2020-09-01 18:15:48 Gbr will have maintenance again 2020-09-01 19:24:17 ok 2020-09-01 20:34:45 Oh, TIL: https://docs.docker.com/engine/reference/run/#pid-settings---pid 2020-09-01 20:35:13 allows you to 'join' containers, ie, when you want to troubleshoot running containers and don't want to install all kinds of tools there 2020-09-01 20:35:19 (sidecar idea) 2020-09-01 20:39:19 you can also do similar with nsenter, but it requires you to know a lot more about how the system is operating 2020-09-01 20:39:32 nod 2020-09-02 09:08:36 ikke: Could you restart the asgen container? :) 2020-09-02 09:09:07 Cogitri: And I assume fetch the latest version? 2020-09-02 09:10:02 Ah yes, I changed the asgen config 2020-09-02 09:10:14 So the docker-compose repo should be updated 2020-09-02 09:10:18 right 2020-09-02 09:10:25 Will do that in about an hour 2020-09-02 10:01:54 Okie, thanks 👍 2020-09-02 10:37:28 Cogitri: apstream-generator has been restarted (and the repository updated) 2020-09-02 10:56:12 Thank you :) 2020-09-02 13:58:47 Sorry, but would it be possible to get an armv7 container so I can debug the webkit2gtk failures? 2020-09-02 14:00:23 Cogitri: I can try now. is that some MR or already in git repo 2020-09-02 14:04:40 It's in master 2020-09-02 14:04:53 Currently doesn't work, I think for some reason it detects armv7 as aarch64 2020-09-02 14:04:59 But not sure 2020-09-02 14:05:51 just started build 2020-09-02 14:09:11 Thanks! :) 2020-09-02 15:23:35 Cogitri: build pass on armv7 lxc 2020-09-02 15:23:52 Huh 2020-09-02 15:23:58 Of 2.29.91-r0, right? 2020-09-02 15:24:27 9b09c5117dcccee35e6c8387d68aeec904a8da77 I think 2020-09-02 15:24:51 Master updated to 2.29.91 2020-09-02 15:24:58 this is latest in repo 2020-09-02 15:26:20 hmm, git pull didn't worked for some reason, I did it before starting build, and now I tried again and I see update 2020-09-02 15:26:36 ok, will try with this new again 2020-09-02 15:26:59 Oh okie, thanks :) 2020-09-02 15:27:26 d9dc2fed9784f93b24812476b9a70ed98369450b 2020-09-02 16:32:22 Cogitri: build failed, log file is 2MB 2020-09-02 16:32:46 I guess it failed with the same error as on the builder? 2020-09-02 16:32:54 Something about dumpJITMemory 2020-09-02 16:33:21 yes 2020-09-02 16:33:38 ../Source/JavaScriptCore/assembler/LinkBuffer.cpp:380:18: error: 'dumpJITMemory' is not a member of 'JSC' 2020-09-02 16:34:47 Did you checkout my MR? That was broken 😅 2020-09-02 16:35:07 I wanted a build straight off master 2020-09-02 16:35:14 ah, you have another MR not merged? 2020-09-02 16:36:01 yes, I run abuild on master branch 2020-09-02 16:36:02 I had a MR which introduced that error in an effort to fix armv7 but closed it because it didn't fix the original issue 2020-09-02 16:37:08 understand. this is C++ which I don't understand (only superficially) 2020-09-02 16:44:26 My $WORK is C++ and I don't understand it either, soooooo :D 2020-09-02 16:45:00 heh 2020-09-02 16:45:52 https://tpaste.us/DDqa 2020-09-02 16:45:55 :) 2020-09-02 17:16:43 Cogitri: now I understand why is around so much bad C++ programs ;) 2020-09-02 17:17:43 It's rather easy to mess things up in C++, yes 2020-09-02 17:19:01 I can tell you story about that but not here, #alpine-offtopic if you are interested to hear 2020-09-03 06:32:06 clandmeter: another security upgrade for gitlab (or rather, the previous one was botched): https://about.gitlab.com/releases/2020/09/02/security-release-gitlab-13-3-3-released/ 2020-09-03 06:59:51 nod, seen it. 2020-09-03 13:28:07 