2023-07-01 01:59:22 <psykose> ikke: not sure why armv7 fails with no logs
2023-07-01 03:11:14 <psykose> and it also poofed :D
2023-07-01 03:20:35 <psykose> seems to not reconnect
2023-07-01 06:24:00 <ikke> had to reboot it
2023-07-01 06:24:51 <psykose> weird
2023-07-01 06:30:29 <ikke> a lot of 0 bytes in /var/log/messages
2023-07-01 06:32:13 <ikke> memory usage was quite high
2023-07-01 06:32:19 <ikke> between 192 and 256G
2023-07-01 06:32:42 <ikke> at the end just 128G though
2023-07-01 06:34:21 <ikke> wait..
2023-07-01 06:34:25 <ikke> again unresponsible :/
2023-07-01 06:35:31 <ikke> unresponsive*
2023-07-01 06:39:09 <psykose> this sounds like a machine bug
2023-07-01 06:39:13 <psykose> and/or kernel
2023-07-01 06:39:33 <psykose> and the fact it goes up/down in a loop makes me thing it's either rebooting itself or stuck on something
2023-07-01 06:39:37 <psykose> the former we had before right
2023-07-01 06:40:29 <ikke> it doesn't reboot on its own
2023-07-01 06:41:28 <ikke> https://tpaste.us/D7ZX 
2023-07-01 06:42:01 <psykose> bleh
2023-07-01 06:42:03 <psykose> which kernel is it
2023-07-01 06:42:27 <ikke> 6.1.30
2023-07-01 06:44:22 <ikke> Can try edge perhaps 
2023-07-01 06:45:16 <psykose> 3.18 is the same kernel
2023-07-01 06:45:28 <psykose> both .36
2023-07-01 06:45:48 <psykose> i can't think of anything sadl
2023-07-01 06:58:43 <ikke> meant linux-edge
2023-07-01 07:00:11 <ikke> https://tpaste.us/BJnk 
2023-07-01 07:13:11 <psykose> ah
2023-07-01 07:14:13 <psykose> can't find much relevant
2023-07-01 07:16:34 <ikke> even with lxc stopped :.
2023-07-01 07:17:00 <ikke> I still have access the the serial console, but no network
2023-07-01 07:26:22 <psykose> yeah this is weird :(
2023-07-01 07:26:45 <ikke> tried to disable services, see if that helps
2023-07-01 07:28:52 <ikke> upgraded kernel to .36
2023-07-01 07:31:12 <ikke> hmm
2023-07-01 07:31:20 <ikke> during upgrade, it hangs again :/
2023-07-01 07:33:03 <ikke> I have to continue later :(
2023-07-01 07:36:49 <psykose> is ok
2023-07-01 11:59:24 <ikke> running on .36 now, see how stable it is
2023-07-01 11:59:30 <ikke> (builders are not started yet)
2023-07-01 12:05:33 <psykose> we'll see if it dies instantly
2023-07-01 12:05:51 <ikke> It didn't so far
2023-07-01 12:05:55 <ikke> started the containers again
2023-07-01 12:07:53 <ikke> armv7 edge aports is corrupt
2023-07-01 12:08:03 <ikke> HEAD is empty
2023-07-01 12:08:41 <psykose> git fsck it i guess
2023-07-01 12:09:01 <psykose> or the classic of rm -r .git/objects and resync
2023-07-01 12:09:10 <psykose> or just the zero-size ones
2023-07-01 12:09:37 <ikke> so far only HEAD and index were corrupt, waiting for fsck to finish
2023-07-01 12:11:05 <psykose> hows armhf
2023-07-01 12:11:57 <ikke> mqtt-exec not running
2023-07-01 12:12:07 <ikke> repo seems fine
2023-07-01 12:17:12 <psykose> sweet
2023-07-01 12:17:24 <ikke> Hope the issue is fixedd
2023-07-01 12:17:39 <ikke> need to check 3.18 aarch64
2023-07-01 12:17:56 <psykose> one of those rare times a small kernel upgrade 'fixed' it
2023-07-01 12:17:57 <ikke> corrupt as well
2023-07-01 12:18:02 <psykose> i mean it was just looping forever before..
2023-07-01 12:18:12 <psykose> even straight on reboot
2023-07-01 12:18:18 <psykose> but we'll see
2023-07-01 12:18:41 <ikke> yeah
2023-07-01 12:20:44 <ikke> ok, all builders should run again
2023-07-01 12:43:54 <ikke> distfiles.a.o needed to be added to known_hosts for armv7
2023-07-01 13:20:12 <ikke> Load average: 103.53 84.29 60.24
2023-07-01 13:21:15 <ikke> Load average: 212.48 119.62 74.04
2023-07-01 13:24:57 <psykose> aye
2023-07-01 14:51:52 <ikke> gitlab was upgraded to 15.11.10
2023-07-01 14:53:56 <psykose> !
2023-07-01 14:53:58 <psykose> ~
2023-07-01 15:10:19 <mps> ikke: we don't have access to dev lxcs on arm?
2023-07-01 15:10:41 <ikke> mps: You said you have a backup of your containers?
2023-07-01 15:10:59 <ikke> I don't have access to the containers on che-bld-1, so I cannot sync them
2023-07-01 15:11:03 <mps> I have aarch64 only and armv7 but not edge
2023-07-01 15:11:36 <ikke> is there somewhere I can access them so that I can sync them?
