2022-07-02 10:12:25 <ikke> redeployed gitlab-test to prepare for the gitlab upgrade to 15.0
2022-07-02 17:48:06 <psykose> ikke: fyi riscv64 has been stuck for days
2022-07-02 18:37:24 <ikke> https://tpaste.us/X04j 
2022-07-02 18:49:17 <ikke> psykose: it's running again
2022-07-02 18:50:50 <psykose> :)
2022-07-03 12:58:35 <mps_> interesting, build of riscv64 e2fsprogs pass in lxc container but fail on builder
2022-07-03 13:07:26 <ikke> gitlab upgraded to 15.0
2022-07-03 13:57:20 <ikke> clandmeter: I cannot deploy the reserved server on equinix, as I cannot select an OS
2022-07-03 13:57:41 <ikke> os -> no results found
2022-07-03 17:38:11 <ptrc> the new gitlab upgrade shows more options to regular users, like "close merge request" or adding labels, all of them erroring out (of course)
2022-07-03 17:39:30 <psykose> the labels are funny because they aren't shown but if you click the label icon from a minimised right side panel it then pops up the ui to add them lol
2022-07-03 17:47:22 <ikke> fun
2022-07-03 17:50:09 <psykose> the diff clicking for more context is also not very functional now, as you can't click show all for the whole file
2022-07-03 17:50:16 <psykose> typical gitlab
2022-07-03 17:54:01 <ikke> Oh, there is no more 'show all context' 
2022-07-03 17:58:59 <psykose> mhm
2022-07-03 19:52:14 <psykose> ikke: are there any stray world packages on the rv64 builder? there seems to be either pcsc-lite-dev or something similar that stuck around
2022-07-03 19:54:23 <ikke> let me check
2022-07-03 19:55:14 <ikke> .makedepends-cosign=20220627.200310 
2022-07-03 19:55:49 <ikke> which indeed pulled in pcsc-lite-dev
2022-07-03 19:58:33 <psykose> haha
2022-07-03 19:58:49 <psykose> that was what it was stuck on during that frozen panic
2022-07-03 19:58:59 <psykose> so i guess if you kill them in such a state you have to double check that after
2022-07-03 19:59:13 <ikke> right
2022-07-03 19:59:26 <ikke> there was an git index.lock file left behind as well which I had to remove
2022-07-03 19:59:29 <psykose> wonder how many things accidentally built with udev support on riscv now
2022-07-03 20:00:40 <ikke> :D
2022-07-03 20:00:49 <psykose> :)
2022-07-03 20:01:22 <ikke> I just added s390x-ci to monitoring
2022-07-03 20:01:32 <ikke> and acpid is chronically crashed
2022-07-03 20:02:31 <minimal> ikke: does s390x actually have ACPI?
2022-07-03 20:02:43 <ikke> minimal: no idea
2022-07-03 20:02:48 <minimal> I don't think so
2022-07-03 20:03:02 <ikke> alpine just starts the service by default
2022-07-03 20:03:27 <minimal> right but maybe it is crashing because s390x doesn't support ACPI?
2022-07-03 20:03:38 <ikke> that's very likely
2022-07-03 20:03:44 <ikke> never looked into it
2022-07-03 20:04:11 <minimal> I seem to remember raising a MR to disable acpid for another arch (armv7?) in the past
2022-07-03 20:05:10 <minimal> ikke: https://gitlab.alpinelinux.org/alpine/aports/-/issues/13700
2022-07-03 20:05:22 <minimal> "As indicated in #12290 (closed) by myself the RPI (and indeed many/most ARM systems) do not support ACPI and so if acpid is run it will exit (aka "crash")."
2022-07-03 20:05:52 <ikke> so apparently it's already merged into alpine-conf
2022-07-03 20:06:25 <minimal> yeah to setup-alpine
2022-07-03 20:06:55 <ikke> so should be fixed for new installations, right?
2022-07-03 20:08:57 <minimal> assuming its in the version of alpine-conf shipped
2022-07-04 09:56:11 <ikke> usa5 (rv64 builder) is stuck again
2022-07-04 09:56:24 <ikke> Need to work on the new equinix infra
2022-07-04 09:57:22 <ikke> again when building cosign apparently
2022-07-05 07:10:01 <clandmeter> ikke: ok let me check with ed
2022-07-05 07:10:12 <clandmeter>  I’m still on holidays
2022-07-05 16:41:02 <ikke> secdb is pretty popular
2022-07-05 16:41:49 <ikke> https://i.imgur.com/A1NobOO.png
2022-07-05 16:42:36 <psykose> probably the multiple secfix edits over the few minutes and multiple scrapes getting triggered
2022-07-05 16:43:44 <ikke> Upgraded Zabbix to 6.2
2022-07-05 16:43:47 <ikke> bleeding edge :)
2022-07-05 16:44:23 <ikke> FYI: https://zabbix.alpinelinux.org/zabbix.php?action=dashboard.view&dashboardid=7
2022-07-05 16:47:26 <psykose> that is a lot of traffic
2022-07-05 16:49:19 <ikke> nld3 you mean?
2022-07-05 16:52:18 <psykose> yea
2022-07-05 16:54:41 <ikke> yup
2022-07-05 16:54:46 <ikke> mostly the mirror
2022-07-05 17:11:38 <clandmeter> ikke: please open a ticket for the issue and mention Ed on it
2022-07-05 17:36:42 <ikke> ok
2022-07-06 10:18:41 <ikke> clandmeter: They resolved it, but sadly I cannot select Alpine :)
2022-07-07 10:21:03 <mps> hehe https://www.phoronix.com/scan.php?page=news_item&px=Systemd-Creator-Microsoft
2022-07-07 10:21:23 <mps> finally :D
2022-07-07 10:42:52 <psykose> what does this have to do with alpine infra
2022-07-07 10:54:17 <mps> hah
2022-07-07 10:54:41 <mps> for me it have but I will not tell
2022-07-07 11:52:46 <ptrc> https://gitlab.alpinelinux.org/anabellebailey55/7qasearch.net/-/issues/1
2022-07-07 11:52:52 <ptrc> 🤔
2022-07-07 12:05:04 <ikke> P00f
2022-07-07 21:37:13 <ikke> clandmeter: _alice setup a new aarch64 gitlab runner host on the new (temporary) server we got from ARM
2022-07-07 21:37:35 <ikke> sorry, that was meant to be addressed to both clandmeter and alice
2022-07-07 21:41:16 <psykose> hurrah
2022-07-07 21:41:19 <psykose> when's the swap
2022-07-07 21:41:38 <ikke> it's already active
2022-07-07 21:41:46 <psykose> :)
2022-07-07 21:41:48 <psykose> awesome
2022-07-07 21:41:54 <psykose> will let you know if i see any issues
2022-07-07 21:42:05 <ikke> only did aarch64, other arches will follow
2022-07-07 21:57:48 <martijnbraam> actual servers from ARM?
2022-07-08 03:11:30 <clandmeter> ikke: thx
2022-07-08 04:18:32 <ikke> martijnbraam: yes, but they indicated that they will be moving away from sponsored bare-metal servers
2022-07-09 00:27:23 <ptrc> another quirk with the latest gitlab version: when trying to filter commits by author, it puts user display name into author= parameter rather than username
2022-07-09 00:29:03 <ptrc> ah, nevermind; it works properly, i just clicked on an alpine developer with seemingly 0 commits to aports
2022-07-09 07:30:51 <mps> now gitlab assigns to me MRs I create though I'm not maintainer. Is this bug or new feature?
2022-07-09 07:52:38 <ikke> psykose also mentioned it, not sure why it happens
2022-07-09 07:53:01 <psykose> you have to unselect yourself before pressing submit
2022-07-09 07:54:22 <mps> ok
2022-07-09 08:54:11 <ikke> https://gitlab.com/gitlab-org/gitlab/-/issues/287636
2022-07-09 08:54:24 <ikke> and follow-up: https://gitlab.com/gitlab-org/gitlab/-/issues/365891
2022-07-09 08:55:05 <ikke> So will eventually be solved again
2022-07-09 09:51:40 <psykose> neat
2022-07-09 09:54:05 <ikke> I use glab to create MRs, so that will leave it unassigned as well
2022-07-09 09:54:35 <ikke> although, not sure as I mostly tend to create MRs for my own packages
2022-07-09 09:55:02 <psykose> probably works
2022-07-10 04:50:06 <ikke> Hmmm. 🤔
2022-07-10 05:59:18 <ikke> For some reason the new arm CI server reboots each night
2022-07-10 17:19:10 <ikke> new CI host rebooted again :/
2022-07-10 17:21:16 <ikke> serial console does not reveal a lot either
2022-07-10 17:21:19 <ikke> it just seems to reboot
2022-07-10 17:41:52 <mps> what is console, BMC
2022-07-10 18:01:37 <ikke> I mean, I see the boot process
2022-07-10 18:01:42 <ikke> but nothing indicating why it's rebooted
2022-07-10 18:01:57 <ikke> It starts with initializing the firmare
2022-07-10 18:50:34 <mps> but no history?
