i need to install the alpine installation disk in the usbkey
2019-01-21 03:10:18 <aGaTHoS> and still have the freebsd installation disk etc..
2019-01-21 03:10:47 <aGaTHoS> multibootusb
2019-01-21 03:10:56 <aGaTHoS> if someone can try to make a package
2019-01-21 03:11:05 <aGaTHoS> if it just compiles...
2019-01-21 03:11:16 <aGaTHoS> its python based downloading by hand
2019-01-21 03:11:25 <aGaTHoS> and done python setup.py didnt work for me
2019-01-21 03:25:27 <hanetzer> hey. someone with amd64 alpine, could you provide me the output of gcc -print-multi-os-directory ?
2019-01-21 03:27:08 <jwh> hmph, can't use ip/prefixlen for interfaces config?
2019-01-21 03:29:30 <aGaTHoS> hanetzer: ../lib
2019-01-21 03:35:20 <hanetzer> aGaTHoS: thanks.
2019-01-21 03:35:30 <hanetzer> gcc -v says --disable-multilib, I assume?
2019-01-21 03:42:15 <aGaTHoS> yes
2019-01-21 03:47:42 <hanetzer> kk, thanks.
2019-01-21 04:17:08 <aGaTHoS> apk upgrade gives me one error
2019-01-21 04:17:20 <aGaTHoS> how can i know what is happening does not give me more information
2019-01-21 04:22:31 <aGaTHoS> the Bootstrapping_Alpine_Linux is correct? can i install alpine in another partition without leaving my actual system with that command? and then i suppose doing a chroot...?
2019-01-21 04:23:07 <aGaTHoS> --arch x86 should be x86_64?
2019-01-21 05:07:09 <aGaTHoS> ok finally installed alpine in another partition , fresh and  2 errors when upgrading!! without saying anything else
2019-01-21 05:07:31 <aGaTHoS> should be normal
2019-01-21 06:00:28 <aGaTHoS> but i dont have the package linux-hardened as in the current installation nor boot dir
2019-01-21 06:00:34 <aGaTHoS> how can i install a kernel?
2019-01-21 06:01:02 <aGaTHoS> apk search linux-hardened returns nothing on the chroot of the new partition with alpine recent installation
2019-01-21 06:03:21 <aGaTHoS> maybe i did it all wrong? i have follow the wiki to install from Installing_Alpine_Linux_in_a_chroot
2019-01-21 06:03:38 <aGaTHoS> but replacing the last-stale by directly edge
2019-01-21 06:03:41 <aGaTHoS> stable*
2019-01-21 06:09:07 <aGaTHoS> why linux-hardened is missing in my recent installation?
2019-01-21 07:09:44 <aGaTHoS> what does exactly setup-disk do? can i do manually?
2019-01-21 07:09:56 <aGaTHoS> because thats the reason i think i lack a kernel
2019-01-21 07:10:22 <aGaTHoS> and because if i root setup-disk -m sys -s 0 / from the chroot installation it gives me an error
2019-01-21 07:10:33 <aGaTHoS>  does not seem to be a mount point
2019-01-21 07:10:42 <aGaTHoS> same with /dev/sd*...
2019-01-21 10:24:17 <parazyd> Hi there. Anyone using irssi?
2019-01-21 10:24:27 <parazyd> I keep getting an Unknown command: SCRIPT
2019-01-21 10:24:41 <steven> there's an irssi irc channel as well
2019-01-21 10:24:43 <steven> just sayi
2019-01-21 10:24:47 <steven> ng*
2019-01-21 10:24:53 <parazyd> Yeah but this only happens in alpine :)
2019-01-21 10:25:10 <parazyd> I have irssi-perl installed and perl loaded. Something's up.
2019-01-21 10:44:36 <mps> parazyd: irssi works fine on Alpine, btw command is '/script'
2019-01-21 10:46:05 <parazyd> Not here...
2019-01-21 10:46:07 <parazyd> Unknown command: script
2019-01-21 10:46:20 <parazyd> Is there anything I should be loading before manually?
2019-01-21 10:47:02 <mps> did you loaded perl
2019-01-21 10:47:18 <mps> '/load perl'
2019-01-21 10:47:30 <parazyd> Yep
2019-01-21 10:47:58 <mps> which version of Alpine you use
2019-01-21 10:48:15 <parazyd> edge
2019-01-21 10:48:29 <parazyd> irssi-1.1.2-r0, irssi-perl-1.1.2-r0
2019-01-21 10:48:51 <parazyd> Maybe there's something I'm missing that I should manually install?
2019-01-21 10:49:01 <parazyd> For example adv_windowlist, nickcolors all work.
2019-01-21 10:49:10 <parazyd> It's just that irssi doesn't know about /script.
2019-01-21 10:49:12 <mps> my irssi is on 3.8 (on server) and works
2019-01-21 10:50:41 <mps> I'm using same version, although on stable, didn't tried on edge yet
2019-01-21 10:51:58 <mps> ok, just tried on edge, it works
2019-01-21 10:53:16 <parazyd> Hmm
2019-01-21 10:53:43 <mps> maybe something in your config or startup?
2019-01-21 10:54:18 <parazyd> I'll give it a look.
2019-01-21 10:54:36 <parazyd> Would you mind putting your perl-* packages on a pastebin for me to compare?
2019-01-21 10:55:33 <mps> what packages you mean, scripts?
2019-01-21 10:55:52 <parazyd> apk list -I *perl*
2019-01-21 10:58:50 <mps> it is not interesting, but ok: http://tpaste.us/ee6p
2019-01-21 11:02:01 <parazyd> Thanks!
2019-01-21 11:04:00 <asyd> mps: have you tried /load perl before run /script ?
2019-01-21 11:04:43 <parazyd> It says perl/perl already loaded.
2019-01-21 11:05:50 <mps> asyd: for irssi works, your question is for parazyd, probably
2019-01-21 11:05:57 <asyd> yeah sorry :)
2019-01-21 11:06:20 <mps> np, we all make mistakes :)
2019-01-21 11:08:12 <parazyd> I think perl/fe is missing...
2019-01-21 11:09:37 <asyd> parazyd: https://dpaste.de/06Oh works here
2019-01-21 11:09:55 <parazyd> I don't have it
2019-01-21 11:10:10 <parazyd> Module perl/perl already loaded
2019-01-21 11:10:40 <mps> parazyd: apk info -L irssi-perl
2019-01-21 11:10:47 <parazyd> Ah!
2019-01-21 11:10:49 <mps> see if all is there
2019-01-21 11:10:55 <parazyd> /load fe_perl helped
2019-01-21 11:11:01 <asyd> parazyd: just ran docker -it alpine:3.8 / apk add irssi irssi-perl
2019-01-21 11:11:31 <parazyd> Loaded module perl/fe
2019-01-21 11:11:38 <parazyd> Now /script works
2019-01-21 12:41:58 <aggaz> Hello. I am trying to figure out how to add a swap file to a diskless alpine installation on a Raspberry Pi. I made a swap file on the principal FAT partition and added the corresponding line to fstab but swap space is not added on boot. Swap space is correctly added by a swapon command. I think that fstab is read before the permanent root modifications done with lbu are done, but I am not sure. Any
2019-01-21 12:42:00 <aggaz> suggestions?
2019-01-21 12:42:57 <AinNero> is the 'swap' service added to the active runlevels?
2019-01-21 12:44:50 <aggaz> AinNero: thank you for the suggestion. I am don't think so. I don't really familiar with OpenRC. If I run rc-status I don't see it.
2019-01-21 12:46:01 <aggaz> tryed to start the service. It does not work but because it needs a read-write file system. I am going to try to activate the writing.
2019-01-21 12:48:09 <aggaz> it works now! thank you very much
2019-01-21 12:48:38 <aggaz> I don't really like the idea to have a writable file system... But I suppose I have to live with it.
2019-01-21 13:04:24 <AinNero> its not a technical requirement i guess, just a formal thing
2019-01-21 13:04:50 <AinNero> maybe its the dependency on 'root' that is misplaced
2019-01-21 13:29:14 <aggaz> AinNero: I don't understand... is there something I can do to keep the filesystem read-only?
2019-01-21 13:29:46 <AinNero> 'i guess' - there is probably a way, but i dont know the details
2019-01-21 13:30:06 <aggaz> ok
2019-01-21 13:34:59 <mps> aggaz: create swap partition
2019-01-21 13:36:16 <aggaz> mps: thank you for the suggestion. I kinda like the idea to have all in the FAT partition.
2019-01-21 13:36:29 <AinNero> oh, swap on ro
2019-01-21 13:36:33 <AinNero> yeah, that can't work
2019-01-21 13:37:49 <mps> then you have to live with rw, you want to write to swapfile on fs , partition then must be writeable
2019-01-21 13:38:55 <mps> but having separate swap partition makes swap a little faster
2019-01-21 13:39:34 <aggaz> mps: ok you are making a quite persuasive argument, I think I will do it. :)
2019-01-21 13:40:16 <aggaz> I head that swap in SD cards is not really a good thing. is there any difference in stress over the SD card when using a file instead of a partition?
2019-01-21 13:40:17 <mps> or, you can have /dev/zram swap if you do not need much swapspace
2019-01-21 13:40:59 <mps> having swap on sd cards is not pleasant experience
2019-01-21 13:41:07 <aggaz> mps: never heard of zram... I will take a look at it. Is there any alpine-specific documentation?
2019-01-21 13:42:05 <mps> zram is kernel module and is not specific to Alpine
2019-01-21 13:42:37 <mps> you need to intall util-linux where zramctl command is packaged
2019-01-21 13:42:46 <aggaz> ok
2019-01-21 13:42:57 <mps> and you could install zram-init apk
2019-01-21 13:48:52 <aggaz> Looking at zram, I like it, but I only have 256mb of ram. I am not sure It would be convenient in my case.
2019-01-21 13:52:05 <mps> aggaz: it depends how much swap you need
2019-01-21 13:54:40 <aggaz> mps: not much, just trying to make a reliable configuration that will not blow up if the little ram I have will be full
2019-01-21 13:56:08 <mps> it depends for what you want to use box
2019-01-21 13:57:06 <mps> I have one board with 512MB and most time 70% of that ram is free
2019-01-21 13:57:57 <aggaz> basically server application (no x). a little box to play with tinc and learn about networks, routing and firewalls
2019-01-21 13:58:41 <mps> Probably you don't need swap for such use case
2019-01-21 13:59:20 <aggaz> even if I plan to have the RPi on all the time for years?
2019-01-21 14:00:05 <mps> on armv7 board where I have radio repeater server/service 'free' says it uses only 27MB of RAM
2019-01-21 14:01:53 <mps> on 1GB ram I had development system, although didn't build large packages there
2019-01-21 14:02:08 <aggaz> wow
2019-01-21 14:04:20 <mps> musl doesn't use much memory like glibc
2019-01-21 14:04:58 <mps> but it depends if the applications also doesn't need a lot memory for work
2019-01-21 14:05:30 <mps> openwrt is musl based and it runs on really small memory devices quite fine
2019-01-21 14:13:45 <aggaz> my free says that I am using 90m. I am not sure if I am using musl. As far as I remember alpine provides 2 different versions, right?
2019-01-21 14:14:23 <mps> no, alpine is musl only
2019-01-21 14:15:23 <aggaz> ah ok I was not aware of it
2019-01-21 14:16:48 <mps> I used Debian for more than two decades and when switched to Alpine on ARM SBC's I thought that the 'free' and 'uptime' are faked to lie to user :)
2019-01-21 14:18:31 <mps> it is much faster, btw
2019-01-21 14:19:41 <aggaz> lol. yes I noticed an improvement myself. in my opinion to date alpine is the best distro for old RPis
2019-01-21 14:20:30 <aggaz> It is a shame that it does not support old PPC mac too. The support for these machines is pratically not existent these days. Even debian is not usable anymore.
2019-01-21 14:21:01 <aggaz> Is there any interest for old PPC (POWER4 processors) in the community?
2019-01-21 14:21:43 <odc> aggaz: you're supposed to use NetBSD ;)
2019-01-21 14:22:11 <odc> it's like the official OS for old stuff
2019-01-21 14:22:39 <aggaz> odc: lol. NetBSD is the only one I did not tryed on my PowerMac G5. But I can guarantee you that nor FreeBSD and OpenBSD are good for X11 on it.
2019-01-21 14:22:59 <odc> heh
2019-01-21 14:23:04 <aggaz> The only distro that runs fine is Gentoo but I hate to wait 4 hours to compile gcc
2019-01-21 14:23:11 <odc> you mean you could not even start X ?
2019-01-21 14:23:18 <aggaz> no.
2019-01-21 14:23:28 <aggaz> there are always driver issues
2019-01-21 14:23:32 <aggaz> for graphic cards
2019-01-21 14:23:41 <odc> oh
2019-01-21 14:23:47 <aggaz> or other issues that prevent you to arrive to the graphic card config
2019-01-21 14:24:04 <aggaz> all started with the KMS transition
2019-01-21 14:24:15 <odc> i see
2019-01-21 14:24:16 <aggaz> old cards (especially ati) are not supported in kernel space
2019-01-21 14:24:24 <odc> i can't help tho ^^
2019-01-21 14:24:38 <aggaz> and eith ndivia your only hope is nouveau which is linux specific
2019-01-21 15:46:28 <tya99> so I was having a look at https://wiki.alpinelinux.org/wiki/Setting_up_ZFS_on_LUKS
2019-01-21 15:46:38 <tya99> >  Note: We can't use the live environment from the installation isos, because we'll install the ZFS kernel module and that isn't possible with a read-only /boot (as provided by the iso).
2019-01-21 15:46:57 <tya99> does that mean I cannot use Diskless Mode?
2019-01-21 15:47:58 <tya99> essentially I was thinking of having alpine linux boot on a usb stick, and then mount my zpool of 6 disks (raidz2) on dmcrypt with a cryptkey
2019-01-21 15:48:19 <ncopa> tya99: i think we include zfs kernel module in the alpine-extended iso image
2019-01-21 15:48:21 <ncopa> but not the others
2019-01-21 15:48:28 <tya99> the cryptkey would be stored somewhere in /etc/ (and be inside the encrypted apkprov
2019-01-21 15:48:40 <tya99> ncopa: yeah that's the image i'd be using
2019-01-21 15:49:22 <ncopa> i think it may work
2019-01-21 15:49:24 <tya99> my friend said i should use https://www.joyent.com/smartos but yeah i don't really want to go with IllumoOS based OS
2019-01-21 15:49:44 <tya99> particularly as it seems openzfs is better https://lists.freebsd.org/pipermail/freebsd-fs/2018-December/027085.html
2019-01-21 15:50:18 <tya99> the idea then was to get a superfast ssd and have a bunch of VMs on that
2019-01-21 15:50:27 <tya99> probably ext4 or something like that
2019-01-21 15:50:43 <tya99> but the zpool would be for home and data
2019-01-21 15:51:07 <tya99> passed through 9p virtio
2019-01-21 15:51:47 <tya99> i have physical access to the machine, so it's not like it's running in a datacentre
2019-01-21 15:54:01 <tya99> ncopa: yeah it looks like the extended iso has: zfs-0.7.8-r0.apk zfs-libs-0.7.8-r0.apk zfs-vanilla-4.14.89-r1.apk
2019-01-21 16:05:06 <muhaha> Is there any cron that can support year parameter? Like http://www.nncron.ru/help/EN/working/cron-format.htm  -> Its good for triggering jobs only one time. (Yes, I know 'atd')
2019-01-21 16:55:20 <ncopa> muhaha: i dont know
2019-01-21 16:58:49 <ccx^xmpp> muhaha: not exactly what you are asking for but check out Leah's snooze
2019-01-21 17:11:56 <ccx^xmpp> Also I believe fcron was the kitchen sink of cron implementations, so maybe that.
2019-01-21 17:49:22 <muhaha> @ccx^xmpp: anacron nor fcron can not handle one time jobs... Only nncron, but its not as package in alpine
2019-01-21 17:56:06 <smemsh> hello, does anyone know if there are any special args known to the alpine boot scripts which can be passed as boot params... in particular i'm trying to force a certain kernel module to load
2019-01-21 18:19:23 <ncopa> smemsh: is the module included in the initramfs? if it is, then you can edit the modules=.... line
2019-01-21 18:20:02 <ncopa> if you dont need it in the early boot (to find/mount rootfs) then you could add it to /etc/modules
2019-01-21 19:09:39 <smemsh> ncopa: thanks. i exploded the initramfs; the module (apple-hid) is not present.  this is the installer iso, the module is needed to use the keyboard.  i can maybe remaster the iso with the module added or something, hrm
2019-01-21 19:12:00 <ncopa> you need the module in early boot?
2019-01-21 19:12:21 <smemsh> i need it to use the keyboard after the kernel boots
2019-01-21 19:12:34 <ncopa> smemsh: i think it may make sense to include it in the default initramfs
2019-01-21 19:12:47 <ncopa> can you please create an issue on bugs.alpinelinux.org?
2019-01-21 19:12:50 <smemsh> i think so because it's needed for most backbooks
2019-01-21 19:12:53 <smemsh> macbooks
2019-01-21 19:13:02 <smemsh> newer ones anyways
2019-01-21 19:13:37 <smemsh> sure
2019-01-21 19:16:09 <elxa> ncopa: is something blocking this https://github.com/gliderlabs/docker-alpine/pull/465 ?
2019-01-21 19:16:24 <smemsh> i don't know if the kernel knows enough to load it automatically though.  maybe will need modules= as you suggest but i don't know how to edit this, it boots through grub apparently and i have to manually tell it the kernel and initrd path, there's no config file apparently
2019-01-21 19:16:46 <smemsh> grub just brings me to a prompt with no menu items
2019-01-21 19:17:28 <smemsh> i believe because it is just loading the generic efi stub on the iso
2019-01-21 19:20:20 <ncopa> i think it should be autoloaded if its available
2019-01-21 19:20:21 <smemsh> i see, i can use the line from the syslinux.cfg and give it to the kernel argv that grub will pass
2019-01-21 19:21:21 <smemsh> not sure because i need to use /etc/modules on debian also
2019-01-21 19:22:30 <ncopa> elxa: the plan for that is to create official minirootfs images from edge
2019-01-21 19:22:51 <ncopa> and that needs to be done as soon 3.9.0 release is out
2019-01-21 19:23:43 <elxa> ok cool, looking forward to that :)
2019-01-21 19:28:19 <smemsh> hm, no apple-hid is not in the debian initramfs either so perhaps you're right it will autodetect if present
2019-01-21 20:49:41 <smemsh> actually debian does have hid-apple.ko (instead of apple-hid) in its initramfs
2019-01-21 23:31:07 <ncopa> smemsh: should be included in the just released 3.9.0 rc3
2019-01-22 11:09:13 <Processus42> I'm not sure if my message from 1 min ago got sent (the Matrix bridge dc'ed me)
2019-01-22 11:09:55 <AinNero> Processus42: it didnt
2019-01-22 11:10:17 <Processus42> Thanks. Here it is: Hello there, what's the official/recommended way to create a user on an image built with mkimage.sh ?
2019-01-22 11:11:24 <AinNero> i recommend using an apkovl for that
2019-01-22 11:14:39 <Processus42> That's the only way I can think of right now. In this case I have either to include an entire /etc/passwd file, or set-it up through a service started at boot. The first solution is less dirty than the second. But I still find it a bit dirty. Am I missing something, or is it "ok" to do so ?
2019-01-22 11:16:29 <AinNero> i agree that its a bit hacky
2019-01-22 11:16:48 <AinNero> but if you want to ensure consistent name->uid mapping, you need to do sth like that anywahys
2019-01-22 11:16:53 <AinNero> *anyways
2019-01-22 11:19:43 <Processus42> Yep. That's why I think the first solution is the "best" (Otherwise I'd have to create a skel directory for this specific user, which means copying files just for a "proper" setup).
2019-01-22 11:20:43 <Processus42> I'll go with this solution then, thanks :)
2019-01-22 13:55:04 <thresh> hi.  what's the packaging policy for the daemons that listen on a privileged port on a first package install with regards autolaunching the service or not?
2019-01-22 14:11:45 <thresh> also can I use versioned depends (e.g. a module that requires a non-specific version of main package)?
2019-01-22 14:12:39 <thresh> like I can have multiple nginx 1.15.8 packages , -r0, -r1, -r2 etc, but those are guaranteed to be API/ABI compatible, so the modules (e.g. geoip) will have a depends="nginx-1.15.8" ?
2019-01-22 14:14:31 <thresh> asking cause this doesnt seem to work, as I think pkgrel is glued into pkgver in a binary package?
2019-01-22 14:14:52 <thresh> obvious solution would be to do a virtual provides, but that's kinda kludgy
2019-01-22 14:17:49 <mps> thresh: something like maybe could work: depends="$pkgname-lib=$pkgver-r$pkgrel"
2019-01-22 14:18:28 <thresh> well except I don't want to depend on -r*, but just on $pkgver
2019-01-22 14:19:02 <mps> then try to remove $pkgrel
2019-01-22 14:21:12 <thresh> that's how it's now
2019-01-22 14:21:49 <thresh> pkginfo for both my nginx/geoip modules, and apk failure to install: https://gist.github.com/thresheek/15d3c69bc2905fac2e6089d7459d2660
2019-01-22 14:26:21 <muhaha> Alpine printf does not handle printf -v VAR ? :/
2019-01-22 14:27:22 <muhaha> ah its busybox version
2019-01-22 14:29:22 <AinNero> muhaha: VAR=$(printf ...)
2019-01-22 14:32:52 <muhaha> thing get complicated... lol  user=foobar; body=$(printf 'Text for %s' $user); curl -X  POST http://endpoint--data-binary "body=$body"
2019-01-22 14:33:45 <thresh> did a virtual provides/depends and now it works, but kludgy
2019-01-22 14:33:46 <thresh> +provides="$pkgname-$pkgver"
2019-01-22 14:33:55 <thresh> +depends="nginx-$_main_version"
2019-01-22 14:33:56 <thresh> :/
2019-01-22 17:14:40 <PureTryOut[m]> where can I download the Alpine 3.9-rc* iso's?
2019-01-22 17:16:17 <klesk> http://dl-cdn.alpinelinux.org/alpine/v3.9/
2019-01-22 17:17:48 <klesk> in the release folder for your architecture
2019-01-22 17:29:19 <PureTryOut[m]> thanks!
2019-01-22 17:30:49 <PureTryOut[m]> The standard images should support EFI right?
2019-01-22 17:43:41 <klesk> i think efi support is included since 3.7?
2019-01-22 17:43:46 <klesk> https://wiki.alpinelinux.org/wiki/Bootloaders
2019-01-22 18:26:16 <PureTryOut[m]> Yeah but it doesn't work on the standard 3.8.2 images
2019-01-22 18:26:25 <PureTryOut[m]> it just goes to a Grub rescue prompt
2019-01-22 18:26:35 <PureTryOut[m]> it did indeed work for 3.9-rc3 though
2019-01-22 18:38:56 <clandmeter> PureTryOut[m], yes grub config generation is inplemented
2019-01-22 18:42:05 <PureTryOut[m]> after installing, grub's cmdline has `root=/dev/sdxY` rather than `root=UUID=<UUID>`. this causes problems booting when unplugging the usb, is this done on purpose?
2019-01-22 20:19:53 <Rapture> anybody have a tutorial on setting up syslog on an alpine docker image? I tried to follow https://wiki.alpinelinux.org/wiki/Syslog but after installing syslog-ng, rc-update and the rest don't work (command not found).
2019-01-22 20:23:53 <Rapture> tried both busybox and syslog-ng packages
2019-01-22 21:48:30 <kahiru> i dont think youre supposed to run syslog inside docker
2019-01-22 21:48:33 <kahiru> along other stuff
2019-01-23 08:47:02 <steven> how long does it usually take to get new packages into the official repo's?
2019-01-23 08:49:23 <_ikke_> steven: Can be quite quick, but currently it's a bit busy due to the new release that is being prepared
2019-01-23 08:49:49 <steven> gotcha, ok. thanks
2019-01-23 08:49:59 <steven> was just wondering since there is no response on github :)
2019-01-23 08:51:21 <steven> so some kind a feedback within a couple weeks would be a reasonable assumption _ikke_ ?
2019-01-23 08:51:38 <_ikke_> Certainly
2019-01-23 08:52:21 <steven> that's all I needed to know, thanks!
2019-01-23 09:20:26 <muhaha> Anyone tried to use at in docker?
2019-01-23 09:21:11 <muhaha> Seems that atd forks to background and there is no possibility to keep it in foreground..
2019-01-23 09:26:13 <muhaha> uh, seems that /usr/sbin/atd -f is working
2019-01-23 19:25:35 <g1eb> hi all! I have a question about installing leveldb on alpine https://pkgs.alpinelinux.org/package/edge/testing/x86/leveldb
2019-01-23 19:26:40 <g1eb> i need to add this package in a docker running alpine image but seems like the leveldb cant be found https://dpaste.de/bEbQ
2019-01-23 19:27:53 <ncopa> g1eb: its because it is in testing repository
2019-01-23 19:28:15 <g1eb> ahh shoot, means it's not mainstream yet?
2019-01-23 19:28:19 <_ikke_> nope
2019-01-23 19:28:33 <g1eb> is there any way to get it though? add the cdn for testing repos?
2019-01-23 19:29:45 <ncopa> exactly
2019-01-23 19:30:08 <_ikke_> are you on dege?
2019-01-23 19:30:26 <ncopa> according https://dpaste.de/bEbQ, no
2019-01-23 19:31:20 <g1eb> no, here's my Dockerfile  http://vpaste.net/NkYmS
2019-01-23 19:31:41 <_ikke_> combining stable and unstable is not really supported, you might run into version conflicts
2019-01-23 19:32:33 <g1eb> you would recommend to switch to edge? how do i do that?
2019-01-23 19:33:03 <g1eb> i mean, im fine using edge instead if that would help me solve this dependency issue
2019-01-23 19:33:55 <_ikke_> use FROM alpine:edge
2019-01-23 19:34:06 <g1eb> oh it's that easy, thanks!
2019-01-23 19:34:36 <_ikke_> and then add the testing repo to /etc/apk/repositories
2019-01-23 19:35:06 <_ikke_> Note that edge might break once and a while
2019-01-23 19:35:46 <g1eb> can i just do echo echo 'https://pkgs.alpinelinux.org/package/edge/testing/x86/' >> /etc/apk/repositories ?
2019-01-23 19:36:02 <asyd> change you FROM alpine to alpine:edge?
2019-01-23 19:36:13 <_ikke_> g1eb: yes
2019-01-23 19:36:25 <_ikke_> asyd: has already been said :-)
2019-01-23 19:36:39 <_ikke_> g1eb: leave out the arch though
2019-01-23 19:36:49 <_ikke_> hmm
2019-01-23 19:36:53 <_ikke_> that url is not correct
2019-01-23 19:37:12 <_ikke_> use this: http://dl-cdn.alpinelinux.org/alpine/edge/testing/
2019-01-23 19:37:13 <asyd> my bad
2019-01-23 19:39:12 <g1eb> thanks _ikke_! building the image now :)
2019-01-23 19:39:54 <g1eb> out of curiousity, what's the reason dependencies in testing repos dont go into main repo?
2019-01-23 19:40:17 <g1eb> im not sure how long leveldb has been around but you would think it should be stable by now right?
2019-01-23 19:41:22 <_ikke_> It doesn't receive much attention apparently
2019-01-23 19:41:35 <_ikke_> on average one commit per year
2019-01-23 19:41:37 <_ikke_> https://git.alpinelinux.org/aports/log/testing/leveldb
2019-01-23 19:42:54 <_ikke_> g1eb: apparently the package does not have a maintainer
2019-01-23 19:44:41 <g1eb> okay that makes sense if it's not actively maintained, maybe just ported once and that's it
2019-01-23 19:45:14 <_ikke_> Yeah, it received some updates once in a while, but nothing more
2019-01-23 19:45:27 <g1eb> on a related subject, we were able to find that package but im getting a build error when i run docker build
2019-01-23 19:45:49 <g1eb> just at a glance, is it related to alpine/ leveldb denendency, here's the log https://dpaste.de/TtZo
2019-01-23 19:46:12 <g1eb> i mean, i've seen this biuld error before and though getting leveldb would fix it
2019-01-23 19:48:09 <_ikke_> you mean the missing xlocale.h?
