| *** jroll08 is now known as jroll0 | 00:51 | |
| *** liuxie is now known as liushy | 08:51 | |
| *** sfinucan is now known as stephenfin | 11:38 | |
| *** vsevolod_ is now known as vsevolod | 12:10 | |
| vsevolod | Sorry if this is wrong place to ask. Read-the-docs hook for jjb/jenkins-job-builder project seems not working since about Sep 2023. | 12:22 |
|---|---|---|
| vsevolod | Probably, should be done by this project. https://zuul.opendev.org/t/openstack/builds?job_name=trigger-readthedocs-webhook All run were succesful, but no results are visible. | 12:23 |
| vsevolod | This documentation last updated at Aug 2023: https://jenkins-job-builder.readthedocs.io/en/latest/ | 12:24 |
| fungi | vsevolod: i seem to recall there was an announcement a while back about rtd needing changes to api authentication, though that probably would only impact on-demand builds and not periodic ones | 12:34 |
| fungi | we don't maintain rtd and don't publish documentation for our own projects there, so first-hand knowledge is likely limited | 12:35 |
| fungi | vsevolod: are you one of the jjb maintainers trying to figure out why rtd builds broke, or a jjb user trying to report this problem to the jjb maintainers so they'll look into it? | 12:36 |
| vsevolod | I am jjb maintainer | 12:37 |
| fungi | got it. do you have access to the rtd account hosting the jjb docs? | 12:37 |
| vsevolod | Nope | 12:37 |
| fungi | do you happen to know who does? it's possible you'll need their help to look at build logs in rtd or update anauthentication credential | 12:38 |
| vsevolod | I know who previous maintainer is. Not sure he have access. | 12:39 |
| vsevolod | Should I email him and ask about this? | 12:39 |
| fungi | that would be a good place to start. the trigger-readthedocs-webhook job, as its name implies, just makes a webhook call to rtd to signal that it should perform a rebuild, but i don't think it receives any diagnostics or status. rtd's webhook may not even tell you that credentials you're passing are wrong (perhaps in order to cut down on brute-force guessing attempts) | 12:42 |
| fungi | most diagnosing will have to happen at the rtd side | 12:43 |
| vsevolod | Got it, thank you. I will email previous maintainer. | 12:43 |
| fungi | the only other thing i can think to suggest is to make sure the docs are still actually buildable, but you presumably have a docs build job which does that already | 12:44 |
| vsevolod | Yes, looks like this. | 12:45 |
| vsevolod | I think, this is the job: https://zuul.opendev.org/t/openstack/build/a542b5e6366547cd9ec08a5df5882c5c | 12:46 |
| vsevolod | https://zuul.opendev.org/t/openstack/builds?job_name=openstack-tox-docs&project=jjb/jenkins-job-builder | 12:46 |
| vsevolod | fungi: I have found RTD builds. Results are publicly available. And it's failing, with description why: https://readthedocs.org/projects/jenkins-job-builder/builds/ | 13:24 |
| vsevolod | I will read RTD docs to find out why. | 13:24 |
| fungi | vsevolod: aha, so they added a new requirement for projects. i guess that went into effect around the end of last year | 13:27 |
| fungi | good to know! | 13:27 |
| vsevolod | yep | 13:28 |
| opendevreview | Clark Boylan proposed opendev/system-config master: Update gitea to 1.21.8 https://review.opendev.org/c/opendev/system-config/+/913686 | 14:42 |
| clarkb | the etherpad 2.0.0 update is fairly involved from an image building perspective. But I think the end result should be an improvement. At least they do multi stage builds now which is nice | 14:53 |
| clarkb | I'm not seeing any upgrade process notes. I suspect that the 2.0 release is largely there because they chagned the deployment and build tooling but the database doesn't appear to have changed | 14:53 |
| fungi | infra-prod-base ran successfully out of the periodic pipeline, so the failure associated with the osuosl mirror does indeed seem to have been temporary and i think 908359 is fully deployed now (i should make a point of checking for evidence of backups when i get a moment between meetings) | 14:56 |
| clarkb | ++ you can fuse mount them and check they look correct | 14:57 |
| fungi | it's mainly the db backup, i think | 14:58 |
| clarkb | https://review.opendev.org/c/opendev/system-config/+/913608/1 is a followup to the centos 7 cleanup change. I realized that I wasn't properly clearing out the obs centos 7 stuff and discovered some other distro content and did a small refactor of the rm block | 14:58 |
| clarkb | fungi: ya that should fuse mount too | 14:58 |
| clarkb | or you can extract the single file directly | 14:58 |