2022-07-10 18:54:49 <ikke> I connected this morning, and I see everything that happened since
2022-07-10 18:55:48 <ikke> https://tpaste.us/a0XB 
2022-07-10 18:55:58 <ikke> here you see when it was booted, and then it started booting again
2022-07-10 18:57:36 <mps> hm, have no idea what this means, indicates something about RAM but ...
2022-07-10 19:15:48 <Hello71> i'll put $5 on overheating
2022-07-10 19:49:28 <ikke> Hmm, cannot find any temperature sensors in /sys/class
2022-07-10 19:51:28 <ikke> ah, found something
2022-07-10 19:53:10 <ikke> I have 2 hwmon sensors that are called SoC temperature
2022-07-10 19:53:46 <ikke> but seems to just output 50000
2022-07-10 19:56:20 <ikke> same output with `sensors`
2022-07-10 19:56:48 <ikke> but system is mostly idle now, so maybe wait for more load
2022-07-10 20:49:37 <mps> after 8 years I upgraded home server from armv7 to aarch64, it was easy with alpine ;)
2022-07-11 05:05:22 <ikke> The SoC temperature just remains constant
2022-07-11 05:05:32 <ikke> So seems like it's not working
2022-07-11 05:26:23 <mps> ikke: does this machine have IPMI interface
2022-07-11 05:43:44 <ikke> Not that I have access to 
2022-07-11 05:43:53 <ikke> It's hosted at equinix
2022-07-11 05:48:35 <mps> ah, you don't have access to host, just some VMs
2022-07-11 05:52:11 <ikke> It's bare metal 
2022-07-11 05:52:31 <ikke> But I just get access to the serial console 
2022-07-11 05:53:42 <mps> but you can login as root?
2022-07-11 06:02:46 <ikke> Yes, I have full access to the OS
2022-07-11 06:03:24 <mps> is there /dev/ipmi0 file
2022-07-11 06:06:09 <ikke> I'd have to check later 
2022-07-11 10:18:50 <ikke> Oh, I do see temperature fluctuations now
2022-07-11 10:21:39 <ikke> mps_: /dev/ipmi0 does not exist
2022-07-11 10:21:54 <ikke> oh, sorry, it does
2022-07-11 10:25:54 <ikke> checking ipmitool sel elist now
2022-07-11 10:30:01 <ikke> hmm, seems to be just rebooting..
2022-07-11 15:32:03 <paradxum> Good morning all, I tried to make a quick update to the docs at: https://wiki.alpinelinux.org/wiki/Finding_the_fastest_mirror and don't have perms. I'm happy to send the quick edit here if you like, or someone could give me perms to update it myself, eaither way is ok with me. The second script hangs in some instances currently.
2022-07-11 15:34:16 <ikke> paradxum: what is your username there?
2022-07-11 15:34:35 <paradxum> eric.schultz (sorry, handles don't match)
2022-07-11 15:34:59 <ikke> no worry
2022-07-11 15:35:19 <ikke> Can you try now?
2022-07-11 15:42:41 <paradxum> Yes, trying now.
2022-07-11 15:44:11 <paradxum> Perfect, thank you, it's all updated!
2022-07-11 15:51:47 <mps> ikke: I didn't used ipmi tool for some time and forgot exact command params to list events and errors
2022-07-11 15:52:05 <ikke> mps: I used ipmi sel elist last 5t0
2022-07-11 15:52:14 <ikke> but it stopped at the 01-05
2022-07-11 15:52:37 <mps> that could depend on BMC firmware
2022-07-11 15:53:00 <mps> different ones needs different options and parameters
2022-07-11 15:53:42 <mps> my method is/was usually 'try and see what works'
2022-07-11 15:56:20 <psykose> ikke: do you have a small guide somewhere on how to start the rust bootstrap for s390x and the ability to give me access to whatever to do it
2022-07-11 17:06:31 <ikke> psykose: from what I understood from last time from nmeum, we should start with running bootstrap.sh on x86_64
2022-07-11 17:06:39 <ikke> with s390x as target
2022-07-11 17:07:02 <nmeum> boostrap.sh will build you a cross compilation environment
2022-07-11 17:07:13 <ikke> right
2022-07-11 17:07:14 <nmeum> to cross compile from x86_64 to your desired architecture
2022-07-11 17:07:27 <ikke> Are all blockers solved?
2022-07-11 17:07:37 <nmeum> and bootstrap.sh will already take care of cross-compiling rust for you
2022-07-11 17:07:37 <psykose> they have been 'solved' on us actually doing it for almost 2 months
2022-07-11 17:07:38 <ikke> for rust on s390x?
2022-07-11 17:07:45 <ikke> ok
2022-07-11 17:08:29 <nmeum> the problem with boostrap.sh is just that cross-compiling of packages breaks regularly so one always needs to fix things along the way ':D
2022-07-11 17:08:41 <psykose> currently it's still the meson things
2022-07-11 17:08:42 <ikke> heh, right
2022-07-11 17:08:46 <nmeum> or disable packages which are not needed for rust
2022-07-11 17:08:47 <psykose> good opportunity for me to fix it as i go
2022-07-11 17:08:48 <psykose> heh
2022-07-11 17:17:59 <ikke> psykose: will you give it a try, or should I?
2022-07-11 17:25:13 <psykose> i suppose i will, from tomorrow when i'm back home, unless you want to snipe yourself into doing it :p
2022-07-11 17:27:34 <ikke> Would probably be tomorrow as well
2022-07-11 17:30:20 <psykose> i was going to start it last week but got sniped by some travels i had planned for a while, crazy how prescheduled dates just sneak up on you all of a sudden..
2022-07-11 17:30:42 <ikke> :D
2022-07-12 15:13:00 <panekj> Ariadne: https://codeberg.org/alpine/alpine-wiki/issues/6
2022-07-12 15:13:15 <Ariadne> i don't want to deal with this right now :D
2022-07-12 15:13:33 <panekj> I understand :)
2022-07-12 15:13:59 <Ariadne> ...those logos sure have an aliexpress feel to them though
2022-07-12 15:14:54 <panekj> it was changed from this https://codeberg.org/repo-avatars/36054-748761d616a7bf1d6849e904aee637dc
2022-07-12 15:15:06 <Ariadne> 404
2022-07-12 15:16:28 <panekj> https://usercontent.irccloud-cdn.com/file/f01QYoRM/image.png
2022-07-12 15:16:44 <panekj> don't ask me what's that, I have no idea
2022-07-12 15:23:25 <psykose> wtf did i just read
2022-07-12 19:06:22 <psykose> ikke: i seem to have crashed my x86_64 nld5-dev1 container
2022-07-12 19:06:58 <psykose> the rv64 one on the same machine works, so i guess it broke somehow from the xz -T0 i was running
2022-07-12 19:07:20 <ikke> it's still running
2022-07-12 19:07:41 <psykose> yeah, but i can't reach it
2022-07-12 19:07:44 <ikke> lots of distccd processes
2022-07-12 19:07:53 <psykose> those are always there cause it's one per thread
2022-07-12 19:08:13 <psykose> (and idle until i use them)
2022-07-12 19:08:31 <ikke> I can ping it, and ssh seems to respond
2022-07-12 19:08:48 <psykose> ssh -l demon -p 22047 nld5-dev1.alpinelinux.org # host unreachable
2022-07-12 19:08:49 <psykose> hmm
2022-07-12 19:09:28 <ikke> Oooh :/
2022-07-12 19:09:45 <ikke> that might be my fault
2022-07-12 19:10:33 <psykose> ah :)
2022-07-12 19:10:37 <psykose> thankfully i ran stuff in screen
2022-07-12 19:10:50 <ikke> for some reason, there was no dhcp entry for that address
2022-07-12 19:10:54 <ikke> so I used it for another container
2022-07-12 19:11:06 <psykose> hah
2022-07-12 19:11:11 <ikke> but for some reason it received a different address anyway
2022-07-12 19:11:38 <ikke> ok, fixed
2022-07-12 19:11:56 <psykose> fixed, thanks
2022-07-12 19:12:03 <psykose> using it to build chromium tarballs for upgrades in 3.16
2022-07-12 19:12:10 <ikke> ok
2022-07-12 19:12:22 <ikke> that machine is busy :P
2022-07-12 19:12:36 <psykose> tarball is done now
2022-07-12 19:13:01 <ikke> bootstrap is still running
2022-07-12 19:14:43 <psykose> ~
2022-07-12 19:15:36 <ikke> ok, now my container gets the expected address, but I lost connection as well (and wasn't using screen or tmux)
2022-07-12 19:15:44 <ikke> but for some reason, it continues to build :D
2022-07-12 19:15:52 <psykose> oof
2022-07-12 19:15:56 <ikke> probably until ssh notices the client is gone
2022-07-12 19:19:07 <ikke> ok, now starting it _with_ tmux :)
2022-07-12 22:57:12 <psykose> ikke: i wonder what the load/usage looks like for the new arm machine for aarch64 CI? it's roughly half the speed of the previous setup (or even less)
2022-07-13 04:24:51 <ikke> psykose: https://zabbix.alpinelinux.org/history.php?action=showgraph&itemids%5B%5D=33724
2022-07-13 04:25:13 <psykose> how many cores are there
2022-07-13 04:25:37 <ikke> In total 160, CI vm has 32
2022-07-13 04:26:13 <ikke> Looks like there are 32 cores 100% utilized on the host
2022-07-13 04:28:32 <psykose> ah, but wasn't that machine not used for anything else
2022-07-13 04:29:04 <ikke> This is the only thing running on it atm
2022-07-13 04:31:17 <ikke> The load on the VM is quite high
2022-07-13 04:31:22 <ikke> 44 atm
2022-07-13 04:32:03 <ikke> wondering if it has to do with NUMA
2022-07-13 04:38:03 <psykose> probably just extra jobs, dunno
2022-07-13 04:38:07 <ikke> but we don't have anything setup for that on usa9 either
2022-07-13 04:38:24 <ikke> it's running 2 jobs at
2022-07-13 04:38:26 <ikke> atm
2022-07-13 04:38:31 <psykose> though it's extremely slow in this configuration and chromium takes roughly 3 hours
2022-07-13 04:39:11 <ikke> rust and gcc-cross-embedded
2022-07-13 04:39:30 <ikke> Trying to figure out what's difference
2022-07-14 19:12:27 <ncopa> anyone looking at build-edge-x86?