2023-07-01 15:11:40 <mps> np, just asked to not create armv7 edge one. now I will create it
2023-07-01 15:12:39 <mps> problem is on m1 apple, qemu-user doesn't work always when emulating arm32
2023-07-01 15:13:05 <mps> so I pulled my old arm32 chromebook and setting  on it
2023-07-01 15:15:04 <mps> and have more and more cables, connectors, adapters and other things on desk :)
2023-07-01 15:17:13 <ikke> The builder at equinix has dmvpn again, so no more fiddling with ipv6
2023-07-01 15:19:15 <mps> nice
2023-07-01 15:29:57 <psykose> 3.18 x86 seems stuck in pull
2023-07-01 15:32:31 <ikke> nope
2023-07-01 15:39:35 <psykose> nwo gone
2023-07-01 15:41:41 <ikke> gone?
2023-07-01 15:41:51 <ikke> it's just sitting idle
2023-07-01 15:44:45 <psykose> it 'looked' stuck before
2023-07-01 15:44:49 <psykose> now it's idle
2023-07-01 16:02:53 <ikke> build-3-17-aarch64 is a thing
2023-07-01 16:03:11 <ikke> 
ACTION hopes nothing gets accientally mixed up

2023-07-01 16:31:01 <ikke> something new appeares
2023-07-01 16:31:30 <ikke> test
2023-07-01 16:32:37 <psykose> :)
2023-07-01 16:32:39 <psykose> passed
2023-07-01 16:49:24 <mps> algitbot: retry master
2023-07-01 17:09:22 <mps> https://build.alpinelinux.org/buildlogs/build-edge-aarch64/community/py3-sybil/py3-sybil-4.0.1-r2.log
2023-07-01 17:11:43 <ikke> "Assertion error: "
2023-07-01 17:13:42 <mps> I'm impatient to see aarch64 edge finish community this evening (or night at least)
2023-07-01 17:17:11 <ikke> Seems like it fails upstream as well: https://app.circleci.com/pipelines/github/simplistix/sybil/341/workflows/b13c5171-48c0-4b16-9e7d-10e488e18477/jobs/1725 🤦
2023-07-01 17:18:25 <mps> huh, disable check for now?
2023-07-01 17:19:02 <ikke> Just that specific check
2023-07-01 17:19:28 <mps> I have aversion to touch python pkgs
2023-07-01 17:20:20 <psykose> i already did that 30 minutes ago
2023-07-01 17:24:16 <mps> nice
2023-07-01 22:22:41 <mps> algitbot: retry master
2023-07-02 03:45:09 <mps> algitbot: retry master
2023-07-02 03:56:59 <mps> algitbot: retry master
2023-07-02 04:58:13 <mps> algitbot: retry master
2023-07-02 10:39:06 <mps> algitbot: retry master
2023-07-03 05:11:18 <psykose> riscv64 is probably stuck
2023-07-03 05:16:10 <ikke> again
2023-07-03 05:16:41 <ikke> go has a tendency of hanging on rv64
2023-07-03 05:18:34 <ikke> https://tpaste.us/ggjy 
2023-07-03 05:19:41 <psykose> yea
2023-07-03 05:19:49 <psykose> not sure if it's some go compiler race or a qemu issue
2023-07-03 05:20:19 <ikke> me neither
2023-07-03 05:20:29 <ikke> all processes are in FUTEX_WAIT
2023-07-03 05:20:32 <ikke> and one zombie
2023-07-03 20:56:22 <psykose> ikke: s390x seems stuck
2023-07-04 08:09:27 <ikke> They're going to move the arm servers today to a location that should have a more stable network connection 
2023-07-04 09:03:13 <mps> 'more stable' or 'stable'
2023-07-04 11:30:44 <ncopa> spam: https://gitlab.alpinelinux.org/alpine/infra/mirrors/-/issues/587
2023-07-04 11:31:20 <ikke> deleted it
2023-07-04 11:31:47 <ikke> That's something we should get rspamd to catch
2023-07-04 11:32:16 <ikke> Once the bayes classifier is active, I think it should catch more
2023-07-04 11:33:03 <ikke> my local rspamd instance rejected it
2023-07-04 11:34:06 <ikke> but the one on smtp.a.o didn't have autolearning enabled and nothing manually classified, so it's not active (requires 200 ham and 200 spam classifications before it's active)
2023-07-04 11:48:28 <ikke> large user spamwave incoming
2023-07-04 13:39:11 <psykose> been goin for long
2023-07-04 13:43:47 <ikke> Yeah, but these are mass added from similar domains 
2023-07-04 13:43:54 <ikke> I've been bulk removing them 
2023-07-04 13:45:35 <psykose> ooh there's been like 3k overnight
2023-07-04 13:45:43 <psykose> was like 9.5 when i went to bed
2023-07-04 13:45:44 <psykose> now 12k
2023-07-04 13:45:56 <ikke> Yeah
2023-07-04 13:46:30 <ikke> I just have a shell + jq snippet that removes latest users from a domain
2023-07-04 15:42:25 <ikke> lots of accounts from a single domain, purging them all
2023-07-04 16:05:21 <psykose> @gmail.com ?