ikke: Pushed another update to asgen's config. Could you also please send me the log of the container before that? I want to report to upstream that for some reason it only includes icons for newly built packages in the daily tarballs 2020-09-03 13:28:48 ok 2020-09-03 15:40:39 ikke: https://gitlab.alpinelinux.org/Cogitri/aports/-/jobs/198982 CI says nothing should be built and the source branch isn't deleted for that job :) 2020-09-03 15:41:19 "fatal: master...HEAD: no merge base" 2020-09-03 15:51:04 git fetch is taking a long time 2020-09-03 15:54:51 Cogitri: ah, 275 commits not helping :) 2020-09-03 15:59:10 Cogitri: can you rebase your MR? 2020-09-03 16:00:39 Sure 2020-09-03 16:01:01 I unmarked it as WIP temporarily to do it from the interface 2020-09-03 16:01:14 Ah okie 2020-09-03 16:02:21 It's taking some time though apparently 2020-09-03 16:07:21 Not sure why it's taking so long. I see the rebase process still working 2020-09-03 16:10:23 finally finished 2020-09-03 16:11:10 Heh, takes some time go rebase all of those comments I guess :D 2020-09-03 16:11:18 Takes a minute or two locally as well 2020-09-03 16:11:25 heh 2020-09-03 16:11:34 Now it's actually building something 2020-09-03 16:11:44 I've increased the amount of fetched commits from 50 to 500 2020-09-03 16:12:01 Do you expect this to finish within 4h? 2020-09-03 21:24:57 `git pull` is taking forever again... 2020-09-03 22:20:27 Seems like MRs don't get updated either when force pushing to them 2020-09-04 09:39:26 I think `sh -x` is enabled on CI? https://gitlab.alpinelinux.org/Cogitri/aports/-/jobs/199277#L8079 is that by design? 2020-09-04 09:40:31 Only for you :P 2020-09-04 09:40:45 You can remove the CI_DEBUG* variable in your project 2020-09-04 09:41:06 I added that to troubleshoot it not building any packages 2020-09-04 10:20:24 Oh :D 2020-09-04 10:20:30 Okie, thanks 👍 2020-09-04 18:31:29 https://youtu.be/W5n_5Idlxvo 2020-09-04 18:58:37 Elpain :) 2020-09-04 19:00:00 el Pain 2020-09-06 07:16:26 what are different in usa4 lxc (ARM) and x86_64? on usa4 (armv7 and aarc64 lxcs) yesterday I tested build of iwd and check pass but failed on x86_64 2020-09-06 07:16:48 and it passed fine on all builders 2020-09-06 07:17:01 now I have same with ell library 2020-09-06 07:17:36 and, btw iwd check failed on all CIs 2020-09-06 07:28:33 usa4 is not x86_64? 2020-09-06 07:28:48 ikke: yes, I know 2020-09-06 07:29:04 So I don't understand your question 2020-09-06 07:29:11 but I don't know where is x86_64 lxc 2020-09-06 07:29:40 nld5 2020-09-06 07:29:58 at least, developer containers are there 2020-09-06 07:30:03 ikke: check passed on usa4 lxcs but fail on nld5 2020-09-06 07:30:21 Ok, and what check is failing? 2020-09-06 07:30:33 dhcp6 test 2020-09-06 07:30:48 no, sorry 2020-09-06 07:30:53 uuid test 2020-09-06 07:31:00 on lxc 2020-09-06 07:31:17 dhcp6 failed on s390x builder 2020-09-06 07:31:27 I'm talking about ell now 2020-09-06 07:31:41 iwd passed all builders yesterday 2020-09-06 07:32:52 ell dhcp6 check passed on armhf and x86 builders but failed on s390x 2020-09-07 08:07:11 nld5 run 4.19.80-0-vanilla kernel 2020-09-07 08:23:05 It's still running alpine 3.10 2020-09-07 08:24:19 ahm, that is why building ell on it failed, I presume 2020-09-07 08:24:37 Does it require a newer kernel? 