2022-07-14 19:12:32 <ncopa> otherwise i'm gonna restart it
2022-07-14 19:12:58 <ikke> was just fixing i9t
2022-07-14 19:13:07 <ncopa> i figured :)
2022-07-14 19:13:30 <ncopa> good. thanks
2022-07-14 19:37:43 <ikke> Still reachable here
2022-07-15 05:50:44 <ncopa> got this on build-edge-aarch64: cp: can't stat '/home/buildozer/.abuild/alpine-devel@lists.alpinelinux.org-616ae350.rsa.pub': No such file or directory
2022-07-15 05:50:54 <ncopa> i copied it from /etc/apk/keys
2022-07-15 06:01:46 <ikke> 👍
2022-07-15 10:01:11 <mps> hmm `>>> ERROR: ubdsrv: devel_v4 is not a valid version` says abuild -r
2022-07-15 10:02:22 <mps> 0.20220711 could be solution?
2022-07-15 10:08:25 <mps> hmm, this could be better to wait when we upgrade linux-headers to 5.19.x
2022-07-15 10:09:34 <mps> I hope no one will rush to add ubdsrv to aports, this is my task
2022-07-15 10:10:06 <mps> https://github.com/ming1/ubdsrv
2022-07-15 10:10:12 <panekj> *speedruns making ubdsrv aport*
2022-07-15 10:10:37 <mps> panekj: what 'speedruns' means
2022-07-15 10:11:26 <panekj> speedrunning means finishing game with certain limitations as fast as possible
2022-07-15 10:11:55 <mps> aha
2022-07-15 10:12:10 <mps> it was intended as joke for me?
2022-07-15 10:12:23 <panekj> yes
2022-07-15 10:12:34 <mps> nice, thanks :)
2022-07-15 10:12:39 <panekj> https://en.wikipedia.org/wiki/Speedrun
2022-07-15 10:12:40 <alpine-meetbot> [WIKIPEDIA] Speedrun | "A speedrun is a playthrough of a video game, or section of a video game, with the goal of completing it as fast as possible. Speedruns often follow planned routes, which may incorporate sequence breaking, and might exploit glitches that allow sections to be skipped or completed more quickly than intended..."
2022-07-15 10:14:40 <mps> panekj: for some reason your irc client didn't sent *bold* text to me
2022-07-15 10:15:26 <panekj> it was supposed to be cursive, but my barely awake brain forgot about /me
2022-07-15 10:16:07 <mps> aha, fine. my irssi set bold with '*'
2022-07-15 10:17:13 <panekj> yep, I'm poisoned with markdown syntax
2022-07-15 10:17:41 <mps> I see ;)
2022-07-15 10:31:23 <ncopa> idea for build.a.o status page: change favicon by status. if all builders are idle: add a green "ok" emblem, if builders are busy: add a yellow "busy"/"working" emblem, if build failure add red 'failure' emblem.
2022-07-17 16:32:49 <psykose> ikke: the x86 CI seems roughly 5x slower than usual for the past few days
2022-07-17 16:38:17 <ikke> Hmm, interesting. Will take a look 
2022-07-17 16:52:41 <ikke> psykose: nothing suspicious going on
2022-07-17 16:52:47 <psykose> strange
2022-07-17 16:53:02 <ikke> load is 0 atm
2022-07-17 17:01:50 <psykose> unrelatedly, could you remind me of the thread/ram count per ci arch and builder arch
2022-07-17 17:04:55 <ikke> x86_64-ci 48 cores 64G mem
2022-07-17 17:05:05 <ikke> same of x86
2022-07-17 17:05:27 <ikke> arm* is 32 cores / 64G mem
2022-07-17 17:05:31 <ikke> (ci)
2022-07-17 17:05:58 <ikke> ppc64le ci is 64 cores / 128G mem
2022-07-17 17:06:09 <psykose> arm* includes aarch64?
2022-07-17 17:06:11 <ikke> yes
2022-07-17 17:06:21 <psykose> okay, makes sense for some of the oom stuff, now builders..
2022-07-17 17:06:36 <ikke> s390x-ci is 8 cores / 32G
2022-07-17 17:06:59 <ikke> x86_64 builder is 32 cores / 64G
2022-07-17 17:07:06 <ikke> some for x86
2022-07-17 17:07:30 <psykose> the issue with some lto ooms for x86_64 on ci is that it's not 2G/thread, but i'm wondering why it now fails on aarch64-build with 6 jobs..
2022-07-17 17:07:59 <psykose> that also explains why the x86_64 builder is slower than ci hah, 32 vs 48
2022-07-17 17:08:28 <ikke> arm*/aarch64 host is 80 cores / 256G memory
2022-07-17 17:09:02 <ikke> ppc64le is 64 cores / 128G
2022-07-17 17:09:18 <ikke> s390x 8 cores / 32G
2022-07-17 17:09:45 <psykose> how are the cores/memory shared per build container?
2022-07-17 17:10:21 <ikke> each container can use everything available
2022-07-17 17:10:34 <ikke> (on aarch / arm we limit containers to a single NUMA domain)
2022-07-17 17:10:35 <psykose> ah, that's why they're so fast normally
2022-07-17 17:10:44 <psykose> ok, makes sense
2022-07-17 17:10:49 <psykose> no idea why the hell 6 jobs oom then
2022-07-17 17:11:17 <ikke> with firefox?
2022-07-17 17:11:20 <psykose> yes
2022-07-17 17:11:49 <psykose> on ci it's the opposite and x86_64 fails without limits; since it's not 2G per thread, but with the 6 cap it's a very large number per thread so i would guess it should always be fine..
2022-07-17 17:12:13 <psykose> it's one of those weird things where they oom without using any memory but the per-thread limits still matter
2022-07-17 17:12:47 <ikke> well, one thing to take into account is that the CI vms take memory as well
2022-07-17 17:12:56 <ikke> they're running on the same host
2022-07-17 17:13:31 <psykose> it's not about the used memory funnily :) if you watched a full graph it would never go above 20% and oom
2022-07-17 17:14:31 <psykose> well, hopefully 4 works, otherwise i have no clue why this stuff is so broken
2022-07-17 17:15:12 <ikke> the aarch64 ci vm got OOM'd
2022-07-17 17:15:52 <psykose> oh, huh
2022-07-17 17:15:58 <psykose> what was the actual usage
2022-07-17 17:16:30 <ikke> This is what zabbix shows: https://zabbix.alpinelinux.org/history.php?action=showgraph&itemids%5B%5D=32593
2022-07-17 17:17:32 <ikke> oh, that was already a while ago
2022-07-17 17:17:34 <ikke> btw
2022-07-17 17:17:43 <ikke> No recen't OOMs
2022-07-17 17:17:59 <psykose> is used actually used or '''used''' from free()
2022-07-17 17:18:29 <ikke> It's more than free reports
2022-07-17 17:18:52 <ikke> probably used + caches and buffers
2022-07-17 17:18:54 <psykose> right, meant it backwards, as in inverse of 'free'
2022-07-17 17:18:54 <psykose> yeah
2022-07-17 17:19:09 <psykose> so not actual used
2022-07-18 10:26:22 <ncopa> ikke: do you know if the build-edge-a* containers are on same numa domain as build-3-16-a*?
2022-07-18 10:26:35 <ikke> let me check
2022-07-18 10:26:43 <ncopa> how is numa domain selected?
2022-07-18 10:26:50 <ikke> In the lxc config
2022-07-18 10:26:52 <ikke> it's per arch
2022-07-18 10:27:03 <ikke> so all aarch64 will be on the same NUMA domain for example
2022-07-18 10:27:23 <ikke> /etc/lxc/aarch64.common.conf
2022-07-18 10:27:27 <ikke> same for other arches
2022-07-18 10:28:01 <ikke> aarch64 and armhf share the same NUMA domain
2022-07-18 10:33:23 <ncopa> ok. makes sense
2022-07-18 10:48:06 <ncopa> im thinking about the arm hardware future. what if we used a mac mini m1 for aarch64?
