2024-09-02 09:01:20 <clandmeter> ikke: im still having issues with dmvpn
2024-09-02 09:01:24 <clandmeter> to che
2024-09-02 09:01:42 <clandmeter> should we check the mtu over mpls?
2024-09-02 09:42:43 <clandmeter> ikke: what was the cmd to purge a pkg from cdn?
2024-09-02 09:43:48 <clandmeter> found it
2024-09-02 09:44:23 <ikke> ftr: https://gitlab.alpinelinux.org/alpine/infra/repo-tools#quick-usage
2024-09-02 09:45:41 <clandmeter> ah ok
2024-09-02 09:45:48 <clandmeter> cely: ^
2024-09-02 09:47:55 <clandmeter> so this will also purge all subpkgs
2024-09-02 09:47:59 <clandmeter> i guess :)
2024-09-02 09:51:05 <ikke> Yes, with --origin
2024-09-02 09:51:46 <clandmeter> i was not aware we had this tool
2024-09-02 09:51:50 <ikke> And repo-tools is available in aports 
2024-09-02 09:51:56 <clandmeter> i am just looking a little deeper
2024-09-02 09:52:14 <clandmeter> it uses the apkindex i guess
2024-09-02 09:52:40 <clandmeter> https://gitlab.alpinelinux.org/alpine/go/-/blob/master/repository/apkindex.go?ref_type=heads
2024-09-02 09:56:33 <ikke> Yes, it does
2024-09-02 11:16:23 <clandmeter> ikke: can you delete the logfile from builder? 
2024-09-02 11:16:36 <clandmeter> seems x86* also affected
2024-09-02 11:17:33 <ikke> Not right now 
2024-09-02 15:15:02 <ikke> clandmeter: I don't see a large log file, where should it be, in distfiles?
2024-09-02 15:16:22 <clandmeter>  /var/cache/distfiles/buildlogs/build-edge-loongarch64/community/datovka/datovka-4.23.8-r0.log
2024-09-02 15:16:46 <clandmeter> i hear x86* has the same issues
2024-09-02 15:16:56 <clandmeter> i patched it, but could be the log is still large
2024-09-02 15:17:01 <ikke> -rw-r--r--    1 1000     1000      943.1K Sep  2 14:45 datovka-4.24.1-r0.log
2024-09-02 15:17:17 <clandmeter> yes thats the latest build
2024-09-02 15:17:24 <clandmeter> it does not keep ollder?
2024-09-02 15:18:48 <ikke> The rsync job would remove older files
2024-09-02 15:18:58 <ikke> though, not buildlogs..
2024-09-02 15:28:46 <clandmeter> was expecting it to not delete until so many days or so
2024-09-02 15:28:52 <clandmeter> not sure how that logic is applied
2024-09-05 06:28:57 <mps> https://qemu-project.gitlab.io/qemu/about/removed-features.html => 'Nios II CPU (removed in 9.1)'
2024-09-05 06:29:28 <mps> ncopa: we have to remove it in upgrade?
2024-09-05 07:03:44 <ncopa> mps: how do you mean? it is upstream qemu that removes it, not us
2024-09-05 07:04:03 <mps> yes
2024-09-05 07:04:26 <mps> so we also have to remove subpkg
2024-09-05 07:04:46 <mps> 'have' = 'must'
2024-09-05 07:05:59 <ikke> Yes, so we indeed need to remove the subpackage 
2024-09-05 07:06:18 <mps> ok
2024-09-05 07:07:39 <ncopa> yes, we need to remove the subpkg
2024-09-05 07:08:43 <mps> ok ok
2024-09-05 07:09:01 <mps> I already built 9.1 and testing a bit
2024-09-05 16:27:46 <mps> bug: /usr/sbin/setup-devd: .: line 6: can't open '/lib/libalpine.sh': No such file or directory
2024-09-05 16:27:55 <mps> it is in /usr/lib/
2024-09-05 16:27:59 <ikke> mps: should be fixed in about 15 minutes
2024-09-05 16:28:15 <mps> ikke: nice, thanks
2024-09-05 16:28:19 <ikke> Not me
2024-09-05 16:28:21 <ikke> craftguy
2024-09-05 16:28:27 <ikke> https://dev.alpinelinux.org/~clandmeter/other/forum.alpinelinux.org/
2024-09-05 16:28:33 <mps> whoever, thanks
2024-09-05 16:28:34 <ikke> sorry: https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/71449
2024-09-05 16:28:58 <mps> for now I will fix it in source locally, so I'm not in a hurry
2024-09-06 04:44:10 <cely> I would appreciate it if someone could restart the build of community/ldc on the x86_64 builder
2024-09-06 04:45:03 <cely> I enabled more tests for ldc in 3400fccdb but those were mostly for aarch64 and loongarch64
2024-09-06 04:45:50 <cely> For x86_64, that commit should enable just 4 additional tests, the ones matched by `druntime-test-gc`
2024-09-06 04:47:06 <cely> Comments in the apkbuild says those tests "randomly fails due to timeout", but they passed in CI, and it didn't take this long
2024-09-06 04:47:29 <cely> So, something is probably different on the builder
2024-09-06 04:48:08 <ikke> What could be different that would take tests to take longer / timeout
2024-09-06 04:48:12 <cely> If they continue to fail, then `druntime-test-gc` probably has to be skipped again
2024-09-06 04:48:49 <cely> I don't think they timeout, but they take longer instead
2024-09-06 04:49:25 <ikke> I can't strace, but I don't see a lot happening,
2024-09-06 04:49:39 <cely> Is it running ctest?