2023-07-04 16:06:18 <ikke> queensmails.com, and other domains as well
2023-07-04 16:06:49 <ikke> lots of <email>@<randomid>.<domain>
2023-07-04 16:07:09 <ikke> @stolts.browndecorationlights.com 
2023-07-04 16:12:38 <psykose> yea
2023-07-04 16:12:53 <ikke> almost below 10k
2023-07-04 16:14:24 <ikke> <10k now
2023-07-04 16:15:55 <ikke> I noticed it because I saw a lot of confirmation mails going through rspamd
2023-07-04 16:26:24 <ikke> psykose: back to 9.5k :)
2023-07-04 16:26:27 <psykose> :D
2023-07-04 16:28:44 <ikke> https://imgur.com/a/6XTSMz4
2023-07-04 16:29:19 <psykose> zoink
2023-07-04 16:29:48 <ikke> They made it a bit too easy this time :D
2023-07-04 16:30:53 <psykose> there's a bunch you probably missed in between
2023-07-04 16:31:11 <psykose> since there's hundreds of @gmail just looking at it right now
2023-07-04 16:33:29 <ikke> yes, I focused on ones with the obvious spam domains
2023-07-04 16:33:47 <mps> algitbot: retry 3.18-stable
2023-07-04 16:34:22 <mps> algitbot: retry master
2023-07-04 16:34:39 <mps> hm
2023-07-04 16:35:07 <psykose> what do you expect that to do
2023-07-04 16:36:16 <mps> looks like some builders are stuck
2023-07-04 16:36:20 <psykose> nope
2023-07-04 16:36:44 <psykose> though if they were that doesn't kill them
2023-07-04 16:38:18 <psykose> well, mozjs102 might be
2023-07-04 16:40:16 <ikke> not anymore
2023-07-04 16:41:06 <ikke> afk
2023-07-04 20:28:14 <ikke> https://docs.gitlab.com/ee/user/admin_area/moderate_users.html#automatically-delete-unconfirmed-users
2023-07-04 20:31:34 <ikke> oh, premium, lesigh
2023-07-04 20:54:28 <ikke> psykose: deleting users under my nose :P
2023-07-04 20:54:39 <psykose> hey i left you a stack of /watch right below
2023-07-04 20:55:35 <ikke> (dlv) p len(issues)
2023-07-04 20:55:38 <ikke> 2919
2023-07-04 20:55:40 <ikke> 3k issues
2023-07-04 20:56:04 <ikke> single user
2023-07-04 20:57:10 <psykose> which one lol
2023-07-04 20:58:29 <ikke> mila west
2023-07-04 20:58:34 <ikke> the one that was just deleted
2023-07-04 21:01:39 <psykose> ah
2023-07-04 21:01:41 <psykose> well
2023-07-04 21:01:47 <psykose> i have deleted like 50 "mila west"'s
2023-07-04 21:02:19 <ikke> lol
2023-07-05 08:26:39 <clandmeter> ikke: are we online again?
2023-07-05 08:26:41 <clandmeter> i guess not
2023-07-05 08:40:15 <ikke> vigir23 is still unreachable 
2023-07-05 09:43:31 <ikke> Hmm
2023-07-05 09:43:59 <ikke> That means it's connecting to zabbix aga 
2023-07-05 09:44:03 <ikke> Again
2023-07-05 10:31:40 <psykose> hmm/
2023-07-05 10:32:46 <ikke> They are available again
2023-07-05 10:33:10 <ikke> > From our side there should not be any unplanned interruptions and the vigir23 now has IPv4 on the uplink, hope that helps a bit
2023-07-05 10:34:47 <psykose> nice
2023-07-05 10:35:10 <psykose> ci still down
2023-07-05 10:36:19 <ikke> need to check that still
2023-07-05 10:39:09 <ikke> dmvpn is now working :)
2023-07-05 10:39:36 <ikke> mps: you should be able to reach you containers now via dmvpn
2023-07-05 10:39:49 <ikke> 172.16.27.110
2023-07-05 10:39:52 <ikke> 172.16.27.111
2023-07-05 10:41:25 <psykose> nice
2023-07-05 10:44:08 <mps> ikke: thank you
2023-07-05 10:49:14 <ikke> time on the CI servers is behind, so certs are not accepted
2023-07-05 10:51:40 <ikke> psykose: all runners should be availabel again
2023-07-05 10:51:47 <psykose> poggers
2023-07-05 10:52:44 <psykose> i pressed 'retry' and it was 'queued for zero seconds' and worked
2023-07-05 10:52:52 <psykose> nah this is a bit too good can you fuck it up a bit please
2023-07-05 10:53:01 <psykose> lets get that to twenty seconds
2023-07-05 10:53:09 <ikke> The latency is a lot better
2023-07-05 10:53:19 <psykose> yea
2023-07-05 10:53:33 <psykose> they're pretty fast
2023-07-05 10:53:34 <psykose> neat
2023-07-05 10:53:54 <ikke> It doesn't have to go via a gre tunnel anymore for ipv4
2023-07-05 10:54:41 <psykose> :)
2023-07-05 15:15:52 <ikke> nice, another gitlab security releasae
2023-07-05 15:16:30 <psykose> awesome
2023-07-05 19:13:16 <ikke> So what builders are we now going to use? mqtt-exec on che-bld-1 seems to have failed on all builders
2023-07-05 19:13:25 <ikke> so it's not active at the moment
2023-07-05 19:17:19 <psykose> probably keep current with vigir disabled, keep ci there
2023-07-05 19:17:23 <psykose> in a bit we'll see how stable
2023-07-06 07:17:13 <clandmeter> ikke: che is ok now?
