| *** bauzas_ is now known as bauzas | 00:17 | |
| *** bauzas_ is now known as bauzas | 00:49 | |
| *** bauzas_ is now known as bauzas | 03:01 | |
| *** bauzas_ is now known as bauzas | 03:50 | |
| *** bauzas_ is now known as bauzas | 04:26 | |
| *** bauzas_ is now known as bauzas | 05:32 | |
| *** bauzas_ is now known as bauzas | 05:45 | |
| *** bauzas_ is now known as bauzas | 06:30 | |
| *** bauzas_ is now known as bauzas | 06:38 | |
| *** ralonsoh_away is now known as ralonsoh | 07:09 | |
| *** bauzas_ is now known as bauzas | 07:33 | |
| *** bauzas_ is now known as bauzas | 08:04 | |
| *** bauzas_ is now known as bauzas | 10:07 | |
| *** bauzas_ is now known as bauzas | 10:15 | |
| *** bauzas_ is now known as bauzas | 11:31 | |
| elodilles | hi, today a lot of stein-eol patches have merged, so I'm planning to run the eol cleanup script ~2-3 hours later today if that is OK ( https://review.opendev.org/q/topic:stein-eol+is:merged ) | 14:52 |
|---|---|---|
| frickler | elodilles: should be fine afaict | 14:53 |
| elodilles | thx o/ | 14:53 |
| frickler | elodilles: regarding https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/891628 I think that this may become a very long list of fixes needed, so I'm thinking about following corvus' idea of force-merging that change | 14:56 |
| frickler | that would lead to config-errors in all affected projects, with the benefit of being able to see them all at once instead of one by one only | 14:57 |
| frickler | fungi: clarkb: ^^ what do you think about that? | 14:58 |
| elodilles | frickler: that is possible that there are many projects where the job is still used, so maybe the force-merge approach is the easiest | 14:59 |
| clarkb | would reverting undo the errors? If so that seems like a good outlet we can use | 15:01 |
| elodilles | i hoped that publish-openstack-sphinx-docs was really used on already eol'd pike, queens, rocky branches, and those are mostly deleted and only some of them should be still opened (those which are not listed under releases repo's deliverables) | 15:02 |
| elodilles | but i fear that it is still lingering on some 'master' branch as well... | 15:03 |
| frickler | clarkb: so force-merge, make a list of the errors, revert, fix the errors, revert the revert? | 15:03 |
| clarkb | frickler: or maybe just revert if the errors become problematic for some reason. We dno't need to revert unless necessary. But it is a out we can use | 15:04 |
| frickler | clarkb: yes, that's what I wanted to suggest, then we're on the same page | 15:04 |
| clarkb | sounds like a plan | 15:04 |
| frickler | so if I get a second +2 on that patch, I'll go ahead with it | 15:06 |
| clarkb | looking now | 15:06 |
| frickler | note the job deletion is stacked on top, but needs a rebase after the regex amendmends | 15:07 |
| frickler | I'll do that once this patch is merged, avoids needing to rebase both | 15:07 |
| clarkb | done | 15:07 |
| clarkb | oh I did only the first. Reviewing the second now | 15:08 |
| clarkb | gerrit renders the diff on the second one oddly. But lgtm | 15:09 |
| opendevreview | Merged openstack/openstack-zuul-jobs master: Remove rocky branch from periodic-stable templates https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/891628 | 15:11 |
| opendevreview | Dr. Jens Harbott proposed openstack/openstack-zuul-jobs master: Clean up rocky branch filters https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/891629 | 15:20 |
| frickler | I count 22 errors https://zuul.opendev.org/t/openstack/config-errors?name=Template+Not+Found | 15:23 |
| frickler | noonedeadpunk: ^^ maybe you can take a look at cleaning up the openstack-ansible projects in that list? | 15:24 |
| frickler | or find someone in your team to do it? | 15:24 |
| frickler | so certainly a good thing we didn't try to go through all of those one by one | 15:25 |
| frickler | but also only queens or older affected, which is a good thing I guess | 15:25 |
| * noonedeadpunk wonders why pike/queens/ocata repos are still there | 15:25 | |
| noonedeadpunk | they all should have been EOLed | 15:26 |
| frickler | noonedeadpunk: likely because of missing release automation | 15:26 |
| noonedeadpunk | I will reach releases team with EOL of Stein regarding these leftovers | 15:26 |
| noonedeadpunk | I think I was writing to IRC once about that but either missed answer or message was lost | 15:27 |
| noonedeadpunk | Will check on that | 15:27 |
| frickler | noonedeadpunk: those branches were not created by release automation, so they cannot be eoled automatically. elodilles may provide more context maybe | 15:28 |
| noonedeadpunk | thanks for the ping | 15:28 |
| frickler | meh, the job removal triggers even more errors in https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/891629 | 16:54 |
| fungi | as the sweater unravels | 17:17 |
| elodilles | the script cleaned up the following repos in today's run: https://paste.opendev.org/show/bagmzLMWsWAw5n5UYsW5/ | 18:00 |
| *** bauzas_ is now known as bauzas | 18:00 | |
| fungi | thanks! | 18:05 |
| fungi | that's a bunch | 18:06 |
| fungi | awesome | 18:06 |
| elodilles | yepp. 11 less periodic-stable fail mail, hopefully | 18:10 |
| fungi | currently it's down around 39 failure e-mails, so yes i'll be curious to see how much it drops | 18:11 |