2022-07-18 10:48:20 <ncopa> wouldnt work for 32 bit though
2022-07-18 10:50:44 <ikke> clandmeter asked telmich if he could host a server there
2022-07-18 10:51:36 <ncopa> i saw
2022-07-18 12:33:05 <clandmeter> a mac mini with 8GB?
2022-07-18 12:34:05 <ncopa> if that is enough. i was thinking with 16G
2022-07-18 12:34:58 <clandmeter> not sure 16 would be enough
2022-07-18 12:37:56 <clandmeter> m1 is nice, but the rest of the hw is limited and no way to upgrade it :(
2022-07-18 12:38:36 <clandmeter> i would like to have one as desktop :)
2022-07-18 12:46:03 <ncopa> they are nice. I have been using a mbp for my aarch64 development machine
2022-07-18 12:54:31 <ncopa> it takes approx as long time to compile kernel in a vm on m1 than it does on the arm monster machine, so i thought we could maybe offload it with a mac mini m1.
2022-07-18 12:55:03 <ncopa> it is silent, so it could maybe even be located in my apartment
2022-07-18 16:02:02 <ncopa> interesting. i can no longer set topic in #alpine-devel channel.  Cannot perform action: You're not a channel operator.
2022-07-18 16:03:50 <ncopa> also not in #alpine-linux. can someone help me change 3.16.0 -> 3.16.1 in the channel topics?
2022-07-18 17:09:35 <ikke> ncopa: you can ask chanserv to make you op
2022-07-18 17:09:59 <ikke> Not sure why, but apparently auto op is not working / no longer active
2022-07-18 17:10:41 <ikke> ncopa: anyway, done
2022-07-19 03:01:58 <tomalok> MR for updating cloud images in alpine-mksite - https://gitlab.alpinelinux.org/alpine/infra/alpine-mksite/-/merge_requests/49 - when someone has a chance
2022-07-19 04:22:23 <ikke> merged, and pushed to production
2022-07-19 06:20:37 <psykose> armhf/ppc64le look stuck on py3-gls
2022-07-19 09:51:20 <psykose> ikke: are the mailing lists entirely down currently?
2022-07-19 10:20:36 <ikke> Not that I'm aware of
2022-07-19 14:39:59 <ncopa> can I create a user 'ncopa' and try build rust for s390x directly on usa2-dev1 with docker?
2022-07-19 14:40:45 <ncopa> hm, not very much disk space there though
2022-07-19 14:49:22 <psykose> how much is there
2022-07-19 14:54:22 <ikke> ncopa: we have the ci host with docker already 
2022-07-19 14:54:33 <ikke> Should have enough space 
2022-07-19 14:57:27 <ncopa> there are /dev/mapper/vg1-lv_var  393G  354G   22G  95% /var
2022-07-19 14:57:41 <ncopa> but /home (on lv_root) has only
2022-07-19 14:57:51 <ncopa>   /dev/vg0/lv_root        4.3G  425M  3.7G  11% /
2022-07-19 14:57:55 <ncopa> 3G free
2022-07-19 14:58:17 <psykose> ah, indeed a bit short
2022-07-19 14:58:19 <ncopa> ikke: what is the ip/hostname for ci host?
2022-07-19 15:01:56 <ikke> s390x-ci.alpinelinux.org
2022-07-19 15:02:17 <ncopa> thanks!
2022-07-19 15:08:55 <ncopa> building it there now
2022-07-19 15:36:15 <ikke> And I'll clean up USA2
2022-07-20 04:21:04 <ikke> clandmeter: ^
2022-07-20 09:05:36 <clandmeter> ikke: ?
2022-07-20 09:16:31 <ikke> Nld3 having <50G :-)
2022-07-20 10:48:10 <clandmeter> we need to really get those 3 new servers up and running
2022-07-20 10:48:18 <clandmeter> its a waste of resources like this
2022-07-20 10:48:42 <clandmeter> but i know its me who is limited in time :(
2022-07-20 11:06:40 <ncopa> I'd like to move mdev-conf package's sources to its own git repo. Do we want it under gitlab.a.o/ncopa/mdev-conf or gitlab.a.o/alpine/mdev-conf? I suppose the latter?
2022-07-20 11:07:09 <ikke> Yes, if we see it as an official alpine package, i'd use the latter indeed
2022-07-20 11:07:16 <ikke> s/package/project
2022-07-20 11:07:16 <alpine-meetbot> ikke meant to say: Yes, if we see it as an official alpine project, i'd use the latter indeed
2022-07-20 11:47:15 <ncopa> https://gitlab.alpinelinux.org/alpine/mdev-conf
2022-07-20 11:47:55 <ikke> 👍
2022-07-20 11:58:53 <mps> later, I think
2022-07-20 13:44:17 <clandmeter> ncopa: ping
2022-07-20 13:45:38 <ncopa> clandmeter: pong
2022-07-20 14:00:40 <clandmeter> ncopa: :)
2022-07-20 14:01:03 <clandmeter> ikke: i think you killed my runner that was configged for docker-abuild :)
2022-07-20 14:01:36 <ikke> 😶
2022-07-20 14:01:40 <clandmeter> ikke: i guess we currently do not have a multi arch runner?
2022-07-20 14:01:58 <clandmeter> with qemu user setup
2022-07-20 16:09:29 <ikke> FYI: just received an e-mail there will be some maintenance on the s390x servers tomorrow (16.00ET), taking ~2 hours, so they will be temporarily be unavailable
2022-07-21 00:13:01 <tomalok> MR for updating https://alpinelinux.org/cloud images -- https://gitlab.alpinelinux.org/alpine/infra/alpine-mksite/-/merge_requests/50
2022-07-21 04:33:50 <ikke> tomalok: merged and pushed
2022-07-21 09:35:22 <clandmeter> ikke: i guess we currently do not have a multi arch runner?
2022-07-21 09:36:58 <ikke> Not yet I'm aware of 
2022-07-21 09:37:09 <ikke> Do you recall where we had that? 
2022-07-21 09:57:47 <psykose> s390x ci seems to have https://gitlab.alpinelinux.org/messense/aports/-/jobs/776366 docker daemon stuff again
2022-07-21 10:16:17 <ikke> psykose: the funny part is that it's inside a docker container where it gives this message
2022-07-21 10:16:23 <psykose> :p
2022-07-21 10:16:32 <psykose> it's a random flake pretty sure, but yeah it's funny
2022-07-21 10:18:17 <ikke> Perhaps some kind of timeout due to load? 
2022-07-21 10:19:16 <ncopa> i think we can remove the stuff i did on the s390x machine
2022-07-21 10:19:16 <psykose> on the socket call? possible
2022-07-21 10:19:16 <psykose> it's the same kind of think with the random other timeouts on ppc especially
2022-07-21 10:19:16 <psykose> thing*
2022-07-21 10:23:43 <ncopa> i cleaned up after my s390x rust work on s390x-ci.a.o
2022-07-21 10:59:21 <ikke> clandmeter: Apparently there are still 2 online
2022-07-21 10:59:28 <ikke> for s390x and ppc64le
2022-07-21 11:23:49 <clandmeter> ikke: i think my old one was multi arch
2022-07-21 11:24:04 <clandmeter> and we have some servers that are multi arch like for riscv64
2022-07-21 11:24:09 <clandmeter> not sure they are also doing CI
2022-07-21 11:37:40 <ikke> They are at least connected to the docker-abuild project
2022-07-21 11:37:53 <ikke> But I have no clue where your multi arch runner ran
2022-07-21 11:47:43 <ikke> hmf
2022-07-21 15:58:11 <ncopa> where are build-edge-riscv64 running?
2022-07-21 15:58:51 <ikke> nld5
2022-07-21 15:58:56 <ikke> sorry
2022-07-21 15:58:58 <ikke> usa5
2022-07-21 15:59:04 <ikke> Need to reboot it, it seems stuck again
2022-07-21 16:00:51 <ncopa> im optimistic about rust for riscv64
2022-07-21 16:01:07 <ncopa> has built for a few hours now
2022-07-21 16:02:26 <ikke> holds on tight
2022-07-21 16:02:33 <ikke> usa5 is booting now
2022-07-21 16:03:33 <ncopa> ok. thanks. I wanted to copy over the cross-compiled binaries so i have it prepared
2022-07-21 16:04:01 <ikke> note that it's using docker, not lxc
2022-07-21 16:04:11 <ncopa> oh. ok
2022-07-21 16:04:56 <ncopa> as long as i can copy in the binaries....
2022-07-21 16:05:02 <ikke> that's no problem
2022-07-21 16:05:50 <ikke> ncopa: you can copy them to /var/cache/distfiles
2022-07-21 16:05:55 <ikke> which is available in the container
2022-07-21 16:06:09 <ikke> /srv/compose/aports-build
2022-07-21 16:07:09 <ikke> ncopa: sorry, /srv/compose/aports-build/storage/distfiles
2022-07-21 16:10:02 <ncopa> stepping out for a couple of hours...