2024-09-06 04:50:07 <ikke> https://tpaste.us/yYVl 
2024-09-06 04:51:58 <cely> Yeah, running ctest, and running druntime-test-gc-release
2024-09-06 04:52:00 <ikke> https://tpaste.us/YYQz 
2024-09-06 04:52:21 <ikke> I think it's deadlocked
2024-09-06 04:52:37 <cely> So, `druntime-test-gc-release` is a bit problematic on the builder
2024-09-06 12:39:55 <ncopa> ikke: can you help me with posting a toot of the new releases?
2024-09-06 12:41:06 <ikke> Yes
2024-09-06 12:43:29 <ncopa> thank you!
2024-09-10 10:57:09 <mps> heh, interesting use of alpine https://github.com/zhovner/OneFileLinux
2024-09-12 13:39:33 <clandmeter> ikke: can you silence the runner?
2024-09-12 13:39:43 <clandmeter>  It’s turned off
2024-09-12 13:39:54 <ikke> 'shared-runner clandmeter (loongarch64) offline', right?
2024-09-12 13:40:06 <ikke> Just pause it, or remove it?
2024-09-12 13:40:18 <clandmeter> I already paused it some time ago
2024-09-12 13:40:26 <ikke> hmm, then it should not alert
2024-09-12 13:40:33 <clandmeter>  but it still screams my name :)
2024-09-12 13:40:37 <ikke> :D
2024-09-12 13:40:51 <clandmeter>  I’m only on mobile
2024-09-12 13:41:12 <clandmeter> clandmeter should be the desktop
2024-09-12 13:49:46 <ikke> For the new-style registered runners I need to explicitly look at the paused flag
2024-09-12 14:02:11 <ikke> clandmeter: I've removed the runner. If you want to use it again, you have to reregister it ;-)
2024-09-12 17:23:41 <clandmeter> That’s ok
2024-09-13 09:48:58 <ikke> Lol
2024-09-13 09:49:12 <ikke> (but also what)
2024-09-13 10:20:24 <ikke> probably a network blip
2024-09-13 10:28:36 <ikke> seems like docker crashed
2024-09-13 14:55:24 <mps> does alpine really wants usr-merge???
2024-09-13 15:08:19 <ikke> mps: It will be most likely in the next FHS, and we tend to follow that
2024-09-13 15:22:40 <mps> my vote is against this
2024-09-13 15:24:05 <ikke> The TSC already agreed to implement it
2024-09-13 15:24:44 <mps> nice, good reason for me to learn macos and move there 
2024-09-15 07:30:51 <mps> oh, do we share /var/cache/distfiles with builder and developers lxcs on pioneer?
2024-09-15 07:32:15 <mps>  /dev/sda1 on /var/cache/distfiles type ext4 (rw,relatime)
2024-09-15 07:32:35 <mps> just removed all logs there
2024-09-15 07:32:57 <mps> :/
2024-09-15 08:42:58 <ikke> Builders are a separate host 
2024-09-15 11:06:10 <ikke> clandmeter: it works to de-assign the elastic ip address from the mirror servers, traffic is automatically rerouted to one of the other servers
2024-09-15 11:07:46 <ikke> upgrading  ^
2024-09-15 11:13:23 <ikke> https://tpaste.us/6KQD 
2024-09-15 11:30:32 <ikke> upgrading usa.t1 now
2024-09-17 18:19:06 <mps> distfiles.alpinelinux.org is not accessible
2024-09-17 18:19:31 <mps> ah now it is
2024-09-17 18:19:54 <mps> algitbot: retry master
2024-09-17 18:22:20 <mps> algitbot: retry master
2024-09-22 10:00:40 <cely> I think the build-edge-loongarch64 is stuck on testing/perl-anyevent-riperedis, most likely in check()
2024-09-22 10:19:51 <ikke> checking
2024-09-22 10:25:48 <cely> Thanks
2024-09-23 19:14:58 <ncopa> i have forgotten the IP address of the loongarch64 builder
2024-09-23 19:15:25 <ncopa> and I see only riscv64 machines in /var/lib/misc/dnsmasq.leases 
2024-09-23 19:21:43 <ncopa> i found it
2024-09-24 14:54:03 <cely> I think testing/perl-anyevent-riperedis can be terminated again on build-edge-loongarch64, i've committed a workaround, and it shouldn't block the builder like that again
2024-09-24 15:11:02 <ikke> done
2024-09-24 15:16:01 <cely> Thanks
2024-09-25 00:26:42 <omni> some CI build-x86_64 issue?