2023-07-06 07:17:21 <ikke> clandmeter: yes
2023-07-06 07:17:30 <clandmeter> better than before?
2023-07-06 07:28:11 <ikke> I think so. Latency appears more stable and lower ( but still around 30ms from zabbix)
2023-07-06 07:28:51 <ikke> Bandwidth I have not tested yet 
2023-07-06 14:15:05 <ikke> started the 3.16 and 3.15 builders on che-bld-1 again and disabled the other builders that are on usa-bld-1
2023-07-06 14:16:39 <psykose> ~
2023-07-06 14:18:37 <ikke> supervise-daemon[54300]: /usr/sbin/corerad, pid 13988, exited with return code 1
2023-07-06 14:18:39 <ikke> supervise-daemon[54300]: respawned "/usr/sbin/corerad" too many times, exiting
2023-07-06 14:19:06 <ikke> No indication why it crashes
2023-07-06 14:19:56 <psykose> you can run it with the same args to get what it prints
2023-07-06 14:20:00 <psykose> doesn't print it itself tho
2023-07-06 14:20:04 <psykose> unless logged somewhere
2023-07-06 14:20:26 <ikke> Is it possible to log stdout with supervisor-daemon?
2023-07-06 14:21:04 <psykose> output_log=
2023-07-06 14:21:10 <psykose> to some file
2023-07-06 14:24:26 <ikke> the file does not appear
2023-07-06 14:30:56 <ikke> needed to make the file writeable
2023-07-06 14:32:39 <ikke> but it still doesn't write there
2023-07-06 14:33:43 <ikke> ah, it's stderr
2023-07-06 14:34:55 <ikke> that's better
2023-07-06 14:35:42 <ikke> psykose: thanks
2023-07-06 14:51:10 <psykose> yeah stderr is error_log
2023-07-06 14:51:25 <psykose> can make them the same file, all it's doing is > 2> on the process really
2023-07-06 14:52:06 <ikke> I just error_log
2023-07-06 14:52:13 <ikke> there is nothing being sent to stdout
2023-07-06 14:53:31 <psykose> yea
2023-07-06 23:09:03 <psykose> ikke: s390x stuck on go again :D
2023-07-07 05:50:48 <psykose> hm, seems to hang twice
2023-07-07 05:51:58 <ikke> yarn is doing something
2023-07-07 05:52:16 <ikke> 400% cpu usage
2023-07-07 05:53:21 <ikke> not sure if it's anything useful :P
2023-07-07 05:55:03 <psykose> don't think i've seen it literally hang
2023-07-07 10:21:27 <ikke> no output for ages, so it's  gone
2023-07-07 10:56:19 <ikke> oh, apparently they are importing issues in gitlab
2023-07-07 10:58:24 <ikke> That's how they can create so many issues in such a short time
2023-07-07 12:39:22 <ikke>  Oh https://gitlab.com/gitlab-org/gl-security/security-engineering/security-automation/spam/spamcheck
2023-07-07 20:46:59 <psykose> nice
2023-07-08 01:21:30 <psykose> ikke: should pass after another kick without the ui build i guess
2023-07-08 05:21:50 <ikke> hmm, oom
2023-07-08 05:24:28 <psykose> the exact same thing passes right before
2023-07-08 05:24:38 <psykose> same for 3.18
2023-07-08 05:24:41 <psykose> sometimes it just be like that tho
2023-07-08 05:24:56 <psykose> it's one of the uhh either the 4g address limit or the vm.max_map_count thing
2023-07-08 05:25:08 <psykose> probably the former in this case
2023-07-08 05:25:48 <psykose> is vault passing at least
2023-07-08 05:27:13 <psykose> yee
2023-07-08 09:51:29 <mps> I have dilemma, people asks to add linux kernel for apple silicon with GPU enabled and mesa pkg with support for it
2023-07-08 09:51:55 <mps> I have both for long time and both are fine, no big issues
2023-07-08 09:52:58 <mps> so question is to push them to aports or not? anyone want to 'encourage' this
2023-07-08 09:54:11 <mps> more and more people buy apple silicon to run linux on them so it would be nice if we have these things in alpine for them 'ready'
2023-07-08 09:54:26 <mps> most other distros already enabled these
2023-07-08 09:57:19 <psykose> sounds more like something to put in postmarketos
2023-07-08 09:58:31 <mps> people whom I know on asahi community wants to run alpine, no one mentioned pmOS
2023-07-08 09:58:57 <psykose> pmos is alpine just with some kernel/support packages if one wants to run it that way
2023-07-08 09:59:03 <psykose> it's the same alpine repos with some extra stuff
2023-07-08 09:59:28 <mps> it is not same and you know this very well I think
2023-07-08 09:59:31 <psykose> we explicitly reject any attempt to put xyz-whatever kernels in alpine forever, so i'm not sure why asahi is some magic exception
2023-07-08 09:59:34 <psykose> yeah it is the same
2023-07-08 10:00:00 <mps> because people wants alpine
2023-07-08 10:00:51 <psykose> and said people that want alpine should know that all the device-specific alpine-related stuff is in pmos