2022-07-21 16:10:15 <ikke> alright
2022-07-22 06:49:52 <tmhoang> ikke: there was a planned reboot of s390x servers during this week. It seems it has been done. Sorry for not letting you know before hands.
2022-07-22 06:50:17 <tmhoang> please help me to verify if all services are correctly started 
2022-07-22 06:58:16 <tmhoang> got an update from toolchains team here yesterday, thanks a lot for pushing the rust work on s390x ncopa :) 
2022-07-22 07:02:49 <ikke> tmhoang: we got an email from Martha, so all is good 
2022-07-22 08:11:35 <ncopa> tmhoang: thanks! wouldn't be possible without the help from IBM
2022-07-22 08:14:26 <ncopa> i think the riscv64 builder deadlocked
2022-07-22 08:14:57 <ncopa> has been stuck since yesterday. i did a tail -f /srv/compose/aports-build/storage/distfiles/buildlogs/build-edge-riscv64/community/yggdrasil/yggdrasil-0.4.4-r1.log but no progress within hours
2022-07-22 08:15:33 <ncopa> im gonna kill the build
2022-07-22 10:16:07 <clandmeter> ncopa: maybe makes sense to update qemu if it runs in qemu user
2022-07-22 10:25:05 <ikke> clandmeter: the host still runs alpine 3.14, maybe we should upgrade it?
2022-07-22 10:25:18 <clandmeter> yeah 
2022-07-22 10:25:27 <clandmeter> i gues that will bump also qemu
2022-07-22 10:26:02 <clandmeter> and maybe fix some of the locks we are seeing 
2022-07-22 10:26:44 <ikke> The host still has issues where it locks up regularly
2022-07-22 10:27:34 <ikke> dmesg output: https://tpaste.us/X04j
2022-07-22 10:27:58 <ikke> I'll do it this afternoon
2022-07-22 10:29:09 <clandmeter> lock as in system itself or a build?
2022-07-22 10:29:26 <clandmeter> we had one epyc system that was unstable, is it this one?
2022-07-22 10:29:36 <ikke> clandmeter: system itself
2022-07-22 10:29:38 <ikke> yes
2022-07-22 10:30:12 <clandmeter> right
2022-07-22 10:30:33 <clandmeter> holidays are coming and this will give me some extra free time
2022-07-22 10:30:43 <clandmeter> ill try to focus on the equinix stuff
2022-07-22 10:31:02 <clandmeter> are you using any of the MS resources?
2022-07-22 10:31:09 <ikke> clandmeter: the new arm server we have has been rebooting regularly as well since I installed it
2022-07-22 10:31:18 <ikke> asked equinix about it, but they see nothing
2022-07-22 10:31:21 <ikke> clandmeter: not atm
2022-07-22 10:31:27 <clandmeter> hmm
2022-07-22 10:31:39 <ikke> clandmeter: I was thinking maybe setting up a server for nightly bootstrap builds
2022-07-22 10:31:42 <clandmeter> so they gave us an unstable server
2022-07-22 10:31:58 <ikke> clandmeter: I suppose, but I cannot find any indication as to why it rebooted
2022-07-22 10:32:16 <ikke> according to their systems, it has an uptime >10 days
2022-07-22 10:32:35 <ikke> https://zabbix.alpinelinux.org/history.php?action=showgraph&itemids%5B%5D=33718
2022-07-22 10:32:36 <clandmeter> let me prepare an email to ARM and see if we can get those systems they mentioned
2022-07-22 10:32:42 <ikke> nod
2022-07-22 10:33:16 <clandmeter> i dont really understand that graph :)
2022-07-22 10:33:30 <clandmeter> it means it has an uptime of 4h?
2022-07-22 10:34:01 <clandmeter> how can equinix know the uptime?
2022-07-22 10:34:22 <ikke> Through the BMC
2022-07-22 10:34:22 <clandmeter> i think they can only do that if the OS has a call home?
2022-07-22 10:34:30 <ikke> "Based on the records from the BMC it seems like the server has been up for the past 11 days as was shared by Joel. "
2022-07-22 10:34:44 <clandmeter> hmm
2022-07-22 10:34:44 <clandmeter> ok
2022-07-22 10:34:50 <ikke> But it might be that it shows how long it was powered on
2022-07-22 10:35:08 <clandmeter> i wonder if they mix bmc and call home
2022-07-22 10:35:39 <clandmeter> if it has rebooted, just copy them the uptime info :)
2022-07-22 10:35:43 <ikke> "Because whatever is triggering these reboots, it doesnt seem to be getting logged on the BMC or on the API. "
2022-07-22 10:36:00 <clandmeter> you installed it manually right?
2022-07-22 10:36:07 <ikke> yes
2022-07-22 10:36:17 <ikke> so no cloud-init stuff
2022-07-22 10:36:21 <clandmeter> right so it does not have the scripts installed for api
2022-07-22 10:36:37 <ikke> They are aware I installed it manually
2022-07-22 10:37:27 <ikke> clandmeter: re that graph: everything it goes down it means the server rebooted
2022-07-22 10:37:31 <ikke> everytime*(
2022-07-22 10:45:30 <clandmeter> Yes but it didn’t show the down part :) I need to adjust the time I guess
2022-07-22 10:46:59 <ikke> right
2022-07-22 10:51:41 <ikke> I've added extra monitoring for usa5 to notice when it's no longer responding (the zabbix agent not sending data)
2022-07-22 10:55:48 <ikke> clandmeter: Is it worthwhile to move the rv64 builder to nld5 (after removing the mirror there)?
2022-07-22 10:55:58 <ikke> or do we wait for the new infra?
2022-07-22 11:05:01 <clandmeter> i remembered you wanted to move it to lxc right?
2022-07-22 11:21:55 <clandmeter> ikke: did you already start using compose based on golang?
2022-07-22 11:22:53 <ikke> Yes
2022-07-22 11:23:06 <clandmeter> its in the repos?
2022-07-22 11:23:18 <ikke> docker-cli-compose
2022-07-22 11:23:25 <clandmeter> ah
2022-07-22 11:23:44 <clandmeter> strange name
2022-07-22 11:23:57 <ikke> It's because it's a plugin 
2022-07-22 11:24:15 <clandmeter> it is?
2022-07-22 11:24:20 <ikke> Rather then a standalone binary 
2022-07-22 11:24:22 <ikke> Yes
2022-07-22 11:24:23 <panekj> you use it as `docker compose`
2022-07-22 11:24:31 <clandmeter> i used it standalone
2022-07-22 11:24:35 <panekj> but you can use it as standalone
2022-07-22 11:24:38 <ikke> Yes
2022-07-22 11:25:05 <clandmeter> i just copied it to usr/local/bin from github :)
2022-07-22 11:25:23 <clandmeter> it seems a lot faster
2022-07-22 11:25:27 <clandmeter> kinda nice
2022-07-22 15:43:59 <ikke> clandmeter: what was the state of the alpinelinux image for azure?
2022-07-22 20:31:57 <psykose> rebooted deu7
2022-07-22 20:32:23 <ikke> Ah, that's why my ssh sessions are disconnected
2022-07-22 20:34:30 <psykose> :)
2022-07-22 20:34:38 <psykose> back up
2022-07-22 20:34:57 <psykose> i hope that doesn't lose you anything
2022-07-22 20:35:02 <psykose> maybe i should give you some warning
2022-07-22 20:35:28 <ikke> No, didn't loose anything
2022-07-22 20:35:45 <ikke> But a heads-up would've been nice
2022-07-22 20:36:09 <psykose> will do the same for deu1 in a bit
2022-07-22 20:36:11 <psykose> updates first
2022-07-22 20:42:51 <ikke> so is algitbot
2022-07-22 20:43:10 <psykose> yup
2022-07-22 20:43:23 <psykose> maybe i should look at the cgit stuff again, but doubt it would be fixed
2022-07-22 20:43:39 <ikke> ppc64le is still gone
2022-07-22 20:44:38 <psykose> let it have its nice vacation
2022-07-22 21:40:55 <ikke> ibm confirmed they need to do maintenance
2022-07-22 21:42:24 <ikke> probably the entire weekend
2022-07-22 22:32:31 <psykose> what machine is dev.a.o on? it has a cgit with 30 workers on it but it's not where the actual cgit it
2022-07-22 22:40:05 <psykose> yeah, the cgit there points to nothing
2022-07-23 02:03:22 <ptrc> hi um, i think https://gitlab.alpinelinux.org/alpine/infra/infra/-/issues/8468 should be revisited
2022-07-23 02:05:28 <psykose> let me guess
2022-07-23 02:05:40 <psykose> you can put anything on the mqtt page
2022-07-23 02:05:44 <psykose> lmao
2022-07-23 02:05:51 <psykose> wouldnt be surprised
2022-07-23 02:06:50 <ptrc> not only that
2022-07-23 02:06:55 <ptrc> you can put anything on -commits
2022-07-23 02:08:42 <ptrc> > mosquitto_pub -L mqtts://msg.alpinelinux.org/build/build-3-16-x86_64/errors -m '{"reponame":"community","hostname":"some-prefix","logurl":"https:\/\/example.com\/your-url-here","pkgname":"something"}'
2022-07-23 02:09:49 <ptrc> aaand publishing to rsync/ does stuff too; or at least prints "uploaded" on -commits
2022-07-23 02:10:03 <psykose> lmao
2022-07-23 02:10:09 <ptrc> well
2022-07-23 02:10:10 <psykose> yep
2022-07-23 02:10:49 <psykose> don't think the uploaded does anything, it's just the notification for upload finished
2022-07-23 02:10:53 <psykose> but yeah anything can be put there
2022-07-23 02:11:07 <psykose> fun stuff
2022-07-23 02:11:39 <ptrc> only a matter of time someone buys a typoed domain and publishes "errors" that redirect to fake alpine gitlab /s
2022-07-23 02:13:15 <psykose> mckay is that you ?!