2020-09-07 08:24:49 I think yes 2020-09-07 08:25:19 pkcs8 in kernel is needed for iwd so I think same is also for ell 2020-09-07 08:25:42 even for build with check enabled 2020-09-07 08:26:00 only for build it can be done on old kernels 2020-09-07 10:34:13 Would be nice if we can upgrade our builders to 3.12 2020-09-07 10:34:32 But I'm a bit uncertain about ppc64le/s390x 2020-09-07 10:36:11 do we have info how many users are there who use this arches 2020-09-07 10:40:45 No, not really 2020-09-07 10:40:49 We don't track users :P 2020-09-07 10:42:02 I think I track one for s390x, tmhoang ;) 2020-09-07 19:23:02 ikke: Did you restart the asgen container? 2020-09-07 19:23:53 I thought I did, right? 2020-09-07 19:24:24 5 days ago: "12:37:28 @ikke │ Cogitri: apstream-generator has been restarted (and the repository updated) " 2020-09-07 19:27:38 Oh whoops, forgot the message, wasn't sure anymore since the generated HTML didn't really look much different (I enabled appstream data validation so I had expected errors all over the place from faulty appstream data) 2020-09-07 19:29:29 I did not remove the older appstream data though 2020-09-07 19:29:54 Well, if it doesn't take up too much space that's not a problem :) 2020-09-07 20:02:39 3.7GB atm 2020-09-10 19:04:36 ikke: can you restart asgen again? Generating metadata for more arches now that apk-polkit-rs is available on them 2020-09-10 21:54:28 Also, what country are the CI machines? Seems like there's some bad CDN on crates.io in some countries that causes 206s 2020-09-11 05:39:14 ikke: Sorry for the ping, but what country are the CI runners in? :) 2020-09-11 05:39:29 diverse 2020-09-11 05:39:33 Currently in contact with Rust upstream due to https://status.crates.io/incidents/hbrpxrpbsdc8 2020-09-11 05:39:44 Ah, so US&EU? 2020-09-11 05:40:09 s390x is in brazil, even 2020-09-11 05:40:41 Ah yes, we don't have Rust on s390x so I didn't test on that 2020-09-11 05:41:36 but most is in EU+US indeed 2020-09-11 05:43:48 Okie, thanks 👍️ 2020-09-11 05:44:52 Also, can you restart asgen again (and pull the asgen compose repo for updates in the config file)? It's generating metadata for more arches now that apk-polkit-rs is available on them 2020-09-11 07:04:11 Cogitri: I've updated the compose repository and restarted the appstream-generator container 2020-09-11 07:04:30 Thanks 👍 2020-09-11 08:11:07 i think ppc64le is in brazil and s390x in usa 2020-09-11 08:11:14 ah, ok 2020-09-11 08:16:48 do we have an lxc host where dns is properly set up, with dnsmasq and alpin.pw resolving working? 2020-09-11 08:17:02 usa1-dev1 [~]# cat /etc/resolv.conf 2020-09-11 08:17:02 nameserver 1.1.1.1 2020-09-11 08:17:12 will not resolve alpin.pw 2020-09-11 08:19:40 nld3.alpin.pw 2020-09-11 08:20:05 thanks 2020-09-11 08:20:36 has been bothering me as well that it's not consistently working :) 2020-09-11 08:21:07 im having issues with getting ncopa-edge-aarch64.usa1.alpin.pw to resolve 2020-09-11 08:22:51 Does /etc/network/interfaces have hostname $(hostname)? 2020-09-11 08:23:03 it does 2020-09-11 08:23:50 oh... 2020-09-11 08:23:58 i think i know what the issue is 2020-09-11 08:25:51 ifupdown does not respect "hostname $(hostname)" 2020-09-11 08:26:17 ifupdown-ng? 2020-09-11 08:26:22 or ifupdown 2020-09-11 08:30:32 ncopa: I'd like to upgrade our infrastructure to Alpine 3.12 2020-09-11 08:30:38 A lot is still on 3.10 2020-09-11 08:31:43 Which of the CI runners are in the EU? 2020-09-11 08:31:56 x86 at least 2020-09-11 08:32:15 I think that's the only one 2020-09-11 08:32:41 nld7, the others are usa1, usa4, usa7 2020-09-11 08:33:02 and ppc64le / s390x, which are brazil and usa 2020-09-11 08:34:52 Okie, thanks 👍 2020-09-11 09:08:37 ncopa: 0.9 has a fix for that :) 2020-09-11 09:48:47 awesome! 