2023-07-08 10:02:04 <mps> most of these people already run alpine on other machine types and simply want to do same on apple silicon
2023-07-08 10:02:56 <psykose> it is pretty much the same and i already explained why
2023-07-08 10:03:59 <mps> I don't see pmOS here https://github.com/AsahiLinux/docs/wiki/SW%3AAlternative-Distros
2023-07-08 10:04:14 <mps> and alpine is there for long time
2023-07-08 10:04:50 <psykose> because you added the kernel to alpine and put it in the docs, yes
2023-07-08 10:05:41 <mps> yes, people read IRC logs, some sites and find alpine, and asks for help to install alpine
2023-07-08 10:06:10 <mps> pmOS people could make their packages from alpine if they want
2023-07-08 10:06:45 <mps> in alpine we already have m1n1, asahi u-boot and kernel
2023-07-08 10:07:26 <mps> we can even enable/add gpu flavor in current linux-asahi
2023-07-08 12:46:06 <ikke> fun, dendrite is crashing again
2023-07-08 17:27:39 <ikke> :/
2023-07-08 17:38:21 <psykose> there was an update for it iirc
2023-07-08 17:38:24 <psykose> maybe they fixed it
2023-07-08 17:39:14 <ikke> Was just pulling new images
2023-07-08 17:39:57 <ikke> but still crashing
2023-07-08 17:40:22 <ikke> Tried to remove the offending event from the DB, but no luck
2023-07-08 17:41:40 <ikke> I guess I'll have to report it
2023-07-08 18:02:29 <ikke> Someone is creating accounts with all kinds of dot variations of the same gmail account (but the suffix is always the same)
2023-07-08 19:31:02 <ikke> Ok, nice,  got grpc to my custom spamcheck server working
2023-07-08 19:31:06 <psykose> ~
2023-07-08 19:31:08 <ikke> (on gitlab-test)
2023-07-08 19:31:09 <psykose> progress
2023-07-08 19:31:16 <ikke> It just logs for now
2023-07-08 19:36:00 <ikke> ok, enabled it now on gitlab.a.o
2023-07-08 19:37:04 <psykose> lol
2023-07-08 20:35:05 <ikke> no data as of yet :(
2023-07-08 20:37:22 <psykose> like it just doesn't work?
2023-07-08 20:38:03 <ikke> No idea
2023-07-08 20:38:12 <ikke> I tested it on gitlab-test, created an issue, and saw it logged
2023-07-08 20:38:20 <ikke> but maybe no issues were created yet
2023-07-08 20:38:56 <psykose> ah
2023-07-08 20:38:57 <psykose> yeah
2023-07-08 20:38:58 <psykose> no issues
2023-07-09 04:18:27 <psykose> there was one spam one somewhere
2023-07-09 04:18:36 <psykose> see if that got picked up
2023-07-09 04:18:40 <psykose> also riscv stuck
2023-07-09 05:19:13 <ikke> just got 3 messages, none of them spam
2023-07-09 05:30:11 <ikke> so importing issues just bypasses the spamchecker, great
2023-07-09 05:30:59 <psykose> it was one off and on your new sign tool repo
2023-07-09 05:31:06 <ikke> yeah, that one I saw
2023-07-09 05:31:13 <ikke> I mean, got a notification for in email
2023-07-09 05:31:19 <ikke> but nothing in spamcheck
2023-07-09 05:31:21 <psykose> ahh
2023-07-09 05:31:42 <psykose> that one couldn't have been imported, could it
2023-07-09 05:31:46 <ikke> We could block post requests to /import_csv
2023-07-09 05:31:49 <psykose> i thought that was random new repos only
2023-07-09 05:31:58 <psykose> yeah i don't see a reason to allow imports at all i guess
2023-07-09 05:31:59 <ikke> You need to be developer on the project
2023-07-09 05:32:06 <psykose> yeah, so that issue was not an import
2023-07-09 05:32:13 <psykose> but still went thru without being logged in spamcheck?
2023-07-09 05:32:49 <ikke> Just saw 3 issues, all for aports
2023-07-09 05:33:06 <psykose> weirx
2023-07-09 05:33:08 <psykose> weird*
2023-07-09 05:35:16 <ikke> Just created an issue myself manually, and I do see that one
2023-07-09 06:02:14 <ikke> https://gitlab.alpinelinux.org/import_csv :)
2023-07-09 06:06:09 <ikke> I've adjusted it to only match /-/issues/import_csv now
2023-07-09 06:11:33 <psykose> :D
2023-07-09 06:27:42 <psykose> still had to kick rv64
2023-07-09 06:29:52 <ikke> done
2023-07-09 14:37:34 <ikke> AH, seeing spam messages now
2023-07-09 14:38:03 <ikke> and import_csv being blocked :)
2023-07-09 16:59:34 <psykose> :D
2023-07-09 17:25:15 <ikke> "POST /lillyrubio/watch/-/issues/import_csv HTTP/1.1" 403 548 "https://gitlab.alpinelinux.org/lillyrubio/watch/-/issues"
2023-07-09 17:26:51 <ikke> much wow, such empty: https://gitlab.alpinelinux.org/lillyrubio/watch/-/issues
2023-07-09 17:27:05 <psykose> poggers
2023-07-09 17:27:11 <psykose> well that's half the puzzle
2023-07-09 17:27:16 <psykose> getting close
2023-07-09 17:27:21 <psykose> the other half is the non-imported issue spam
2023-07-09 17:27:29 <ikke> yes, will work on that
2023-07-09 17:27:36 <psykose> good work!