2022-07-23 02:15:05 <ptrc> i also wonder if publishing to git/aports/master can trigger jirutka's mirror script
2022-07-23 02:15:17 <ptrc> conveniently, that endpoint accepts a git commit hash
2022-07-23 02:15:27 <psykose> make it build something from 2015
2022-07-23 02:15:54 <ptrc> hey, i don't want to mess up the github aports mirror
2022-07-23 02:16:12 <psykose> well you have two options here
2022-07-23 02:16:23 <psykose> either you don't and we forget about this in 5 minutes
2022-07-23 02:16:28 <psykose> or you do and it's funny and gets fixed
2022-07-23 02:16:37 <psykose> i suppose don't, sure
2022-07-23 02:16:58 <ptrc> ah, no, the script only gets the "repo" field from the payload
2022-07-23 02:17:47 <psykose> :3
2022-07-23 07:10:11 <ikke> psykose: dev.a.o is on one of the x86* builders, where we used to host a lot of these infra before
2022-07-23 07:10:17 <psykose> aha
2022-07-23 07:10:20 <psykose> yeah made sense
2022-07-23 07:10:25 <psykose> i noticed load was full during the build
2022-07-23 07:10:39 <psykose> i've used quite a bit of disk space too as i put the chromium tarballs there
2022-07-23 07:10:47 <psykose> only latest version, so just ~10G at max
2022-07-23 07:10:50 <ikke> ok
2022-07-23 07:11:08 <psykose> but, thankfully, we now have maintained chromium in 3.16
2022-07-23 07:11:09 <ikke> When we have the new equinix infra setup, we can see if we can move stuff further away from there
2022-07-23 07:11:11 <psykose> completely up to date
2022-07-23 07:11:19 <psykose> i was asking to remove the cgit things
2022-07-23 07:11:25 <psykose> they're running the dev container but the things are empty
2022-07-23 07:11:29 <ikke> yeah, makes sense
2022-07-23 07:11:31 <psykose> probably some other cleanup to do
2022-07-23 07:11:39 <ikke> https://git-old.alpinelinux.org/
2022-07-23 07:11:46 <ikke> It still hosts that
2022-07-23 07:11:51 <psykose> are all of these moved
2022-07-23 07:12:33 <ikke> all the acf stuff has been moved
2022-07-23 07:12:42 <ikke> docs has been moved
2022-07-23 07:12:59 <ikke> most of the alpine-* repos have been moved
2022-07-23 07:13:47 <ikke> All the rest I don't know
2022-07-23 07:20:04 <psykose> well
2022-07-23 07:20:06 <psykose> user/ doesn't matter
2022-07-23 07:20:14 <psykose> hosted is ancient
2022-07-23 07:20:24 <psykose> testing/ is a test
2022-07-23 07:20:46 <psykose> and that's about it, you can compare the rest from bare git.a.o
2022-07-23 07:20:48 <ikke> clandmeter: what do you think about removing the mirror from nld5
2022-07-23 07:20:51 <psykose> some of the abuild stuff is missing
2022-07-23 07:21:03 <psykose> er alpine
2022-07-23 07:32:06 <ikke> atm, we have 40G free on the nld3 mirror :/
2022-07-23 07:32:59 <ikke> 3 days ago was when ncopa did the releases?
2022-07-23 07:33:32 <ikke> went from 55G to 40G
2022-07-23 08:08:54 <psykose> do you have a graph of the past 2 hours
2022-07-23 08:12:09 <ikke> https://zabbix.alpinelinux.org/history.php?action=showgraph&itemids%5B%5D=33355
2022-07-23 08:12:55 <ikke> ~3GB freed
2022-07-23 08:13:53 <psykose> sweet
2022-07-23 09:13:23 <psykose> now mesa too
2022-07-23 09:13:43 <ikke> ~4G in total atm
2022-07-23 09:20:35 <psykose> neat
2022-07-24 20:58:28 <Cogitri> Heya, can we increase the maximum attachment size? Unfortunately Rust releases with their entire dependency tree vendored tend to take up a bit of space (<20MB), so the current 10MB limit makes it impossible to release a new release of apk-polkit-rs currently
2022-07-24 20:58:46 <ikke> attachment size?
2022-07-24 20:58:55 <Cogitri> Ah on Gitlab
2022-07-24 20:59:16 <Cogitri> 
ACTION uploaded an image: (22KiB) < https://matrix.org/_matrix/media/r0/download/gnome.org/cdddbf06668335805beb092272344062beeaf019/image.png >

2022-07-24 21:01:13 <ikke> Ok, found it, reasied it to 50M
2022-07-24 21:01:19 <ikke> raised*
2022-07-24 21:01:29 <Cogitri> Thanks 👍️
2022-07-24 22:55:03 <tomalok> ikke: MR for revamping https://alpinelinux.org/cloud -- https://gitlab.alpinelinux.org/alpine/infra/alpine-mksite/-/merge_requests/52 -- image filter is working!
2022-07-25 05:34:29 <ikke> https://wwwdev.alpinelinux.org/cloud/
2022-07-25 05:40:23 <ikke> tomalok: https://i.imgur.com/B61jbd0.png dropdown boxes text is truncated at the bottom for me (firefox)
2022-07-25 06:45:41 <ncopa> very nice!
2022-07-25 07:56:42 <ptrc> https://postmarketos.org/blog/2022/07/25/considering-sourcehut/
2022-07-25 07:57:48 <ptrc> some people on #postmarketos are suggesting that maybe alpine could host them instead
2022-07-25 08:02:12 <ikke> Problem is we have no headroom on our ARM ci
2022-07-25 08:04:54 <panekj> don't they use sr.ht for CI already?
2022-07-25 08:04:57 <ptrc> would it be feasible if they somehow bring their own runners?
2022-07-25 08:05:09 <panekj> I presume it would be only git hosting
2022-07-25 08:05:11 <ptrc> panekj: they use gitlab.com runners for merge requests
2022-07-25 08:05:20 <panekj> oh well
2022-07-25 08:05:32 <panekj> ptrc: it's possible to have own runners
2022-07-25 08:06:02 <panekj> although I'm not sure whether global runners can be restricted for specific projects
2022-07-25 08:06:07 <ptrc> yeah, i know, i do that myself; it's more of a "would alpine infra team be okay with that" question
2022-07-25 08:06:30 <ptrc> oh, hm..
2022-07-25 08:07:18 <ptrc> i guess currently  runners need to be allowed globally to run ci in forks
2022-07-25 08:08:51 <ikke> Yes, but you can use tags to limit what runners are being used 
2022-07-25 08:11:01 <clandmeter> ptrc: they suggesting just CI or part of our gitlab instance?
2022-07-25 08:12:46 <ptrc> i imagine pmos is mostly concerned about the newly introduced "5 users per namespace" limit on gitlab.com, so it's mostly about hosting the repositories
2022-07-25 08:13:34 <ptrc> (as in, the post on gitlab.com they linked doesn't mention any new ci limits whatsoever)
2022-07-25 08:19:08 <clandmeter> did they try to apply for the open source program?
2022-07-25 08:19:25 <clandmeter> https://about.gitlab.com/solutions/open-source/join/
2022-07-25 08:20:08 <ptrc> someone from pmos said they already use it
2022-07-25 08:20:46 <clandmeter> ultimate has limitations?
2022-07-25 08:21:09 <clandmeter> or is this not related to gitlab.com hosting?
2022-07-25 08:21:24 <panekj> if they already are using open source plan then what is the problem?
2022-07-25 08:21:32 <ptrc> 09:25:35 <z3ntu> We already use this, otherwise we wouldn't be able to use gitlab now.. but being dependent on the charity of gitlab and whatever their open source program people think is not a great situation to be in 
2022-07-25 08:22:26 <clandmeter> well, thats a problem with any solution
2022-07-25 08:22:31 <clandmeter> or you need to build it yourself
2022-07-25 08:23:34 <ptrc> they apparently talk a bit more in-depth about it on the podcast: https://cast.postmarketos.org/episode/20-Considering-SourceHut-special/
2022-07-25 08:23:39 <clandmeter> they could also take our docker compose stuff and run it on any cloud or bare metal provider
2022-07-25 08:24:35 <clandmeter> tbh, keeping it up2date is the most work, running it is kind of "easy"
2022-07-25 08:25:00 <clandmeter> but i dont want to talk for ikke :)
2022-07-25 08:28:19 <ikke> It doesn't take that much time for me honestly :-)
2022-07-25 08:49:39 <ncopa> we used git send-email workflow in alpine earlier. It was problematic for many contributors.