| fungi | reducing it by a quarter would be great | 18:12 |
| elodilles | hmmm, i've found a weird thing... at least i can't figure out how it is possible... horizon's stable/stein is deleted, but horizon's periodic stable jobs are still running against stein... i don't even see the zuul inventory file where does it come from. this is an example: https://zuul.opendev.org/t/openstack/build/5d3168053f224f1ebe9953d1055a437f | 18:19 |
| fungi | when was the branch deleted? | 18:19 |
| elodilles | let me check it.... (but not today) | 18:20 |
| elodilles | (here it should have a line with a zuul yaml file from horizon's repo: https://zuul.opendev.org/t/openstack/build/5d3168053f224f1ebe9953d1055a437f/log/zuul-info/inventory.yaml#38-43 ) | 18:21 |
| elodilles | fungi: it was deleted in August 16th | 18:22 |
| fungi | looks like it ran with master of horizon: https://zuul.opendev.org/t/openstack/build/5d3168053f224f1ebe9953d1055a437f/log/zuul-info/inventory.yaml#122 | 18:23 |
| fungi | presumably because it couldn't find the stable/stein branch | 18:23 |
| elodilles | :S | 18:25 |
| elodilles | but why does it work differently then other repos? | 18:25 |
| fungi | a very good question. maybe some cache entry didn't get correctly invalidated? clarkb ^ i don't suppose you have any ideas? | 18:26 |
| elodilles | the branch was deleted successfully & i don't see stable/stein on opendev.org/openstack/horizon either, so it is like how it should be | 18:27 |
| fungi | https://opendev.org/openstack/openstack-zuul-jobs/src/branch/master/zuul.d/project-templates.yaml#L2622-L2670 does have branch matchers for stein, but i don't think that should cause zuul to trigger the jobs for those branches when they don't exist | 18:28 |
| elodilles | for all the other repos this works fine ^^^ only triggers a job if the given repo has a stable/stein branch with the 'periodic-stable-jobs' template in it | 18:30 |
| frickler | https://zuul.opendev.org/t/openstack/project/opendev.org/openstack/horizon?branch=stable%2Fstein&pipeline=periodic-stable looks like they may be defined in project-config? | 18:30 |
| fungi | oh, good catch... https://zuul.opendev.org/t/openstack/freeze-job?pipeline=periodic-stable&project=openstack%2Fhorizon&job=openstack-tox-docs&branch=stable%2Fstein | 18:31 |
| elodilles | so is this defined somewhere in opendev.org/openstack/project-config ? (i don't find it :-o) | 18:34 |
| frickler | I was suspecting official-openstack-repo-jobs but that's not it | 18:35 |
| fungi | well, the pipeline is defined in project-config and the horizon project is as well, but yes i'm still not clear what in there is telling zuul to trigger jobs on a nonexistent branch | 18:36 |
| fungi | zuul doesn't have that same job freeze for other projects stable/stein was deleted from, based on spot checks | 18:37 |
| fungi | it definitely seems like a leftover/stale cached entry for that branch, like zuul never found out about the deletion | 18:38 |
| frickler | maybe that's related to the "refs/heads" issue. could it also affect branch deletions? | 18:38 |
| fungi | maybe. what introduced the regression? | 18:38 |
| frickler | I don't know, need to check with corvus | 18:39 |
| fungi | if it ended up in a zuul restart between previous branch deletions and the horizon stein deletion, then that seems suspect yeah | 18:40 |
| elodilles | yepp, the first mail with horizon's failing stable/stein periodic is August 17th, so right after the branch was deleted | 18:44 |
| frickler | other branches deleted on August 16 do not show this issue, though | 18:49 |
| elodilles | only horizon | 18:49 |
| frickler | yes. moving to #opendev to get corvus involved | 18:49 |
| frickler | I cannot find openstack/compute-hyperv in governance, is that some kind of oversight? | 19:00 |
| fungi | it should have been retired when the winstackers team ceased to exist, but maybe that got missed | 19:00 |
| frickler | oh, yes, I just found https://review.opendev.org/c/openstack/governance/+/886880 | 19:01 |
| frickler | that was not as long ago as I expected | 19:01 |
| elodilles | yepp, just wanted to say it should be in legacy.yaml | 19:01 |
| frickler | but https://opendev.org/openstack/compute-hyperv seems untouched | 19:02 |
| frickler | even retirement would usually only empty the default branch, not stable branches, right? | 19:02 |
| fungi | openstack/os-win, openstack/oswin-tempest-plugin and openstack/networking-hyperv also needed retiring | 19:03 |
| fungi | frickler: we'd also remove them from zuul's config though | 19:03 |
| fungi | so jobs would stop being run | 19:04 |
| frickler | ah, that would solve my current issue, right | 19:04 |
| frickler | so moving this topic to #-tc | 19:04 |
| elodilles | it should look like this for example: https://opendev.org/openstack/panko | 19:04 |
| frickler | ack | 19:05 |
| elodilles | i guess this was missed as the team disappeared from community :/ | 19:05 |
| gmann | frickler: fungi: elodilles: yes complete project winstacker is retired and I am working on cleaning up the deps (like nova hyperV driver) etc and repo content | 19:06 |
| elodilles | gmann: ack, thanks! | 19:06 |
| *** bauzas_ is now known as bauzas | 22:11 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!