2023-07-09 17:28:09 <ikke> I found it through rspamd, saw a mail about "imported issues"
2023-07-09 17:28:52 <ikke> otherwise I would not have figured how they added the issues
2023-07-11 12:39:46 <mps> what keys are in alpine-keys
2023-07-11 12:40:18 <ikke> The keys used by the builders, or that have been used in the past
2023-07-11 12:40:22 <ikke> how so?
2023-07-11 12:42:24 <mps> don't see any description except dir hierarchy
2023-07-11 12:45:47 <ikke> The APKBUILD shows what keys belong to what arch
2023-07-11 12:46:31 <ikke> and also, https://alpinelinux.org/releases.json
2023-07-11 12:49:23 <mps> ikke: thanks
2023-07-11 14:22:48 <psykose> ikke: the arm ci should probably not be jobs=24 unless it's shared with the builders
2023-07-11 14:23:47 <ikke> The vms only have 24 cores
2023-07-11 14:24:03 <psykose> ah
2023-07-11 14:24:28 <ikke> I divided the cores over the vms 
2023-07-11 14:24:56 <ikke> Not sure if oversubscribing makes sense since they all build generally at the same time
2023-07-11 14:26:44 <psykose> most builds aren't super linear with using all the cores at once
2023-07-11 14:26:52 <psykose> + sometimes armhf disabled etc
2023-07-13 13:33:19 <ikke> psykose: https://gitlab.alpinelinux.org/alpine/infra/compose/pg-upgrade
2023-07-13 13:34:45 <psykose> aw neat
2023-07-13 13:34:47 <psykose> good work :D
2023-07-13 13:35:40 <ikke> Wanted to test gitlab 16 locally, but need to migrate the DB :D
2023-07-13 13:37:45 <psykose> ~
2023-07-13 14:01:39 <ikke> almost at 12k users again :(
2023-07-13 14:01:48 <ikke> need to clean up all these unverified accounts
2023-07-13 16:45:01 <ikke> upgraded my local instance succesfully to gitlab 16 :)
2023-07-13 16:45:07 <ikke> now going to work on gitlab-test
2023-07-13 17:07:07 <ikke> interesting, this time it's only vigir23 being unreachable
2023-07-14 03:57:24 <psykose> ~
2023-07-14 03:57:26 <psykose> yeah it's a bit flake
2023-07-14 05:03:00 <psykose> riscv64 is also stuck
2023-07-14 10:13:05 <ikke> gitlab-test is now running gitlab 16
2023-07-14 11:29:46 <ikke> Trying to figure out why Gitlab::HTTP has issues with tls SNI, but Net::HTTP does not, and neither does HTTParty (which Gitlab::HTTP uses)
2023-07-14 11:30:05 <ikke> HTTParty.get 'https://version.gitlab.com' -> success
2023-07-14 11:30:28 <ikke> Gitlab::HTTP.get('https://version.gitlab.com') -> self-signed certificate (due to missing SNI)
2023-07-14 11:31:11 <ikke> Net::HTTP.get(URI('https://version.gitlab.com')) -> success
2023-07-14 12:01:13 <ikke> psykose: I think I'm closer. https://gitlab.com/gitlab-org/gitlab/-/blob/v16.0.7-ee/lib/gitlab/http_connection_adapter.rb#L50 returns an URI object where the hostname is an ip address
2023-07-14 12:01:22 <ikke> So that would disable SNI
2023-07-14 13:12:21 <ikke> ttps://gitlab.com/gitlab-org/gitlab/-/issues/413528
2023-07-14 13:12:23 <ikke> https://gitlab.com/gitlab-org/gitlab/-/issues/413528
2023-07-14 13:17:18 <ikke> Ok, there is an option to disable dns rebinding prevention, which fixes the issue
2023-07-14 13:17:23 <ikke> or works around
2023-07-14 13:22:04 <ikke> https://github.com/ruby/net-http/issues/141
2023-07-14 15:39:35 <ikke> One downside of gitlab 16 is that all access tokens have an expiry date now :*
2023-07-14 15:40:11 <ikke> That's going to be really anoying :/
2023-07-14 17:17:35 <ikke> gitlab upgraded to 16.0. MR still appear empty immediately after opening them
2023-07-14 17:19:03 <ikke> merge requests are working
2023-07-14 17:46:13 <psykose> ~
2023-07-14 17:46:32 <psykose> good work :)
2023-07-14 17:59:12 <ikke> If 12 months from new things break, it's because access tokens have expired :.
