| *** ysandeep|out is now known as ysandeep|rover | 03:24 | |
| *** ysandeep|rover is now known as ysandeep|afk | 04:24 | |
| *** ysandeep|afk is now known as ysandeep|rover | 04:45 | |
| opendevreview | daniel.pawlik proposed openstack/ci-log-processing master: Moge container images to quay.io https://review.opendev.org/c/openstack/ci-log-processing/+/843107 | 06:56 |
|---|---|---|
| *** ysandeep|rover is now known as ysandeep|rover|lunch | 08:02 | |
| *** ysandeep|rover|lunch is now known as ysandeep|rover | 08:38 | |
| *** dviroel|out is now known as dviroel | 11:21 | |
| opendevreview | daniel.pawlik proposed openstack/ci-log-processing master: Ignore directories that does not exists or don't have access https://review.opendev.org/c/openstack/ci-log-processing/+/842977 | 11:38 |
| opendevreview | Dr. Jens Harbott proposed openstack/project-config master: Add missing volumes to Grafana AFS page https://review.opendev.org/c/openstack/project-config/+/843133 | 12:16 |
| opendevreview | daniel.pawlik proposed openstack/ci-log-processing master: DNM Add support for Elastic Recheck tool https://review.opendev.org/c/openstack/ci-log-processing/+/843138 | 12:31 |
| opendevreview | yatin proposed openstack/devstack-gate master: Use stable constraints for Tempest venv for stable/train testing https://review.opendev.org/c/openstack/devstack-gate/+/843148 | 14:12 |
| opendevreview | Merged openstack/project-config master: Add missing volumes to Grafana AFS page https://review.opendev.org/c/openstack/project-config/+/843133 | 14:41 |
| opendevreview | Merged openstack/project-config master: Add neutron-vpnaas-stable-maint group https://review.opendev.org/c/openstack/project-config/+/842693 | 15:11 |
| *** ysandeep|rover is now known as ysandeep|out | 15:17 | |
| dansmith | clarkb: I'll say this over here since the qa meeting is still going | 15:49 |
| dansmith | clarkb: I think my problem was staleness in ~/.local | 15:49 |
| dansmith | I use virtfs to mount my home dir from my throwaway VM to my host | 15:49 |
| dansmith | and at some point I got dbcounter installed into ~/.local probably whilst trying to debug the install failure last week | 15:50 |
| dansmith | so it was picking that up, with no entry_points.txt in it instead of the publicly installed one and thus failing to load | 15:50 |
| dansmith | so, even though the vm was fresh, my home dir was unfresh | 15:50 |
| clarkb | ah | 15:51 |
| *** dviroel is now known as dviroel|lunch | 15:58 | |
| *** jpena is now known as jpena|off | 16:51 | |
| *** dviroel|lunch is now known as dviroel | 16:58 | |
| opendevreview | yatin proposed openstack/devstack-gate master: Switch legacy-tempest-neutron-full-stable to focal https://review.opendev.org/c/openstack/devstack-gate/+/843176 | 17:18 |
| fungi | CN = wiki.openstack.org | 17:46 |
| fungi | Not After : May 24 23:59:59 2023 GMT | 17:46 |
| fungi | #status log Replaced manually-obtained X.509 HTTPS cert on legacy wiki.openstack.org server | 17:46 |
| fungi | clarkb: ^ thanks! | 17:47 |
| clarkb | perfect | 17:47 |
| fungi | huh, statusbot is mia | 17:47 |
| fungi | i'll look into that now | 17:47 |
| elodilles | clarkb fungi : i'm planning to do some EOL'd branch cleanup if that is not interfere with your work now ^^^ (i guess, not, but wanted to be sure) | 17:55 |
| fungi | #status log Replaced manually-obtained X.509 HTTPS cert on legacy wiki.openstack.org server | 17:56 |
| opendevstatus | fungi: finished logging | 17:56 |
| fungi | elodilles: should be fine, thanks for letting us know! | 17:56 |
| elodilles | ++ | 17:57 |
| clarkb | ya I have nothing that would interfere | 17:59 |
| opendevreview | Merged openstack/project-config master: Fix python list for requirements propose-updates job https://review.opendev.org/c/openstack/project-config/+/842853 | 18:18 |
| opendevreview | Clark Boylan proposed openstack/project-config master: Install setfacl on test nodes https://review.opendev.org/c/openstack/project-config/+/843210 | 20:05 |
| opendevreview | Merged openstack/project-config master: Build and publish wheel mirror for Ubuntu Jammy https://review.opendev.org/c/openstack/project-config/+/842841 | 20:13 |
| elodilles | clarkb fungi : today these were deleted: https://paste.opendev.org/show/b35CjByPMENT1poLDjez/ | 20:16 |
| fungi | thanks elodilles! | 20:17 |
| elodilles | the newton and ocata needed to be deleted half-manually, as those series are marked already EOL | 20:17 |