| opendevreview | Merged zuul/zuul-jobs master: prepare-workspace-git: Add ability to define synced pojects https://review.opendev.org/c/zuul/zuul-jobs/+/887917 | 14:59 |
| corvus | that zuul-jobs change touches a lot of jobs, so i'll monitor for breakage | 14:59 |
| fungi | thanks! | 15:00 |
| clarkb | in theory it is a noop for us because we don't set the flag that it adds, but ya basically every job runs that role | 15:00 |
| corvus | don't usually have to wait this long for someone to upload a change.... | 15:10 |
| corvus | okay i've seen a job run path that point with the new code | 15:18 |
| corvus | s/path/past/ | 15:18 |
| fungi | yay! | 15:20 |
| noonedeadpunk | fungi: well, it feels we'd need to reach OFTC folks regarding #openstack-freezer channel, as osn is not responsive | 16:40 |
| fungi | thanks for the heads up, i'll look into how we ask for it | 16:43 |
| clarkb | corvus: if you get a second do you understand: [e: cb5f2676f4d440408d600c3501b685fc] No matching parents for job devstack and change <Change 0x7fb69dc0db50 openstack/kuryr-kubernetes 913419,1> | 18:23 |
| clarkb | corvus: that change is trying to run devstack + tempest jobs on a chagne to stable/2024.1. Devstack doesn't haev a stable/2024.1 branch yet so there is no explicit branch match but shouldn't it fallback to master to find teh definition of `devstack` to use? | 18:24 |
| clarkb | it does fallback to master when matching another parent job in the inheritance path: [e: cb5f2676f4d440408d600c3501b685fc] Variant <Job devstack-tempest branches: None source: openstack/tempest/zuul.d/base.yaml@master#1> matched <Change 0x7fb69dc0db50 openstack/kuryr-kubernetes 913419,1> | 18:25 |
| clarkb | but it seems to get to the `devsatck` job and gives up due to no matches | 18:26 |
| clarkb | I'm not able to find any explicit branch matchers on these jobs that would prevent fallback to master either | 18:28 |
| corvus | clarkb: there's no fallback behavior for branch matchers; tempest is a branchless project so it has no branch matcher; devstack has branches, so each gets an implied branch matcher; to overcome this branches will need to match, or use pragmas to set branch equivalencies | 18:51 |
| clarkb | ah ok now that you've said that I think devstack may have historically had pragmas related to this | 18:54 |
| clarkb | gmann: ^ anyway sounds like you should be branching devsatck before everything else | 18:55 |
| corvus | or branch it first | 18:55 |
| clarkb | and then this problem can be avoided | 18:55 |
| corvus | yeah that | 18:55 |
| gmann | ok, clarkb corvus thanks | 18:56 |
| gmann | I think we will be branching devstack soon. will check with elodilles when we are ready | 18:56 |
| elodilles | it's strange because we used to do like this in previous releases up till now :-o (devstack and grenade branched when everything else have branched already) | 19:16 |
| clarkb | elodilles: yes I think devstack used pragmas before | 19:17 |
| clarkb | elodilles: so another option would be to add that pragma in the interim | 19:17 |
| clarkb | `git log -p` then look for pragma | 19:17 |
| clarkb | basically it said master was an implied branch which I htink is the fallback behavior gmann and I expected | 19:18 |
| fungi | looks like it temporarily used one in 2022 for openstacksdk's feature/r1 branch | 19:19 |
| clarkb | ya but I think it would've applied to all devstack jobs | 19:20 |
| fungi | not finding any, unless it was in a different repo than openstack/devstack | 19:21 |
| clarkb | the one that was added for openstacksdk feature/r1 would have applied to all devstack jobs | 19:21 |
| clarkb | and since master was listed it would be an implied matcher for projects without a master branch right now | 19:21 |
| clarkb | er without a matching branch | 19:22 |
| fungi | yeah, but it was removed when openstacksdk merged feature/r1 into master | 19:22 |
| fungi | a couple years ago | 19:22 |
| clarkb | ah ok | 19:22 |
| fungi | i mean i'm not finding any other uses of pragmas in openstack/devstack history (either on master or recent stable branches) | 19:23 |
| clarkb | right I'm just pointing out that that pragma may have made stable brnach cutting "easier" in the past | 19:23 |
| clarkb | not that there were other pragmas | 19:23 |
| fungi | right, i agree, i'm just not finding evidence that there were any | 19:25 |
| hashar | fungi: congratulations on releasing git-review 2.4.0 :) | 22:07 |
| fungi | hashar: thanks for all the help, of course! | 22:08 |
| hashar | m;-] | 22:08 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!