2022-07-25 08:50:37 <ncopa> we spent much more time then on dealing with badly sent patches or helping users set up their email accounts than we do today with gitlab, even if the number of contributors probably are 10x or more these days
2022-07-25 13:06:09 <tomalok> ikke: tested on chrome and ipad safari - will take a look on ff today
2022-07-25 14:36:10 <tomalok> ikke: seems to render okay for me on FF?  not entirely sure whatyou mean by "dropdown boxes text is truncated at bottom" -- here's what i see -- https://dev.alpinelinux.org/~tomalok/cloud-ff-1.jpg
2022-07-25 14:39:01 <ikke> I added a screenshot 
2022-07-25 14:41:04 <ikke> It's not when you open it, just in the closed state. For you, it looks alright I see. 
2022-07-25 14:42:59 <mps> ah, tomalok is now alpine developer
2022-07-25 14:45:45 <tomalok> ff v91.11.0esr here on macOS if that makes any difference
2022-07-25 14:48:32 <ikke> tomalok: might be font related 
2022-07-25 14:49:17 <tomalok> check the javascript console just in case?
2022-07-25 14:50:39 <ikke> Will do
2022-07-25 14:51:04 <tomalok> fonts should all come from CSS, too...
2022-07-25 14:52:22 <tomalok> could just be something that pops up with drop downs if there's none currently on the site, wouldn't have been seen til now
2022-07-25 15:08:20 <ikke> tomalok: on my desktop it looks alright
2022-07-25 19:11:12 <Hello71> the height is fixed to 2.25em, which is ... not good
2022-07-25 19:46:16 <ikke> Hello71: apparently that's part of the theme
2022-07-25 19:46:21 <Hello71> right
2022-07-25 20:03:31 <ikke> ok, seems like ppc64le builders are back again now
2022-07-25 21:13:29 <tomalok> Hello71: what's the location of that in CSS file?  What would an appropriate height value be?  (I could, in theory, fix via this MR)
2022-07-25 21:13:45 <Hello71> well it's from upstream pure theme
2022-07-25 21:14:00 <tomalok> aha
2022-07-25 21:14:10 <ikke> pure-min.css:609
2022-07-25 21:15:41 <tomalok> we could edit or simply override it with a different value
2022-07-25 21:17:17 <tomalok> my pure-min.css only goes up to a very long line 11...
2022-07-25 21:17:30 <ikke> Oh, probably auto-formatted
2022-07-25 21:17:58 <tomalok> couple CSS warnings an errors in that file too, fwiw
2022-07-25 21:19:56 <ikke> There are new versions of pure available
2022-07-25 21:20:15 <ikke> This is 0.6.0, there is 0.6.2
2022-07-25 21:20:23 <ikke> But also 1.x and 2.x
2022-07-25 21:20:36 <ikke> https://github.com/pure-css/pure/releases
2022-07-25 22:03:16 <tomalok> i wonder how "drop in" ready 1.x and 2.x are...  it's potentially a larger impact across the whole site...
2022-07-25 22:04:06 <psykose> ikke: none of the build log urls work
2022-07-25 22:27:51 <psykose> actually seems to only be an issue for me
2022-07-25 22:47:44 <tomalok> what if we <link>'d purecss from a CDN url instead of serving it up locally?
2022-07-25 23:09:49 <tomalok> tried it locally with <link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/purecss@2.1.0/build/pure-min.css" integrity="sha384-yHIFVG6ClnONEA5yB5DJXfW2/KC173DIQrYoZMEtBvGzmf0PKiGyNEqe9N6BNDBH" crossorigin="anonymous"> and i'm not seeing any differences.
2022-07-25 23:12:30 <Hello71> it's the same in master
2022-07-25 23:12:48 <Hello71> better to use local files for privacy (and speed)
2022-07-25 23:13:27 <tomalok> i've pasted the contents in the local pure-min.css file, and the CSS linter is happier now
2022-07-25 23:14:31 <tomalok> grids-responsive-min.css is still 0.6.0
2022-07-25 23:15:40 <tomalok> updating that too
2022-07-25 23:16:45 <tomalok> .conf-schedule-day in styles.css seems to have an extra . just before the closing }
2022-07-25 23:43:00 <tomalok> ikke: PR updated to also update Pure to 2.1.0 -- everything looks fine to me -- hopefully this fixes what you saw?
2022-07-27 07:28:44 <mps> what variable is for abuild top dir? one where APKBUILD is
2022-07-27 07:29:02 <panekj> APORTSDIR?
2022-07-27 07:29:13 <mps> panekj: thank you
2022-07-27 07:29:32 <panekj> wait, you mean directory that has APKBUILD?
2022-07-27 07:30:14 <mps> yes, and APORTSDIR doesn't work
2022-07-27 07:30:28 <ptrc> startdir
2022-07-27 07:30:50 <mps> APORTSDIR is top level for aports
2022-07-27 07:31:24 <mps> nice, looks like startdir works
2022-07-27 07:31:28 <mps> thanks
2022-07-27 07:32:45 <mps> btw, I upgraded kernels for apple silicon to latest asahi branch, also m1n1 and u-boot 
2022-07-27 07:33:27 <mps> short guide for upgrade is here https://arvanta.net/alpine/upgrade-kernel-m1/
2022-07-27 07:33:46 <mps> if anyone of you use alpine on apple silicon
2022-07-27 07:33:55 <panekj> i wish
2022-07-27 07:35:11 <mps> nowadays Air macbook could be bought for about 1000 euros with educational discount
2022-07-27 07:48:20 <ncopa> mps: qemu 7 does not work with my m1 macbook with more than 3G memory
2022-07-27 07:48:36 <ncopa> i used to have 8G for the vm
2022-07-27 07:52:01 <mps> ncopa: on macos?
2022-07-27 07:52:11 <ncopa> yup. macos 12.5
2022-07-27 07:52:23 <ncopa> not sure if it was introduced with macos 12.5 or qemu 7
2022-07-27 07:52:25 <mps> yes, I noticed that
2022-07-27 07:53:03 <mps> with previous macos it worked fine with 8GB for VM even
2022-07-27 07:53:19 <ncopa> yup
2022-07-27 07:53:24 <mps> I think this is macos problem
2022-07-27 07:53:47 <mps> but didn't looked much because I rarely boot to macos
2022-07-27 07:54:39 <mps> but I noticed that when I tested booting alpine in qemu from disk partition and not VM image
2022-07-27 07:55:49 <mps> will ask later people who runs linux in qemu
2022-07-27 08:21:00 <mps> ncopa: maz (qemu develeper iirc) told me to remove 'highmem=off'
2022-07-27 08:21:25 <mps> could you try without this option
2022-07-27 08:28:22 <ncopa> i get error without that option
2022-07-27 20:44:16 <mps> is there a way to check if shell script is in posix standard
2022-07-27 20:46:08 <clandmeter> mps: https://github.com/koalaman/shellcheck/issues/1527
2022-07-27 20:46:38 <ikke> mps: note that ash supports more then just posix, not sure if that's what you are targeting
2022-07-27 20:47:06 <ikke> There is also checkbashisms
2022-07-27 20:48:05 <mps> ash, ofc
2022-07-27 20:48:22 <ikke> shellcheck has no real support for ash
2022-07-27 20:48:42 <ikke> (becausse ash has many variants, as you can enable or disable features in bb ash for example)
2022-07-27 20:49:04 <clandmeter> ash is almost posix
2022-07-27 20:49:08 <mps> yes, but I think we have posix compliant one
2022-07-27 20:49:28 <clandmeter> i think its just 2 or 3 things it adds on top?
2022-07-27 20:49:38 <clandmeter> iirc
2022-07-27 20:49:41 <mps> idk
2022-07-27 20:49:56 <ikke> certain parameter substitutions at least
2022-07-27 20:50:01 <ikke> ${a/b/c}
2022-07-27 20:50:52 <clandmeter> yes 
2022-07-27 20:50:54 <mps> just posted patch to u-boot and noticed that it need bash, and I don't like idea to add bash to makedeps
2022-07-27 20:51:00 <clandmeter> and local usage 
2022-07-27 20:51:10 <ikke> mps: shellcheck will warn you 
2022-07-27 20:51:16 <ikke> if you use /bin/sh as shebang
2022-07-27 20:51:36 <mps> aha, lets try
2022-07-27 20:51:37 <ikke> clandmeter: set -o pipefail as well
2022-07-27 20:51:47 <clandmeter> right
2022-07-27 20:51:50 <clandmeter> thats 3 :)
2022-07-27 20:51:55 <clandmeter> but i guess there are more
2022-07-27 20:52:04 <ikke> probably, don't know from the top of my head
2022-07-27 20:52:30 <ikke> #shellcheck disable=SC3040 is almost muscle memory :P
2022-07-27 20:53:00 <mps> ikke: it passed with '/bin/sh'
2022-07-27 20:54:08 <ikke> re ash: https://github.com/koalaman/shellcheck/issues/1841
2022-07-27 20:54:35 <ikke> oh, that one has no other comments
2022-07-27 20:55:09 <ikke> mps: is that expected or not?