| elodilles | and now i realized that i did a mistake: requirements' newton branch was deleted, as the newton-eol tagging patch was merged: https://review.opendev.org/c/openstack/releases/+/838571 | 20:20 |
| elodilles | but the newton-eol was not created, as newton was EOL | 20:20 |
| elodilles | so now i think we lost the history of requirements' stable/newton :S | 20:21 |
| elodilles | fungi: can we do something about this? ^^^ | 20:21 |
| clarkb | elodilles: so the change to releases for newton-eol merged but the tag wasn't created? | 20:26 |
| clarkb | I think if you know what the old HEAD was you can still tag it | 20:27 |
| clarkb | or create the branch again off that HEAD, tag it, then delete the branch | 20:27 |
| clarkb | I see d94785ceb702587131ce98a1b9f05b2869a66180 locally but I have no idea how old that is | 20:27 |
| clarkb | you have until 04:17 when the git gc runs | 20:29 |
| clarkb | elodilles: does your script record the HEADs before you delete them? | 20:30 |
| elodilles | clarkb: the tagging is 'skipped' by the job if the branch is in EOL state (and newton is in EOL) | 20:32 |
| fungi | yeah, but does the script record the head commit id when it's deleting a branch? | 20:33 |
| clarkb | ok. I think a few things should be changed before the next round of deletions. The script that delets should ensure that whatever branch HEAD is deletes has a corresponding tag. And you should record what those heads are | 20:33 |
| elodilles | clarkb: so, though openstack/releases repository contains the correct hash for newton-eol of requirements branch, it was not tagged yet, as i wanted to tag it manually | 20:33 |
| clarkb | elodilles: oh can you share that? | 20:33 |
| clarkb | but I think what I would do is make stable/newton again from that sha right now | 20:34 |
| clarkb | then go back through your process | 20:34 |
| fungi | d94785ceb702587131ce98a1b9f05b2869a66180 refs/remotes/origin/stable/newton | 20:34 |
| elodilles | clarkb: here we have the hash: https://review.opendev.org/c/openstack/releases/+/838571 | 20:34 |
| fungi | that's the last one i had for it as well | 20:34 |
| opendevreview | Merged openstack/project-config master: Install setfacl on test nodes https://review.opendev.org/c/openstack/project-config/+/843210 | 20:34 |
| clarkb | elodilles: ya so in gerrit create a new branch for stable/newton with that set as the ref | 20:34 |
| clarkb | then you should be able to go back and tag it and then remove the branc hagain | 20:35 |
| clarkb | But also ensuring that you don't delete branches without a tag by comparing the tag and the branch head sha and recording every sha head for each branch deleted is a good idea as future improvements to the tooling | 20:35 |
| clarkb | elodilles: you should have permissions to do that via the gerrit web ui. Are you planning to do the branch creation as suggested? | 20:36 |
| elodilles | clarkb: OK, so now i'm creating stable/newton with hash d94785ceb702587131ce98a1b9f05b2869a66180 | 20:36 |
| clarkb | yes that is what I would do then start the process over again | 20:37 |
| elodilles | clarkb: Error 403 (Forbidden): update for creating new commit object not permitted | 20:37 |
| elodilles | so gerrit does not allow that :S | 20:37 |
| clarkb | well I think that is saying the commit has alredy been deleted? | 20:38 |
| clarkb | the gc runs daily at 04:17 UTC but maybe when you tell gerrit to delete a branch it prunes automatically | 20:38 |
| elodilles | :S | 20:38 |
| clarkb | "creating new commit object not permitted" specifically doesn't seem like a branch creation problem but a I cannot create that commit for you implying it is gone | 20:39 |
| clarkb | in that case someone will probably have to push it back to gerrit | 20:39 |
| clarkb | fungi: ^ unless you haev better ideas | 20:40 |
| clarkb | that commit is still in the giteas which is what I expected | 20:43 |
| clarkb | (just not pointed to) | 20:43 |
| clarkb | I'm more wary of running git show on the running repo behind gerrit | 20:43 |
| clarkb | elodilles: how did you try to create the branch? | 20:44 |
| elodilles | clarkb: here, with 'create new' >>> https://review.opendev.org/admin/repos/openstack/requirements,branches | 20:45 |
| clarkb | running as my admin account I get `fatal: not permitted: create on refs/heads/stable/newton`. I can try escalating privs even further | 20:46 |