2019-01-23 19:49:31 <g1eb> yeah, is that something related to using alpine:edge?
2019-01-23 19:49:47 <g1eb> i don't think so, i think i've seen this error before i changed to alpine
2019-01-23 19:51:03 <g1eb> OB
2019-01-23 19:51:08 <g1eb> oops
2019-01-23 19:53:21 <jwh> 
ACTION sighs

2019-01-23 19:53:33 <jwh> sshd not starting for no reason
2019-01-23 19:53:34 <_ikke_> 
ACTION tries to console jwh

2019-01-23 19:53:38 <dalias> is there a python-wxtools package for alpine?
2019-01-23 19:54:04 <jwh> it actually stopped working while I was logged into it, just junked new connections, nothing in logs, now its died entirely
2019-01-23 19:54:05 <_ikke_> https://pkgs.alpinelinux.org/packages?name=python-wxtools&branch=edge
2019-01-23 19:54:47 <dalias> obviously i can apk search...
2019-01-23 19:54:58 <dalias> i was wondering if it's named differently or buried in something else
2019-01-23 19:55:38 <_ikke_> not all python libs are getting packaged
2019-01-23 19:57:19 <dalias> *nod*
2019-01-23 19:57:49 <dalias> trying pip, tried that before but had the wrong package name i think
2019-01-23 20:24:52 <dalias> opencv package seems currently non-installable. it depends on outdated libavcodec etc
2019-01-23 20:25:26 <_ikke_> it's in testing, so probably it needs to be bumped
2019-01-23 20:57:13 <g1eb> hi guys, running into a different error on build with alpine:edge, missing something called versioneer when getting requirements for wheel, any ideas?
2019-01-23 20:57:30 <g1eb> to be exact im getting an error when: Getting requirements to build wheel: finished with status 'error'
2019-01-23 20:58:01 <g1eb> seems to be happening in python, when importing the 'versioneer' package
2019-01-23 20:59:29 <g1eb> this is a requirement for the pandas package
2019-01-23 20:59:59 <g1eb> im trying to make an isolated dockerfile for testing purposes
2019-01-23 21:29:27 <g1eb> this is my Dockerfile now, isolated to include pandas only for testing purposes: http://vpaste.net/dCmqT
2019-01-23 22:22:15 <feuerteufel> Hi
2019-01-23 22:23:28 <tmh1999> g1eb : I don't have any problem with your dockerfile, not running docker though, just exec the commands
2019-01-23 22:23:40 <tmh1999> feuerteufel: Hi
2019-01-23 22:24:16 <feuerteufel> Is there a reason why apk try's to downgrade the vanilla kernel from 4.19.15 to 4.19.7 in edge?
2019-01-23 22:24:36 <tmh1999> there was a mistake, please check it in few hours or tomorrow
2019-01-23 22:24:58 <tmh1999> current 3.9 rc4 and edge are affected. 3.9 rc3 is fine
2019-01-23 22:26:27 <feuerteufel> So it's better to wait?
2019-01-23 22:29:07 <tmh1999> I'm afraid so
2019-01-23 22:29:31 <feuerteufel> Ok, thank's
2019-01-23 22:40:29 <g1eb> tmh1999 thanks for checking, strangely the error was specific to pandas version 0.22, does not happen for newer versions
2019-01-24 05:41:15 <unrznbl[m]> am I right that alpine doesn't have cross-compiler packages? like arm binutils/etc for x86 hosts?
2019-01-24 05:41:18 <unrznbl[m]> so build your own?
2019-01-24 05:42:13 <_ikke_> unrznbl[m]: alpine does verry little cross-compilation itself
2019-01-24 05:43:08 <unrznbl[m]> I wasn't thinking too clearly but obviously postmarketOS has arm toolchain in a chroot that is alpine I think. :p So maybe I can snag one from there. :)
2019-01-24 05:43:29 <_ikke_> what pacakge are you looking for?
2019-01-24 05:43:36 <_ikke_> alpine has armhf and armv7
2019-01-24 05:44:53 <unrznbl[m]> oh. armv7 would probably work.
2019-01-24 05:45:27 <_ikke_> armv7 is new though
2019-01-24 05:45:44 <_ikke_> https://pkgs.alpinelinux.org/packages?name=&branch=edge&arch=armv7
2019-01-24 05:45:45 <unrznbl[m]> we'll see... I'll go find it, try it. thanks!
2019-01-24 05:46:34 <unrznbl[m]> oh. right. Yeah I need a toolchain for building. S'all good. I have to build one custom anyway for something else. :) Can probably just re-use that one. :)
2019-01-24 06:04:11 <unrznbl[m]> out of odd curiosity, has anyone ever written something that helps translate lists of debian packages into alpine package names?
2019-01-24 07:14:21 <mepholic> _ikke_: that's interesting
2019-01-24 07:14:34 <mepholic> I was doing my own armv7 builds of alpine for a while
2019-01-24 11:32:47 <mps> anyone knows of Alpine based Redmine docker or lxc image. Asked that last night on #alpine-offtopic already, so repeating here in hope on better luck
2019-01-24 11:38:11 <hiro> mps: so now that it's easy to make docker packages, the only problem is that it becomes impossible to distribute and then find them?
2019-01-24 11:46:20 <mps> well, I didn't tried to make docker images, yet. have seen Redmine in dockers but some other distributions, tried to find same in Alpine docker but didn't succeded and had a hope that someone here knows if there are any
2019-01-24 12:00:21 <_ikke_> mps: We run redmine in an lxc container, but I'm not aware of any docker images
2019-01-24 12:05:46 <mps> _ikke_: what OS? Alpine?
2019-01-24 12:06:08 <clandmeter> we as in alpine :)
2019-01-24 12:06:22 <clandmeter> so yes we only run alpine in our infra
2019-01-24 12:07:21 <mps> but Redmine is not packaged in aports
2019-01-24 12:07:36 <clandmeter> we use bundler
2019-01-24 12:07:46 <clandmeter> we packaged it before
2019-01-24 12:07:55 <clandmeter> but ruby is a nightmare for packages
2019-01-24 12:07:59 <clandmeter> packagers
2019-01-24 12:08:02 <mps> I remeber that is was in aports
2019-01-24 12:08:19 <clandmeter> ruby dependency hell
2019-01-24 12:09:19 <clandmeter> we are looking into a proper way to integrate docker in our current infra.
2019-01-24 12:09:38 <mps> never tried anything serious with ruby. is the bundler complicated to use for novice to ruby
2019-01-24 12:10:03 <clandmeter> no just follow the offical steps from redmine
2019-01-24 12:10:52 <mps> thanks, will look and try. Now I'm running Redmine in Debian lxc on Alpine host
2019-01-24 12:12:05 <mps> gitlab is also in ruby
2019-01-24 12:14:57 <clandmeter> for smaller projects gitea is much easier to setup/maintain.
2019-01-24 12:15:35 <clandmeter> it just does not scale.
2019-01-24 12:15:36 <mps> right, but I need Redmine for my customer not for me
2019-01-24 12:16:28 <mps> for my work I use gitolite, small and simple
2019-01-24 14:31:14 <tkzv> Hello. Is this the place to ask about "testing" branch? There's a package hdf5-1.8.20-r0, which can't work with new file format. Can I do something to update the repository to the most recent 1.10.4?
2019-01-24 14:31:31 <tkzv> https://pkgs.alpinelinux.org/package/edge/testing/x86_64/hdf5
2019-01-24 14:34:21 <ncopa> tkzv: you can send us a patch, but we are currently prioritizing the 3.9.0 release so it may take a while
2019-01-24 14:36:51 <tkzv> ncopa: A patch to hdf5/APKBUILD ? I can do it. And how long is "a while"? A week? A month?
2019-01-24 14:48:53 <ncopa> Depends on how much work it require of us to  review or fix the patch
2019-01-24 15:50:45 <tkzv> @ncopa: Best case, the patch would only have to change the hash and the version, like this: https://git.alpinelinux.org/aports/commit/?id=cd84da54e924e8581d2ff08a650684f9ced7ac31 I'll get back to you when I finish testing it.
2019-01-24 16:04:34 <ncopa> tkzv: abump hdf5-1.10.4 && git format-patch -1 --stdout | tpaste
2019-01-24 16:10:20 <tkzv> https://wiki.alpinelinux.org/wiki/Aports_tree seems to have a wrong URL for the aports tree. https://git.alpinelinux.org/aports works, git://dev.alpinelinux.org/aports doesn't.
2019-01-24 16:19:23 <Rapture> getting a aws-cli (missing): when trying to install it on tomcat:8.5-jre8-alpine image. Little confused as to why. https://pkgs.alpinelinux.org/packages?name=aws-cli&branch=edge
2019-01-24 16:21:46 <tkzv> @ncopa: abuild gives a very uninformative ">>> ERROR: hdf5: all failed" during building. This is going to take some time.
2019-01-24 16:22:01 <ncopa> tkzv: i got it built here, but it breaks ABI
2019-01-24 16:22:06 <Rapture> nvm, edge/testing and not main
2019-01-24 16:22:17 <ncopa> so a handful packages will need to be rebuilt
2019-01-24 16:22:44 <_ikke_> How do you find that out? soname bummp>?
2019-01-24 16:23:48 <ncopa> checkapk
2019-01-24 16:23:54 <_ikke_> ah ok
2019-01-24 16:24:25 <ncopa> http://tpaste.us/4yvl
2019-01-24 16:25:01 <ncopa> when /usr/lib/libsomething.so.N changes it is normally an ABI breakage
2019-01-24 16:26:38 <tkzv> They did change binary format.
2019-01-24 16:46:05 <ncopa> tkzv: i pushed update
2019-01-24 16:46:48 <tkzv> @ncopa: THANKS!
2019-01-24 19:38:19 <unrznbl[m]> any thoughts on how to "pip install" into python2.7 site-packages when you also have python3 installed? seems like normally there is a pip3 and pip2 for each and pip is a symlink to one or the other.
2019-01-24 19:55:03 <smemsh> unrznbl[m]: if you use the right pip, ie the one associated with your py2.7, it should work correctly
2019-01-24 19:57:36 <unrznbl[m]> yeah, that's what I thought, but I don't have one! Oh wait... I think I might know why... I did a pip3 install upgrade of pip. I'll start over by removing pythons and such. :)
2019-01-24 19:58:13 <unrznbl[m]> maybe the upgrade clobbered /usr/bin/pip which was v2.
2019-01-24 20:42:52 <smemsh> ncopa: thanks for the hid-apple addition.  unfortunately the keyboard still does not work.  this module was required for working keyboard under debian, so i'm at a loss to explain.  the new kernel in 3.9-rc5 (4.19.17) corrupts the framebuffer too (supposed to be fixed in 4.20) but i will figure out a way to boot and what the real needed module is (possibly applespi?) for keyboard and let you know
2019-01-24 20:51:14 <aggaz> I have alpine on a RPi1B. After boot rc-status always shows me that acpid is crashed. I can not rebbot the service because "Failed to stop busybox acpid". I do not understand if this is a problem or if it is normal for a RPi. Any idea?
2019-01-24 21:03:08 <_ikke_> aggaz: try rc-service zap acpid
2019-01-24 21:03:12 <_ikke_> and then start it again
2019-01-24 21:06:28 <chrisrx> Hi there. I've got a newbie question: on a 64 bit machine, can I install gcc-32bit ? I can do that on my opensuse box, but I haven't seen any relevant package in alpine
2019-01-24 21:06:29 <chrisrx> Thanks
2019-01-24 21:07:14 <_ikke_> Alpine does not ship any cross compilers afaik
2019-01-24 21:07:35 <chrisrx> Ah ok
2019-01-24 21:07:47 <chrisrx> So if I wanted to compile in 32 bit, I'd have to be running on a 32bit proc?
2019-01-24 21:08:13 <_ikke_> I think a 32-bit chroot would even work
2019-01-24 21:08:35 <chrisrx> Interesting! Any idea how I would go about getting that up and running?
2019-01-24 21:08:46 <_ikke_> alpine itself is using lxc
2019-01-24 21:09:07 <chrisrx> (I haven't used a chroot since 2006 when I had an amd64 processor but most debian packages were for 32 bits only)
2019-01-24 21:10:47 <aggaz> _ikke_: thank you for your reply. Indeed I can stop acpid with "rc-service acpid zap", the following "rc-service acpid start" runs fine, but "rc-status" still show acpid as crashed.
2019-01-24 21:11:14 <_ikke_> can you check if the process is running?
2019-01-24 21:11:49 <aggaz> you mean like with a ps ?
2019-01-24 21:12:15 <aggaz> I can not see with ps aux
2019-01-24 21:12:24 <aggaz> it seems is not rnning
2019-01-24 21:14:29 <aggaz> if I stop acpid with zap, and the run it on the shell ("acpid -d") I get "acpid: /dev/input/event0: No such file or directory"
2019-01-24 21:15:07 <aggaz> is it possible that RPi1B is missing some pieace of hardware to use acpid?
2019-01-24 21:19:43 <smemsh> so i915.modeset=0 fixes the framebuffer corruption i see in 3.9-rc5, but then xorg will not work.  external keyboard does work.  i think spi_pxa2xx_pci.ko and spi_pxa2xx_platform.ko are needed to get the native keyboard to work.  i will try making my own installer iso with kernel 4.20 and these additional drivers so i can test
2019-01-24 21:20:15 <smemsh> aggaz: acpi is a pc thing, probably the rpi doesn't have it
2019-01-24 21:20:56 <jwh> the 64bit one maybe
2019-01-24 21:21:07 <aggaz> smemsh: thank you
2019-01-24 21:21:26 <jwh> as armv8 does have uefi and acpi (finally)
2019-01-24 21:22:47 <aggaz> uefi? does it mean that we will not need u-boot anymore?
2019-01-24 21:23:09 <jwh> no idea how or if any vendors are implementing it
2019-01-24 21:24:19 <aggaz> anyway, I suppose I can just take off acpid from the runlevel.
2019-01-24 21:24:38 <aggaz> and live without it
2019-01-24 21:25:18 <jwh> hmmm, chances of getting alpine booted on this exynos?
2019-01-24 21:25:37 <jwh> between 0 and 100, 0 being lol no and 100 being barely
2019-01-24 21:27:06 <mps> aggaz: on my arm32 boxes acpid works fine
2019-01-24 21:27:40 <aggaz> jwh: If I understand correctly, Exynos processors are used on Samsung Galaxy S III (read from wikipedia) and if I remember correctly postmarketOS runs on that phone and it is alpine based
2019-01-24 21:29:21 <aggaz> mps: ummm. Do you think it is possible that it is a RPi1B related problem or do you think something is wrong with my configuration?
2019-01-24 21:30:16 <jwh> yeah
2019-01-24 21:30:20 <mps> never worked with RPis, but in my experience on different arm32 it always worked
2019-01-24 21:30:29 <jwh> I guess the real test is if armhf will just work
2019-01-24 21:30:56 <jwh> then I can go about x compiling
2019-01-24 21:31:29 <jwh> otherwise it's going to be a faff to bring it up
2019-01-24 21:32:03 <mps> I had armhf on few boxes and worked but I usually build customized kernels for my use cases
2019-01-24 21:32:37 <jwh> armhf is alright for the exynos armv7 chips right?
2019-01-24 21:32:54 <mps> now I upgraded to armv7 and just one is left on armhf, but it will be upgraded when I found some time
2019-01-24 21:33:45 <aggaz> mps: ok thank you for your input. Now I am confused... I am not even sure if I really need acpid on the RPi
2019-01-24 21:34:08 <mps> I have exynos 5800 chromebook (samsung) and run on it armhf for more then two years, month ago upgraded to armv7 edge
2019-01-24 21:35:10 <mps> aggaz: if you want to control power button or some other events it is really useful
2019-01-24 21:35:27 <aggaz> I don't even have a butto :D
2019-01-24 21:35:42 <mps> on my chromebook acpid works even for suspend-to-ram
2019-01-24 21:36:21 <aggaz> it looks like I do not even have a device for acpid to play with
2019-01-24 21:36:34 <aggaz> (missing /dev/input/event0)
2019-01-24 21:37:00 <aggaz> in my understanding it can be either a kernel issue or a lack of hardware. correct me if I am wrong
2019-01-24 21:38:35 <jwh> mps: hm
2019-01-24 21:38:42 <mps> huh, didn't worked with devices which don't have /dev/input/eventX so not sure, but could be
2019-01-24 21:43:05 <mps> a year ago I've built real acpid to test acpi on some arm32 because in Alpine is busybox applet, not acpid like in other 'big' distros
2019-01-24 21:43:21 <jwh> hm
2019-01-24 21:43:43 <jwh> linux-vanilla or custom, question is if all the drivers are there for vanilla
2019-01-24 21:44:07 <mps> on arm's I use custom kernels
2019-01-24 21:44:20 <jwh> yeah, I've always used custom ones because its a nightmare
2019-01-24 21:45:05 <mps> because linux-vanilla will not work on chromebooks
2019-01-24 21:45:59 <jwh> guess I should at least see if it boots
2019-01-24 21:46:28 <mps> I've booted some SBC's with vanilla but vanilla is not have enabled/configured drivers and options I need
2019-01-24 21:46:50 <jwh> yeah thats normally the problem
2019-01-24 21:47:33 <jwh> also my intel stick broke, crc32.ko has magically disappeared
2019-01-24 21:48:09 <mps> Armbian have kernels customized for some SBC's
2019-01-24 21:53:41 <Tetsumaki> upgrade to 3.9 with success ;) thanks ncopa and many other contributers
2019-01-24 21:56:45 <jwh> umm
2019-01-24 21:56:58 <jwh> hm
2019-01-24 22:06:04 <aggaz> well, thank you all for the suggestions. see ya
2019-01-24 22:51:12 <chrisrx> I have a weird issue in an alpine docker image ("alpine:latest", an official release)
2019-01-24 22:51:22 <chrisrx> There's a symbolic link which doesn't seem to be found
2019-01-24 22:51:34 <chrisrx> But switching to a non-alpine image works well
2019-01-24 22:51:40 <chrisrx> Any idea how that could happen?
2019-01-25 07:31:31 <aGaTHoS> i have changued the craphics card and now xorg dont founds any screen
2019-01-25 07:32:19 <aGaTHoS> also sway does not work
2019-01-25 07:32:28 <aGaTHoS> shoudl work with nouveau drivers?
2019-01-25 08:17:00 <Necktwi> ls
2019-01-25 08:18:07 <asyd> No such file or directory
2019-01-25 08:18:21 <Necktwi> is there a ninja musl patch for st.st_mtim.tv_nsec
2019-01-25 08:19:51 <aGaTHoS> it seems noveau does not work for me neither nv
2019-01-25 08:20:07 <aGaTHoS> nv does not have support for new cards
2019-01-25 08:20:12 <aGaTHoS> and nouveau probably the same
2019-01-25 08:20:34 <aGaTHoS> anyone is using propietary nvidia drivers?
2019-01-25 08:20:40 <aGaTHoS> and can help me to install?
2019-01-25 08:22:28 <Necktwi> where can i find alpine source files of various software like ninja?
2019-01-25 08:23:15 <aGaTHoS> source files?
2019-01-25 08:23:37 <aGaTHoS> you mean builds scritps?
2019-01-25 08:24:04 <Necktwi> does build scripts include patches for source files?
2019-01-25 08:24:13 <aGaTHoS> yes
2019-01-25 08:24:28 <Necktwi> where can i find them?
2019-01-25 08:24:36 <aGaTHoS> but the sources are just downloaded from official packages websites
2019-01-25 08:24:53 <Necktwi> yeah i got that
2019-01-25 08:25:02 <aGaTHoS> in the git alpine repo
2019-01-25 08:25:16 <Necktwi> i'm looking for ninja patches
2019-01-25 08:26:14 <aGaTHoS> git.alpinelinux.org
2019-01-25 08:26:28 <aGaTHoS> aports
2019-01-25 08:28:02 <Necktwi> yeah browsing it but can't find any patches for ninja
2019-01-25 08:28:59 <aGaTHoS> cant help you
2019-01-25 08:48:54 <aGaTHoS> it seems there is a sgrrand/alpine-pkg-glibc for using progs compiled against glibc, im try this to see if can install nvidia drivers but how can i install this package??
2019-01-25 09:01:50 <aGaTHoS> it does not work
2019-01-25 09:01:56 <aGaTHoS> nor libc6-compat
2019-01-25 09:11:29 <aGaTHoS> ok plenty full of errors trying to compile the nvidia drivers
2019-01-25 09:15:26 <aGaTHoS> http://sprunge.us/P4J081
2019-01-25 09:15:31 <aGaTHoS> can anyone help me with this?
2019-01-25 09:22:37 <AinNero> aGaTHoS: paste is not helpful
2019-01-25 09:22:53 <AinNero> are those vim codes?
2019-01-25 09:23:33 <aGaTHoS> ?
2019-01-25 09:24:18 <AinNero> what you just posted in the paste
2019-01-25 09:24:54 <aGaTHoS> ive pasted
2019-01-25 09:24:56 <aGaTHoS> the test
2019-01-25 09:25:01 <aGaTHoS> not the nvidia compilation output
2019-01-25 09:25:40 <aGaTHoS> http://sprunge.us/gLgJcE
2019-01-25 09:26:21 <aGaTHoS> this is
2019-01-25 09:30:12 <AinNero> standard alpine toolchain?
2019-01-25 09:32:36 <aGaTHoS> i supose
2019-01-25 09:36:52 <AinNero> where does, for example the structleak_plugin.so come from?
2019-01-25 09:37:01 <AinNero> symbol not found sounds like the file is present
2019-01-25 09:38:02 <AinNero> oh, its shipped as source with linux-*-dev
2019-01-25 09:38:17 <aGaTHoS> i supose
2019-01-25 09:38:25 <aGaTHoS> it asked for the kernel sources
2019-01-25 09:38:34 <aGaTHoS> and i installed linux-hardened-dev
2019-01-25 09:38:59 <AinNero> ... do we even ship those anymore?
2019-01-25 09:39:14 <aGaTHoS> do you even ship kernels?
2019-01-25 09:39:20 <AinNero> yeah, but not hardened
2019-01-25 09:39:23 <AinNero> for a year now
2019-01-25 09:39:30 <aGaTHoS> because for having a kernel and sources i have to uncomment 3.6 repo
2019-01-25 09:39:45 <aGaTHoS> in edge there is no kernel
2019-01-25 09:39:59 <aGaTHoS> ah there is no hardened...
2019-01-25 09:40:04 <aGaTHoS> which is the name?
2019-01-25 09:40:18 <AinNero> .. so you were mixing 3.6 stuff with what? edge?
2019-01-25 09:40:32 <AinNero> there is linux-vanilla
2019-01-25 09:40:42 <AinNero> and linux-vanilla, respectively
2019-01-25 09:40:49 <AinNero> *linux-vanilla-dev
2019-01-25 09:41:10 <aGaTHoS> i remember before reinstall i tried to update kernel and from edge searching for linux-* didnt found anything
2019-01-25 09:41:26 <aGaTHoS> so i've installed from a livecd
2019-01-25 09:41:35 <aGaTHoS> then update repos
2019-01-25 09:41:37 <AinNero> you were likely wrong, i think the people would notice if they dont have a kernel
2019-01-25 09:41:37 <aGaTHoS> to edge
2019-01-25 09:41:47 <aGaTHoS> and make a apk upgrade
2019-01-25 09:41:56 <aGaTHoS> but if i want to install sources
2019-01-25 09:42:07 <aGaTHoS> have to uncomment the previous repo
2019-01-25 09:42:16 <aGaTHoS> you say there is vanilla
2019-01-25 09:42:18 <aGaTHoS> ok
2019-01-25 09:42:22 <aGaTHoS> what i have to do
2019-01-25 09:42:30 <aGaTHoS> uninstall linux-hardened
2019-01-25 09:42:35 <aGaTHoS> and install linux-vanilla?
2019-01-25 09:43:04 <AinNero> in theory, yes, but i dont know what your repositories file looks like
2019-01-25 09:43:37 <aGaTHoS> main community and testing from edge now
2019-01-25 09:43:45 <aGaTHoS> and i see linux-vanilla
2019-01-25 09:43:53 <AinNero> also, linux-hardened did not get the meltdown/kpti patches
2019-01-25 09:44:01 <aGaTHoS> so its safe to delete linux-hardened and install linux-vanilla?
2019-01-25 09:44:25 <AinNero> yes, the apk install should change your boot config accordingly
2019-01-25 09:44:42 <AinNero> you use syslinux?
2019-01-25 09:45:01 <aGaTHoS> mmm
2019-01-25 09:45:01 <AinNero> as bootloader?
2019-01-25 09:45:06 <aGaTHoS> i have zfs
2019-01-25 09:45:11 <aGaTHoS> not now
2019-01-25 09:45:13 <aGaTHoS> efi
2019-01-25 09:45:15 <aGaTHoS> refind
2019-01-25 09:45:53 <AinNero> oh, there are ongoing bugs related to that https://bugs.alpinelinux.org/issues/9903
2019-01-25 09:45:55 <aGaTHoS> cant uninstall linux-hardened because of zfs-hardened
2019-01-25 09:46:13 <aGaTHoS> and if i uninstall this is gonna remove the zfs module, well not from memory right
2019-01-25 09:46:43 <AinNero> not from memory, but you would need to restart soon-isn
2019-01-25 09:46:47 <AinNero> *soon-ish
2019-01-25 09:47:29 <aGaTHoS> gonna make it now
2019-01-25 09:47:58 <AinNero> uhm
2019-01-25 09:48:00 <AinNero> another note
2019-01-25 09:48:15 <AinNero> grub-mkconfig + ZFS + mkinitfs doesn't work together currently
2019-01-25 09:48:30 <aGaTHoS> i dont use grub
2019-01-25 09:48:39 <AinNero> ah right, no problem then
2019-01-25 09:48:42 <aGaTHoS> syslinux or now refind
2019-01-25 09:48:54 <AinNero> dont know much about EFI stuff
2019-01-25 09:49:10 <aGaTHoS> well now installing linux-vanilla... why are installed all firmwares?
2019-01-25 09:49:38 <aGaTHoS> its just like grub or syslinux but with fat partition
2019-01-25 09:50:06 <aGaTHoS> and executed in bios
2019-01-25 09:50:24 <AinNero> ... no BIOS, thats the purpose of EFI
2019-01-25 09:50:39 <aGaTHoS> well i mean is executed in the motherboard
2019-01-25 09:50:42 <AinNero> i know that, but like, my testing machine with EFI all were crap
2019-01-25 09:50:47 <AinNero> so no EFI support from me
2019-01-25 09:50:52 <aGaTHoS> not installed in the firs hd sector
2019-01-25 09:51:10 <aGaTHoS> i just installed refind
2019-01-25 09:51:18 <aGaTHoS> is like an efi grub or syslinux
2019-01-25 09:51:34 <aGaTHoS> just edit a file and put very similar entries like in grub or syslinux
2019-01-25 09:51:39 <aGaTHoS> also it scans
2019-01-25 09:51:44 <AinNero> stop
2019-01-25 09:52:04 <aGaTHoS> ?
2019-01-25 09:52:11 <AinNero> i know that, but like, my testing machine with EFI all were crap
2019-01-25 09:52:14 <AinNero> so no EFI support from me
2019-01-25 09:52:20 <AinNero> no need to explain me the stuff
2019-01-25 09:52:20 <aGaTHoS> ok
2019-01-25 09:53:16 <aGaTHoS> its really needed to install every firmware with kernel?