2023-07-14 20:08:08 <ikke> So the automaintainer says it doesn't find any maintainers on the merge requests
2023-07-14 21:31:19 <psykose> seems to work again though
2023-07-14 21:31:22 <psykose> i guess you fixed it
2023-07-14 21:33:56 <ikke> Nope, didn't change anything 
2023-07-14 21:34:16 <psykose> :)
2023-07-14 21:34:23 <psykose> love it when that happens :D
2023-07-14 21:34:31 <ikke> Was just debugging / checking the logs 
2023-07-20 09:57:30 <mps> anyone have rust pkgs from version 1.67 and up somewhere
2023-07-20 09:58:14 <mps> hope is that someone didn't cleaned archive
2023-07-20 09:59:47 <psykose> ikke: could you kick ppc64le
2023-07-20 10:22:46 <ncopa> i can kick it
2023-07-20 10:23:26 <psykose> thanks
2023-07-21 09:01:56 <clandmeter> cz mirror will have a short downtime tomorrow night. i don't expect anybody to notice it. but just in case..
2023-07-21 19:23:18 <ikke> someone misconfigured their spam email address generator :D
2023-07-21 19:23:27 <ikke> {lucia|mariana|lorena|jimena|dolores|brenda|silvana|renata}{1|2|3|4|5|6|7|8|9}@mamerto.online
2023-07-21 20:35:58 <mps> psykose: why we need to rebuild linux-edge with every new gcc
2023-07-21 21:59:43 <psykose> for out of tree modules
2023-07-22 00:52:34 <psykose> the kernel build system requires `gcc --version | head -n1` to match for its own build and any later built external modules
2023-07-22 00:52:39 <psykose> been that way forever
2023-07-22 00:53:25 <psykose> so if you upgrade gcc and gcc --version changes and you try to build an out of tree module, it fails and just says it expected the previous version
2023-07-22 08:20:08 <mps> sorry, my question was not clear, and not technical. I know this. I meant to ask do we as distro provider need to do this. but 'morning is wiser than evening' and I think now all is ok
2023-07-22 08:20:44 <mps> and I forgot that we build -dev subpkg :) thanks for reminding me
2023-07-22 19:12:43 <Hello71> psykose: only if you use gcc plugins, right?
2023-07-22 19:13:23 <psykose> you mean like if you disable all of them in kernel config it won't check entirely later? probably
2023-07-22 19:13:27 <psykose> idk who disables that tho
2023-07-22 19:13:28 <psykose> didn't test
2023-07-22 19:14:33 <Hello71> I mean CONFIG_GCC_PLUGINS=n
2023-07-22 19:14:45 <psykose> yeah
2023-07-22 19:14:47 <Hello71> I guess if you use clang? :p
2023-07-22 19:15:00 <psykose> well you can also just =n it since i'm not aware of any plugins actually in use..
2023-07-22 19:15:02 <psykose> they're all dev stuff
2023-07-22 19:15:17 <Hello71> also hardening
2023-07-22 19:15:25 <psykose> aha right
2023-07-22 19:15:28 <psykose> the oneee thing
2023-07-22 19:19:27 <mps> gcc plugins are disabled in linux-edge
2023-07-22 19:20:09 <mps> when I find free time I'll check all this again
2023-07-23 19:28:16 <psykose> mps: yeah i checked
2023-07-23 19:28:28 <psykose> the warning is printed, but the actual error is only for gcc plugin
2023-07-23 19:28:33 <psykose> so edge doesn't have to be rebuilt after all
2023-07-23 19:28:33 <psykose> :D
2023-07-23 19:54:31 <mps> psykose: thanks for info. you saved me some time ;)
2023-07-24 19:09:46 <clandmeter> gitlab going down? or is that another host?
2023-07-24 19:09:57 <clandmeter> deu1-dev1
2023-07-24 19:10:05 <ikke> That's another host
2023-07-24 19:10:57 <ikke> hosting cgit, algitbot, mirrors.a.o, chat.a.o and secdb
2023-07-24 19:11:10 <ikke> deu7 is wg.a.o
2023-07-24 19:11:19 <clandmeter> ok, seems they need to move it to another host
2023-07-24 19:11:43 <ikke> yes, I see
2023-07-24 20:07:44 <ikke> clandmeter: fyi, they'll try to live migrate, in which case it should remain online
2023-07-25 08:34:11 <mps> my riscv64 lxc doesn't answer to ping
2023-07-25 08:35:19 <mps> it is not started?
2023-07-25 08:52:54 <ikke> It's not set to auto start. Started it now
2023-07-25 08:53:54 <mps> ikke: thank you
2023-07-25 12:41:02 <psykose> ikke: did you ever enable that gitlab cache
2023-07-25 12:48:13 <ikke> Yes, it's active 
2023-07-25 12:51:06 <psykose> sweet
2023-07-26 03:33:01 <psykose> ikke: could you start my rv container too
2023-07-26 04:44:32 <ikke> done
2023-07-26 04:58:55 <psykose> thanks
2023-07-26 05:01:25 <ikke> i think it was because lxc was started before qemu-binfmt
2023-07-26 05:01:49 <ikke> added rc_need=qemu-binfmt to /etc/conf.d/lxc
2023-07-26 05:03:36 <psykose> makes sense
2023-07-26 09:51:45 <mps> do we have automation script/method to rebuild dependent pkgs when upgrading some of base pkgs?
2023-07-26 09:52:20 <mps> or simply rebuild all
2023-07-26 09:53:04 <mps> skimming trough glib revdep shows about 450 pkgs
2023-07-26 09:54:06 <mps> though in my test upgrading glib doesn't require rebuild of dependent pkgs
2023-07-26 09:57:07 <ikke> ap revdeps + apkgrel -a
2023-07-26 09:58:38 <mps> literally this?