2020-09-12 10:27:21 Is ppc64le stuck? 2020-09-12 10:39:42 hmm, it's building emacs 2020-09-12 10:41:15 Waiting for git... 2020-09-12 10:42:17 ups :P 2020-09-12 11:54:40 ikke: did asgen crash? Seems like it only generated aarch64 :/ 2020-09-12 11:55:47 Cogitri: it's up 2020-09-12 11:55:58 Huh 2020-09-12 11:57:55 The job gave a segfault though 2020-09-12 11:58:05 Segmentation fault 2020-09-12 11:58:08 run-parts: /etc/periodic/daily/run-generator: exit status 139 2020-09-12 12:02:31 Ugh :/ 2020-09-12 12:03:27 Can you maybe set `"validateMetainfo": false` in asgen-config.json in `Features`? 2020-09-12 12:03:40 Then it should hopefully not segfault 2020-09-12 12:03:58 IN the root? 2020-09-12 12:04:04 oh, soryr 2020-09-12 12:04:35 done 2020-09-12 12:06:07 Thanks 2020-09-12 12:06:11 Needs a restart I think 2020-09-12 12:06:16 I did 2020-09-12 12:06:37 Great, thanks :) 2020-09-14 09:25:31 I think the ppc64le edge runner is gone? 2020-09-14 09:35:20 emacs again 2020-09-14 09:36:03 :/ 2020-09-14 14:05:22 Oh derp, asgen is crashing because I only cherry-picked the patch for it only on edge and we run it on 3.12 :D 2020-09-14 14:09:53 Actually nevermind, was switching up Dockerfiles ^^ 2020-09-14 15:07:19 oh 2020-09-14 15:09:37 Seems like it's crashing in some other place now 2020-09-14 16:00:57 Fun, after installing the -dbg packages it works :D 2020-09-15 20:00:58 clandmeter: Someone (third-party) responded to our gitlab issue, found the commit that introduced the behavior (maintainers only able to rebase on public projects) 2020-09-16 06:20:20 ikke: ref? 2020-09-16 06:21:18 and goodmorning :) 2020-09-16 06:57:25 clandmeter: good morning 2020-09-16 06:57:28 https://gitlab.com/gitlab-org/gitlab/-/issues/211714 2020-09-16 07:28:32 good morning 2020-09-16 09:09:01 who is our ppc64le contact? 2020-09-16 09:09:10 i think we should upgrade our ppc64le build machine 2020-09-16 09:09:14 its running alpine 3.9 2020-09-16 09:09:58 Yeah, I'd like to upgrade all our infra 2020-09-16 09:10:10 but I'm not sure if ppc64le / s390x needs special care 2020-09-16 09:10:36 it shouldnt, but we may need someone that can help us rescue it in case it does not boot up 2020-09-16 09:11:21 yeah 2020-09-16 09:11:33 Last time, we had an e-mail conversation with Gustavo and mksully 2020-09-16 09:23:21 maybe walbon on #alpine-devel 2020-09-16 17:48:57 ikke: So it seems like asgen works in a 3.12 container (at least it didn't have any hiccups on my CI when I tried reproducing the SIGSEGVS), could you maybe restart the container again, hopefully it should just work then 2020-09-17 12:31:39 Ok, added support for GNOME Software to fetch the appstream data now. As such I changed asgen's config to not export in date-versioned dirs. Could you please purge the volume thingie for asgen, pull the latest version and restart it again? Then we can drop the alpinelinux-appstream-data package 2020-09-17 13:51:19 Will do. Please remind me tonight 2020-09-17 14:27:24 Okie 👍️ 2020-09-17 17:00:59 Cogitri: can I discard the validateMetadataInfo setting? 2020-09-17 20:20:29 Yup 2020-09-18 18:58:20 ikke: Forgot to remind you yesterday about asgen so here's your ping :) 2020-09-18 19:38:55 Cogitri: `Current branch master is up to date.` 2020-09-18 19:39:01 You did not update the composer project? 