2019-01-25 09:53:20 <AinNero> no
2019-01-25 09:53:32 <aGaTHoS> its installing 87 packages as dependencies
2019-01-25 09:53:48 <AinNero> you can pick the firmware you need and add it to world
2019-01-25 09:53:54 <aGaTHoS> all linux-firmware-*
2019-01-25 09:54:00 <AinNero> then the other firmwares will be uninstalled
2019-01-25 09:54:24 <aGaTHoS> i have not specified one why is installing me all?
2019-01-25 09:54:49 <AinNero> because thats the default
2019-01-25 09:55:46 <AinNero> install a specific linux-firmware-*, and it will uninstall all you have not manually installed
2019-01-25 09:57:38 <aGaTHoS> do i need someone?
2019-01-25 09:58:43 <AinNero> depends on your hardware
2019-01-25 09:58:58 <AinNero> if you dont know which firmware blobs you need, just dont touch the firmwares and leave all installed
2019-01-25 09:59:20 <aGaTHoS> makes no sense
2019-01-25 10:00:05 <AinNero> if you dont know which firmware you need, you should probably leave them installed as it is per default
2019-01-25 10:00:40 <aGaTHoS> well is done gonna restart and try to compile again
2019-01-25 10:03:55 <aGaTHoS> im back, now a different error
2019-01-25 10:04:23 <aGaTHoS> failed to execute /sbin/ldconfig
2019-01-25 10:05:07 <aGaTHoS> but it seems have compiler
2019-01-25 10:05:09 <aGaTHoS> compiled
2019-01-25 10:06:04 <aGaTHoS> wow
2019-01-25 10:06:09 <aGaTHoS> they are even loaded in memory
2019-01-25 10:06:32 <AinNero> /sbin/ldconfig is part of musl-utils
2019-01-25 10:07:21 <aGaTHoS> they are installed
2019-01-25 10:07:37 <aGaTHoS> now i have nvidia kernel module loaded and also nouveau
2019-01-25 10:07:46 <aGaTHoS> should remove nouveau i suppose?
2019-01-25 10:08:26 <aGaTHoS> ohh segfault
2019-01-25 10:08:33 <aGaTHoS> when trying to run xorg
2019-01-25 10:09:29 <aGaTHoS> failed to load glamorgl
2019-01-25 10:09:43 <aGaTHoS> is this necesary for nvidia or just a module xorg tries to load?
2019-01-25 10:10:03 <AinNero> i dont know
2019-01-25 10:10:09 <aGaTHoS> it seems it loads nouveau
2019-01-25 10:10:19 <aGaTHoS> even if i have put autoload false
2019-01-25 10:10:24 <aGaTHoS> and commented this module
2019-01-25 10:10:37 <aGaTHoS> i think have to use a xorg.conf traditional file
2019-01-25 10:11:23 <aGaTHoS> which more libs i need
2019-01-25 10:11:41 <aGaTHoS> the nvidia drivers have it all? where thwy put their opengl libraries etc?
2019-01-25 10:12:38 <aGaTHoS> http://sprunge.us/2pvPSD
2019-01-25 10:12:46 <aGaTHoS> my xorg.log.0
2019-01-25 10:13:37 <aGaTHoS> trying sway with wayland ways
2019-01-25 10:14:18 <aGaTHoS> Error relocating /usr/lib/libEGL.so.1: __strtod_internal: symbol not found
2019-01-25 10:15:08 <aGaTHoS> should i link
2019-01-25 10:15:25 <aGaTHoS> libEGL* to libEGL_nvidia.*?
2019-01-25 10:15:49 <aGaTHoS> i dont know which libraries has installed the nvidia driver appart from kernel module
2019-01-25 10:16:10 <aGaTHoS> buty i suppose i have to link lib*gl* to the nvidia ones
2019-01-25 10:16:12 <aGaTHoS> right?
2019-01-25 10:17:14 <aGaTHoS> i remember years ago in the past when distros still dont automatically install nvidia drivers when i download compile ans install the nvidia drivers by hand i had to chose between nvidia gl libearies and mesa
2019-01-25 10:18:06 <aGaTHoS> weel time to read the nvidia drivers manual thought
2019-01-25 10:21:18 <aGaTHoS> maybe that linker error
2019-01-25 10:21:20 <aGaTHoS> ...
2019-01-25 10:22:39 <aGaTHoS> says after its compiled the linker links with the closed-source portion
2019-01-25 10:22:46 <aGaTHoS> and that step it failed
2019-01-25 10:22:52 <aGaTHoS> need a glibc linker?
2019-01-25 10:28:40 <aGaTHoS> seems all the egl libs are from nvidia installation
2019-01-25 10:30:54 <aGaTHoS> so i dont have to link anything
2019-01-25 10:40:20 <aGaTHoS> sway does not work weston does not work xorg does not work
2019-01-25 10:40:24 <aGaTHoS> no graphics
2019-01-25 10:43:33 <aGaTHoS> which package belongs weston-launch?
2019-01-25 10:48:51 <AinNero> thats searchable here: https://pkgs.alpinelinux.org/contents
2019-01-25 10:52:16 <aGaTHoS> how i do effective me added to a group without logoff?
2019-01-25 10:53:36 <aGaTHoS> weston-launch does nothing
2019-01-25 10:53:45 <aGaTHoS> sway gives error xorgs segsfaults
2019-01-25 10:53:49 <aGaTHoS> fail
2019-01-25 10:57:04 <aGaTHoS> http://sprunge.us/QSfw5V
2019-01-25 10:57:13 <aGaTHoS> this is my compilation error
2019-01-25 10:57:35 <aGaTHoS> the driver is even compiled and installed
2019-01-25 10:57:47 <aGaTHoS> which makes this part? this ldconfig?
2019-01-25 11:00:56 <aGaTHoS> and xorg says failed to load module "nvidia" (module doesnt support this OS, 0)
2019-01-25 11:00:58 <aGaTHoS> :S
2019-01-25 11:02:03 <aGaTHoS> http://sprunge.us/BY65Nd
2019-01-25 11:02:13 <aGaTHoS> if someone can help me...
2019-01-25 11:04:13 <mps> aGaTHoS: https://www.youtube.com/watch?v=iYWzMvlj2RQ
2019-01-25 11:07:40 <aGaTHoS> i think instead of /sbin/ldconfig it should use /usr/glibc-compat/sbin/ldconfig
2019-01-25 11:07:58 <aGaTHoS> maybe i can move the muslc one temporary out and link this one
2019-01-25 11:08:07 <aGaTHoS> and recompile the driver
2019-01-25 11:08:46 <mps> aGaTHoS: look at void linux, try to boot it and see if it works
2019-01-25 11:09:16 <mps> if it works then look at their build files for nvidia
2019-01-25 11:09:40 <aGaTHoS> im in text mode
2019-01-25 11:10:09 <aGaTHoS> cant follow this link and coping to the mobile hadn't worked
2019-01-25 11:10:20 <aGaTHoS> tell me the title of the video so i can search for it in the mobile
2019-01-25 11:10:49 <mps> that was joke, famous 'nvidia f*ck you'
2019-01-25 11:11:59 <aGaTHoS> :S it does not appear me
2019-01-25 11:12:08 <aGaTHoS> gonna try to copy well again
2019-01-25 11:13:38 <aGaTHoS> ok gonna enter here from mobile
2019-01-25 11:13:39 <mps> void linux have nvidia proprietary drivers
2019-01-25 11:13:59 <mps> void linux is also musl based
2019-01-25 11:14:05 <aGaTHoS> i know
2019-01-25 11:14:26 <aGaTHoS> but i didnt know they have propietary drivers
2019-01-25 11:14:46 <mps> so, you can look at their build scripts and patches
2019-01-25 11:14:58 <aGaTHoS> but first have to install something...
2019-01-25 11:15:11 <aGaTHoS> because again as allways my previous system has crashed
2019-01-25 11:15:25 <aGaTHoS> this time an external hd in which i had the linux partition
2019-01-25 11:16:10 <aGaTHoS> ok gonna get void linux or freebsd and when i have graphical interface gonna research their nvidia build
2019-01-25 11:16:24 <aGaTHoS> do they also use a build system i suppose?
2019-01-25 11:16:32 <aGaTHoS> not precompiled packages right?
2019-01-25 11:17:00 <aGaTHoS> but first gonna make this trick
2019-01-25 11:17:18 <aGaTHoS> move ldconfig and link to libc-compat one
2019-01-25 11:17:27 <aGaTHoS> to see if i can trick the nvidia installer
2019-01-25 11:19:28 <aGaTHoS> ok
2019-01-25 11:19:31 <aGaTHoS> this time no errors
2019-01-25 11:19:38 <aGaTHoS> lets see if works
2019-01-25 11:21:50 <aGaTHoS> still segsfaults
2019-01-25 11:22:52 <aGaTHoS> can i know to which module belongs the seg fault adress?
2019-01-25 11:24:59 <aGaTHoS> module doesn't support this OS, 0
2019-01-25 11:25:06 <aGaTHoS> why xorg.log says that
2019-01-25 11:25:08 <aGaTHoS> ?
2019-01-25 11:25:59 <mps> void linux have their package scripts called xbps, and their build system are in alpine repo, just 'apk search xbps'
2019-01-25 11:27:02 <aGaTHoS> so what should i do with xbp
2019-01-25 11:27:04 <aGaTHoS> s
2019-01-25 11:29:05 <mps> read and look if you could build nvidia on Alpine
2019-01-25 11:30:56 <aGaTHoS> do you think it could be efi related? im reading that some oses need the module to be signed
2019-01-25 11:31:14 <aGaTHoS> and that error of the module does not support this os...
2019-01-25 11:31:37 <aGaTHoS> i dont thi nk i have secure boot enabled
2019-01-25 11:32:12 <aGaTHoS> is not that then the module even will not be loaded
2019-01-25 11:32:14 <mps> it shouldn't be efi related, but who knows what are done in closed software
2019-01-25 11:41:35 <FreeBDSM> hello
2019-01-25 11:41:48 <FreeBDSM> I want to use certbot on alpine
2019-01-25 11:42:12 <_ikke_> I'm using certbot on alpine
2019-01-25 11:42:18 <FreeBDSM> but unfortunately there is no 'nginx' certbot plugin available in the list of alpine packages
2019-01-25 11:42:35 <FreeBDSM> I know I could use certonly, but I'd prefer the automatism
2019-01-25 11:42:36 <_ikke_> Do you need one?
2019-01-25 11:42:38 <FreeBDSM> yeah
2019-01-25 11:42:54 <_ikke_> I use certonly with a post script to restart nginx
2019-01-25 11:43:28 <FreeBDSM> with nginx plugin certbot knows how to alter nginx vhost config files
2019-01-25 11:43:46 <_ikke_> I have a simple conf I can include to expose the right directory
2019-01-25 11:44:52 <FreeBDSM> huh?
2019-01-25 11:45:09 <_ikke_> http://tpaste.us/qKYO
2019-01-25 11:45:15 <_ikke_> that's all the nginx config you need
2019-01-25 11:45:28 <FreeBDSM> no, unfortunately not
2019-01-25 11:45:36 <_ikke_> It works for me :)
2019-01-25 11:45:53 <FreeBDSM> certbot can force https
2019-01-25 11:46:03 <FreeBDSM> it will alter http rule to redirect to https
2019-01-25 11:46:14 <_ikke_> sure
2019-01-25 11:46:23 <FreeBDSM> and it will create a https rule pointing to cert files
2019-01-25 11:46:24 <_ikke_> That's really simple as well
2019-01-25 11:46:40 <FreeBDSM> that's not simple if you do it manually
2019-01-25 11:47:02 <_ikke_> I do it manually
2019-01-25 11:47:03 <FreeBDSM> because of course you can manually change it if you have 1 host, but what if you have 50 of them?
2019-01-25 11:47:42 <_ikke_> If you have 50 hosts, you will need to manage more then just the LE part
2019-01-25 11:48:01 <_ikke_> You tend to already have a mechanism to manage the config anyway
2019-01-25 11:49:34 <FreeBDSM> 50 hosts, each serves 2 fqdns, so you suggest me to alter 100 files manually? do not want to.
2019-01-25 11:49:52 <_ikke_> No, I don't suggest that :)
2019-01-25 11:50:19 <FreeBDSM> I want certbot with nginx plugin, why can I have it on other OSes, but not on my tiny lovely alpine?
2019-01-25 11:50:32 <_ikke_> FreeBDSM: Because no one bothered to package it
2019-01-25 11:50:49 <clandmeter> where can i find that plugin?
2019-01-25 11:51:25 <FreeBDSM> https://github.com/certbot/certbot/tree/master/certbot-nginx probably this
2019-01-25 11:52:12 <_ikke_> The alpine package does a pyton setup.py on the certbot project, that apparently does not install any plugins (which will probably need to go in a subpackage anyway)\
2019-01-25 11:52:29 <clandmeter> ok so its shipped with certbot itself
2019-01-25 11:52:33 <_ikke_> yes
2019-01-25 11:52:40 <FreeBDSM> yes and no
2019-01-25 11:52:41 <_ikke_> you need to go into the subdir and run setup.py there as well
2019-01-25 11:53:14 <FreeBDSM> it can be installed via `sudo apt install python-certbot-nginx` on debians
2019-01-25 11:53:29 <_ikke_> FreeBDSM: He means the source
2019-01-25 11:54:01 <FreeBDSM> yeah, the sources, unfortunately, seem to live in a subdir of the git repo
2019-01-25 11:54:05 <clandmeter> i think FreeBDSM has a point
2019-01-25 11:54:10 <clandmeter> we should probably ship them
2019-01-25 11:54:17 <clandmeter> i thought we actally already did
2019-01-25 11:54:20 <_ikke_> I never said we shouldn't for the record
2019-01-25 11:54:40 <clandmeter> i didnt say you did ;-)
2019-01-25 11:54:52 <FreeBDSM> can confirm both of you :)
2019-01-25 11:55:07 <_ikke_> :)
2019-01-25 11:55:42 <clandmeter> hmm i think we need to refactor that apkbuild
2019-01-25 11:55:44 <_ikke_> There are a lot of plugins, would we package all of them?
2019-01-25 11:56:48 <FreeBDSM> I guess nginx and maybe apache would cover 99% cases
2019-01-25 12:00:36 <_ikke_> proof of concept: http://tpaste.us/NK5l
2019-01-25 12:08:56 <FreeBDSM> I see 3.9 and edge branches on pkgs.alpinelinux.org, but alpinelinux.org doesn't know of 3.9 release (yet)
2019-01-25 12:09:04 <FreeBDSM> is it to be released soon?
2019-01-25 12:09:38 <_ikke_> yes
2019-01-25 12:09:44 <_ikke_> currently at rc-6
2019-01-25 12:09:46 <_ikke_> -5
2019-01-25 12:13:54 <clandmeter> _ikke_, the problem is deps
2019-01-25 12:14:47 <clandmeter> i need to go
2019-01-25 12:45:51 <FreeBDSM> just fyi: `apt -y install python-certbot-nginx` will install this list of packages on debian https://paste.ee/p/RoPPS
2019-01-25 15:08:16 <FreeBDSM> why don't you want to make apk commands have synonyms?
2019-01-25 15:08:22 <FreeBDSM> apk add == apk install
2019-01-25 15:08:48 <FreeBDSM> apk del == apk remove == apk uninstall == apk deinstall
2019-01-25 15:11:15 <_ikke_> His has come up earlier, not sure what hapened though
2019-01-25 15:12:40 <FreeBDSM> is there no pip package for python3 in alpine?
2019-01-25 15:13:30 <FreeBDSM> ah, the binary name is pip3
2019-01-25 15:13:37 <FreeBDSM> and it is provided by python3
2019-01-25 15:13:39 <FreeBDSM> got it
2019-01-25 15:14:05 <_ikke_> https://git.alpinelinux.org/aports/tree/main/py2-pip/APKBUILD#n3
2019-01-25 15:14:08 <_ikke_> right
2019-01-25 15:15:10 <FreeBDSM> TT no luck: python package `certbot-nginx` fails to install under both python2 and python3 due to `Could not find a version that satisfies the requirement cffi!=1.11.3,>=1.8 (from versions: )` `No matching distribution found for cffi!=1.11.3,>=1.8`
2019-01-25 15:15:52 <_ikke_> did you install py3-cffi?
2019-01-25 15:16:15 <FreeBDSM> well, that's weird
2019-01-25 15:16:27 <FreeBDSM> if I have to use pip - it should pull packages from pip
2019-01-25 15:16:52 <FreeBDSM> but yeah, luckily there are py2-cffi, py-cffi and py3-cffi available in apk repo
2019-01-25 15:17:28 <_ikke_> If you want to officially package it, all dependencies should come from aports
2019-01-25 15:20:18 <FreeBDSM> downgraded back to python2, installed py2-cffi, py2-pip, doing `pip install certbot-nginx` and it still fails with the same error
2019-01-25 15:20:33 <FreeBDSM> doing `pip list` and seeing cffi 1.10.0
2019-01-25 15:20:42 <_ikke_> Why are you doing it that way?
2019-01-25 15:20:56 <FreeBDSM> because there's no nginx-certbot in apk repo
2019-01-25 15:21:13 <FreeBDSM> and I know it's a python program so it is available in pips
2019-01-25 15:22:03 <_ikke_> not necessarily :-)
2019-01-25 15:22:19 <_ikke_> Not all python programs are available via pypi
2019-01-25 15:22:24 <FreeBDSM> yes, but `pip search` found it
2019-01-25 15:22:50 <_ikke_> the error you are getting is a pip/python error though, very little to do with alpine
2019-01-25 15:23:14 <FreeBDSM> yeah, but isn't this ridiculous? http://i.imgur.com/9l4ppdS.png
2019-01-25 15:24:33 <_ikke_> Not sure why that happens
2019-01-25 15:26:27 <FreeBDSM> and I'm not even sure if that's pip's/python's issue or alpine's
2019-01-25 15:29:20 <smemsh> in order to make a new install iso with a particular custom package build, i will need to host my own repository with the rebuilt package and supply it to mkimage.sh --extra-repository, yes?
2019-01-25 15:30:42 <FreeBDSM> I've tried updating pip from pip (succeeded for both py2 and py3) and then install certbot-nginx package, the errors now differ: python2 failed because the lack of gcc and python3 failed because `No working compiler found`, so some progress
2019-01-25 16:12:58 <clandmeter> FreeBDSM, can you add an issue on bugs.a.o regarding certbot?
2019-01-25 16:14:10 <FreeBDSM> clandmeter: I will, once I'll have some spare time. I'm trying to just install and run from pip and turns out it's something to do with MUSL and toolkit
2019-01-25 16:35:55 <unrznbl[m]> I seem to be having some problems with compiling against libintl/gettext... can't link dgettext and libintl_gettext. I checked a debian machine and I see dgettext in libc-2.28.so but don't see that symbol anywhere in /lib on alpine. I checked and I have gettext, gettext-dev, gettext-libs all installed. On debian I don't see libintl_gettext anywhere so not sure exactly what is going on. I am trying to build gnu arm
2019-01-25 16:35:56 <unrznbl[m]> toolchain from scratch (dgettext) and weechat (libintl_gettext).
2019-01-25 16:36:21 <unrznbl[m]> i understand that alpine uses musl instead of libc so maybe that's part of the problem to get around?
2019-01-25 16:45:00 <misterunknown> unrznbl[m]: AFAIK it's not possible, since musl-libc does not support that (https://wiki.musl-libc.org/functional-differences-from-glibc.html#iconv)
2019-01-25 16:46:02 <unrznbl[m]> cool. thanks for that reference! I'm not too familiar with all of glibc and musl. Yet. ;)
2019-01-25 16:47:31 <misterunknown> #metoo, but I learned a bit about, since using alpine linux
2019-01-25 16:48:10 <FreeBDSM> I've figured the necessary build deps to install certbot-nginx via pip, is this list of any value to you as maintainers?
2019-01-25 16:48:54 <_ikke_> perhaps a list of dependencies
2019-01-25 16:51:15 <unrznbl[m]> interesting. weechat is packaged already but at least the version I am trying requires libintl, which gets me into this mess. Hmm.
2019-01-25 16:51:22 <unrznbl[m]> I was trying to build weechat with python3 so I could use wee-slack :p
2019-01-25 16:53:19 <unrznbl[m]> so... does it seem possible to build a gnu arm toolchain for alpinelinux with musl? Maybe some config option to disable or not use libintl.
2019-01-25 16:55:04 <misterunknown> FreeBDSM: Have you ever tried to package it yourself? Shouldn't be too hard. You can take a look at existing packages, e.g. https://github.com/alpinelinux/aports/blob/master/community/py3-async-timeout/APKBUILD
2019-01-25 16:57:24 <FreeBDSM> misterunknown: I don't know, I have a list of deps, but they are mixed together (build deps and runtime deps)
2019-01-25 16:57:26 <unrznbl[m]> packaging libintl? Yeah... that sounds pretty good. I have played around with aports a bit for pmos.
2019-01-25 16:57:36 <unrznbl[m]> oops. wrong thread. ;)
2019-01-25 16:58:08 <misterunknown> FreeBDSM: Nothing unusual. You can have either in your APKBUILD file.
2019-01-25 16:59:30 <misterunknown> unrznbl[m]: Didn't meant you. I don't really know how to solve your problem – maybe it's possible to compile without libintl support?
2019-01-25 17:00:07 <unrznbl[m]> misterunknown: yeah, maybe have to mangle weechat and gcc/binutils/newlib for the cross-tools. Thanks. :)
2019-01-25 17:00:23 <FreeBDSM> misterunknown: I hardly understand how any of that works. In my understanding - if someone builds stuff for me - all I need is a resulting binary and runtime deps right? so I'd need to figure out what each of packages from the list (build-base, gcc, openssl-dev, libffi, libffi-dev, python2-dev, py2-virtualenv) is just a build dep and thus not needed to be listed as a dependency for package (when it's binary is built into package)
2019-01-25 17:06:41 <mps> unrznbl[m]: libintl is packaged in Alpine
2019-01-25 17:08:58 <unrznbl[m]> mps: ah. thanks. Not sure why I didn't see that yet. I added it and still have the same problem with weechat. I'll dig into it more another time. Thanks again. :)
2019-01-25 17:09:21 <unrznbl[m]> I can see that libintl package installed a .so that has the symbols I want... so probably just a gcc option problem :)
2019-01-26 13:35:08 <xrs> hey guys, I'm setting up an alpinelinux system with i3. When starting Xorg server
2019-01-26 13:36:28 <xrs> hey guys, I'm setting up an alpinelinux system with i3. When starting Xorg server i3 will start but my build it keybard and mouse will not work. Any help is really appriciated :-)
2019-01-26 13:38:54 <xrs> I read the information of the wiki and think I have install the right packages .. I'm using a think pad
2019-01-26 13:40:32 <mps> xrs: did you installed and started udev
2019-01-26 13:41:12 <xrs> mps, yes
2019-01-26 13:41:27 <mps> sorry, eudev I mean
2019-01-26 13:43:57 <xrs> mps, eudev is installed, what is the service na�me
2019-01-26 13:44:01 <xrs> I can not start it
2019-01-26 13:44:05 <xrs> with rc-update
2019-01-26 13:44:22 <xrs> or rc-service
2019-01-26 13:45:07 <mps> to enable it at boot you have to 'rc-update add udev'
2019-01-26 13:45:24 <mps> rc-service is to stop/start by hand
2019-01-26 13:45:49 <xrs> yep, udev service is running and enabled
2019-01-26 13:46:18 <mps> actually udev should be started in sysinit runlevel
2019-01-26 13:46:49 <xrs> ok, just changed it to sysinit
2019-01-26 13:47:04 <mps> do you have xf86-input-libinput installed
2019-01-26 13:47:09 <xrs> yep
2019-01-26 13:47:38 <xrs> I followed the manuals and installed (hopefully) all necessary packages
2019-01-26 13:47:53 <mps> huh, then the problem could in kernel modules, i.e. not loaded appropriate drivers
2019-01-26 13:48:23 <xrs> i3 works with external mouse
2019-01-26 13:48:50 <xrs> I checked the xorg output and solved all errors and warnings, but the problem remains
2019-01-26 13:49:12 <mps> are there some errors in /var/log/Xorg.0.log which can say something about problem
2019-01-26 13:50:05 <xrs> This is what I mean :) There are no more errors or warnings. Except lots of infos concerning devices..
2019-01-26 13:50:34 <mps> and, your keyboard and mouse works in console mode?
2019-01-26 13:51:11 <xrs> there is an intersting hint which states: config/udev: Adding input device ... No input driver specified, ingorning this device
2019-01-26 13:51:24 <xrs> yes, both work in console mode
2019-01-26 13:51:46 <xrs> so kernel mods should not be the problem, right?
2019-01-26 13:52:13 <mps> right, something with libinput or udev, probably
2019-01-26 13:52:22 <xrs> thos hints of Xorg.�0.log are a lot, and from other systems I know they are mostly harmless
2019-01-26 13:54:50 <mps> you can try to check with evtest but not sure if it help to solve problem but you will know if there are problem with drivers
2019-01-26 13:56:39 <xrs> okay, thank you, I'll try it. could it be something with the xorg.conf?
2019-01-26 13:57:07 <xrs> I tried without, after with, also made changes by commenting out and using the standard.. nothing helped
2019-01-26 13:57:21 <mps> of course, could be, but shouldn't in default install
2019-01-26 13:58:25 <xrs> evtest is nice.. trackpoint works
2019-01-26 13:58:59 <xrs> keyboard, too
2019-01-26 13:59:55 <xrs> mps, okay, wait, there is an EE in Xorg.0.log: Failed to load module "evdev" (module does not exist)
2019-01-26 14:00:08 <xrs> interesting?
2019-01-26 14:00:51 <mps> do you use Alpine kernel or some other
2019-01-26 14:01:19 <xrs> okay, just figured out one package was missing
2019-01-26 14:01:41 <xrs> xf86-input-evdev
2019-01-26 14:01:56 <xrs> mps, it works now!
2019-01-26 14:02:29 <mps> ah, so, it didn't install automatically?
2019-01-26 14:02:31 <xrs> I'm sorry, somehow this got under the radar, thank you for your help
2019-01-26 14:02:48 <xrs> let me check, I have documented my steps
2019-01-26 14:03:51 <xrs> strange, I have it my list of installed packages
2019-01-26 14:04:04 <xrs> and used this manual: https://wiki.alpinelinux.org/wiki/Alpine_setup_scripts#setup-xorg-base
2019-01-26 14:04:43 <xrs> everything seems fine .. I guess, somewhere I made a little mistake. Again, thank you for your time. the conversation helpd :-)
2019-01-26 14:06:20 <mps> xrs: good, Alpine is for those who are ready to debug and not for those who want 'one click install' ;)
2019-01-26 14:07:15 <xrs> mps, for now, I really like alpinelinux.. if you are part of the community, thanks for this wonderful system
2019-01-26 14:08:34 <mps> xrs: we all are part of the community, and looks like we have one more member right now ;)
2019-01-26 14:10:08 <xrs>  :)
2019-01-26 15:36:20 <emersont1> hello, i don't know if you are aware, but the channel's log is being posted to github here https://github.com/grobe0ba/logs/
2019-01-26 15:36:40 <_ikke_> logs are considered public anyway
2019-01-26 15:38:18 <emersont1> it's a bit of a grey area
2019-01-26 15:45:18 <_ikke_> irc log services have existed for a long time
2019-01-26 15:45:32 <_ikke_> though now with the gdpr, many seem to have stopped
2019-01-26 16:37:22 <aspyct> hi all :) I need to automatically mount a usb drive when plugged in to my alpine box. After a short lookup, it seems autofs, udev and co are not available on alpine. any suggestion on how I might do that?
2019-01-26 16:44:38 <_ikke_> udev should be available
2019-01-26 16:45:03 <_ikke_> it was at least
2019-01-26 16:47:51 <_ikke_> apparently eudev
2019-01-26 16:53:58 <aspyct> _ikke_: thanks, will check it out!