2023-07-26 10:13:39 <ikke> Not literally 
2023-07-26 10:16:27 <mps> I would bet that someone have ready script for this
2023-07-26 10:18:57 <mps> though I upgraded glib locally and for now didn't noticed any problem with dependent pkgs
2023-07-27 02:51:49 <psykose> there's no new glib version though
2023-07-27 02:52:00 <psykose> it also doesn't need any rebuilds
2023-07-27 04:19:31 <psykose> glib is abi-stable every release, and the odd-versions (2.77) are unstable, to be clear
2023-07-27 12:13:47 <mps> I can confirm that also glib 2.77 ABI is stable
2023-07-27 12:18:48 <psykose> the abi is always stable
2023-07-27 12:18:55 <psykose> the actual library is not
2023-07-27 12:22:12 <mps> well, running edge so ...
2023-07-27 12:28:48 <psykose> if you mean you built it and use it then yeah go for it
2023-07-27 12:29:01 <psykose> if you mean upgrade to 'random unstable library prerelease' in aports, then no
2023-07-27 12:29:47 <mps> yes, locally ofc
2023-07-27 12:29:53 <psykose> yeah works fine for that
2023-07-27 12:30:01 <psykose> it's only the usual stable-abi-forward-new-symbols
2023-07-27 12:30:25 <psykose> so there's a chance that if you downgrade things built against 2.77 won't work on 2.76 (some define macro used a different symbol, etc)
2023-07-27 12:31:47 <mps> right
2023-07-27 12:33:05 <mps> but some 'smart' upstream software require this
2023-07-27 12:34:16 <psykose> is it telegram-desktop? :p
2023-07-27 12:34:23 <psykose> they need 2.77 glib+glibmm
2023-07-27 12:34:24 <psykose> annoying
2023-07-27 12:35:21 <mps> :)
2023-07-27 12:35:32 <mps> 'smart' ;p
2023-07-27 12:36:14 <psykose> hehe
2023-07-27 18:23:04 <psykose> ikke: do you see the recent spam to -devel
2023-07-27 18:23:09 <psykose> (just to check the rspamd tuning)
2023-07-27 18:24:59 <ikke> Last thing I've seen is an empty reply to the wiki relicense thread
2023-07-27 18:25:25 <ikke> My local spam filter is apparently blocking it
2023-07-27 18:33:03 <psykose> yep
2023-07-27 18:33:11 <psykose> you mentioned that before and told me to tell you when it shows up
2023-07-27 18:33:12 <psykose> :)
2023-07-27 18:33:17 <psykose> there's been a few
2023-07-27 18:33:19 <psykose> on -infra too
2023-07-27 18:34:40 <ikke> Yeah, I'm learning / removing them now
2023-07-27 18:39:04 <ikke> psykose: where in -infra?
2023-07-27 18:39:16 <psykose> oh oops
2023-07-27 18:39:17 <psykose> hmm
2023-07-27 18:39:29 <psykose> i confused it with mirrors
2023-07-27 18:39:30 <psykose> haha
2023-07-27 18:39:32 <ikke> heh
2023-07-27 18:39:33 <psykose> nvm
2023-07-31 18:37:02 <mps> strange `ap revdep perl-xs-parse-sublike-dev` gives 'sh: .: line 5: can't open '/../../*/lemmy/APKBUILD': No such file or directory'
2023-07-31 18:37:20 <ikke> you need to run it in the specific repo diir
2023-07-31 18:37:22 <ikke> dir
2023-07-31 18:37:28 <ikke> for example main/ or community/
2023-07-31 18:41:21 <mps> I run it in testing
2023-07-31 18:42:19 <ikke> I see, let me check
2023-07-31 18:42:32 <mps> and whatever pkg I use msg is sam
2023-07-31 18:42:36 <mps> same*
2023-07-31 18:43:17 <ikke> ah, lemmy-ui is the culprit
2023-07-31 18:44:01 <mps> '_translations_commit="$( . $startdir/../../*/lemmy/APKBUILD; echo $_translations_commit )"' ;-)
2023-07-31 18:44:21 <ikke> yesd
2023-07-31 18:44:31 <ikke> let me fix that
2023-07-31 18:49:07 <ikke> mps: if you pull, it should be fixed
2023-07-31 18:49:49 <mps> yes it is fixed
2023-07-31 18:49:52 <mps> thanks
2023-07-31 19:19:31 <ncopa> this may need some more time from CI https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/49346
2023-07-31 19:19:44 <ncopa> how can I give more CI time?
2023-07-31 19:20:18 <pj> iirc, the user had to set that themselves in CI settings
2023-07-31 19:20:30 <ikke> yes, or I do it for them :P
2023-07-31 19:20:58 <pj> that's also a possibility
2023-07-31 19:21:25 <ikke> ncopa: you can do that in the CI/CD settings on their fork
2023-07-31 19:22:03 <ikke> I've changed it, but the pipeline needs to be restarted before it takes effect
2023-07-31 19:22:05 <pj> https://gitlab.alpinelinux.org/USER/aports/-/settings/ci_cd
2023-07-31 19:22:11 <ikke> indeed
2023-07-31 19:22:32 <ncopa> thanks!