2020-09-18 19:41:48 I've pulled the latest docker image and recreated the containers (with pruned volumes) 2020-09-18 19:43:10 Nop, composer didn't have changed 2020-09-18 19:43:23 Only docker changed :) 2020-09-18 19:44:04 nod 2020-09-18 19:44:24 Thanks for looking into it 2020-09-18 19:44:47 Running the cron now 2020-09-18 19:44:51 (manually) 2020-09-18 19:53:12 Seems like stuff appears on a.a.o, so that's good 2020-09-19 08:45:35 Cogitri: the job failed with some 'no space left on device' error, but the host has plenty of space free 2020-09-19 09:25:01 Huh, that's weird, it worked fine on my VPS 2020-09-19 09:26:00 https://tpaste.us/RzZ5 2020-09-19 09:38:42 Maybe you could docker exec into the container and make sure that writing files really does work? 2020-09-19 09:42:08 THis was execed into the container 2020-09-19 10:08:34 ikke: Ah, I mean with another program :) 2020-09-19 14:17:25 Seems like asgen did generate data for all arches though (at least on edge) 2020-09-19 14:18:40 I can write to /cache/cache/tmp 2020-09-19 14:35:18 Hm, weird 2020-09-19 14:55:19 Actually, maybe the icon tarballs get too big. Maybe you could try setting the `cached` property of icons to false and `remote` to true? 2020-09-21 07:38:37 ikke: mirrors are having issues? 2020-09-21 10:07:48 clandmeter: You mean these alerts? 2020-09-21 10:35:18 ikke: yes 2020-09-21 10:38:27 Some mirror have become quite unstable lately 2020-09-21 10:38:34 aarnet hasn't been syncing for some time 2020-09-21 10:52:56 if uk is not uptodate it means one of our higher tier servers is probably full 2020-09-21 10:56:03 nld3 2020-09-21 10:56:52 /srv/mirror is at 95% 2020-09-21 10:58:04 1006.9G 904.3G 51.4G 95% /srv/mirror 2020-09-22 05:46:43 ^ This one is announced 2020-09-22 06:01:55 morning 2020-09-22 06:02:00 we have a problem with our CI 2020-09-22 06:02:03 https://gitlab.alpinelinux.org/J0WI/aports/-/jobs/211596#L52 2020-09-22 06:02:30 problem is that it gave green light when it should have given error 2020-09-22 06:03:41 ime, building in docker, lxc and native are different 2020-09-22 06:04:10 that was not the issue here 2020-09-22 06:04:13 something fails on docker but works in lxc 2020-09-22 06:04:24 it failed to merge the merge request 2020-09-22 06:04:46 ah 2020-09-22 06:05:17 if it fails to checkout the git branch, it should return failure and a red flag. not a green flag that indicates that everything is ok 2020-09-22 06:05:41 currently it give green flag because there was nothing to build if it fails to check out the branch 2020-09-22 06:08:46 that is why I dislike to merge MRs which are not clean and don't rebase cleanly 2020-09-22 06:09:12 i did rebase the MR yesterday 2020-09-22 06:09:24 and pushed it to the MR 2020-09-22 06:09:27 so it was clean 2020-09-22 06:09:35 and it still doesn't merge cleanly? 2020-09-22 06:09:54 it did merge cleanly, which is why it got merged to master 2020-09-22 06:10:30 problem was that CI for some reason failed to check out the branch, and did not return error 2020-09-22 06:10:46 problem is that CI gave green light when it shouldnt 2020-09-22 06:10:56 so it got merged on CI success 2020-09-22 06:11:13 aha, now understand 2020-09-22 06:12:21 hmm, then something doesn't work as it should 2020-09-22 06:14:34 that is what i am saying :) 2020-09-22 08:06:23 ncopa: will look into it. I certainly expected it to error out on such events 2020-09-22 13:44:20 Puhhh 2020-09-23 10:47:33 clandmeter: seems like uk is alright, for some reason the script that fetches the data has issues 2020-09-23 16:01:39 Ah ok 2020-09-23 21:46:15 clandmeter: our gitlab issue is getting some traction now :) 2020-09-24 09:35:40 ikke: yes ive seen it. 2020-09-24 09:35:52 i guess we are still having issues with deadlocks on gitlab? 