2019-01-26 16:54:16 <TBB> udev alone won't do it though
2019-01-26 16:54:36 <TBB> well, kinda can, just not right
2019-01-26 16:55:53 <TBB> but basically udev + udisks can do it right, you just might need some extra configuration for policykit
2019-01-26 17:03:30 <aspyct> ah, didn't know about udisks. good tip!
2019-01-26 17:47:21 <hanetzer> yeah, udisks is useful
2019-01-26 17:55:19 <aspyct>  `udisksctl status` fails with "Error connecting to the udisks daemon: Could not connect: No such file or directory". anyone knows what's wrong?
2019-01-26 17:55:40 <aspyct> what's the daemon I should start?
2019-01-26 17:58:41 <_ikke_> Hmm, I don't see a daemon in the package
2019-01-26 18:01:37 <aspyct> (I should probably state that I'm using a raspberry pi)
2019-01-26 18:06:54 <_ikke_> aspyct: so there is a binary called udisks-daemon, it just has no open-rc service definition
2019-01-26 18:07:37 <aspyct> what's the full path? I can't seem to find it
2019-01-26 18:08:15 <_ikke_> https://pkgs.alpinelinux.org/contents?file=udisks-daemon&path=&name=&branch=edge
2019-01-26 18:08:50 <_ikke_> If I read the man-page of that daemon, I get the idea that it relies on systemd/dbus
2019-01-26 18:09:02 <_ikke_> "administrators should never need to start this daemon as it will be automatically started by dbus-daemon(1) whenever an application calls into the org.freedesktop.UDisks service."
2019-01-26 18:09:17 <aspyct> ah, maybe I need another package then
2019-01-26 18:09:34 <_ikke_> dbus I guess
2019-01-26 18:09:43 <aspyct> let's try
2019-01-26 18:10:28 <aspyct> indeed!
2019-01-26 18:10:29 <aspyct> thx
2019-01-26 18:10:44 <aspyct> dbus needs to be installed and the service started, then all works fine
2019-01-26 18:11:01 <_ikke_> good
2019-01-26 18:11:01 <aspyct> maybe it should be added as a dependency to the udisks2 package?
2019-01-26 18:11:30 <_ikke_> aspyct: Could you open an issue about it on bugs.a.o? Then someone can take a look at it
2019-01-26 18:12:03 <aspyct> sure!
2019-01-26 18:14:46 <TBB> hm hm. thought I'd give slim another try, but it doesn't seem to play well together with i3
2019-01-26 18:15:47 <TBB> the desktop session lives for a second, then terminates and goes back to slim login, and the window manager stays on the background eating up 100% of a core
2019-01-26 18:19:41 <TBB> fortunately, I don't really -need- a graphical login manager, the session works just fine from commandline with startx
2019-01-26 18:21:08 <TBB> oh, and udisks is usable without dbus via command line, so that's probably why dbus isn't a dependency... alpine's dependency policy seems to be "no deps unless absolutely necessary"
2019-01-26 18:23:11 <TBB> ah, seems your fresh bug report proves otherwise :)
2019-01-26 18:24:01 <aspyct> well I might have missed something, feel free to point it out. until 2h ago I used alpine only in my docker containers :)
2019-01-26 18:24:41 <aspyct> could be that something else was off too
2019-01-26 18:24:44 <TBB> I've been off Alpine for about 3 weeks due to my work project where I used it exclusively ending at the end of last year
2019-01-26 18:25:22 <aspyct> it looks pretty cool and clean I have to say!
2019-01-26 18:25:26 <TBB> and now that I'm setting up a workstation at home so I can keep/start contributing I notice I've forgotten a lot in such a short time it's not even funny :D
2019-01-26 18:25:34 <aspyct> lol
2019-01-26 18:26:01 <TBB> I mean, in December I could've just given you a straight-up complete response on how to make usb media automountable
2019-01-26 18:26:27 <TBB> provide you with the configs and all... now I no longer remember the details, and of course all my notes on the topic...
2019-01-26 18:26:36 <TBB> ...are at the office where I no longer have access.
2019-01-26 18:27:10 <aspyct> ah well
2019-01-26 18:27:24 <aspyct> it's an opportunity for me to take some notes as well :) did you just leave that job?
2019-01-26 18:28:39 <TBB> well, let's put it this way: the project reached release status and the end customer froze the release, next release is planned for sometime next year and they're not willing to pay for development during the freeze
2019-01-26 18:29:05 <TBB> which is going to be funny because they won't have anyone developing it
2019-01-26 18:29:46 <TBB> but we basically replaced an aging and bloated Red Hat based product with a lean mean Alpine machine
2019-01-26 18:30:25 <TBB> let me tell you, it was an effing painful transition but in the end I'd say it paid off :)
2019-01-26 18:30:30 <aspyct> freeze for a year? that doesn't sound like a good planning decision...
2019-01-26 18:30:48 <aspyct> well I'm using fat distros for work as well, centos mostly
2019-01-26 18:30:55 <aspyct> but that's really because we need openshift
2019-01-26 18:31:14 <aspyct> I'd be happy to go for leaner distros, but I have to say openshift is pretty comfortable...
2019-01-26 18:31:14 <TBB> ah well, I'm just a consultant, decisions good or bad are for other people to make. and I got a good 6 years of work out of it so no complains from me, although I would've loved to continue working on it
2019-01-26 18:31:39 <TBB> yeah, there are reasons to use RHEL/CentOS
2019-01-26 18:32:00 <aspyct> if I may (and if no NDA is involved...), what's the product?
2019-01-26 18:32:31 <TBB> can't really talk about it due to the reason you mentioned
2019-01-26 18:33:24 <aspyct> classic :)
2019-01-26 18:33:38 <aspyct> too bad you didn't reach release though, I know the feeling
2019-01-26 18:33:49 <mps> TBB: what's wrong with the slim in your case, it works for me on different boxes without any issue
2019-01-26 18:33:59 <aspyct> I once had a project suddenly cancelled. the morning we were working, the afternoon we were... off, I guess
2019-01-26 18:34:40 <TBB> mps, that's what I'm trying to figure out, I've got i3 start in .xinitrc and slim runs it just like it should but the session dies after a second
2019-01-26 18:35:34 <mps> also, I start awesome in .xinitrc and have no issues. could be something with i3
2019-01-26 18:35:43 <TBB> aspyct, well, I was responsible for the OS side of things and it had to be ready before the rest of the stuff... so what I did had to be at release level before the rest of the guys
2019-01-26 18:36:45 <mps> and my wife have xfce also with slim as XDM, no issues
2019-01-26 18:37:22 <TBB> mps, I've done this same thing with lxdm and that combo used to work. but I wouldn't mind switching (back) to slim
2019-01-26 18:40:33 <mps> TBB: however you want, but if you think that the slim have a bug please fill issue because I'm interested to keep it healthy on Alpine
2019-01-26 18:40:47 <TBB> ah, you're the maintainer?
2019-01-26 18:41:30 <mps> no, I'm not, but use it and want it to stay in Alpine
2019-01-26 18:42:31 <mps> even rebuilt it without consolekit and polkit dependecies for me
2019-01-26 21:18:50 <succfemboi> Is it possible to install alpine on an armv7 chromebook?
2019-01-26 21:19:50 <_ikke_> alpine started to build for armv7, but I'm not sure if a chromebook needs any special provisions
2019-01-26 21:22:26 <succfemboi> I can't really find much about installing distros on that type of device, especially on arm. One can run crouton on chromeos which generates a chroot but I can't run any browser within it
2019-01-26 21:22:54 <jwh> oh, this is relevant to my interests
2019-01-26 21:23:07 <jwh> is there any intention of doing something similar to alarm?
2019-01-26 21:23:28 <succfemboi> Idk what alarm is
2019-01-26 21:23:36 <jwh> archlinuxarm
2019-01-26 21:23:40 <succfemboi> oh
2019-01-26 21:24:23 <mps> succfemboi: I have it on armv7 chromebook, samsung exynos peach pi
2019-01-26 21:25:11 <succfemboi> I've used archlinuxarm on a raspberry pi, but I did boot after a while probably because it's rolling release.
2019-01-26 21:26:03 <jwh> I meant more the packaged kernels and tarballs to get going
2019-01-26 21:26:04 <succfemboi> I have a chromebook Samsung exynos5 (flattened device tree)
2019-01-26 21:26:24 <jwh> I was about to start getting alpine on my exynos odroid, so I was thinking about creating something similar
2019-01-26 21:27:01 <succfemboi> Might make it easier to install and run then when properly packaged so yeah
2019-01-26 21:27:04 <jwh> yea
2019-01-26 21:27:17 <mps> succfemboi: exynos5 is it exynos5800
2019-01-26 21:27:31 <succfemboi> Idk
2019-01-26 21:27:41 <jwh> shouldn't be difficult to knock up a script that can generate something to dd either
2019-01-26 21:28:23 <mps> succfemboi: do you know arhlinuxarm url for your model
2019-01-26 21:30:54 <jwh> I'll have to get a proper uart for this, should probably have got a board with video at least heh
2019-01-26 21:31:58 <succfemboi> It's a Samsung Chromebook XE303C12
2019-01-26 21:32:33 <succfemboi> Last I remembered the Debian installation instructions for it looked a little more complicated than usual
2019-01-26 21:34:14 <mps> succfemboi: one with 2GB of RAM?
2019-01-26 21:34:36 <succfemboi> yes
2019-01-26 21:35:29 <mps> I have second one, with 4GB, but they are mostly same and installation is not problem if you follow archlinuxarm guide
2019-01-26 21:36:21 <mps> my first installation was archlinuxarm on that chromebook, but soon converted to Alpine
2019-01-26 21:37:48 <succfemboi> But maybe a direct install path would be easier
2019-01-26 21:38:03 <succfemboi> *method
2019-01-26 21:38:30 <succfemboi> But how similar is alpine to arch actually?
2019-01-26 21:41:43 <mps> afaik, no one is made ready and signed Alpine kernel for chromebooks so strait install will not work. You need signed kernel if you didn't flashed u-boot on it which doesn't need signed kernels
2019-01-26 21:43:26 <succfemboi> I have developer mode enabled and could simply disable signing on boot
2019-01-26 21:45:31 <mps> are you sure? did you tried that?
2019-01-26 21:48:45 <succfemboi> No, but the archlinuxarm instructions tell: ` crossystem dev_boot_USB=1 dev_boot_signed_only=0 `
2019-01-26 21:50:43 <mps> right, but they have signed kernel in tar ball to download, and they say 'dd if=xxxx of=/dev/sdx'
2019-01-26 21:50:55 <mps> that in if= is signed kernel
2019-01-26 21:51:27 <mps> developer mode only allow you to boot from external media
2019-01-26 21:52:58 <mps> there were good guide on the net where one man from the samsung wrote good guide about that. but the site is down some time for maintainance
2019-01-26 21:53:26 <succfemboi> I thought on arch one could flash a custom firmware on chromebooks already
2019-01-26 21:53:32 <mps> here it is, when it comes online again: https://s-opensource.org/author/javiermartinezsamsungcom/
2019-01-26 21:54:10 <succfemboi> Since when was it down?
2019-01-26 21:54:23 <mps> I had a similar hope, but till I found nv-uboot (non verified) that didn't worked
2019-01-26 21:55:48 <mps> don't know, don't follow it, just found in my bookmarks for samsung chromebooks, because I have another one (Samsung chromebook plus) with aarch64 Alpine install
2019-01-26 21:57:30 <mps> eh, yes, forgot to add, Alpine is a lot faster on these chromebooks than the archlinux or debian, I tried it both
2019-01-26 22:05:20 <succfemboi> I don't mind if I can only boot from an SD-card for now
2019-01-26 22:07:48 <mps> then it is simplest and fastest method is to follow archlinuxarm guide, IMO
2019-01-26 22:07:50 <clandmeter> FreeBDSM, i pushed an updated certbot to edge with nginx plugin. can you check if it works?
2019-01-26 22:12:30 <succfemboi> But can't it be on alpine rather?
2019-01-26 22:13:34 <mps> replace archlinux root filesystem with another one, Alpine or Debian if you like
2019-01-26 22:14:59 <succfemboi> and that would work?
2019-01-26 22:15:48 <succfemboi> How would I be able to configure audio or trackpad on alpine?
2019-01-26 22:16:31 <mps> it worked for me ;) but you have to manually set boot services, console and other what sysadm have to do from the tar.gz root filesystem
2019-01-26 22:17:38 <mps> follow Alpine wiki how to install X. setup-xorg iirc
2019-01-26 22:25:23 <succfemboi> What's actually with package support on alpine arm? Is there Thunderbird available for that architecture?
2019-01-26 22:26:57 <_ikke_> Seems to not even be available for x86_64
2019-01-26 22:43:59 <clandmeter> I think we tried in the past but it was complicated...
2019-01-26 22:54:15 <_ikke_> clandmeter: Are you referring to the branch, where you *have* to select 1?
2019-01-26 22:55:12 <clandmeter> _ikke_, ?
2019-01-26 22:55:42 <clandmeter> I'm talking about Thunderbird
2019-01-26 23:16:38 <FreeBDSM> clandmeter: sure
2019-01-26 23:17:48 <FreeBDSM> oh, to the edge...
2019-01-26 23:22:04 <FreeBDSM> clandmeter: are you sure? https://pkgs.alpinelinux.org/packages?name=*certbot*&branch=edge&arch=x86_64
2019-01-26 23:23:30 <FreeBDSM> but I see it is available for x86 arch
2019-01-26 23:25:11 <jwh> hm, no dehydrated, perhaps I should add it
2019-01-26 23:35:13 <FreeBDSM> clandmeter: no, unfortunately I am unable to test that, as there's not x86_64 build of this package available and in docker hub image i386/alpine:edge is based on alpine 3.7, not 3.9
2019-01-26 23:41:01 <FreeBDSM> filed https://bugs.alpinelinux.org/issues/9925
2019-01-26 23:45:34 <FreeBDSM> some build logs are missing: https://build.alpinelinux.org/buildlogs/build-edge-x86/community/certbot-postfix/certbot-postfix-0.30.1-r0.log https://build.alpinelinux.org/buildlogs/build-edge-x86/community/certbot-apache/certbot-apache-0.30.1-r0.log https://build.alpinelinux.org/buildlogs/build-edge-x86/community/certbot-nginx/certbot-nginx-0.30.1-r0.log why's that?
2019-01-27 11:19:51 <jayaura> Hi,  I would like to try out Alpine linux on my laptop. What is the recommended download media? I would prefer if I have all the packages necessary with me before hand and not download on the fly during installation
2019-01-27 11:30:44 <mps> jayaura: extended iso, but you will have to add packages from the net, because even extended are doesn't have much packages
2019-01-27 11:32:59 <jayaura> mps: I see
2019-01-27 11:33:03 <jayaura> thanks for the info
2019-01-27 11:39:46 <jayaura> mps: does the installation media contain an offline copy of the installation handbook ?
2019-01-27 11:41:33 <mps> no, look at wiki.alpinelinux.org, and keep in mind that not everything docemnted
2019-01-27 11:42:56 <mps> Alpine strive to be minimal on install media and on installation. man pages are in separate packages, for example
2019-01-27 11:44:16 <jayaura> mps: Hmm thats going to be a problem. I need to connect to wifi for internet access. can I just install a minimal version from the extended media and later deal with further configuration? Or is it that internet access is mandatory for installation process ?
2019-01-27 11:45:37 <mps> Internet access is not mandatory for basic install, but you will need it for anything else
2019-01-27 11:47:06 <mps> I didn't tried last two release to install from from ISO but I think they have drivers and needed packages for wifi, but I'm not sure
2019-01-27 12:00:29 <jayaura> mps: thanks I'll try it out. mine is all intel so i wouldnt need any extra drivers I hope
2019-01-27 12:03:28 <mps> intel boards is usually easiest to install, but if you have exotic hardware could be a problem. but you will didn't know until you try
2019-01-27 12:06:33 <jwh> hm, if I install linux-vanilla, it's not going to boot on this exynos is it?
2019-01-27 13:24:30 <jayaura> mps: around? having issues with booting uefi. archlinux wiki explicitly asks to run efibootmgr to add a boot entry, but alpine wiki does not have this step. I just restarted to come here. any experience with efi ?
2019-01-27 13:26:21 <mps> afaik, efi is problematic on stable and older versions of Alpine, but it is added to 3.9_rc5 I think, and some people reported that it works. I didn't tried it though
2019-01-27 13:27:12 <mps> someone else here could be of the help, or try to ask on #alpine-devel channel
2019-01-27 13:28:23 <jayaura> okay let try some renaming hacks. I'll ask there if it doesnt work. Or maybe switch to grub
2019-01-27 13:28:34 <jayaura> thank you
2019-01-27 13:30:09 <mps> in 3.9_rc5 grub is already set. v3.9 rc5 is next stable and we hope that it will be officially stable in next few days
2019-01-27 13:31:12 <mps> there are ISO's with v3.9_rc5 to download
2019-01-27 13:32:03 <mps> look here: http://dl-cdn.alpinelinux.org/alpine/v3.9/releases/x86_64/
2019-01-27 14:18:43 <terra1> I'm trying to boot banana-pi (allwinner A20) for first time but I stuck at "Starting kernel .."
2019-01-27 14:18:48 <terra1> https://paste.ubuntu.com/p/ZNrTNsyBKT/
2019-01-27 16:16:00 <mps> terra1: you need to make kernel, modules and dtb's probably. I did it for my allwinner boards, A20
2019-01-27 16:16:42 <mps> or extract them from some other distros, Armbian ones are good
2019-01-27 16:31:39 <FreeBDSM> clandmeter: hey, have you seen my messages?
2019-01-27 16:56:23 <terra1> mps: I actually tried armbian kernel too. Same issue.
2019-01-27 16:57:43 <terra1> it seems bootlin toolchain is not compatible with alpine..
2019-01-27 16:57:53 <terra1> .../bin/arm-buildroot-linux-musleabihf-as: __rawmemchr: symbol not found
2019-01-27 17:07:53 <mps> you don't need toolchain to install alpine on A20 (and similar boards) just put root FS and kernel, modules, dtbs, and u-boot of course
2019-01-27 17:09:39 <mps> I tested Armbian kernels and u-boot on A20 with Alpine, but I prefer to build my kernels for these boards and I'm making it on Debian which is on lxc container
2019-01-27 17:23:42 <terra1> mps: I'm trying to compile u-boot first.
2019-01-27 17:24:59 <terra1> I have already working armbian setup by I changed root installation with ALpine except /lib/modules and /boot but didn't work
2019-01-27 17:25:13 <terra1> then I'm starting from scratch
2019-01-27 17:25:29 <mps> it is simple, maybe I have some short guide how to install Alpine on Banana pi, will look later
2019-01-27 17:26:09 <terra1> so wheres u-boot-with-spl for Banana-pi?
2019-01-27 17:26:52 <mps> build it yourself from the source
2019-01-27 17:27:30 <terra1> yes, that is the one I'm tryin to do with bootlin toolchain.
2019-01-27 17:27:41 <terra1> And it seems it is not that simple at all.
2019-01-27 17:28:22 <aggaz> terra1: if I remember correctly I had a similar issue with an Orange Pi One. In my case the problem was that the kernel was not printing on /dev/tty. I solved by adding something on the kernel line of the bootloader, can not find the exact line but was someting to say were to show the console
2019-01-27 17:29:48 <mps> here is the wiki guide (not my): https://wiki.alpinelinux.org/wiki/DIY_Fully_working_Alpine_Linux_for_Allwinner_and_Other_ARM_SOCs
2019-01-27 17:29:50 <terra1> aggaz: thanks.. I also suspected that. Maybe I have to change tty pins on board.
2019-01-27 17:30:47 <terra1> mps: Sorry but that was the one of the most unuseful howtos that I encoured in long time.
2019-01-27 17:31:27 <terra1> I'm not sure that writer what was thinking..
2019-01-27 17:31:52 <mps> hmm, it is far from perfect but not un-useful, IMO
2019-01-27 17:32:22 <aggaz> If you can read uboot stuff it means that your tty pins are correct. I thin the line should be something like "console=tty1" where you specify kernel modules
2019-01-27 17:32:53 <terra1> maybe. Doesn't mention a single bit about Alpine downloads for any ARM boards and how use them on sunxi.
2019-01-27 17:33:03 <aggaz> on my RPi this option is set on cmdline.txt bit I think that for generic arm the file name is different
2019-01-27 17:34:21 <mps> boot parameters could be set in /boot/extlinux/extlinux.conf
2019-01-27 17:34:30 <aggaz> thats it
2019-01-27 17:34:59 <mps> here is one of my boards: APPEND root=/dev/mmcblk0p2 rw rootwait console=${console}
2019-01-27 17:35:39 <aggaz> where is the variable ${console} defined?
2019-01-27 17:35:47 <aggaz> in the same file (just curious)
2019-01-27 17:35:56 <terra1> aggaz:  same question
2019-01-27 17:36:48 <aggaz> anyway, if you, terra1, have the same line and can not see the kernel booting, I would try to substitude ${console} with tty1 or tty0 and so on
2019-01-27 17:37:26 <terra1> you mean ttyS0 ttyS1 probably..
2019-01-27 17:37:43 <terra1> that are the physical ones.
2019-01-27 17:38:11 <aggaz> well I don't know :) I read "console=tty1" on my RPi and I was suggesting that :)
2019-01-27 17:40:32 <mps> RPi have different bootloader, not u-boot, and I didn't worked with RPis
2019-01-27 17:41:08 <aggaz> yes but I think that we are talking about kernel parameters here, and thay should be the same
2019-01-27 17:41:55 <mps> Alpine tar.gz for armhf have packaged u-boot for u-boot-sunxi-with-spl.bin under /u-boot/Cubieboard2
2019-01-27 17:42:49 <terra1> mps: I saw it but  /u-boot/Cubieboard has same file
2019-01-27 17:42:51 <mps> they are not, IIRC RPis have /dev/ttyAMA0 as serial console while A20 have /dev/ttyS0
2019-01-27 17:43:19 <mps> this first is for A10, I think
2019-01-27 17:43:56 <aggaz> I was not aware of that
2019-01-27 17:44:50 <terra1> mps: yes it is a10
2019-01-27 17:44:51 <mps> I didn't tried these u-boot's but just know they are there, and dtb's are in tar.gz, look for them
2019-01-27 17:45:40 <mps> theoretically it should work, but (to repeat myself) I didn't tried, so cannot be sure
2019-01-27 18:07:49 <terra1> I tried all possible tty options but same...
2019-01-27 18:24:05 <mps> have you got anything on serial console
2019-01-27 18:47:19 <terra1> mps: https://paste.ubuntu.com/p/gbMNG52RyX/
2019-01-27 18:52:01 <clandmeter> FreeBDSM, yes i see you msgs. edge and v3.9 have new certbot. Sadly i cannot help you with docker images.
2019-01-27 18:53:04 <mps> terra1: hmm, probably something missing in initramfs
2019-01-27 18:54:17 <AinNero> terra1: what am i looking at?
2019-01-27 18:55:37 <AinNero> im surprised there is a sane Allwinner support, they are in my shitlist for lack of mainline support
2019-01-27 18:56:31 <mps> AinNero: Allwinners works fine, I have some and with mainline kernels
2019-01-27 18:56:57 <AinNero> the boot log you showed looks like a working manual boot
2019-01-27 18:57:10 <aggaz> terra1: I see that your append line uses as separator both "," and " ". Is that correct?
2019-01-27 18:57:41 <AinNero> aggaz: the , separator is part of the modules and console options
2019-01-27 18:58:30 <mps> modules= should have "" around, IIRC
2019-01-27 18:58:38 <aggaz> AinNero: oh yeah I get it now
2019-01-27 18:59:06 <AinNero> mps: does not need to
2019-01-27 18:59:19 <AinNero> personally i discourage using quotes in the kernel command line, its not shell script
2019-01-27 18:59:53 <mps> aha, ok
2019-01-27 19:01:12 <mps> terra1: I have one Banana Pi somewhere here, will look for free sd card and try to boot it using Alpine kernel to see does it works
2019-01-27 19:01:33 <aggaz> terra1: reading about this stuff,new memories are emerging in my brain. I think that adding the console parameter as I suggested worked with a previous alpine version but not with the new/current one
2019-01-27 19:02:02 <terra1> mps: thanks for support :)
2019-01-27 19:02:39 <terra1> aggaz: i think it won't matter because i also tried without any console parameter
2019-01-27 19:04:20 <mps> what is the kernel version, I noticed 4.19.xx have some problems with console
2019-01-27 19:04:21 <terra1> i was using armbian's u-boot+spl but now it is time to dd sdcard with Alpine's u-boot
2019-01-27 19:04:37 <terra1> it is 3.9-rc5
2019-01-27 19:05:08 <mps> 3.9, isn't it to old
2019-01-27 19:05:11 <AinNero> terra1: how did you get the log of the device? which baudrate?
2019-01-27 19:05:20 <terra1> Arbbian was also using 4.19 which was fine wirh ubuntu
2019-01-27 19:05:35 <terra1> AinNero: 115200
2019-01-27 19:06:38 <AinNero> actually, i just noticed, since when do arm machines have ttyS[0-3] ?
2019-01-27 19:07:16 <AinNero> tty0 is vga i guess, so i dont see a console= option that could print on serial
2019-01-27 19:08:35 <AinNero> afaik the ttyS* are a aspect of x86 hardware
2019-01-27 19:08:43 <mps> terra1: you didn't attached serial console? you are using hdmi?
2019-01-27 19:09:06 <aggaz> I found a note of my installation of Alpine on Orange Pi, I added console=ttyS0,115200n8
2019-01-27 19:09:20 <mps> AinNero: Allwinners have ttySX
2019-01-27 19:09:46 <aggaz> I think I was using alpine 3.8, and then with 3.7
2019-01-27 19:09:52 <aggaz> the 3.8 did not work
2019-01-27 19:10:05 <AinNero> mps: interesting
2019-01-27 19:10:16 <aggaz> sorry I meant "i was using 3.7 and 3.8 did not work anymore"
2019-01-27 19:11:07 <terra1> mps: I attached to onboard uart
2019-01-27 19:11:10 <mps> 'cat /proc/consoles' -> 'ttyS0                -W- (EC p a)    4:64'
2019-01-27 19:12:40 <mps> terra1: ok, that 'paste' is on your terminal where you attached serial cable
2019-01-27 19:16:58 <terra1> mps: otherwise it would be quite interesting to see u-boot on hdmi
2019-01-27 19:17:20 <clandmeter> terra1, did you try /dev/ttyAMA0?
2019-01-27 19:17:48 <terra1> clandmeter: no.. I'm reinstalling now
2019-01-27 19:19:27 <mps> clandmeter: Allwinners doesn't use ttyAMAX but ttySX
2019-01-27 19:23:33 <aggaz> A question: Are the names for serial ports showed in /dev directory of a working linux the same names that the kernel expect in the append line?
2019-01-27 19:24:05 <aggaz> Looking here: https://www.kernel.org/doc/html/v4.14/admin-guide/serial-console.html I don't see mentions of names like "ttyAMA0" or others of the like
2019-01-27 19:25:16 <mps> ttyAMA0 is for some ARM boxes
2019-01-27 19:28:13 <terra1> spl: mmc init failed with error: -123
2019-01-27 19:28:47 <terra1> Cubieboard2 u-boot file even doesn't see mmc
2019-01-27 19:30:47 <terra1> clandmeter: which toolchain you suggest to compile u-boot on Alpine?
2019-01-27 19:34:37 <mps> terra1: which armhf tar file you are trying to install? alpine-uboot-3.8.2-armhf.tar.gz?