2022-07-27 20:55:20 <ikke> https://github.com/koalaman/shellcheck/issues/853
2022-07-27 20:55:42 <mps> ikke: I think so
2022-07-27 20:56:00 <ikke> clandmeter: apparently ash supports [[ ]] as well
2022-07-27 20:57:22 <ikke> https://github.com/koalaman/shellcheck/issues/853#issuecomment-357281107
2022-07-27 20:58:12 <mps> hmm, shellcheck is not on aarch64
2022-07-27 20:59:08 <ikke> right, we only have ghc for x86_64 atm
2022-07-27 20:59:49 <ikke> https://github.com/koalaman/shellcheck/releases/tag/v0.8.0
2022-07-27 21:00:30 <mps> yes, I see on page
2022-07-29 16:36:48 <panekj> is there any way to find latest package version without apk? preferably via API
2022-07-29 16:37:08 <panekj> does pkgs.a.o has some kind of functionality like that?
2022-07-29 16:37:17 <ikke> pkgs.a.o does not have an API
2022-07-29 16:37:47 <panekj> I feared that
2022-07-29 16:38:18 <ikke> panekj: simple way that works for now is to extract an APKINDEX
2022-07-29 16:38:27 <ikke> That has all the info you need
2022-07-29 16:39:10 <ikke> https://gitlab.alpinelinux.org/alpine/go/-/blob/master/repository/apkindex.go#L49
2022-07-29 16:41:27 <panekj> yeah, I know, I'm just trying to help someone while arguing with people who suggest absolutely dumb ideas
2022-07-29 16:41:46 <ikke> heh
2022-07-29 16:42:02 <ikke> There were plans to replace aports-turbo
2022-07-29 16:42:09 <ikke> But that still needs to be rolled out
2022-07-29 16:42:18 <panekj> is there something written or in progress?
2022-07-29 16:42:30 <ikke> It is something written
2022-07-29 16:42:50 <ikke> Though, it was also still missing some featured that were deemed necessary
2022-07-29 16:42:51 <panekj> is it on gitlab?
2022-07-29 16:43:00 <ikke> gitlab.com
2022-07-29 16:43:05 <panekj> (:
2022-07-29 16:43:29 <ikke> https://gitlab.com/postmarketOS/apkbrowser
2022-07-29 16:43:31 <panekj> some time go I started rewriting aports-turbo (in rust, yes, kill me)
2022-07-29 16:44:23 <mps> rust?! better java then :D
2022-07-29 16:44:27 <ikke> I've heard rust was not that great for the web
2022-07-29 16:44:37 <panekj> for basic things it's ok
2022-07-29 16:44:45 <mps> rust is reserved for kernel drivers ;p
2022-07-29 16:44:48 <panekj> pkgs.a.o isn't super advanced
2022-07-29 16:45:08 <ikke> nope
2022-07-29 16:45:48 <panekj> I would do it in Go and it would take me much less time, but given that Go won't enter `main`, and rust will, I decided to give rust a try (:
2022-07-29 16:46:10 <mps> seriously, I have fear that I have again to start learn rust
2022-07-29 16:46:16 <ikke> I still do everything in go :p
2022-07-29 16:46:22 <ikke> I don't care it's not in main
2022-07-29 16:46:28 <mps> go is fine 
2022-07-29 16:46:50 <panekj> go is... special
2022-07-29 16:47:14 <ikke> You basically need to loose any qualms you have :P
2022-07-29 16:47:40 <mps> I stopped to be member of community when they added CoC, though I was early user
2022-07-29 16:47:42 <panekj> I also dislike that Go didn't have proper SQLite driver that didn't require CGO until like year ago
2022-07-29 16:48:03 <ikke> Not sure that's really the fault of Go though
2022-07-29 16:48:18 <panekj> it's fault of Go they can't link to C statically (:
2022-07-29 16:48:42 <panekj> (CGO is also very terrible)
2022-07-29 16:49:00 <ikke> yeah, I stay away from cgo
2022-07-29 16:49:06 <mps> for service things nowadays I use crystal
2022-07-29 16:51:58 <panekj> I like things that are just a single file (:
2022-07-29 16:53:27 <mps> crystal compiles single binary
2022-07-29 16:54:21 <mps> but they didn't have risc-v support yet
2022-07-29 16:55:03 <panekj> can it embed data in resulting binary?
2022-07-29 16:55:33 <mps> I think so, though didn't tried
2022-07-29 16:58:57 <mps> i have some grumbles in mind, i.e. why we don't move flat aports repo
2022-07-29 17:02:23 <ncopa> panekj: curl --silent https://dl-cdn.alpinelinux.org/alpine/edge/main/x
2022-07-29 17:02:23 <ncopa> 86_64/APKINDEX.tar.gz | tar -zx -O APKINDEX 
2022-07-29 17:02:32 <ncopa> panekj: curl --silent https://dl-cdn.alpinelinux.org/alpine/edge/main/x86_64/APKINDEX.tar.gz | tar -zx -O APKINDEX 
2022-07-29 17:06:01 <ncopa> curl --silent https://dl-cdn.alpinelinux.org/alpine/edge/main/x86_64/APKINDEX.tar.gz | tar -zx -O APKINDEX | awk -F: '$1=="P" {p=$2} $1=="V" {v=$2} $0=="" {print p,v}' 
2022-07-29 17:11:04 <panekj> thanks, I was expecting some awk magic :P
2022-07-29 17:47:43 <mps> Linus Torvalds switched to apple silicon, I hope aarch64 will be better supported in kernel from now
2022-07-29 17:48:19 <panekj> is there anything that lacks in aarch64 linux?
2022-07-29 17:48:42 <mps> speed should be better at first
2022-07-29 17:49:32 <mps> and memory management should be better
2022-07-29 17:50:23 <panekj> from what I know he doesn't code anymore, only reviews patches so I would not expect improvements from him
2022-07-29 17:50:40 <mps> some programs and drivers should be PAGE SIZE independent
2022-07-29 17:51:15 <mps> right, but he can yell on buggy things
2022-07-29 17:52:50 <mps> this is example https://lists.zx2c4.com/pipermail/wireguard/2022-July/007764.html
2022-07-29 18:03:58 <Newbyte> mps where did you read that?
2022-07-29 18:04:39 <Newbyte> never mind, I see the email you linked
2022-07-29 18:04:59 <Newbyte> doesn't particularly soudn like he "switched" to it
2022-07-29 18:05:03 <Newbyte> s/soudn/sound/
2022-07-29 18:05:03 <alpine-meetbot> Newbyte meant to say: doesn't particularly sound like he "switched" to it
2022-07-29 18:05:06 <Newbyte> but still interesting to see
2022-07-29 18:19:19 <mps> Newbyte: friend on IRC told he helping Linus to switch asahi kernel with fedora distro on apple M2
2022-07-29 18:20:13 <mps> friend is one who made debian runing on apple silicon
2022-07-30 04:41:44 <psykose> gonna restart deu1/deu7 in a bit
2022-07-30 09:34:31 <mps> !37061
2022-07-30 09:34:49 <mps> this should be moved to main
2022-07-30 09:35:09 <mps> PureTryOut: ^ what you think about this
2022-07-30 09:38:10 <psykose> what part of u-boot needs it
2022-07-30 09:39:24 <mps> psykose: most arm64 u-boot needs BL31 from arm-trusted-firmware
2022-07-30 09:39:41 <psykose> actually disregard, wrong channel
2022-07-30 09:40:03 <mps> and arm-trusted-firmware need these tools to build BL31
2022-07-30 09:41:11 <mps> psykose: if the people are good every channel is ok, if the people are bad then all channels are wrong
2022-07-30 09:42:07 <panekj> the point of channels is to keep specific topics in specific channels (: so that when someone needs something that is a devel topic, they don't have to search through alpine-infra
2022-07-30 09:42:43 <mps> panekj: in theory yes, but practice is different
2022-07-30 09:42:56 <panekj> in practice it is very much alike
2022-07-30 09:43:22 <psykose> sorry could you take this to #alpine-philosophy
2022-07-30 09:43:31 <psykose> unless you want to discuss infrastructure
2022-07-30 09:44:17 <mps> psykose: no, I started here and I will continue, if you don't have anything to say just ignore msgs or me
2022-07-30 09:48:32 <psykose> that is a lot of words to say 'i am scared of talking in the actual -devel channel', but you do you
2022-07-30 09:49:01 <mps> psykose: yes, you are right about this
2022-07-30 09:50:08 <mps> I'm on this channel only because I'm member of infra team, else I will not be here
2022-07-31 18:43:48 <ikke> cleaned up distfiles on usa2