2020-09-24 09:36:15 but looking at the dev channel, it seems a more general issue overall? 2020-09-24 09:48:03 clandmeter: I don't hear a lot of issues 2020-09-24 09:48:25 I've only seen it once myself that a rebase in interface actually hung 2020-09-24 09:52:47 Gitlab tends to be rather slow at rebasing or opening MRs sometimes 2020-09-24 09:53:16 Sometimes Gitlab needs like 2 or 3 minutes to create a MR 2020-09-28 15:54:58 mps: there are 2 containers on nld3 2020-09-28 15:55:53 ikke: I have two lxc on usa4 and one x86_64 but forgot where it is 2020-09-28 15:56:16 Right, but those are just dev containers, right? 2020-09-28 15:56:23 and I had one bare VM on linode, but I think clandmeter removed it on my request 2020-09-28 15:56:33 yea 2020-09-28 15:56:35 ikke: right 2020-09-28 15:57:27 I looked some months ago I have seen mailmain3 have some scripts to import data from some other ML systems 2020-09-28 15:59:18 yeah, but sr.ht has it's own format iirc 2020-09-28 16:02:04 I don't know what is sr.ht format but if we look at DB schema (and data) maybe we can make script which can export it to some 'standard' format 2020-09-28 16:03:14 yes 2020-09-28 16:23:55 This is the schema: https://tpaste.us/prM4 2020-09-28 16:28:58 hmm, not small, simple 2020-09-28 16:29:19 and these webhooks complicate things, imo 2020-09-28 16:29:33 We can ignore that 2020-09-28 16:30:02 We just need all subscribers and posts 2020-09-28 16:31:18 good, we have to look how these data can be 'integrated/imported' to mailman 2020-09-28 16:32:03 I think the easiest way is probably to generate mbox / maildirs 2020-09-28 16:32:18 I thought about same 2020-09-28 16:32:31 ie, the opposite of how we imported it 2020-09-28 16:33:24 but we will need to import users subscribed to new ML system 2020-09-28 16:34:08 yes, that would require something dedicated 2020-09-28 16:34:41 the `email` table seems to contain all posts 2020-09-28 16:34:50 I have some experience in exporting postgresql data with perl 2020-09-28 16:37:48 if I don't have much $day_job work I would ask you for complete DB dump to try to export some things 2020-09-28 16:38:09 ncopa: clandmeter: Any objections? ^ 2020-09-28 16:39:11 nope 2020-09-28 16:39:37 i would love somebody to look into the ml 2020-09-28 16:40:39 clandmeter: I'm thinking about it from time when you asked me to try different solutions 2020-09-28 16:42:07 even talked with my son to write ML software with Mojolicious (perl web framework) but he don't have time now and will not have in near future 2020-09-28 21:41:46 i rather just go with something that people already know 2020-09-28 21:43:29 trading one opinionated ML software for a different opinionated ML software just means a different group of devs is pissed off 2020-09-29 05:48:23 ikke: no objection. thank you for working on it mps 2020-09-29 06:33:51 ncopa: np, I would like to get 'proper' ML back on alpine 2020-09-29 06:35:00 Ariadne: don't worry, also I wouldn't like opionionated anything in alpine 2020-09-29 18:46:41 git push is stalling on gitlab 2020-09-29 18:57:29 hmm 2020-09-29 21:39:58 git push is still very slow :) 2020-09-30 08:26:57 ikke: do you get new mirror request emails? 2020-09-30 08:33:28 clandmeter: yes /o\ 2020-09-30 08:33:56 im adding them 2020-09-30 14:51:59 ikke: i added them 2020-09-30 14:52:11 we need to inform them and let them double check the data