2019-01-27 19:35:30 <terra1> mps: alpine-uboot-3.9.0_rc5-armhf.tar.gz
2019-01-27 19:42:30 <mps> terra1: ok, this u-boot from 'u-boot/Cubieboard2/u-boot-sunxi-with-spl.bin' doesn't work BananaPi, just tried it
2019-01-27 19:43:04 <mps> but one you tried from Armbian works
2019-01-27 19:43:58 <mps> If you are willing to try I could post to tpaste my notes, which is not perfect, but could help
2019-01-27 19:44:20 <mps> or to mail if you preffer
2019-01-27 19:44:51 <mps> just pm me, if you like to post your mail address
2019-01-27 19:45:11 <mps> but faster and easier will be tpaste
2019-01-27 19:58:09 <terra1> mps: https://github.com/alpinelinux/aports/blob/master/testing/mame/APKBUILD
2019-01-27 19:58:40 <terra1> maintainer is me, so you know my email. Thanks.
2019-01-27 20:03:17 <terra1> mps: oh, of course tpaste or something will be useful
2019-01-27 20:03:30 <terra1> for me and anyone who interested
2019-01-27 20:05:19 <mps> ah, you are terra, :) I see 1 after nick which looked strange to me why are we have another terra :)
2019-01-27 20:06:04 <mps> http://tpaste.us/jXOn
2019-01-27 20:06:33 <mps> that is just my notes and do not follow it blindly
2019-01-27 20:08:32 <mps> sorry for now, I have meeting with local HAM (radio amateur) group
2019-01-27 20:08:38 <terra1> mps: I see, thanks.
2019-01-27 20:09:08 <terra1> did you extract u-boot+spl from armbian image?
2019-01-27 20:09:22 <terra1> cross-compile option seems problematic on alpine
2019-01-27 20:10:45 <mps> have to go to meeting, will be on chat but not on my working machine, so cul
2019-01-27 20:11:07 <terra1> mps: ok thank you anyway.
2019-01-27 20:38:44 <FreeBDSM> clandmeter: that's not a problem, I'll build one up myself, I just needed someone to build x86_64 version of certbot-nginx and now I see that fcolista did, will test it now
2019-01-27 20:49:57 <FreeBDSM> clandmeter, fcolista, thank you both! just tested certbot-nginx and it worked juuuust fine, the error is finally gone! thanks a lot!
2019-01-27 22:06:29 <mps> terra1: looked at boot A20, I think the problem is in initramfs, I remember that I had to tweak it when prepared armhf to boot in qemu
2019-01-27 22:09:59 <terra1> mps: you used armbian initramfs then?
2019-01-27 22:10:50 <mps> no, i took kernel, dtbs, and copied /lib/modules/$kver to Alpine rootfs
2019-01-27 22:11:40 <mps> and u-boot from their filesystem
2019-01-27 22:12:19 <terra1> how do you first boot without initramfs
2019-01-27 22:13:11 <mps> that was more then a year ago, after that I build kernel, modules and u-boot from sources in lxc container with minimal debian installation with their cross-tools
2019-01-27 22:14:47 <mps> I took armhf Alpine miniroot.tar.gz and extracted it on rootfs (sd card) and copied rest
2019-01-27 22:16:07 <mps> if you want I can make kernel, modules and u-boot in tar and put it to you to download
2019-01-27 22:16:32 <mps> I have all that, just have to make tarball
2019-01-27 22:17:03 <terra1> I tried so many cross-compilers on https://musl.cc/ to compile u-boot.. no one able to finish
2019-01-27 22:18:53 <mps> cross-build tools is my next todo for Alpine :)
2019-01-27 22:21:15 <terra1> I was using linaro on before switchin to ALpine 3 years ago.
2019-01-27 22:21:25 <terra1> *on Arch
2019-01-27 22:22:46 <mps> any cross tools should work, easiest way is to run it under some of container, docker or lxc
2019-01-27 22:23:51 <mps> I worked more than 20 years with so it was naturally easy for me to run it under container
2019-01-27 22:24:14 <mps> s/with so/with debian so/
2019-01-27 22:28:38 <terra1> yes, container is anot her option but collecting build sdk on debian or redhat like distros is a bit mess.
2019-01-27 22:29:06 <terra1> I like ALpine's build environment and I use a chroot for that purpose
2019-01-27 22:31:12 <mps> I used chroots but novadays switched to lxc
2019-01-27 22:31:17 <terra1> I'm making, updating packages in that chroot
2019-01-27 22:31:34 <mps> tried arm in qemu but it is too slow
2019-01-27 22:31:44 <terra1> :)
2019-01-27 22:33:46 <mps> I have physical arm32 with 4GB RAM and 8 CPUs for building armhf/armv7 programs
2019-01-27 22:34:36 <mps> not fast, but quite fine because I can tolerate it's slowness especially slow sd cards
2019-01-27 22:43:03 <terra1> mps: rockchip or slogic ?
2019-01-27 22:43:12 <terra1> *amlogic
2019-01-27 22:44:16 <terra1> btw, I just noticed that kernel doesn't log on serial port it switches to hdmi after loadin kernel
2019-01-27 22:44:29 <terra1> so my problem solved somehow
2019-01-27 22:44:59 <terra1> I only used /usr/lib/linux-u-boot-next-bananapi_5.69_armhf/u-boot-sunxi-with-spl.bin from Armbian image
2019-01-27 22:45:11 <terra1> everything else is ALpine
2019-01-27 22:47:46 <terra1> now I'm watching Alpine login prompt on my Banana-pi
2019-01-27 22:50:49 <aggaz> terra1: finally! :)
2019-01-27 22:51:12 <aggaz> so basically you can not see the kernel booting but at least you can se the login?
2019-01-27 22:52:20 <terra1> initialization appears on hdmi port not serial
2019-01-27 22:52:38 <terra1> after u-boot it switches to hdmi
2019-01-27 22:53:02 <mps> terra1: very nice to hear
2019-01-27 22:53:29 <mps> so probably missing serial driver in initramfs?
2019-01-27 22:54:18 <terra1> it tooks much longer than usual things I do but it helped a lot to refresh my embedded system memories
2019-01-27 22:54:30 <terra1> mps: exactly
2019-01-27 22:55:20 <terra1> exactly --> most probably
2019-01-27 22:55:26 <mps> I remember that I had to tweak initramfs, but forgot what exactly, and didn't wrote notes about
2019-01-27 22:55:41 <aggaz> are you using alpine's kernel?
2019-01-27 22:55:51 <terra1> yes ALpine kernel
2019-01-27 22:56:29 <mps> so, kernel works, only missing part is u-boot for Bananapi
2019-01-27 22:56:42 <terra1> yes
2019-01-27 22:56:57 <aggaz> does it mean that these drivers are for some reason missing in the current alpine distribution?
2019-01-27 22:57:24 <terra1> orange-pi could be next but I like banana-pi more because of sata port
2019-01-27 22:59:43 <mps> aggaz: looking in config-vanilla for armhf, console drivers are there, in kernel and not as modules
2019-01-27 23:00:06 <mps> but not sure, till I find time to 'debug' it
2019-01-27 23:00:40 <aggaz> I am curious because I had a similar problem with orange pi one
2019-01-27 23:01:32 <aggaz> because of this I then switched to debian, but still I would like to resolve it :)
2019-01-27 23:02:53 <aggaz> if the drivers are there, could it be something about the dts files that are not pointing to the seial console in the proper way?
2019-01-27 23:03:45 <mps> aggaz: shouldn't be, because Alpine didn't modified anything in dts
2019-01-27 23:03:48 <terra1> I was mistaken about u-boot and hdmi.. u-boot debug messages also appears on hdmi
2019-01-27 23:04:31 <terra1> I remember that this feature only exist on serial
2019-01-27 23:05:26 <mps> last time I tried  to install from the scratch is more then a year, so forgot how I did it exactly. novadays i just copy working/archived installations
2019-01-27 23:05:27 <terra1> probably this feature added later or i remember wrong.
2019-01-27 23:11:48 <terra1> usb keyboard is not working
2019-01-27 23:12:12 <terra1> maybe a insufficient power problempower
2019-01-27 23:12:25 <terra1> I have to deal wirh this later..
2019-01-27 23:12:56 <terra1> it is late and I have to go to sleep
2019-01-27 23:13:04 <terra1> byes all.
2019-01-27 23:13:16 <mps> good night
2019-01-27 23:14:05 <aggaz> Well, I'll go too, good night
2019-01-28 10:54:00 <Andreetje> Hi, i've got some problems using a proxy..... Normal install with setup-alpine, enter proxy... then it fails... using direct connection, afterwards changing using proxy, no problems at all
2019-01-28 10:55:55 <Andreetje> ssl_client....wrong_version number error
2019-01-28 11:29:57 <_ikke_> Andreetje: Might be because it uses busybox wget or something similar at that point, not sure htough
2019-01-28 11:30:17 <Andreetje> looks like it.
2019-01-28 11:31:56 <Andreetje> is it an bug or an feature :)
2019-01-28 11:32:23 <_ikke_> Hard to tell honestly, but feel free to create an issue at bugs.alpinelinux.org with the details
2019-01-28 11:35:46 <clandmeter> Andreetje, which alpine version?
2019-01-28 11:36:25 <Andreetje> 3.8.2 Virt
2019-01-28 11:36:43 <clandmeter> you boot from iso and run installer?
2019-01-28 11:36:49 <Andreetje> i even tried the 3.9 Rc and the "standard"
2019-01-28 11:36:58 <Andreetje> yes just boot and setup-alpine
2019-01-28 11:37:05 <clandmeter> 3.9 didnt make a diff?
2019-01-28 11:37:12 <Andreetje> nope.
2019-01-28 11:37:45 <clandmeter> ok please add an issue to bugs.a.o as _ikke_ suggested.
2019-01-28 11:38:17 <Andreetje> could it be that during setup-alpine the repo's are being listed... that fails (wget ? )
2019-01-28 11:39:21 <clandmeter> im not 100% sure i understand your flow.
2019-01-28 11:39:46 <clandmeter> you first try it with proxy but it fails. afterwards you try without proxy and it works.
2019-01-28 11:39:59 <clandmeter> after install finish and switch back to proxy it works?
2019-01-28 11:40:02 <Andreetje> yes
2019-01-28 11:40:10 <Andreetje> i can reproduce it ;)
2019-01-28 11:40:27 <Andreetje> setup-apkrepos is being called from setup-alpine...
2019-01-28 11:40:32 <clandmeter> so the only time it fails is when you are using proxy inside setup-alpine
2019-01-28 11:40:42 <Andreetje> when using proxy and setup-apkrepos it fails....
2019-01-28 11:40:43 <Andreetje>  :)
2019-01-28 11:42:13 <clandmeter> i dont have a proxy here to quickly test
2019-01-28 11:42:35 <Andreetje> wget being used in the setup-apkrepos is the "problem" :)
2019-01-28 11:43:05 <clandmeter> why would that be the problem?
2019-01-28 11:43:26 <_ikke_> clandmeter: might be an incompatibility with their proxy
2019-01-28 11:43:37 <Andreetje> that looks like it
2019-01-28 11:43:44 <Andreetje> squid 3 proxy
2019-01-28 11:43:47 <_ikke_> Andreetje: can you try to directly wget to the proxy?
2019-01-28 11:43:54 <_ikke_> with busybox wget
2019-01-28 11:44:10 <clandmeter> Andreetje, what happens if you add wget before you run the installer?
2019-01-28 11:44:40 <Andreetje> same error : 501 error
2019-01-28 11:44:56 <_ikke_> 501 seems to be an http error, not ssl
2019-01-28 11:45:09 <Andreetje> http & https use the same proxy
2019-01-28 11:45:33 <_ikke_> 501 is not implemented
2019-01-28 11:45:44 <_ikke_> is it a socks proxy?
2019-01-28 11:45:54 <Andreetje>  squid3 proxy
2019-01-28 11:46:03 <clandmeter> are you sure its wget that is the issue?
2019-01-28 11:46:17 <clandmeter> not apk?
2019-01-28 11:46:31 <_ikke_> might be, that's what I'm trying to find out
2019-01-28 11:46:33 <Andreetje> apk works with proxy
2019-01-28 11:46:48 <Andreetje> wget connects to proxy and then gives a 501 error
2019-01-28 11:47:02 <clandmeter> what is wget doing?
2019-01-28 11:47:22 <clandmeter> what are the args
2019-01-28 11:47:43 <Andreetje> i don't know the exact args of setup-apkrepos
2019-01-28 11:48:50 <Andreetje> wget --spider -q -T 5  ${url%/}/edge/main/$arch/APKINDEX.tar.gz)
2019-01-28 11:49:43 <clandmeter> ok
2019-01-28 11:51:14 <clandmeter> and after install you can run setup-apkrepos fine?
2019-01-28 11:51:22 <Andreetje> no
2019-01-28 11:51:42 <clandmeter> ok so its just setup-apkrepos
2019-01-28 11:51:50 <Andreetje> looks like it.
2019-01-28 11:52:20 <Andreetje> and that is called after the setup-proxy in the setup-alpine script.
2019-01-28 11:53:36 <clandmeter> it will loop over all mirrros
2019-01-28 11:53:50 <clandmeter> does it happen on a specific one?
2019-01-28 11:54:41 <clandmeter> i guess --spider only fetches the header
2019-01-28 11:56:28 <Andreetje> i'm not that deep into it ;)
2019-01-28 11:57:17 <Andreetje> there is an MIRRORS_URL=https://mirrors.alpinelinux.org/mirrors.txt
2019-01-28 11:57:26 <Andreetje> in the source of setup-apkrepos
2019-01-28 11:57:38 <Andreetje> with wget i can't even get that file
2019-01-28 11:58:09 <Andreetje> so either an problem with my squid proxy in combination with wget or an wget proxy problem :)
2019-01-28 12:02:55 <clandmeter> you said it works after install?  but now you say it wont even download  that file?
2019-01-28 12:03:15 <Andreetje> apk works after install with proxy
2019-01-28 12:03:36 <Andreetje> the setup-apkrepos does not work at all with proxy
2019-01-28 12:03:53 <Andreetje> apk works with and without proxy
2019-01-28 12:05:43 <clandmeter> can you download the mirrors file afterwards?
2019-01-28 12:07:35 <Andreetje> wget https://mirrors.alpinelinux.org/mirrors.txt
2019-01-28 12:07:36 <Andreetje> nope
2019-01-28 12:08:39 <clandmeter> does it work when you apk add wget?
2019-01-28 12:09:28 <Andreetje> Yeah, good tip
2019-01-28 12:10:19 <Andreetje> the default included wget does not work, apk add wget works
2019-01-28 12:11:14 <clandmeter> so if you apk add wget before you run the install it should work?
2019-01-28 12:11:48 <Andreetje> no that easy because the repositories file points to the iso by default.
2019-01-28 12:24:23 <pax_rhos> hello. Is there a reason why 'ca-certificates' package doesn't contain 'update-ca-trust' binary file?
2019-01-28 12:24:46 <pax_rhos> ncopa: maybe you, as a maintainer of that package, could tell?
2019-01-28 12:25:52 <pax_rhos> ah, nvm, that's probably a centos-only thing
2019-01-28 12:27:08 <clandmeter> pax_rhos, would that be similar like update-ca-certificates?
2019-01-28 12:28:54 <pax_rhos> clandmeter: yeah, probably, trying now
2019-01-28 14:35:41 <ncopa> update-ca-trust sounds like an equivalent to update-ca-certificates
2019-01-28 15:18:08 <Andreetje> ncopa : still there ? :)
2019-01-28 15:18:32 <ncopa> Andreetje: im here
2019-01-28 15:18:47 <Andreetje> seen my "problem" ?
2019-01-28 15:19:41 <Andreetje> the wget used in/by busybox does not work correct with squid proxy so... setup-alpine with proxy fails...
2019-01-28 15:38:51 <Ai9zO5AP> hi
2019-01-28 15:39:07 <Ai9zO5AP> how could I install openjdk on alpine ? thank you
2019-01-28 15:39:12 <ncopa> Andreetje: but it works with GNU wget?
2019-01-28 15:39:19 <ncopa> Ai9zO5AP: apk add openjdk8
2019-01-28 15:40:04 <Andreetje> @ncopa : Yep
2019-01-28 15:40:42 <ncopa> does gnu wget use the squid proxy for https?
2019-01-28 15:40:51 <ncopa> or does it simply ignore the proxy setting?
2019-01-28 15:41:05 <Andreetje> works with proxy
2019-01-28 15:41:25 <Andreetje> sorry must go, can explain more tomorrow
2019-01-28 15:41:29 <Andreetje> Thnx !
2019-01-28 15:41:40 <ncopa> can you file an issue on bugs.a.o?
2019-01-28 15:41:45 <Andreetje> Sure
2019-01-28 15:41:50 <ncopa> with the details, error messages and strace output
2019-01-28 15:41:53 <ncopa> or similar
2019-01-28 15:41:55 <ncopa> thanks!
2019-01-28 15:41:58 <Andreetje> i'll try ;)
2019-01-28 15:42:10 <Andreetje> bye
2019-01-28 15:43:17 <Ai9zO5AP> ncopa: are you sure that the pkg exist ?
2019-01-28 15:44:51 <ncopa> Ai9zO5AP: yes, but i guess it depends on which architecture you are using
2019-01-28 15:45:07 <ncopa> ncopa-macbook:~$ apk search openjdk8 | tpaste
2019-01-28 15:45:07 <ncopa> http://tpaste.us/gMOy
2019-01-28 15:47:03 <Ai9zO5AP> ncopa: I just install standard and when I type apk add openjd8 i get required by: world[openjdk8]
2019-01-28 15:47:13 <Ai9zO5AP> should I add some repos
2019-01-28 15:51:09 <Ai9zO5AP> 29;2~I should add edge
2019-01-28 15:51:33 <ncopa> i think you need add community repo
2019-01-28 15:56:20 <Ai9zO5AP> ncopa: I see thank you very much
2019-01-28 15:57:36 <pax_rhos> ncopa: clandmeter: `update-ca-certificates` seems to be broken :( https://i.imgur.com/K2WPLma.png - this is how this command works on a debian-based docker container. And this is what I get on alpine: `WARNING: ca-certificates.crt does not contain exactly one certificate or CRL: skipping`. The cert is the same. Why does alpine seem to look for ca-certificates.crt? that's not my cert's name!
2019-01-28 15:59:06 <pax_rhos> I copied the file with that name, now there are 2 warnings: the old one + `WARNING: Skipping duplicate certificate in file ca-cert-ca-certificates.pem`
2019-01-28 16:28:24 <pax_rhos> https://github.com/gliderlabs/docker-alpine/issues/30 just feel the pain of the users about update-ca-certificates: they seem to experience the very same problem as I do. Judging by the comments in that ticket - it looks like the issue was patched around Alpine 3.2 but was re-introduced later. Could someone, please, take a look at it?
2019-01-28 16:30:58 <Ai9zO5AP> if I have something like "can't execute '/home/green/Freenet/bin/wrapper-linux-x86-64' " and the url exist , what should I do please
2019-01-28 16:31:20 <AinNero> Ai9zO5AP: look what interpreter/dynamic linker exists
2019-01-28 16:31:25 <AinNero> /what/if/
2019-01-28 16:32:06 <AinNero> if its a binary, run `file` on it, and look if the interpreter listed there exists
2019-01-28 16:33:52 <AinNero> wild guess: its the interpreter path for glibc, and you are missing libc6-compat, but thats just a wild guess
2019-01-28 16:37:27 <Ai9zO5AP> AinNero: I will try and see , thanks a loty
2019-01-28 16:45:44 <Ai9zO5AP> AinNero: thank you, I make it work
2019-01-28 19:58:23 <maldridge> I am using v4 dhcp, and I'd like to have static nameservers in addition to those supplied by dhcp.  is this easily achievable?
2019-01-28 19:59:27 <ncopa> maldridge: yes. you can either configure all servers to use dhcp, and let the dhcp server provide the static ip for a given MAC address
2019-01-28 19:59:53 <ncopa> or you can have a dedicated range for static addresses, that you exclude from the dhcp pool
2019-01-28 20:00:06 <maldridge> please re-read, I want static DNS servers, not static DHCP
2019-01-28 20:00:52 <ncopa> oh
2019-01-28 20:00:56 <maldridge> if I can get openresolv then that will do what I want I think
2019-01-28 20:01:23 <ncopa> i think we have an openresolv package
2019-01-28 20:01:35 <ncopa> and i think i have the same issue, for openvpn use
2019-01-28 20:01:49 <maldridge> there is one, not sure if it will automagically plumb in to the config from ifupdown
2019-01-28 20:02:07 <ncopa> it may be somewhat tricky though
2019-01-28 20:02:15 <maldridge> welp, lunch locally; I'll look into this after lunch
2019-01-28 20:02:56 <ncopa> glibc and bsd's stub resolvers will normally try first dns server in resolv.conf and then go to next
2019-01-28 20:03:20 <ncopa> i musl all servers listed in resolv.conf will be asked and are expected to return same result
2019-01-28 20:38:03 <dalias> any idea why my py3-sip package would lack the /usr/bin/sip program?
2019-01-28 20:38:19 <dalias> pkgs.alpinelinux.org says it's there but my installed package doesn't have it
2019-01-28 20:43:56 <dalias> somehow there's a broken py3-sip-4.19.13-r0 in addition to the one i should be getting py3-sip-4.19.5-r0
2019-01-28 20:45:43 <ncopa> seems like we have 2 different py3-sip
2019-01-28 20:46:11 <ncopa> one provided by community/py-sip and one provided by community/py3-sip
2019-01-28 20:48:03 <ncopa> dalias: the /usr/bin/sip program comes with py-sip, which has 2 subpackages, py2-sip and py3-sip
2019-01-28 20:48:37 <mepholic> :thinking:
2019-01-28 20:51:18 <smemsh> i found more kernel modules beyond hid-apple needed for macbook pro to use all the hardware (including keyboard needs spi drivers), not in vanilla kernel.  what is the policy for this, are they just added on request, or perhaps laptops are not the target of alpine so leave custom kernels? or, i can submit the list of ones i found in bug
2019-01-28 20:51:59 <clandmeter> smemsh, that sounds like a bug, we should fix that.
2019-01-28 20:52:18 <clandmeter> i think ncopa is using a macbook
2019-01-28 20:53:47 <ncopa> smemsh: yes, i think we want fix it if you know what modules needed
2019-01-28 20:54:38 <ncopa> i think the newer macbook pros may not work with linux
2019-01-28 20:55:03 <clandmeter> can a macbook pro boot from usb-c?
2019-01-28 20:55:32 <ncopa> iirc the keyboard is not on usb bus, and i dont know if there are any keyboard drivers at all for linux
2019-01-28 20:55:38 <smemsh> they do work, with caveats.  4.20 is required for the framebuffer not to scramble, i915.modeset=0 is required with 4.19
2019-01-28 20:55:41 <dalias> ncopa, ah
2019-01-28 20:56:02 <ncopa> dalias: its still a bug, we should not have both py-sip and py3-sip
2019-01-28 20:56:13 <clandmeter> smemsh, which pro version do you have?
2019-01-28 20:56:27 <ncopa> i have an older, late 2015
2019-01-28 20:56:41 <clandmeter> i think mine is '17
2019-01-28 20:56:48 <ncopa> what i read, the newer ones with some crypto hardware, does not work at all
2019-01-28 20:57:05 <clandmeter> wasnt there some workaround now?
2019-01-28 20:57:09 <ncopa> you can disable the secure boot thing, but then you dont get access to the nvme at all
2019-01-28 20:57:33 <ncopa> that is what I read at least
2019-01-28 20:57:44 <clandmeter> ive never tried it
2019-01-28 20:57:55 <clandmeter> im still trying to get used to a macbook
2019-01-28 20:58:10 <clandmeter> i really dislike the keyboard.
2019-01-28 20:58:12 <smemsh> the module delta i found so far is mostly for power management stuff intel_rapl, x86_pkg_temp_thermal, intel_powerclamp, intel_uncore.  the keyboard needs spi_pxa2xx_pci and _platform, and maybe apple_bl is needed.  i have not compared against compiled in list yet.  also efi_pstore and efivars would be useful
2019-01-28 20:59:04 <smemsh> the newest 2018 ones yes have some issue with the initial bootloader not working because it's not signed or something.  2017 however does work
2019-01-28 20:59:24 <ncopa> clandmeter: i love the keyboard
2019-01-28 20:59:35 <clandmeter> yes i know
2019-01-28 20:59:41 <clandmeter> you have a nice one
2019-01-28 20:59:46 <ncopa> clandmeter: oh, its the thin, 12"?
2019-01-28 20:59:48 <clandmeter> the new one sucks
2019-01-28 21:00:02 <clandmeter> new macbook have diff keyboards
2019-01-28 21:00:04 <clandmeter> less travel
2019-01-28 21:00:12 <ncopa> right
2019-01-28 21:00:37 <clandmeter> im thinking of switching to x1
2019-01-28 21:00:44 <smemsh> i also have a late 2015 which works fine with debian for long time and excellent power management.  the 2017 one has power management issues and you have to turn off thunderbolt controllers or it sucks power
2019-01-28 21:00:53 <clandmeter> but thats one expensive keyboard :)
2019-01-28 21:01:36 <clandmeter> ncopa, yours is still with a backlight logo?
2019-01-28 21:02:05 <smemsh> the low travel of the new keyboards is great for speed but the clicking is obnoxiously loud
2019-01-28 21:02:24 <clandmeter> its opposite for me.
2019-01-28 21:02:33 <clandmeter> its makes even more errors for me.
2019-01-28 21:02:43 <clandmeter> i want to hit them too fast...
2019-01-28 21:03:58 <ncopa> yes i have backlight logo
2019-01-28 21:04:03 <smemsh> you have to not think of it as percussion anymore but you're just moving your fingers slightly and more like gliding them
2019-01-28 21:04:18 <clandmeter> i also have an xps15, but the keyboard is also... not perfect.
2019-01-28 21:04:35 <ncopa> i bought apple magic keyboard for my linux desktop too
2019-01-28 21:04:39 <ncopa> bluetooth
2019-01-28 21:04:43 <clandmeter> when you are used to mech keys, the rest sucks.
2019-01-28 21:04:48 <ncopa> its great
2019-01-28 21:08:30 <smemsh> ok i will submit an issue with list of modules for macbook pro, it may not be complete but should be close.  also kernel 4.20 is needed for framebuffer to work but that will likely have to wait until a longterm 4.20 on kernel.org appears, i gather
2019-01-28 21:09:59 <clandmeter> smemsh, do you have some ref regarding the power issues and pro17?
2019-01-28 21:19:20 <smemsh> clandmeter: this might be it, there are others on there: https://github.com/Dunedan/mbp-2016-linux/issues/24
2019-01-28 21:26:59 <maldridge> ncopa: I guess I can make my other DNS server recurse to work around this non-standard behavior, but its a bit annoying
2019-01-28 21:30:44 <ncopa> yeah, i think that woudl be the best
2019-01-28 21:30:54 <ncopa> and yes, it is a bit annoying, but it is by design
2019-01-28 21:31:09 <maldridge> so now I just need to figure out how to make the machine DNS static
2019-01-28 21:40:19 <maldridge> is there a decent guide for that?
2019-01-28 22:41:30 <mps> on aarch64 'apk add linux-vanilla-dev' says: ERROR: linux-vanilla-dev-4.19.18-r0: BAD signature
2019-01-28 22:42:09 <mps> //etc/apk/repositories line http://dl-cdn.alpinelinux.org/alpine/v3.9/main
2019-01-29 01:09:27 <maldridge> ncopa: does the resolver go in order?  i.e. my first resolver should always work, but if it doesn't I'd like to have a backup?
2019-01-29 01:09:35 <maldridge> I've got a bootstrap problem at the moment...