2024-09-25 05:10:38 <ikke> omni: I don't see anything in particular going wrong?
2024-09-25 08:43:07 <omni> ikke: it was perhaps just temporary, but occurred in https://gitlab.alpinelinux.org/alpine/aports/-/jobs/1530526 and https://gitlab.alpinelinux.org/alpine/aports/-/jobs/1530546
2024-09-25 09:05:01 <ikke> omni: looks like an issue with docker hub
2024-09-26 05:27:36 <ikke> Seems like an issue with ipv6 ^
2024-09-26 05:27:41 <ikke> can ping them over ipv4
2024-09-26 05:30:03 <ikke> https://status.equinix.com/incidents/7gjtwr1hdky8
2024-09-26 05:43:49 <ikke> I made the icmp ping loss check for nld-bld-1/2 a bit more stable
2024-09-26 06:30:47 <ncopa> I have deleted one of the scaleway riscv64 CI runners
2024-09-26 06:32:42 <ncopa> i wonder if I can delete the other too?
2024-09-26 06:33:14 <ikke> Did you remove the runner from gitlab as well?
2024-09-26 06:36:21 <ncopa> yes
2024-09-26 06:36:50 <ncopa> i think i can delete the other scaleway instance. we have 3 riscv64 builder, which should be enough
2024-09-26 06:37:17 <ncopa> I can add it back if needed in the future
2024-09-26 06:40:12 <ikke> Nod
2024-09-26 07:45:38 <ncopa> both scaleway instances are gone now
2024-09-26 07:45:48 <ncopa> and I deleted the runners in gitlab
2024-09-26 15:19:36 <iggy>  is it possible to scale the scaleway instances when needed? I'm not sure how long they take to spin up, but a little bit of lag before jobs start shouldn't be too bad? Do they even bill them like that?
2024-09-26 15:20:14 <ikke> There is nothing automated for that
2024-09-26 15:21:11 <ikke> Maybe with docker-machine, but no idea if that support scaleway
2024-09-26 15:21:30 <ikke> scaleway does bill per usage
2024-09-26 15:33:32 <ncopa> everything is possible
2024-09-26 15:33:41 <ncopa> the impossible just takes longer time
2024-09-26 15:39:10 <ncopa> im thinking of how to create disk images from CI. what is easiest? give a gitlab runner real root access so loopback devices can be mounted and do `apk add --root ...` or run a docker build to create a docker image?
2024-09-26 15:39:57 <ncopa> from docker image we can export to tar and do tar2ext4 something, or extract the files under fakeroot and then mkfs.ext4 with a directory input
2024-09-26 15:40:18 <ncopa> the problem is to run the install scripts and triggers
2024-09-28 06:32:01 <ikke> clandmeter: can we remove apkbrowser from .attic on gbr2? It's a duplicate and takes up quite some space
2024-09-28 13:45:02 <clandmeter> i guess anything in .attic can go
2024-09-28 13:45:31 <ikke> Just wanted to make sure there isn't anything in there you wanted to keep for some reason
2024-09-29 20:06:33 <clandmeter> ikke: do you know when releases.json updates edge arches?
2024-09-29 20:08:47 <ikke> Manually here: https://gitlab.alpinelinux.org/alpine/infra/alpine-mksite/-/blob/master/alpine-releases.conf.yaml?ref_type=heads
2024-09-29 20:19:44 <clandmeter> ok, so we will only add loongarch when testing is finished?
2024-09-29 20:20:03 <ikke> I suppose so
2024-09-29 20:20:18 <ikke> CI is a bit broken as well since I switched it to the new builders
2024-09-29 20:20:36 <ikke> repo from the new builder
2024-09-29 20:22:08 <clandmeter> a bit?
2024-09-29 20:22:31 <ikke> Packages for testing fail because it cannot find the testing repo
2024-09-29 20:22:43 <clandmeter> right
2024-09-29 20:22:48 <clandmeter> as expected
2024-09-29 22:02:53 <clandmeter> i switched pkgs to new version
2024-09-29 22:11:13 <clandmeter> the old one should be at pkgs-old.a.o
2024-09-30 08:05:59 <ncopa> looks nice
2024-09-30 21:46:04 <ptrc> pkgs.a.o doesn't seem to work with contents anymore?
2024-09-30 21:46:13 <ptrc> oh
2024-09-30 21:46:19 <ptrc> it has main repo selected by default
2024-09-30 21:46:28 <ptrc> this is silly