2019-01-29 01:17:38 <dalias> maldridge, resolver is in parallel
2019-01-29 01:18:00 <maldridge> hrm, that is a problem, since in some cases an external resolver can't answer for my zone
2019-01-29 01:18:35 <dalias> you can't use multiple nameservers that disagree about if/how domains are defined
2019-01-29 01:19:24 <maldridge> right
2019-01-29 01:19:38 <maldridge> which is rapidly meaning that musl's resolver is unusable to me, and I'll need to run unbound somewhere
2019-01-29 01:20:13 <maldridge> (which doesn't actually clean up any of the problems of bootstrapping such an environment)
2019-01-29 01:20:54 <dalias> (note: it would be broken with glibc too, just in subtler ways. if your local nameserver failed to respond in time, you'd get a positive "does not exist" result rather than a "lookup failed" result
2019-01-29 01:22:52 <maldridge> while you are correct, I can almost guarantee that the local resolver can answer faster by virtue of having everything local and in memory
2019-01-29 01:25:11 <dalias> right. it's more like if you have a network blip or it gets rebooted or crashes or something
2019-01-29 01:25:26 <dalias> in which case the right answer is an error not does-not-exist
2019-01-29 01:25:31 <maldridge> I'm not entirely sure how you'd have a network blip on localhost, but sure
2019-01-29 01:25:52 <dalias> indeed, it's less likely on localhost
2019-01-29 01:26:07 <dalias> but you could have swap-to-hell if you have swap enabled
2019-01-29 01:26:46 <maldridge> nope
2019-01-29 01:26:51 <maldridge> oomkiller bites immediately
2019-01-29 01:30:20 <maldridge> it doesn increasingly seem though that my only option is HA unbound
2019-01-29 06:39:42 <samof76> Hi. I am totally new here.
2019-01-29 06:40:02 <samof76> I have been using alpine in my dev environment
2019-01-29 06:40:34 <samof76> But one of the org concerns while moving to production is that, does alpine come with LTS?
2019-01-29 07:08:55 <samof76> Does alpine have LTS releases?
2019-01-29 07:09:37 <maldridge> samof76: you seem new to IRC; in general you may need to idle for quite some time for someone to answer your inquiry, largely because they may be asleep on the other side of the planet
2019-01-29 07:09:56 <maldridge> this channel in particular is quite quiet
2019-01-29 07:10:12 <maldridge> but to give a non-official answer to your question, I run alpine in prod and I'm reasonably happy with it
2019-01-29 07:10:35 <maldridge> (though I also run void in prod, which is rolling release; its all about what you are confident in managing)
2019-01-29 07:12:13 <TBB> no LTS in the RHEL sense of it or even the Ubuntu sense of it. that said, stable releases have a lifespan longer than, say, Ubuntu unstables
2019-01-29 09:40:29 <lloydz> Hi!
2019-01-29 09:40:47 <consus> Hi
2019-01-29 09:41:51 <lloydz> need help install alpine linux on pc engines APU2 board. This system has only serial console and i can't change the iso to use serial console
2019-01-29 09:42:21 <consus> Hm
2019-01-29 09:42:41 <consus> You can just install alpine in a qemu and tarball the resulting FS over the console
2019-01-29 09:42:47 <consus> Or USB
2019-01-29 09:42:50 <consus> Or whatever
2019-01-29 09:43:01 <lloydz> .. on the system is allready a llinux system, so i can add a new partition and install via chroot and alpine rootfilesystem package. need help
2019-01-29 09:43:46 <consus> Also syslinux allows you to change the console device
2019-01-29 09:43:56 <consus> Have you tried that?
2019-01-29 09:44:11 <consus> https://wiki.alpinelinux.org/wiki/Enable_Serial_Console_on_Boot
2019-01-29 09:44:51 <lloydz> i don't have knowledge about qemu. i found a wiki article.. yes exactly. but after i change the iso with isomaster, i can't boot
2019-01-29 09:45:52 <lloydz> is there an ready iso with serial console enabled?
2019-01-29 09:47:16 <consus> Hmm
2019-01-29 09:47:20 <consus> append "quiet console=ttyS0,9600 console=tty0"
2019-01-29 09:47:28 <consus> Seems like you can do this an runtime too
2019-01-29 09:47:52 <jwh> you can create images with the vm script
2019-01-29 09:47:55 <jwh> just change the kernel
2019-01-29 09:49:55 <lloydz> yes that is the term, i have to search at syslinux to change it at runtime.
2019-01-29 09:51:14 <lloydz> i found a wiki article for the ALIX board: https://wiki.alpinelinux.org/wiki/Bootstrapping_Alpine_on_PC_Engines_ALIX.3 but it is alpine version 1.x
2019-01-29 09:51:55 <lloydz> ok. thank you
2019-01-29 09:56:09 <pax_rhos> ncopa: hey, you are a maintainer of ca-certificate, could you confirm it is broken for you as well?
2019-01-29 09:58:02 <pax_rhos> oh, I see the version is 'flagged', wonder what that means
2019-01-29 09:58:19 <AinNero> it has updates
2019-01-29 09:58:55 <pax_rhos> all ca-cert* packages are flagged on edge, but not on 3.9
2019-01-29 09:59:02 <pax_rhos> or 3.8
2019-01-29 09:59:19 <pax_rhos> and 3.8 is using an even older version :(
2019-01-29 10:03:12 <clandmeter> pax_rhos, you cannot flag stable releases
2019-01-29 10:03:21 <pax_rhos> oh
2019-01-29 10:03:22 <clandmeter> stable releases only get sec updates
2019-01-29 10:03:38 <pax_rhos> well, it looks like the package is unusable, should I file a bug?
2019-01-29 10:03:49 <clandmeter> if you think its a bug yes.
2019-01-29 10:22:38 <pax_rhos> https://bugs.alpinelinux.org/issues/9935
2019-01-29 10:25:48 <pax_rhos> I hope this issue gets addressed at least before 3.9 gets released, or gets backported
2019-01-29 10:27:50 <AinNero> pax_rhos: https://github.com/gliderlabs/docker-alpine/issues/30
2019-01-29 10:28:02 <AinNero> https://github.com/alpinelinux/aports/commit/3faf2e032027ffa255f9a49f2f17c8e0239173ff
2019-01-29 10:29:10 <AinNero> the bug was talked about yesterday afaik
2019-01-29 10:29:32 <pax_rhos> AinNero: yeah, I posted that :)
2019-01-29 10:29:48 <AinNero> ah right
2019-01-29 12:07:56 <pax_rhos> The same issue is actual for edge branch and ca-certificates-20180924-r3 in it.
2019-01-29 12:43:56 <pax_rhos> who's responsible for the official docker image of alpine? https://hub.docker.com/_/alpine contains '&amp;&amp;' instead of '&&' in an example code block
2019-01-29 12:45:00 <AinNero> See the section "Quick reference"
2019-01-29 12:45:38 <AinNero> if you check the file contents, its correct there
2019-01-29 12:45:38 <pax_rhos> thanks, didn't notice :)
2019-01-29 12:45:49 <AinNero> its a rendering bug on the docker hub
2019-01-29 13:24:41 <markand> hello there
2019-01-29 13:33:11 <steven> hi.
2019-01-29 13:38:34 <misterunknown> Greetings.
2019-01-29 14:37:12 <tmhoang> <misterunknown> Hello
2019-01-29 14:37:21 <tmhoang> markand Hello
2019-01-29 14:37:32 <tmhoang> steven Hello
2019-01-29 15:33:02 <ncopa> pax_rhos: thank you for your bug report, is the problem that the ca certs are aoutdated or is it the warning that is the problem
2019-01-29 15:33:19 <ncopa> or just that it gives different output than debian?
2019-01-29 15:33:29 <ncopa> or that it does not import certs in /usr/local?
2019-01-29 15:33:49 <pax_rhos> it looks like ca-certificates package (of different versions) present in alpines - is just simply broken
2019-01-29 15:34:17 <ncopa> so the cert you add is not imported?
2019-01-29 15:35:05 <pax_rhos> well, the warning tells me that it didn't find the cert
2019-01-29 15:35:25 <pax_rhos> I don't really know how to test it :(
2019-01-29 15:36:27 <ncopa> the warning says it tries to import a cert that is a bundle (not a single cert), so it skips it
2019-01-29 15:37:00 <ncopa> it does not say: /usr/local/share/ca-certificates/mail.amur-cit.ru.crt not found
2019-01-29 15:37:32 <ncopa> i'd like to fix it but i dont understand what does not work
2019-01-29 15:37:42 <ncopa> i can change the code to not print the warning if you want
2019-01-29 15:44:18 <ncopa> pax_rhos: what is the error message you get when you try connect to the server?
2019-01-29 16:47:55 <pax_rhos> ncopa: I'm sorry, I've broken production and was fixing it. Let me try it again.
2019-01-29 16:51:29 <pax_rhos> ncopa: yeah, sorry for confusing you guys with a ticket, turns out I should just have ignored the warning, the import is in fact succeeds was
2019-01-29 16:51:52 <pax_rhos> was^W
2019-01-29 16:54:15 <pax_rhos> ncopa: weird that an even older version (20161130) on debian does output some lines upon success, and throws no warnings
2019-01-29 16:54:25 <pax_rhos> (they are listed in ticket)
2019-01-29 17:13:40 <pax_rhos> ncopa: since you know about certificates, do you, by any chance, know how to make those certs work via php? our application fails to send an email using that mail server with self-signed cert: `[error] [scripts] stream_socket_enable_crypto(): SSL operation failed with code 1. OpenSSL Error messages:
2019-01-29 17:13:40 <pax_rhos> error:14007086:SSL routines:CONNECT_CR_CERT:certificate verify failed:in/base/app/vendor/swiftmailer/swiftmailer/lib/classes/Swift/Transport/StreamBuffer.php at line 103`, I've tried to goolge the problem and stumbled upon this https://github.com/composer/composer/issues/3346 discussion, `php -r "var_dump(openssl_get_cert_locations());"` returns this:  https://paste.ee/p/WxITY
2019-01-29 17:22:09 <unrznbl[m]> are packages built on alpinelinux or on some other system? I was trying to use aports/main/weechat/APKBUILD on an alpine system on x86_64 system but get errors with libintl_gettext.
2019-01-29 17:24:40 <unrznbl[m]> I tried "abuild -r" and do have libintl package installed. (have had these problems with gnu arm toolchain and crosstool-ng)
2019-01-29 17:39:35 <unrznbl[m]> well... nevermind... I altered APKBUILD a bit and it seems to install but fail on a shared object (so) symbol relocation error. Will look at it more later. :) Would be nice to figure out my troubles with NLS/libint. https://bpaste.net/show/d914ca6ce64f
2019-01-29 18:50:27 <ncopa> pax_rhos: the update-ca-certificates tool in alpine is different than debian. does similar but different things
2019-01-29 18:50:38 <ncopa> it is a completely different tool
2019-01-29 18:50:54 <ncopa> and i bet alpine's version does it faster :)
2019-01-29 18:52:27 <ncopa> pax_rhos: does it work if you copy the cert to /etc/ssl/certs and un update-ca-certificates?
2019-01-29 18:52:52 <ncopa> s/un//
2019-01-29 18:54:27 <ncopa> if not, then try make the /etc/ssl/cert.pem a symlink that points to /etc/ssl/certs/ca-certificates.crt
2019-01-29 21:47:25 <ncopa> does this look ok? https://wwwtest.alpinelinux.org/posts/Alpine-3.9.0-released.html
2019-01-29 23:28:27 <qrvpzvb> topic needs update, 3.9 is out!
2019-01-29 23:28:38 <qrvpzvb> any idea why the switch to openssl ?
2019-01-29 23:29:30 <mps> qrvpzvb: it is discussed on mailing list, you can find reasons there
2019-01-29 23:30:18 <mps> I don't have url right now, sorry
2019-01-29 23:31:52 <m4t> yay
2019-01-29 23:32:11 <qrvpzvb> I found it
2019-01-29 23:37:27 <dimon222[m]> nice, happy 3.9 alpine
2019-01-29 23:41:03 <ncopa> thanks!
2019-01-29 23:42:06 <maldridge> well done
2019-01-29 23:42:22 <ncopa> thanks to everyone who has contributed
2019-01-30 00:23:02 <jn__> congratulations on the release!
2019-01-30 00:29:15 <TBB> \o/
2019-01-30 07:34:05 <vir2l> thx for 3.9 release
2019-01-30 07:34:45 <vir2l> everything works for me except for dovecot :-(
2019-01-30 07:35:22 <vir2l> system hangs at openrc boot-time with msg: Caching service dependencies ... UNTIL sysrq key pressed
2019-01-30 07:35:53 <vir2l> does anyone give me a hint where to look first to find the problem?
2019-01-30 07:57:01 <vir2l> one interesting thing is, that i didn't run rc-update to put dovecot into a runlevel. it is installed but not startet and yet it hangs at boot time ??? what happens while running "apk add dovecot" to the system
2019-01-30 08:00:23 <TBB> that's strange. doesn't seem to set itself up to run via rc-update at least.
2019-01-30 08:04:43 <clandmeter> vir2l: do you get a shell prompt?
2019-01-30 08:05:59 <vir2l> no
2019-01-30 08:06:35 <vir2l> normally I force the rc-update during my installer-script
2019-01-30 08:07:21 <vir2l> but I leave dovecot completely away just to see whats going on. so it must be during pre- post- installer while apk add
2019-01-30 08:07:49 <TBB> just checked those, didn't seem to contain such code
2019-01-30 08:08:12 <vir2l> or in the patches that are new in 3.9
2019-01-30 08:08:18 <vir2l> dovecot package
2019-01-30 08:08:56 <vir2l> brb
2019-01-30 08:11:52 <clandmeter> vir2l: it sounds strange.
2019-01-30 08:12:10 <clandmeter> no package should automatically add itself to a runlevel
2019-01-30 08:12:54 <clandmeter> are you sure its dovecot?
2019-01-30 08:13:05 <clandmeter> is this real hardware?
2019-01-30 08:13:43 <clandmeter> i think we have seen issues with sshd starting and not enough entropy
2019-01-30 08:21:53 <mps> clandmeter: sshd starting issue because not enough entropy is not Alpine specific. It is mostly issue in VM's which don't have enough randomness source or some low end boxes
2019-01-30 08:22:56 <clandmeter> yes i know
2019-01-30 08:23:05 <clandmeter> thanks for reminding me
2019-01-30 08:23:12 <mps> haveged and rngd could help to some degree there, and for VM's virtio-rng kernel module
2019-01-30 08:24:44 <mps> I know that you know, but write this above for others who read and don't yet read about this issue
2019-01-30 08:25:25 <clandmeter> im searching the kernel option to enable it again. i cant find it.
2019-01-30 08:25:57 <clandmeter> it would be nice for ppl to be able to add alternative source of entropy without having to go into single mode.
2019-01-30 08:26:35 <clandmeter> i think it should have been added to the relnotes
2019-01-30 08:27:25 <mps> I tested it a little on small arm board, haveged helped to speed up booting but still not fast as before
2019-01-30 08:28:04 <mps> and, not only sshd also noticed problem with haproxy
2019-01-30 08:30:40 <mps> btw, who is responsible for download page, armv7 is missing there
2019-01-30 08:31:02 <clandmeter> random.trust_cpu=on
2019-01-30 08:31:30 <mps> does that work on 4.19 kernel
2019-01-30 08:31:41 <clandmeter> https://www.kernel.org/doc/html/v4.19/admin-guide/kernel-parameters.html
2019-01-30 08:33:00 <mps> then, this is goog candidate to be added to release notes
2019-01-30 08:33:15 <mps> s/goog/good/
2019-01-30 09:20:03 <parazyd> Are there any plans for netifrc on Alpine?
2019-01-30 09:24:29 <clandmeter> not that i know of
2019-01-30 09:24:44 <clandmeter> current debian style does not meet your requirements?
2019-01-30 09:25:59 <parazyd> I'm wondering how to make an entry for a wireguard interface.
2019-01-30 09:26:30 <clandmeter> i can give you an example if you like
2019-01-30 09:26:41 <parazyd> Oh, that would be kind!
2019-01-30 09:26:47 <clandmeter> i am very kind
2019-01-30 09:26:50 <clandmeter> and handsome
2019-01-30 09:26:54 <Lord> :-)
2019-01-30 09:26:54 <parazyd> :3
2019-01-30 09:27:04 <parazyd> I've been using wg-quick to start my interface(s)
2019-01-30 09:27:11 <clandmeter> right
2019-01-30 09:27:15 <clandmeter> you can remove it
2019-01-30 09:27:29 <clandmeter> at least i dont use it.
2019-01-30 09:27:40 <parazyd> Ok
2019-01-30 09:28:14 <parazyd> How would it look in /etc/network/interfaces then?
2019-01-30 09:28:15 <clandmeter> oh crap i removed my wireguard install
2019-01-30 09:28:27 <parazyd> /o\
2019-01-30 09:28:38 <clandmeter> its pretty simple actually
2019-01-30 09:29:02 <clandmeter> from my head
2019-01-30 09:30:19 <clandmeter> you do a regular iface config in interfaces
2019-01-30 09:30:43 <clandmeter> but you append: pre-up ip link add dev wg0 type wireguard
2019-01-30 09:31:20 <parazyd> I see
2019-01-30 09:31:46 <clandmeter> and you can do a post-up with a route
2019-01-30 09:32:25 <clandmeter> i think you also have to do a post-down to remove the interface iirc
2019-01-30 09:32:40 <clandmeter> in general you have 3 extra lines in interfaces
2019-01-30 09:32:48 <parazyd> Okay. Thanks a lot! I'll try it out :)
2019-01-30 09:34:06 <clandmeter> the module will be auto loaded when you add the interface
2019-01-30 09:34:13 <clandmeter> so you dont have to do that manually
2019-01-30 09:34:31 <clandmeter> i should have written that down on the wiki... sorry
2019-01-30 09:34:34 <parazyd> Ha, works :)
2019-01-30 09:34:51 <clandmeter> yes its very nice
2019-01-30 09:34:51 <parazyd> Thank you, handsome op
2019-01-30 09:34:55 <parazyd> :p
2019-01-30 09:34:59 <clandmeter> haha :p
2019-01-30 09:35:56 <clandmeter> wireguard is very nice because of simplicity and performance.
2019-01-30 09:36:42 <mps> parazyd: would you mind to put config on some paste server, so I don't have to reinvent the will. I'm also using wg-quick and think to replace it
2019-01-30 09:37:18 <parazyd> clandmeter: Yeah. I'm using it on a digitalocean droplet to forward to a home server, since at home I don't have a static IP.
2019-01-30 09:37:28 <parazyd> Pretty sweet
2019-01-30 09:37:30 <clandmeter> i modified the package so you are able to remove it. should be in 3.9 (or edge)
2019-01-30 09:37:34 <parazyd> mps: One sec
2019-01-30 09:38:01 <clandmeter> wg-quick pulls in a lot of deps which are actually not needed.
2019-01-30 09:38:11 <parazyd> mps: https://ptpb.pw/5hR9
2019-01-30 09:38:16 <clandmeter> well, needed for wg-quick not regular wg.
2019-01-30 09:38:43 <mps> yes, I'm using wireguard a long time for my private net in server and client modes, simple but works and fast on DSL lines
2019-01-30 09:39:31 <mps> parazyd: thanks
2019-01-30 09:39:36 <parazyd> You're welcome
2019-01-30 09:52:36 <vir2l> back again
2019-01-30 09:52:50 <vir2l> anything new about dovecot?
2019-01-30 09:53:12 <_ikke_> new since / regarding/
2019-01-30 09:53:14 <_ikke_> ?
2019-01-30 09:53:32 <vir2l> was afk for about one hour
2019-01-30 09:54:01 <vir2l> since 3.9 i cannot boot after installing dovecot
2019-01-30 09:54:37 <vir2l> system, hangs at openrc "caching service dependencies" till I hit sysrq key
2019-01-30 09:55:05 <vir2l> apk del dovecot makes system back booting fine
2019-01-30 09:55:50 <vir2l> so it must be somwhere in pre- post- installer scripts or in the patches coming fpr dovecot since 3.9 rc4 till now
2019-01-30 09:56:06 <vir2l> but cant find any mistakes
2019-01-30 09:56:57 <vir2l> also strange ... dovecot is not set to runlevels via rc-update or so...
2019-01-30 09:57:26 <_ikke_> And it's also not present in the package
2019-01-30 09:57:41 <vir2l> what do you mean ?
2019-01-30 09:57:53 <_ikke_> No file that would enable the service
2019-01-30 09:58:26 <_ikke_> eg, /etc/runlevels/*
2019-01-30 09:58:28 <vir2l> this is not the problem, because i force it during my installer script
2019-01-30 09:58:48 <mps> vir2l: on which arch you have issue with dovecot
2019-01-30 09:59:09 <vir2l> isnt it strange, that system does not boot yet without dovecot being in any runlevel ??
2019-01-30 09:59:18 <vir2l> x64
2019-01-30 10:00:16 <_ikke_> vir2l: you said ealier that removing dovecot fixes it, not adding it..?
2019-01-30 10:00:29 <mps> hmm, there could be some issue on armhf, but not on x86_64. it is related to dovecot-pigeon-hole, iirc
2019-01-30 10:00:29 <vir2l> so what I think it has nothing to do with the init-scripts bescause they were not in use
2019-01-30 10:00:47 <vir2l> yes -- apk del dovecot makes the system boot again
2019-01-30 10:02:12 <mps> do you have ssl/tls configured for dovecot
2019-01-30 10:02:59 <vir2l> I would if the plain package would not make trouble ;-)
2019-01-30 10:03:01 <mps> oh, sorry, you wrote it is disabled to boot, now see
2019-01-30 10:04:07 <vir2l> not only disabled ... I need to remove it completely. disable sounds to me like removing form runlevel
2019-01-30 10:04:41 <vir2l> maybe I have some side-effects occuring during my installer
2019-01-30 10:05:17 <vir2l> I will try alpine-setup instead trying to re-produce
2019-01-30 10:05:45 <vir2l> If dovecot works after that it must be an issue in my script
2019-01-30 10:08:54 <mps> ah, you have some custom made installer script?
2019-01-30 10:13:06 <vir2l> yes
2019-01-30 10:13:41 <vir2l> i need to have one, because my server provider does not offer alpine as image
2019-01-30 10:14:23 <vir2l> but its a simple bootstrap with (a lot of) customizations
2019-01-30 10:15:46 <vir2l> openrc verbose mode does not show everything :-(
2019-01-30 10:18:45 <mps> you can try install locally in qemu with your custom installer to debug it
2019-01-30 10:24:21 <Andreetje> vir2l is your script public ?
2019-01-30 10:25:51 <vir2l> Andreetje: I check it and upload it
2019-01-30 10:26:09 <vir2l> post link in a minute
2019-01-30 11:09:51 <Ai9zO5AP> I see that alpine 3.9 block on  Starting busybox crond ... [ ok ]
2019-01-30 11:10:00 <Ai9zO5AP> I upgrade and restart
2019-01-30 11:12:11 <genghiz> Question about alpine policies. There are certain packages which maintain no stable versions. Like gitea, for instance. If a new version is released, will the stable releases get them? Or will I have to be on edge for that?
2019-01-30 11:15:13 <genghiz> Or, say, nextcloud. 3.9 has stabilised on 15.0.2, but will we get point releases? Or is frozen completely frozen, no point releases either?
2019-01-30 11:15:38 <genghiz> (Sorry for the beginner questions, am not very familiar with Alpine)
2019-01-30 11:16:54 <AinNero> genghiz: personall recommendation: If you run production things, avoid things that don't have stable branches
2019-01-30 11:17:38 <genghiz> AinNero It's mostly for personal projects and because I like the project, hence I'm not _very_ fussed with it. Good advice in general though, thanks.
2019-01-30 11:18:07 <Ai9zO5AP> have anyone manage to run alpine 3.9 ?
2019-01-30 11:18:24 <AinNero> Ai9zO5AP: yes
2019-01-30 11:18:52 <Ai9zO5AP> AinNero: It does not block after Starting busybox crond ... [ok]
2019-01-30 11:19:17 <genghiz> I've had no problems, personally (on an lxc container, though)
2019-01-30 11:19:21 <AinNero> not for me
2019-01-30 11:19:42 <AinNero> i upgraded to 3.9 on a physical machine
2019-01-30 11:20:50 <Ai9zO5AP> AinNero: I change the version to 3.9 and after I run apk upgrade --pure but I did not type sync before reboot
2019-01-30 11:21:14 <AinNero> rebooting does a sync per default
2019-01-30 11:21:38 <AinNero> but you could open a ticket, because i think this seems to be a spurious problem with openrc
2019-01-30 11:23:13 <Ai9zO5AP> AinNero: I am going to open it, thank you
2019-01-30 11:23:51 <AinNero> maybe search if a similar issue already exists
2019-01-30 11:24:14 <AinNero> i remember chrony or wpa_supplicant sometimes blocking for a while
2019-01-30 11:32:01 <clandmeter> vir2l: can you add a boot parameter?
2019-01-30 11:37:35 <clandmeter> AinNero: what kind of device?
2019-01-30 11:37:39 <clandmeter> sorry
2019-01-30 11:37:45 <clandmeter> Ai9zO5AP: what kind of device
2019-01-30 11:38:39 <AinNero> clandmeter: ThinkPad X220, badly reproducable
2019-01-30 11:38:46 <AinNero> oh i see
2019-01-30 11:39:09 <clandmeter> i tagged the wrong nick :)
2019-01-30 11:39:26 <clandmeter> im used to have enough with 2 chars
2019-01-30 11:53:44 <Ai9zO5AP> clandmeter: I am running bhyve
2019-01-30 11:54:52 <clandmeter> Ai9zO5AP: what happens if you add boot parameter random.trust_cpu=on
2019-01-30 11:55:21 <Ai9zO5AP> i am going to test it and see
2019-01-30 11:58:24 <Ai9zO5AP> clandmeter: I do not know how can I thank you
2019-01-30 11:58:45 <Ai9zO5AP> clandmeter: it works very well, and I should share with bhyve users
2019-01-30 11:59:10 <clandmeter> Ai9zO5AP: you can probably fix it by adding an entropy daemon
2019-01-30 11:59:44 <vir2l> clandmeter: do you mean kernel parameter ?
2019-01-30 11:59:52 <clandmeter> vir2l: yes
2019-01-30 12:00:10 <clandmeter> vir2l: try adding random.trust_cpu=on
2019-01-30 12:00:40 <vir2l> okay, but do you think it has an impact to openrc ?
2019-01-30 12:01:03 <Ai9zO5AP> clandmeter: done, thank you again
2019-01-30 12:01:21 <clandmeter> Ai9zO5AP: we disabled cpu generated entropy
2019-01-30 12:01:41 <clandmeter> so user should decide if they want to use it.
2019-01-30 12:04:43 <clandmeter> instead of enabling cpu generation you could install Haveged.
2019-01-30 12:08:49 <vir2l> clandmeter: wow
2019-01-30 12:09:18 <vir2l> this looks good ... it shortens the "caching service dependencies..." by far
2019-01-30 12:09:46 <clandmeter> you are probably on some virtual machine which does not have entropy
2019-01-30 12:10:16 <vir2l> what is entropy ??
2019-01-30 12:10:25 <clandmeter> random data
2019-01-30 12:10:32 <clandmeter> important for encryption
2019-01-30 12:11:05 <clandmeter> you need a source to generate proper random data. if you use the cpu and it has a backdoor, it would be not so random.
2019-01-30 12:11:19 <vir2l> as I have a good connection to my server hoster, what have I say to him, so that "entropy" becomes available for his customers
2019-01-30 12:11:39 <clandmeter> you can use a daemon to generate it yourself
2019-01-30 12:11:44 <clandmeter> like i suggested
2019-01-30 12:11:45 <Ai9zO5AP> clandmeter: just install it ?
2019-01-30 12:11:58 <vir2l> and how the hack did you figure out, that it solves my problem ??
2019-01-30 12:11:59 <clandmeter> install it and make it run on boot
2019-01-30 12:12:10 <clandmeter> im super smart
2019-01-30 12:12:13 <clandmeter> and handsome
2019-01-30 12:12:18 <vir2l> indeed you are
2019-01-30 12:12:32 <clandmeter> you guys are not the first to drop in here regarding this issue.
2019-01-30 12:12:48 <vir2l> i work on linux for about 10 years now, and still have not that knowledge
2019-01-30 12:13:09 <clandmeter> alpine tries to be secure by default. it has some consequences...
2019-01-30 12:13:33 <clandmeter> we need to add it to the release notes.
2019-01-30 12:13:40 <clandmeter> i hope users who upgrade will read it :)
2019-01-30 12:13:51 <Ai9zO5AP> i should understand alpine
2019-01-30 12:15:11 <vir2l> I tried it on my shadow server first
2019-01-30 12:15:31 <vir2l> cause my server is my mailbox :-D
2019-01-30 12:15:50 <clandmeter> Ai9zO5AP: apk add haveged && rc-update add haveged
2019-01-30 12:16:47 <clandmeter> then you should be able to remove the kernel cmdline option
2019-01-30 12:17:55 <clandmeter> https://wiki.alpinelinux.org/wiki/Entropy_and_randomness
2019-01-30 12:18:03 <mps> clandmeter: I agree, and thought about adding that last night when 'fall to bed' after release :)
2019-01-30 12:18:34 <mps> I expected these questions soon
2019-01-30 12:19:05 <mps> but was to lazy to wake and again bother ncopa
2019-01-30 12:21:32 <vir2l> last question: is random.trust_cpu=on preferrable for production?
2019-01-30 12:22:24 <vir2l> or can it produce side effects in the qemu scopes?
2019-01-30 12:22:30 <mps> vir2l: you have to decide yourself, but first investigate potential risks
2019-01-30 12:23:19 <mps> in qemu, and some other VM's you can use virtio-rng kernel module
2019-01-30 12:23:28 <vir2l> this seems to me like a bottomless drum
2019-01-30 12:24:17 <clandmeter> its pretty simple, do you trust hw your vm runs on.
2019-01-30 12:24:30 <mps> well, quite well understand you. After more than a two decades with linux sometimes I feel like you
2019-01-30 12:24:32 <vir2l> i trust my hoster :-D
2019-01-30 12:24:33 <clandmeter> if yes, you enable the option. if not you generate it yourself.
2019-01-30 12:25:23 <clandmeter> its not only "the hoster"
2019-01-30 12:25:32 <clandmeter> the hoster will use hw
2019-01-30 12:25:39 <clandmeter> probably intel
2019-01-30 12:26:06 <mps> but running server on the net this 'entropy source' issue is just 'another one brick in the wall'
2019-01-30 12:26:11 <_ikke_> that hw will also be shared
2019-01-30 12:26:22 <vir2l> never asked him
2019-01-30 12:28:03 <clandmeter> its trust code vs trust hw. you can read the source of haveged, you cannot debug the upper hw.
2019-01-30 12:28:19 <vir2l> anyway ... thx guys, i try to finish installation now
2019-01-30 12:28:59 <vir2l> and ... I love alpine :-) Its like coming home after a long adventure searc for the best distro :-)
2019-01-30 12:29:43 <mps> if someone can tolerate somewhat slow boot and think that the entropy is important then I think few minutes (maybe 10) could be acceptable
2019-01-30 12:30:02 <vir2l> I use it for about 2 years now in production as mailserver, and it serves me very very well
2019-01-30 12:32:48 <vir2l> goodbye guys
2019-01-30 12:47:08 <tugrik> hi - forum.alpinelinux.org doesn't exist any more?
2019-01-30 12:47:51 <_ikke_> correct4
2019-01-30 12:48:03 <tugrik> is it coming back, or permanently gone?
2019-01-30 12:48:16 <_ikke_> There are not plans to revive it
2019-01-30 12:48:48 <tugrik> got it, just found a cached page on Google explaining as such, thanks
2019-01-30 13:24:10 <EuroDomenii> Is there an equivalent of the apt vulnerability https://security-tracker.debian.org/tracker/CVE-2019-3462 for apk ?
2019-01-30 13:25:21 <EuroDomenii> In the past, Max Justicz has found https://justi.cz/security/2018/09/13/alpine-apk-rce.html
2019-01-30 13:26:08 <AinNero> that issue was unrelated to redirects
2019-01-30 13:26:46 <mps> EuroDomenii: this issue is was a little overhyped by author, imo
2019-01-30 13:26:48 <AinNero> like apk and apt are different programs, with different approaches
2019-01-30 13:27:07 <AinNero> mps: Security Issues are always overhyped
2019-01-30 13:27:32 <mps> AinNero: right :)
2019-01-30 13:27:41 <AinNero> except in this case, it didn't even get an own domain
2019-01-30 13:27:50 <AinNero> with fancy website
2019-01-30 13:28:22 <AinNero> if you know my blog, i have a rant up about reputation whoring in the ITsec industry
2019-01-30 13:28:44 <EuroDomenii> thanks for reply
2019-01-30 13:28:48 <mps> I looked it earlier but forgot exact url
2019-01-30 13:29:11 <EuroDomenii> So, the apk packages are sanitizing correctly the redirects?
2019-01-30 13:29:58 <clandmeter> EuroDomenii: i dont know if we are vulnerable. The persons who could know are not online atm.
2019-01-30 13:30:12 <mps> EuroDomenii: it is hard to inject package which is not passed official check on Alpine
2019-01-30 13:30:50 <AinNero> EuroDomenii: apk does not use hashsums from repository server like debian does
2019-01-30 13:30:56 <mps> that one could be used only if you install package from untrusted repo
2019-01-30 13:31:25 <clandmeter> EuroDomenii: the best way would be to verify it yourself and let us know.
2019-01-30 13:31:35 <mps> but, if you install package from untrusted repo it is always risky
2019-01-30 13:32:26 <EuroDomenii> Thanks for the tips. I'm quite new to alpine. Anyway, it's may worth the vulnerability mechanism in debian, to check if it applies to alpine
2019-01-30 13:32:52 <clandmeter> EuroDomenii: some insight on that sec issue you mentioned https://git.alpinelinux.org/apk-tools/commit/?id=6484ed9849f03971eb48ee1fdc21a2f128247eb1
2019-01-30 13:33:42 <mps> also, in Debian if someone add package with 'rm -rf /' in post-install you are doomed
2019-01-30 13:34:26 <AinNero> mps: is there a package system that prevents this?
2019-01-30 13:34:50 <mps> AinNero: I doubt, all run as root, afaik
2019-01-30 13:34:52 <ViltusVilks> if you are in Docker then it is not so critical, but depends on the code.. if docker.sock passed, then...
2019-01-30 13:35:37 <AinNero> but that the general question whether you want to run stangers code on your machine
2019-01-30 13:36:25 <mps> right, that is first rule in security on the computers
2019-01-30 13:37:51 <AinNero> but essentially binary linux distributions are just controlled RCE
2019-01-30 13:37:52 <mps> remember weak ssh keys in debian few years ago
2019-01-30 13:39:06 <mps> real minefields are more 'secure' than computer security
2019-01-30 13:39:45 <AinNero> thats bullshit
2019-01-30 13:39:58 <AinNero> you dont get ripped apart from touching a wrong button
2019-01-30 13:40:18 <mps> difference is, in real minefield you risk leg and in computer security field you risk data and what not else
2019-01-30 13:41:22 <mps> oh, yea, forgot that in real minefield you risk something between legs  also :(
2019-01-30 13:42:25 <mps> security is process and not a final solutions, i.e. there is no final solutions
2019-01-30 13:42:31 <EuroDomenii> I've been disconnected
2019-01-30 13:42:37 <EuroDomenii> <@clandmeter> EuroDomenii: i dont know if we are vulnerable. The persons who could know are not online atm.
2019-01-30 13:43:04 <EuroDomenii> So, should I report it somewhere, or this issues is already on the radar
2019-01-30 13:43:20 <EuroDomenii> and apk maintainers will check the similarity of apt exploit?
2019-01-30 13:43:39 <mps> EuroDomenii: irc logs are archived somewhere on alpinelinux.org, but can't remember exact url
2019-01-30 18:34:01 <gr8> hey, just read that you are moving from LibreSSL to OpenSSL? That's very intersting, why did you make that move?
2019-01-30 18:39:17 <AinNero> gr8: http://lists.alpinelinux.org/alpine-devel/6308.html
2019-01-30 18:43:26 <gr8> ok
2019-01-30 19:03:32 <FreeBDSM> hello, ncopa, could you please tell me what file should I point at in my php.ini's `[openssl] openssl.cafile=?` ?
2019-01-30 19:06:04 <FreeBDSM> /etc/ssl/certs/ca-certificates.crt ?
2019-01-30 19:06:08 <_ikke_> sounds good
2019-01-30 19:06:46 <FreeBDSM> well, still doesn't work :(
2019-01-30 19:08:53 <FreeBDSM> https://github.com/docker-library/php/blob/57b41cfc2d1e07acab2e60d59a0cb19d83056fc1/7.2/alpine3.7/fpm/Dockerfile wait, so the image uses libressl instead of openssl?
2019-01-30 19:09:29 <FreeBDSM> disregard that, I use my own image, it uses openssl
2019-01-30 19:16:09 <FreeBDSM> I've tried setting it to /usr/local/share/ca-certificates/mycert.crt, still no luck
2019-01-30 19:16:18 <FreeBDSM> tried to rename .crt to .pem - still no luck :(
2019-01-30 19:19:15 <FreeBDSM> looks like a bug to me :(
2019-01-30 19:19:41 <FreeBDSM> but I'm not entirely sure, as I use an application (based on SwiftMailer) to test that
2019-01-30 19:24:15 <FreeBDSM> also, who's bug it is? alpine? openssl in alpine? ca-certificates in alpine? php7-fpm in alpine? php7-openssl in alpine?
2019-01-30 19:26:12 <FreeBDSM> hooold on, `php -i` returns https://paste.ee/p/6ODPV this about openssl, so is it using OpenSSL or LibreSSL?
2019-01-30 19:27:37 <rafalcpp> kernels are still using grsecurity
2019-01-30 19:27:40 <rafalcpp> are they?
2019-01-30 19:27:58 <rafalcpp> how does Alpine go around the problem of grsec being closed off, the organization buys the patch?
2019-01-30 19:28:36 <AinNero> no, Alpine does not ship grsec anymore
2019-01-30 19:28:49 <rafalcpp> AinNero: since when / why not?
2019-01-30 19:28:55 <AinNero> licensing
2019-01-30 19:29:20 <AinNero> i think 3.7 was last with hardened kernel
2019-01-30 19:29:25 <AinNero> so, a year ago
2019-01-30 19:29:25 <rafalcpp> that's sad
2019-01-30 19:29:33 <rafalcpp> maybe people should create like opengrsec
2019-01-30 19:29:56 <AinNero> dont know. the important patches have been upstreamed anyways
2019-01-30 19:30:00 <rafalcpp> also we could use Linus not being noob about security
2019-01-30 19:30:07 <rafalcpp> AinNero: they are not
2019-01-30 19:30:26 <rafalcpp> just minimal and often broken version of them are, it seems
2019-01-30 19:31:10 <AinNero> your opinions are interesting
2019-01-30 19:32:16 <FreeBDSM> https://bugs.alpinelinux.org/issues/8199 looks like this is my bug
2019-01-30 19:32:57 <AinNero> rafalcpp: if you are so good with security yourself, why dont you compile your own kernel?
2019-01-30 19:33:03 <AinNero> with the patches of your choice?
2019-01-30 19:33:11 <rafalcpp> that example is over a year ago, but https://www.grsecurity.net/an_ancient_kernel_hole_is_not_closed.php
2019-01-30 19:33:27 <mepholic> \o/ passive agressiveness always wins
2019-01-30 19:33:49 <rafalcpp> AinNero: I've been doing that, but I was wondering how other people nowdays go about obtaining the patch
2019-01-30 19:34:29 <mepholic> rafalcpp: pay spender
2019-01-30 19:34:42 <mepholic> or like, sign away your firstborn to him
2019-01-30 19:34:46 <rafalcpp> mepholic: obviously. I wondered how other distributions do it
2019-01-30 19:35:04 <mepholic> they don't
2019-01-30 19:35:08 <mepholic> nobody uses grsec anymore
2019-01-30 19:35:24 <rafalcpp> AinNero: above example, seems to prove attitude of Linus/kernel not being good with security, and example of how it's not true that PaX patches have been upstreamed correctly/fully
2019-01-30 19:35:26 <AinNero> rafalcpp: after reading the grsec post, it reads more like slandering on mainline
2019-01-30 19:35:45 <mepholic> you might be able to find some boutique shops that actually thought it was a good idea to pay spender
2019-01-30 19:36:07 <mepholic> maybe someone like nxp or some random hosting company
2019-01-30 19:36:20 <rafalcpp> mepholic: so they are only ones with secured kernel then?
2019-01-30 19:36:25 <rafalcpp> (on Linux family)
2019-01-30 19:36:31 <mepholic> but due to licensing restrictions now, any smart distribution will not touch grsec with a 40 foot pole
2019-01-30 19:36:47 <rafalcpp> AinNero: you mean that information there is not factual?
2019-01-30 19:37:14 <AinNero> rafalcpp: its the fame whoring approach of security people
2019-01-30 19:37:32 <mepholic> I actually agree with AinNero here
2019-01-30 19:37:41 <mepholic> I'm not saying that grsec has no technical merits
2019-01-30 19:37:41 <AinNero> and its coming from people who make money on selling their 'fixes'
2019-01-30 19:37:54 <rafalcpp> AinNero: who cares.  Do you think it is not factual that this vulnerabilities remained in Linux vanilla for years
2019-01-30 19:37:59 <mepholic> however they need to fix their PR
2019-01-30 19:38:06 <rafalcpp> because if so, then it seems no distribution now ships properly secured kernel for linux
2019-01-30 19:38:21 <rafalcpp> (assuming this didn't drastically changed over last year)
2019-01-30 19:38:46 <mepholic> rafalcpp: to be fair, you should be more concerned about vulnerabilities and backdoors in your hardware then in your kernel
2019-01-30 19:38:55 <mepholic> grsec won't protect against hardware flaws
2019-01-30 19:38:58 <AinNero> rafalcpp: explain the attack vector
2019-01-30 19:39:08 <rafalcpp> mepholic: and hardware will not secure me against physical attacks
2019-01-30 19:39:18 <rafalcpp> yeap, fixing one thing at a time
2019-01-30 19:39:20 <mepholic> that's also not entirely true
2019-01-30 19:39:44 <mepholic> i mean physical attacks are almost impossible to entirely defend against
2019-01-30 19:39:51 <rafalcpp> mepholic: ok I don't consider H&K MP5 "a hardware" ;)
2019-01-30 19:39:58 <mepholic> but at that point you might need to re-evaluate your threat model
2019-01-30 19:40:01 <AinNero> rafalcpp: after having a maintainer-like dayjob for soon 2 years i can totally understand why linus hates security people
2019-01-30 19:40:09 <mepholic> rafalcpp: well, it is
2019-01-30 19:40:21 <rafalcpp> exactly. Scope of my question was about kernel security.  not about HW security, that's another topic
2019-01-30 19:40:44 <mepholic> so is a Smith & Wesson M&P9
2019-01-30 19:41:09 <mepholic> some pretty slick hardware if you ask me
2019-01-30 19:41:28 <FreeBDSM> so, php wins :(
2019-01-30 19:41:30 <rafalcpp> so about kernel secuity, it seems no Linux ships secured kernel. And it seems Linux did not upstream all of protections from PaX
2019-01-30 19:41:40 <rafalcpp> I guess Alpine was last one
2019-01-30 19:42:19 <AinNero> grsec are asses with licensing, so what do you expect?
2019-01-30 19:42:20 <mepholic> rafalcpp: that's only true if you only consider kernels with the grsec patchset applied to be "secure" though
2019-01-30 19:42:23 <mepholic> that's my point
2019-01-30 19:42:58 <mepholic> and realistically
2019-01-30 19:43:21 <mepholic> if that's how you see it, and you have a problem with the fact that "no distros ship secure kernels anymore"
2019-01-30 19:43:30 <mepholic> take it up with the grsec team, not us :)
2019-01-30 19:43:34 <rafalcpp> mepholic: they are certainly removing many attack vectors
2019-01-30 19:43:53 <AinNero> rafalcpp: i know a linux distro that ships grsec per default
2019-01-30 19:44:08 <mepholic> because they're the only people who have any say in this, or have any way to solve the 'problem'
2019-01-30 19:44:21 <rafalcpp> rightm nothing is perfectly secure. Maybe that seL4
2019-01-30 19:44:32 <rafalcpp> AinNero: which one?
2019-01-30 19:44:37 <mepholic> now you're just kidding yourself
2019-01-30 19:44:42 <mepholic> first sentence: correct
2019-01-30 19:44:46 <AinNero> most security-oriented commercial ones
2019-01-30 19:44:49 <mepholic> addendum: not
2019-01-30 19:45:13 <rafalcpp> AinNero: do you know any good ones?
2019-01-30 19:45:16 <AinNero> yes
2019-01-30 19:46:47 <AinNero> on the serious side
2019-01-30 19:46:57 <mepholic> i mean there's like subgraph or whatever too I guess
2019-01-30 19:47:00 <mepholic> but
2019-01-30 19:47:02 <mepholic> ¯\_(ツ)_/¯
2019-01-30 19:47:13 <AinNero> if you are expecting a product that has to be payed for, except for free, you are essentially looking for piracy, rafalcpp
2019-01-30 19:47:22 <AinNero> but whatever
2019-01-30 19:47:49 <AinNero> security folks are sometimes oblivious to actual issues
2019-01-30 19:48:13 <AinNero> insulting head maintainers who rarely even do patches fits the stereotypes of being pre-opinionated
2019-01-30 19:50:12 <AinNero> feel free to fork and do better! would be a nice contribution to mankind
2019-01-30 19:51:03 <AinNero> or take a example on me
2019-01-30 19:51:20 <AinNero> i used to complain on software quite alot. i was a real PITA for maintainers
2019-01-30 19:51:42 <AinNero> but then i figured out, we all would be better off if i'd just invest my hate energy into working out fixes
2019-01-30 19:53:09 <AinNero> sometimes we forget that OSS is about contributing, not making demands
2019-01-30 20:13:21 <mepholic> AinNero++
2019-01-30 20:14:10 <mepholic> as someone who used to be annoying and bitchy to maintainers, and now a maintainer myself
2019-01-30 20:14:10 <dmc> why is this channel +s lol
2019-01-30 20:14:17 <dmc> can't find it with alis
2019-01-30 20:14:26 <mepholic> I can attest to what AinNero said
2019-01-30 20:14:29 <mepholic> :)
2019-01-30 20:14:32 <mepholic> dmc: good q
2019-01-30 20:14:57 <dmc> luckily the topic for ##alpine has "This is not #alpine-linux" :p
2019-01-30 20:29:12 <rafalcpp> fun tidbit, so by advice of #openbsd I will go with grsec (custom patch), and also maybe openbsd.  potentially seL4 if I have too much free time.  (for my security oriented network facing servers)
2019-01-30 20:29:38 <rafalcpp> actually, using custom kernel on Alpine could be a good idea?   Is alpine anyway suited to running internet servers?
2019-01-30 20:31:53 <trfl> very much so, alpine has a smaller attack surface than many other distros and is less likely to break randomly
2019-01-30 20:38:20 <trfl> unrelated question: after upgrading from 3.8 to 3.9, I have some packages that didn't upgrade such as glib-networking (on 2.54.1-r1, 2.56.1-r0 is in repos) and iso-codes (3.79-r0 vs 4.1-r0). Both of these packages were pulled in as dependencies and are not present in /etc/apk/world. How do I figure out what's holding them back?
2019-01-30 21:20:58 <trfl> interestingly, when trying to `apk del iso-codes` it says it can't be removed due to "libxklavier xfce4-settings xfce4", all of which are the latest versions available in main or community
2019-01-30 21:33:23 <gdh> When making a package with a custom check() function, is it necessary to have "options=!check"?
2019-01-30 21:34:04 <trfl> nope! and you'd get a warning either way if apk thinks it is :>
2019-01-30 21:34:35 <gdh> trfl, thank you.
2019-01-30 22:17:25 <dalias> my py3-qt5 package is not providing the PyQt5.sip module:
2019-01-30 22:17:26 <dalias> ModuleNotFoundError: No module named 'PyQt5.sip'
2019-01-30 22:17:30 <dalias> is this expected?
2019-01-31 04:53:54 <KH405_TV> Quick question for yall, how do I upgrade to 3.9 ?
2019-01-31 05:11:32 <KH405_TV> http://dl-cdn.alpinelinux.org/alpine/edge/main
2019-01-31 05:11:32 <KH405_TV> http://dl-cdn.alpinelinux.org/alpine/edge/community
2019-01-31 05:11:32 <KH405_TV> http://dl-cdn.alpinelinux.org/alpine/edge/testing those are my repos and I'm not upgrading to 3
2019-01-31 05:11:36 <KH405_TV> .9
2019-01-31 05:15:28 <dalias> ncopa, i think this commit was buggy: https://git.alpinelinux.org/aports/commit/community/py3-qt5/APKBUILD?id=4e939e86bb7d8950cf8fe32635a22799e8856461
2019-01-31 05:15:44 <dalias> supposedly there's new stuff you have to do builting pyqt5 now that sip was moved out of pyqt
2019-01-31 05:15:56 <dalias> http://pyqt.sourceforge.net/Docs/PyQt5/installation.html#downloading-sip
2019-01-31 08:49:09 <mps> KH405_TV: instead of edge in url put v3.9, http://dl-cdn.alpinelinux.org/alpine/v3.9/main
2019-01-31 09:06:25 <otaku42> moin
2019-01-31 09:07:45 <otaku42> what is the best way to investigate when and why the layout of a set of packages has been changed?
2019-01-31 09:07:55 <otaku42> i've just upgraded from v3.8 to v3.9, and hit the wall because my lxc containers fail to start. the reason for that is a failure in parsing configuration, because a per-container-included common configuration file is no longer available.
2019-01-31 09:08:01 <otaku42> from what i was able to find out so far it seems that the lxc package contents have been reorganised, and some files have been moved to a new package now
2019-01-31 09:08:07 <otaku42> the file in question, alpine.common.conf, has been moved from package lxc to package lxc-templates-legacy-alpine, which for example makes me wonder: why "legacy"? does that mean this stuff will go away soon or something?
2019-01-31 09:22:48 <pax_rhos> ncopa: nope, it doesn't :(
2019-01-31 09:27:40 <clandmeter> KH405_TV: when changing version dont forget to use --available
2019-01-31 09:40:11 <steven> eh just saw the changelogs for 3.9, why did you guys move from libressl back to openssl?
2019-01-31 09:44:08 <_ikke_> libressl broke compatibility
2019-01-31 09:44:25 <_ikke_> and many of the reasons to switch away from openssl have been resolved
2019-01-31 09:45:13 <steven> ah ok
2019-01-31 09:45:16 <steven> makes sense I guess
2019-01-31 09:45:24 <steven> also two weeks in, still waitin _ikke_ :D
2019-01-31 10:20:43 <otaku42> does anyone in here use lxc on alpine, and ran into issues where containers would fail to start after upgrading from alpine 3.8 to 3.9?
2019-01-31 10:21:30 <clandmeter> otaku42: yes its because of new config format
2019-01-31 10:21:42 <clandmeter> "probably"
2019-01-31 10:22:01 <clandmeter> otaku42: alpine containers?
2019-01-31 10:22:13 <otaku42> clandmeter: the config format issue has been sorted out already.
2019-01-31 10:22:59 <otaku42> clandmeter: what i'm facing now is something along the lines of: cgfsng - cgroups/cgfsng.c:__cg_legacy_setup_limits:2410 - Failed to set "devices.deny" to "a"
2019-01-31 10:23:52 <danieli> that doesn't sound right
2019-01-31 10:24:18 <otaku42> clandmeter: that seems to stem from the "lxc.cgroup.devices.deny = a" line in /usr/share/lxc/config/common.conf
2019-01-31 10:24:40 <otaku42> danieli: i agree, but i failed so far to fix this issue by myself.
2019-01-31 10:24:48 <clandmeter> otaku42: i have only done basic testing with lxc
2019-01-31 10:25:02 <clandmeter> so it could be broken in some setups
2019-01-31 10:25:14 <clandmeter> v2 to v3 is a major change.
2019-01-31 10:25:24 <otaku42> clandmeter: and yes, all containers are alpine containers
2019-01-31 10:25:49 <clandmeter> did you add something special in your container config?
2019-01-31 10:26:02 <otaku42> clandmeter: they have been upgraded to v3.9 at the same time as the host has. since 3.9 comes with a new kernel, i have rebooted the host, and that is when the fun started.
2019-01-31 10:26:24 <clandmeter> otaku42: i did the same at home yesterday
2019-01-31 10:26:29 <clandmeter> i think they all run
2019-01-31 10:27:18 <otaku42> clandmeter: nothing special in the config, as far as i can judge. do you guys prefer a special pastebin service?
2019-01-31 10:27:23 <clandmeter> one item of attention is that you need lxc-templates-legacy-alpine
2019-01-31 10:27:35 <clandmeter> but i guess you already added that.
2019-01-31 10:27:54 <otaku42> clandmeter: yes, i found that out already - see my question above (about 1.5h ago)
2019-01-31 10:29:29 <clandmeter> otaku42: you can use whatever pastbin, we have tpaste.us in alpine
2019-01-31 10:29:32 <clandmeter> its our own service
2019-01-31 10:30:29 <clandmeter> otaku42: you have this problem with all containers?
2019-01-31 10:50:20 <otaku42> clandmeter: sorry, was on the phone (doing the debugging besides my office work)
2019-01-31 10:53:32 <otaku42> clandmeter: yes, the issue seems to occur with all containers. in fact i wondered why autostart wouldn't work anymore, which is configured for all containers. i've checked that for one container so far (with the result mentioned above), but guess that the reason is the same for all other containers. didn't double-check that yet, though.
2019-01-31 10:57:42 <otaku42> clandmeter: http://tpaste.us/0LmK
2019-01-31 10:59:57 <otaku42> clandmeter: the lxc-alpine-o42 template is derived from lxc-alpine and takes care of stuff like automatic adress assignment, dns name registration and such stuff. i wouldn't expect the cause of this issue there.
2019-01-31 11:03:14 <clandmeter> otaku42: what happens if you create a a n ew container? can you start that one?
2019-01-31 11:07:46 <otaku42> clandmeter: will try now, but i guess it will fail, too.
2019-01-31 11:08:20 <clandmeter> lxc-create c1-alpine -t alpine
2019-01-31 11:10:59 <otaku42> clandmeter: i have tried both, my custom template and as per your snippet the standard alpine template. same results: won't start.
2019-01-31 11:11:42 <clandmeter> can you start it with foreground arch and paste me the output?
2019-01-31 11:16:34 <otaku42> clandmeter: http://tpaste.us/aRBm
2019-01-31 11:17:27 <clandmeter> cgroup service is started?
2019-01-31 11:20:49 <otaku42> clandmeter: "rc-service cgroups start" says "WARNING: cgroups has already been started"
2019-01-31 11:23:45 <clandmeter> otaku42: tpaste your mounts please
2019-01-31 11:24:26 <otaku42> clandmeter: http://tpaste.us/qKkO
2019-01-31 11:24:38 <clandmeter> hmm
2019-01-31 11:25:00 <clandmeter> can you restart cgroups?
2019-01-31 11:25:41 <clandmeter> hmm thats actually looking good
2019-01-31 11:26:37 <otaku42> clandmeter: restarting cgroups fails, because "service lxc failed to stop"
2019-01-31 11:27:05 <clandmeter> its something cgroup related
2019-01-31 11:27:21 <clandmeter> can you stop lxc?
2019-01-31 11:27:42 <clandmeter> and verify no containers are running
2019-01-31 11:27:57 <otaku42> clandmeter: nope, manually running "rc-service lxc stop" fails.
2019-01-31 11:28:10 <clandmeter> what does lxc-ls -f report?
2019-01-31 11:28:15 <clandmeter> any are running?
2019-01-31 11:28:15 <otaku42> clandmeter: without further explanation, that is.
2019-01-31 11:28:38 <otaku42> clandmeter: nope, all containers are in state "STOPPED"
2019-01-31 11:28:47 <clandmeter> zap the lxc service
2019-01-31 11:29:19 <clandmeter> btw, when you upgraded did you verify all apk new files in etc/?
2019-01-31 11:30:24 <otaku42> clandmeter: http://tpaste.us/NK4l
2019-01-31 11:32:15 <clandmeter> I have a lot more cgroups mounted
2019-01-31 11:32:22 <clandmeter> you are missing them for some reason
2019-01-31 11:32:29 <clandmeter> thats why it complains
2019-01-31 11:32:46 <clandmeter> this is related: https://github.com/lxc/lxc/issues/2268
2019-01-31 11:33:42 <clandmeter> update-conf doesnt show anything?
2019-01-31 11:34:21 <clandmeter> i need to go soon
2019-01-31 11:35:17 <clandmeter> also make sure you upgraded correctly ie use --available when switching to another alpine version. (probably not related).
2019-01-31 12:20:01 <otaku42> clandmeter: didn't check update-conf by now (didn't know that it exists even) - i'm coming from a debian background and i'm used to having a in-upgrade-notice of configuration changes.
2019-01-31 12:20:17 <otaku42> clandmeter: so thanks for the hint, will look at it now
2019-01-31 12:29:32 <otaku42> clandmeter: update-conf showed quite some files, but none of the suggested changes appear to be related. in fact, the cgroups entry in /etc/mounts would have been removed rather than something along these lines been added.
2019-01-31 12:46:17 <otaku42> clandmeter: aha. "apk upgrade --available" did something: http://tpaste.us/YMdQ
2019-01-31 12:47:57 <otaku42> clandmeter: because of the mentioning of kmod i've decided to reboot the box afterwards. now it's not coming up anymore, and since it's remote i can't check that until i return home in a few hours.
2019-01-31 12:48:10 <otaku42> clandmeter: so: ... to be continued ... :-)
2019-01-31 12:48:27 <otaku42> clandmeter: thanks for your help so far
2019-01-31 15:40:00 <trfl> I'm still drawing blanks on the alpine packages that refused to upgrade from 3.8 to 3.9, in case anyone has any ideas
2019-01-31 15:40:23 <trfl> tried an apk fix with all the packages in question (pcre pcre-dev libpcrecpp libpcre32 libpcre16 iso-codes glib-networking) and for all 7 it just prints "APK unavailable, skipped"
2019-01-31 15:41:09 <trfl> and I still don't know why I have these installed in the first place, since they were pulled in as dependencies... but idk by what :p
2019-01-31 15:55:58 <trfl> ...wait, "apk fix -u iso-codes" actually upgraded it! but "apk upgrade -a" didn't... did I break something just now?
2019-01-31 15:56:37 <trfl> (it was stuck on 3.79-r0, now it's on 4.1-r0)
2019-01-31 16:52:51 <vejetaryenvampir> pickfire:
2019-01-31 16:53:18 <vejetaryenvampir> I finally found. :)
2019-01-31 17:47:11 <kahiru> hey, congrats on the release
2019-01-31 18:08:51 <mps> anyone experienced cupsd stuck during boot after upgrade to v3.9
2019-01-31 18:09:10 <AinNero> it might affect multiple services
2019-01-31 18:10:33 <mps> strange for cups, do it needs good enough entropy
2019-01-31 18:11:13 <mps> I have small lag with chrony, but cups pauses boot about two minutes
2019-01-31 18:46:24 <steinex> hi, is znc missing from 3.9.0 repositories?
2019-01-31 18:55:50 <mpmc> steinex: https://pkgs.alpinelinux.org/packages?name=znc&branch=v3.9
2019-01-31 19:03:54 <steinex> got it. thanks.
2019-01-31 19:12:03 <dalias> fcolista & ncopa, would something like this be acceptable? http://ix.io/1zEp .. also, the local copy of sip needs to be installed as PyQt5.sip for pyqt5 to work; right now it's broken. i just did it manually. i'm not sure how it should be done; for compat with old package, sip source could be added as a source file for this package and built with it. alternately PyQt5.sip could be made a separate package this one depends on
2019-01-31 19:12:43 <dalias> there may be other qt subpackages that should be added here too; these are just the ones i hit trying to use python software (Cura)
2019-01-31 19:17:25 <KH405_TV> I understand that I could put 3.9 in my repos, my question was more like isn't edge always supposed to be latest version ?
2019-01-31 19:29:42 <Ai9zO5AP> KH405_TV: but not stable 100%
2019-01-31 19:29:56 <Ai9zO5AP> If I understand the Philo of alpine
2019-01-31 19:31:34 <maldridge> I'm in the process of QA'ing 3.9 before upgrading the fleet at $work.  I'm noticing that starting crond seems to take a really long time now.  Is there something I should be aware of with that?
2019-01-31 19:31:55 <clandmeter> KH405_TV: they are not the same
2019-01-31 19:32:03 <clandmeter> maldridge: entropy
2019-01-31 19:32:53 <maldridge> for cron?
2019-01-31 19:33:36 <clandmeter> probably
2019-01-31 19:33:45 <maldridge> what does cron need that much entropy for?
2019-01-31 19:33:57 <clandmeter> you have so much?
2019-01-31 19:34:10 <maldridge> no, this is a vm
2019-01-31 19:34:20 <clandmeter> https://tpaste.us/vbKM
2019-01-31 19:34:21 <maldridge> but cron shouldn't need *any* entropy
2019-01-31 19:34:44 <jwh> are you sure its cron and not something else?
2019-01-31 19:34:46 <clandmeter> try set that to on
2019-01-31 19:34:49 <clandmeter> see if it helps
2019-01-31 19:34:51 <Frenck> <(*_*)> Well hello there!
2019-01-31 19:35:03 <maldridge> jwh: the last item in the startup is cron
2019-01-31 19:35:26 <maldridge> I can reboot it and see about capturing logs
2019-01-31 19:35:27 <jwh> openrc got me on a similar issue, looked like the previous service
2019-01-31 19:35:27 <Frenck> Just a small question, I just noticed the new docker images are available for 3.9 via the official Docker library. Although, I noticed one of the main new features of 3.9 is missing on Docker... ARMv7 support....
2019-01-31 19:35:46 <jwh> not enough entropy on boot for sshd
2019-01-31 19:35:48 <Frenck> digging into the issues, I see the minirootfs isn't build for armv7, is there any specific reason for it?
2019-01-31 19:36:02 <maldridge> ah, so crond appears to have come up
2019-01-31 19:36:11 <clandmeter> maldridge: try setting that option. if it fixes it, its entropy
2019-01-31 19:36:27 <clandmeter> we have seen many reports last days
2019-01-31 19:36:30 <jwh> that will work if you have aesni though right?
2019-01-31 19:36:30 <mps> Frenck: armv7 is not signed yet
2019-01-31 19:36:40 <jwh> +only
2019-01-31 19:36:50 <maldridge> hmmm, this looks like it may be vboxsf failing to initialize
2019-01-31 19:36:54 <maldridge> that's the last error in the log
2019-01-31 19:37:17 <jwh> it's a pain in the ass, there just isn't enough entropy even on real hardware
2019-01-31 19:37:19 <maldridge> I just find the entropy argument a bit weak, the vm host has hardware rngs filling the pool constantly
2019-01-31 19:37:42 <jwh> that why you need to use the option clandmeter suggested
2019-01-31 19:37:47 <jwh> the kernel won't use the cpu rng
2019-01-31 19:37:56 <jwh> tin foil hats, nsa etc
2019-01-31 19:37:59 <KH405_TV> clandmeter: can you explain ?
2019-01-31 19:38:18 <maldridge> I think I'll troubleshoot the vboxsf failure first, as that for sure wasn't an issue in 3.8.2
2019-01-31 19:38:20 <KH405_TV> I tough edge was kinda like a ''beta''
2019-01-31 19:38:32 <jwh> good shout
2019-01-31 19:38:37 <clandmeter> yes, so they are not the same
2019-01-31 19:38:37 <Frenck> mps: Thanks
2019-01-31 19:38:54 <jwh> fyi I have exactly the same problem wrt entropy on my VMs as well as physical hardware, that kernel option helps where I have aesni
2019-01-31 19:39:08 <clandmeter> stable has a fixed version and only receives sec updated and important fixes.
2019-01-31 19:39:31 <KH405_TV> What about edge?
2019-01-31 19:39:39 <clandmeter> its always changing
2019-01-31 19:39:45 <clandmeter> for the good and the...
2019-01-31 19:40:03 <KH405_TV> My point is, I've got shit that is only available on edge brach, any way I can update to 3.9 ?
2019-01-31 19:40:23 <maldridge> https://imgur.com/a/xhd91nH
2019-01-31 19:40:26 <clandmeter> that makes no sense to me.
2019-01-31 19:40:31 <jwh> you can use specific packages from edge whilst staying on 3.9
2019-01-31 19:40:41 <jwh> but it'll probably break :d
2019-01-31 19:41:16 <KH405_TV> Well I mean, why edge isn't on 3.9?
2019-01-31 19:41:17 <jwh> maldridge: not rebuilt?
2019-01-31 19:41:22 <maldridge> I'm not an alpine expert by any means, but that looks like a bork'd package to me
2019-01-31 19:41:24 <jwh> because its edge not 3.9
2019-01-31 19:41:31 <KH405_TV> What is edge then ?
2019-01-31 19:41:39 <jwh> "master"
2019-01-31 19:41:41 <jwh> if you wil
2019-01-31 19:41:43 <maldridge> KH405_TV: literally the edge of the tree
2019-01-31 19:41:51 <clandmeter> i think you need stuff from testing repo?
2019-01-31 19:41:59 <maldridge> its as far forward as you can go and not be on development branches
2019-01-31 19:42:23 <clandmeter> or what are you missing in stable?
2019-01-31 19:42:36 <clandmeter> maldridge: ive seen that issue i think before.
2019-01-31 19:42:48 <maldridge> I think I just found it
2019-01-31 19:42:59 <maldridge> there's still some 3.8's scattered in build pipeline caches
2019-01-31 19:43:18 <KH405_TV> I'm dumb, i don't get it
2019-01-31 19:43:29 <maldridge> KH405_TV: do you know debian?
2019-01-31 19:43:32 <KH405_TV> Yeah
2019-01-31 19:43:37 <maldridge> its testing
2019-01-31 19:44:01 <maldridge> well, at least it would ahve been the last time I ran testing, I think they refer to the rolling one as "unstable" now
2019-01-31 19:44:19 <mps> maldridge: edge is more like debian unstable
2019-01-31 19:44:33 <clandmeter> KH405_TV: which package are you missing?
2019-01-31 19:44:33 <maldridge> mps: fair, I've been out of the debian world for quite some time now
2019-01-31 19:44:58 <KH405_TV> apk-autoupdate and something from deluge, I don't remember which one
2019-01-31 19:45:31 <clandmeter> yes its from testing
2019-01-31 19:45:37 <clandmeter> testing is never in stable
2019-01-31 19:45:43 <clandmeter> its called testing for a reason
2019-01-31 19:46:02 <clandmeter> you can pin repositories and mix things
2019-01-31 19:46:08 <clandmeter> but you will have no warranty
2019-01-31 19:46:32 <KH405_TV> Yeah, I get that, just weird that the ''testing'' is not the latest version available
2019-01-31 19:46:47 <jwh> what
2019-01-31 19:46:49 <maldridge> why would it be?
2019-01-31 19:46:52 <clandmeter> sorry i dont understnad what you are saying
2019-01-31 19:47:04 <maldridge> testing isn't testing for the software, its testing for alpine
2019-01-31 19:47:34 <clandmeter> testing is in edge, its. edge is rolling so changes apply all the time.
2019-01-31 19:47:55 <clandmeter> my typing sucks
2019-01-31 19:47:58 <KH405_TV> Yeah, i get that, anyway, my point is: Is edge ever gonna be 3.9 ?
2019-01-31 19:48:04 <jwh> whaaaat
2019-01-31 19:48:12 <jwh> edge is already >3.9
2019-01-31 19:48:16 <maldridge> if you want a technical answer then yes
2019-01-31 19:48:19 <maldridge> it was several weeks ago
2019-01-31 19:48:21 <maldridge> you just missed it
2019-01-31 19:48:36 <clandmeter> it will be 3.10 in few months.
2019-01-31 19:48:40 <KH405_TV> Okay, then something is wrong
2019-01-31 19:48:45 <KH405_TV> I'm still on 3.8.1
2019-01-31 19:48:48 <KH405_TV> And on edge
2019-01-31 19:48:51 <jwh> ...
2019-01-31 19:48:52 <maldridge> 
ACTION recalls the scene from spaceballs

2019-01-31 19:49:01 <jwh> hm i'm getting hungry
2019-01-31 19:49:11 <clandmeter> you are mixing things here
2019-01-31 19:49:20 <clandmeter> you have repository lines
2019-01-31 19:49:26 <clandmeter> i think thats what you mean?
2019-01-31 19:49:55 <clandmeter> the content of /etc/apk/repositories
2019-01-31 19:50:06 <KH405_TV> cat /etc/apk/repositories
2019-01-31 19:50:06 <KH405_TV> http://dl-cdn.alpinelinux.org/alpine/edge/main
2019-01-31 19:50:06 <KH405_TV> http://dl-cdn.alpinelinux.org/alpine/edge/community
2019-01-31 19:50:06 <KH405_TV> http://dl-cdn.alpinelinux.org/alpine/edge/testing
2019-01-31 19:50:18 <clandmeter> so you are on edge
2019-01-31 19:50:22 <KH405_TV> Yeah
2019-01-31 19:50:33 <KH405_TV> cat /etc/alpine-release
2019-01-31 19:50:33 <KH405_TV> 3.8.1
2019-01-31 19:51:00 <jwh> tinkerbell:~# cat /etc/alpine-release
2019-01-31 19:51:00 <jwh> 3.9.0
2019-01-31 19:51:02 <jwh> did you upgrade? :D
2019-01-31 19:51:06 <KH405_TV> YEs
2019-01-31 19:51:15 <clandmeter> did you do: apk upgrade -a?
2019-01-31 19:51:23 <clandmeter> or better: apk -U upgrade -a
2019-01-31 19:51:42 <clandmeter> its like a dist-upgrade
2019-01-31 19:51:53 <KH405_TV> Still nothing
2019-01-31 19:52:21 <KH405_TV> I get those lines v3.9.0-30-g9581ca6cee [http://dl-cdn.alpinelinux.org/alpine/edge/main]
2019-01-31 19:52:21 <KH405_TV> v3.9.0-31-g165a574d78 [http://dl-cdn.alpinelinux.org/alpine/edge/community]
2019-01-31 19:52:21 <KH405_TV> v3.9.0-28-gdb71657811 [http://dl-cdn.alpinelinux.org/alpine/edge/testing]
2019-01-31 19:52:35 <KH405_TV> But it ain't upgrading ...
2019-01-31 19:53:08 <clandmeter> apk info -W /etc/alpine-release
2019-01-31 19:53:12 <clandmeter> what does it say?
2019-01-31 19:53:46 <KH405_TV> ERROR: /etc/alpine-release: Could not find owner package
2019-01-31 19:54:07 <clandmeter> what kind of install is this?
2019-01-31 19:55:14 <KH405_TV> What do you mean?
2019-01-31 19:55:28 <clandmeter> how did you install this system?
2019-01-31 19:56:23 <KH405_TV> It's a container running on Proxmox
2019-01-31 19:56:41 <clandmeter> can you "pastebin" the contents of /etc/apk/world ?
2019-01-31 19:57:00 <jwh> KH405_TV: using the turnkey/whatever templates?
2019-01-31 19:57:18 <clandmeter> its probably generated without the use of apk
2019-01-31 19:57:18 <KH405_TV> https://pastebin.com/VJZnXE9r
2019-01-31 19:57:31 <jwh> clandmeter: yeah I was just checking since I have proxmox hosts too
2019-01-31 19:57:49 <clandmeter> can you add alpine-base?
2019-01-31 19:57:59 <clandmeter> apk add alpine-base
2019-01-31 19:58:22 <jwh> mine are ok btw
2019-01-31 19:58:34 <jwh> using the proxmox templates
2019-01-31 19:58:44 <clandmeter> i have no idea what users use to make them.
2019-01-31 19:58:50 <clandmeter> containers are always trouble :)
2019-01-31 19:59:06 <jwh> they're lxc so likely adapted from those
2019-01-31 19:59:18 <KH405_TV> Done
2019-01-31 19:59:35 <clandmeter> so what is in your release file now?
2019-01-31 19:59:40 <jwh> clandmeter: mine work ok :D
2019-01-31 20:00:14 <KH405_TV> 3.8.1
2019-01-31 20:00:25 <KH405_TV> Did update upgrade, still 3.8.1
2019-01-31 20:01:05 <clandmeter> and you add alpine-base?
2019-01-31 20:01:09 <KH405_TV> Yes
2019-01-31 20:01:30 <clandmeter> apk info alpine-base
2019-01-31 20:01:30 <KH405_TV> (4/4) Installing alpine-base (3.9.0-r0)
2019-01-31 20:01:34 <clandmeter> which version does it show?
2019-01-31 20:01:53 <KH405_TV> alpine-base-3.9.0-r0 description
2019-01-31 20:01:57 <clandmeter> ok so the file should have been updated
2019-01-31 20:02:10 <clandmeter> its part of alpine-base
2019-01-31 20:02:42 <KH405_TV> Weird
2019-01-31 20:02:46 <clandmeter> anyway, i dont know how your container is setup, so i cant really help you with that.
2019-01-31 20:02:57 <clandmeter> if apk is happy and  you set repofile to edge, it should be edge.
2019-01-31 20:03:17 <KH405_TV> can I generate a new file ?
2019-01-31 20:03:35 <clandmeter> apk info -L alpine-base
2019-01-31 20:03:39 <clandmeter> does it list it?
2019-01-31 20:04:24 <clandmeter> btw did you modify it?
2019-01-31 20:04:27 <KH405_TV> No
2019-01-31 20:04:31 <KH405_TV> And yes it list it
2019-01-31 20:04:46 <clandmeter> is there a similr file next to it?
2019-01-31 20:04:54 <clandmeter> apk-new
2019-01-31 20:05:23 <clandmeter> etc/alpine-release.apk-new
2019-01-31 20:05:55 <KH405_TV> Yes
2019-01-31 20:06:04 <KH405_TV> 3.9
2019-01-31 20:06:07 <clandmeter> run update-conf
2019-01-31 20:06:33 <KH405_TV> use new ?
2019-01-31 20:06:42 <clandmeter> for this file yes
2019-01-31 20:06:47 <clandmeter> its not installed by apk
2019-01-31 20:06:52 <clandmeter> so it has no trace of it
2019-01-31 20:06:55 <clandmeter> and will not overwrite it
2019-01-31 20:07:07 <clandmeter> nothing in etc will be overwriteen
2019-01-31 20:07:42 <clandmeter> ok i have to move on, good luck.
2019-01-31 20:08:10 <KH405_TV> Thanks alot!
2019-01-31 20:11:09 <clandmeter> your welcome :)
2019-01-31 21:31:05 <amarks> Having clock trouble. hwclock --localtime shows correct time, hwclock --utc shows incorrect time, date shows same as hwclock --utc (claims to be PST), /etc/timezone and /etc/localtime are set appropriately, system is a VM
2019-01-31 21:31:29 <amarks> host time is correct with hw time set to UTC on host
2019-01-31 21:33:15 <amarks> Pacific Standard Time should be UTC-8, but system shows UTC-4 as current time
2019-01-31 21:40:30 <AinNero> could you post output of `date -R` and `readlink /etc/localtime` ?
2019-01-31 21:41:35 <maldridge> clandmeter: do you have docs for the random flag?  I just get unknown key when I try to apply it
2019-01-31 21:42:01 <amarks> Thu, 31 Jan 2019 05:41:36 -0800 (Date -R)
2019-01-31 21:42:10 <amarks> I didn't symlink localtime, I copied
2019-01-31 21:43:15 <AinNero> which tzdata file?
2019-01-31 21:43:30 <amarks> diff /etc/locatime /usr/share/zoneinfo/America/Los_Angeles gives nothing, files are identical
2019-01-31 21:45:36 <AinNero> ye, timezone offset looks like its applied twice
2019-01-31 21:45:43 <AinNero> whats the hypervisor?
2019-01-31 21:45:47 <amarks> bhyve
2019-01-31 21:47:06 <amarks> really weird, seconds are 13 in the future
2019-01-31 21:47:39 <AinNero> bhyve also has hwclock localtime as default
2019-01-31 21:48:00 <amarks> how can I tell alpine that system clock is localtime
2019-01-31 21:48:05 <amarks> but my timezone is PST
2019-01-31 21:48:54 <AinNero> yeah, and bhyve is setting the VM clock to host's local time
2019-01-31 21:49:00 <AinNero> and the guest thinks its utc
2019-01-31 21:49:08 <AinNero> so it applies the timezone offset again
2019-01-31 21:49:21 <AinNero> edit /etc/conf.d/hwclock
2019-01-31 21:51:45 <maldridge> I guess the better question to ask is how to make changes to the kernel cmdline persist in alpine
2019-01-31 21:52:10 <AinNero> maldridge: update-extlinux.conf in /etc, usually
2019-01-31 21:52:38 <amarks> AinNero: Thanks, fixed.
2019-01-31 21:52:59 <maldridge> I see
2019-01-31 21:53:10 <maldridge> I really just need to set that I'm willing to trust the CPU manufacturer
2019-01-31 21:53:25 <maldridge> this is not a very machine friendly interface to automate
2019-01-31 21:53:38 <AinNero> maldridge: the startup delay issue?
2019-01-31 21:54:20 <maldridge> yes
2019-01-31 21:54:20 <AinNero> you could nerf the update-extlinux.conf via overwrite=0 and manage your boot config manually
2019-01-31 21:54:39 <maldridge> right, but then that's more that my automation has to keep up with, and is inherently fragile
2019-01-31 21:55:13 <AinNero> alpine doesn't keep the version numbers in the kernel/initramfs files, so its reasonable to expect that a boot config will never change
2019-01-31 21:55:21 <AinNero> s/files/filenames/
2019-01-31 21:55:51 <maldridge> you're right that the filenames won't change, but the UUID will be different across machines
2019-01-31 21:56:39 <AinNero> how are you automting the stuff?
2019-01-31 21:56:50 <maldridge> at this layer its ansible
2019-01-31 21:57:05 <maldridge> so yes, I could get the UUID out of the machine facts and then template the file
2019-01-31 21:57:57 <AinNero> i see
2019-01-31 21:58:59 <maldridge> I may just deal with degraded performance for a while in the name of "security"
2019-01-31 22:00:00 <AinNero> a BOFH would dynamically detect the value in update-extlinux.conf
2019-01-31 22:00:07 <AinNero> since its actually just a shell script
2019-01-31 22:00:26 <maldridge> a BoFH would set root to /dev/null as the lusers don't matter
2019-01-31 22:00:44 <maldridge> a reasonable sysop would observe that the root disk doesn't move and give a device name
2019-01-31 22:02:23 <AinNero> last time we met you were an void dev, what happened?
2019-01-31 22:03:18 <maldridge> I still am
2019-01-31 22:03:31 <maldridge> but $work has different requirements that void can't easily satisfy right now
2019-01-31 22:04:42 <AinNero> i see
2019-01-31 22:04:55 <maldridge> and I think its a bit foolish to specialize down into a single distro; its always good to know at least 2
2019-01-31 22:05:45 <maldridge> in case you're curious btw, the magic word from $work is "containers"
2019-01-31 22:08:29 <maldridge> I guess I've also reviewed enough patches tagged alpine-expat at this point that I figured it was time to hop over here and take a closer look
2019-01-31 22:52:28 <allenp> x86_64 on qemu host, just updated from 3.8 to 3.9, vm hangs for several minutes at "loading iptables state and starting firewall"
2019-01-31 22:52:40 <allenp> ip6tables is fine, not sure where to look or how to diag
2019-01-31 22:53:38 <allenp> rules for v4 and v6 are identical
2019-01-31 22:54:42 <maldridge> sounds almost like you rresolver isn't available at the time the firewall may be needing dns names
2019-01-31 22:57:14 <allenp> I'm using the googs for both v4 and 6
2019-01-31 22:58:41 <allenp> I have two vm I just did the update on, both configured pretty much the same.  It eventually comes up so not a huge deal just, you know, want to figure out what is going on
2019-01-31 22:59:33 <maldridge> do you use DNS names in your rulesets?
2019-01-31 23:00:04 <allenp> negative, however I see sshguard is listed as crashed via rc-status
2019-01-31 23:00:17 <allenp> although that starts after iptables
2019-01-31 23:00:44 <maldridge> hmm, that was going to be my guess
2019-01-31 23:00:50 <maldridge> as that is enough to hang iptables
2019-01-31 23:02:56 <allenp> ok well this is odd, both VMs show sshguard as "crashed" via rc-status, but they are in fact up and running
2019-01-31 23:14:52 <allenp> well, sideways progress?  purged iptables and now it hangs at "Starting busybox crond"
2019-01-31 23:15:02 <allenp> 
ACTION flips table

2019-01-31 23:15:05 <maldridge> ah
2019-01-31 23:15:13 <maldridge> I suspect it doesn't hang at crond
2019-01-31 23:15:15 <maldridge> but at sshd
2019-01-31 23:15:20 <maldridge> I was just fighting with this
2019-01-31 23:15:47 <allenp> do tell
2019-01-31 23:16:41 <maldridge> in 3.9 you have to explicitly trust the CPU, otherwise RDRAND initialization stalls
2019-01-31 23:16:58 <maldridge> random.trust_cpu=on in the kernel cmdline if you trust the CPU manufacturer
2019-01-31 23:18:31 <allenp> I'm a linux newb, how do I "kernel cmdline" ?
2019-01-31 23:18:52 <maldridge> hit tab during boot when it offers to let you edit
2019-01-31 23:19:03 <maldridge> add that string right after "quiet" near the bottom of the screen
2019-01-31 23:22:27 <allenp> does that edit persist ?
2019-01-31 23:22:31 <maldridge> no
2019-01-31 23:22:42 <maldridge> you'd need to edit /boot/extlinux.conf to persist it
2019-01-31 23:22:51 <maldridge> or some file in /etc I haven't yet found
2019-01-31 23:23:11 <maldridge> I've instead accepted a few seconds of boot delay for now since I'm up against a few deadlines on the project where I discovered this problem
2019-01-31 23:26:25 <allenp> a few seconds would be nice, this is still several minutes
2019-01-31 23:26:34 <allenp> cpu limited perhaps?
2019-01-31 23:26:52 <maldridge> oh
2019-01-31 23:27:24 <maldridge> yeah might be dependent on a number of things if its truly entropy constrained
2019-01-31 23:28:04 <allenp> I wonder what the difference is between this host and the other
2019-01-31 23:33:39 <mps> maldridge: allenp: file is /etc/update-extlinux.conf
2019-01-31 23:34:40 <allenp> mps: maldridge thanks for the help
2019-01-31 23:34:47 <maldridge> np
2019-01-31 23:35:03 <mps> look for default_kernel_opts
2019-01-31 23:36:52 <mps> and after you set parameters run 'update-extlinux' as root, of course
2019-01-31 23:37:25 <mps> or, 'update-extlinux -v' for verbose run
2019-01-31 23:48:17 <steinex> openntpd fails the constraint check against www.google.com because of "tls connect failed: 2a00:1450:4001:808::2004 (www.google.com): ssl verify memory setup failure"
2019-01-31 23:48:25 <steinex> any ideas? ca-certificates-cacert is installed
2019-01-31 23:54:01 <steinex> seems to be the same as https://bugs.alpinelinux.org/issues/9635 - i'm using 3.9.