| elodilles | clarkb: and can you maybe simply tag d94785ceb702587131ce98a1b9f05b2869a66180 with newton-eol? as actually that would be enough and maybe it would be easier to do (or am i wrong?) :/ | 20:48 |
| clarkb | its not easier because then I have to go pull my key out of cold storage and it will be the wrong key | 20:48 |
| clarkb | fungi: any objection to me trying to run gerrit create-branch openstack/requirements stable/newton d94785ceb702587131ce98a1b9f05b2869a66180 as project boostrapper? | 20:49 |
| clarkb | just in case this is a corner case acl problem and not an issue with the underlying git repo? | 20:49 |
| clarkb | I went ahead and ran it and it didn't fail | 20:50 |
| clarkb | elodilles: https://review.opendev.org/admin/repos/openstack/requirements,branches shows it now | 20:50 |
| clarkb | elodilles: https://opendev.org/openstack/requirements/commits/branch/stable/newton | 20:51 |
| elodilles | clarkb: phew. thanks! \o/ | 20:51 |
| clarkb | elodilles: I think you should be good now and can restart the process. I don't understand why we got those permissiosn errors as admin and as a release manager. I thought the acls allowed creating branches | 20:51 |
| elodilles | sorry for the trouble :S | 20:51 |
| clarkb | it may be that beacuse the sha wasn't referred to be anything it needed extra perms | 20:52 |
| clarkb | "gerrit create-branch openstack/requirements stable/newton d94785ceb702587131ce98a1b9f05b2869a66180" is the ssh command I ran as a project boot strapper for the record | 20:52 |
| clarkb | 'fatal: not permitted: create on refs/heads/stable/newton' is what I got when I was an admin and not bootstrapper | 20:56 |
| clarkb | so ya I suspect that because the commit isn't already know by a ref gerrit is maybe not smart enough to know it may still be there | 20:56 |
| elodilles | clarkb: i will postpone the deletion to another day (that was my original plan, too, but i forgot it and deleted the newton branch), as I have to setup a key first to be able to do signed tags | 20:56 |
| clarkb | elodilles: why not update the job that does it to run if the tag doesn't already exist rather than caring about openstack management state? | 20:57 |
| clarkb | what matters is the git repo state not openstacks rules | 20:57 |
| fungi | clarkb: sorry, stepped away for dinner, no objection but sounds like you discovered another gerrit quirk | 20:59 |
| elodilles | clarkb: you are right, maybe that should be done. i was just not sure, to worth it as we have some old stale branch that was somehow missed to be tagged with newton-eol, mitaka-eol and some older, for several years ago | 20:59 |
| clarkb | fungi: no worries. I figured it wasn't going to make anythin worse as I'd already confirmed my local copy of the repo had that as the branch head | 21:01 |
| clarkb | from whenever I last updated | 21:01 |
| clarkb | I'm about to pop out on a bike ride myself | 21:01 |
| elodilles | clarkb: the whole story is related to my mail: http://lists.openstack.org/pipermail/openstack-discuss/2022-May/028550.html | 21:01 |
| clarkb | I think manually tagging is fine too. It just seems like maybe that rule is bad if it allows this siutation to happen so maybe reconsider the rule | 21:02 |
| fungi | i also realize now that i misread one of the questions in that ml post | 21:03 |
| fungi | no, nevermind i didn't misread it | 21:03 |
| elodilles | clarkb: yepp, though i hope similar situation will not happen whenever these older branches are cleaned | 21:03 |
| elodilles | clarkb: thanks again for the help! | 21:03 |
| elodilles | have a nice bike ride! | 21:04 |
| fungi | you had "tag branches with *-eol where that is missing" as one of the actions there, i guess it was just overlooked for stable/newton on openstack/requirements | 21:04 |
| * elodilles needs some sleep anyway | 21:04 | |
| fungi | get some sleep elodilles! this will keep ;) | 21:04 |
| clarkb | please double check the branch exists as expected first so that the 04:17 GC doesn't do something we don't want | 21:05 |
| clarkb | but I think we are good if you confirm | 21:05 |
| elodilles | fungi: thanks o:) | 21:05 |
| elodilles | clarkb: in gitea i can see it, so that looks OK | 21:06 |
| fungi | if d94785c is the correct last known state, i have it here in my local repo as well so can always push the missing commits if really needed | 21:06 |
| clarkb | elodilles: doulbe check at https://review.opendev.org/admin/repos/openstack/requirements,branches too | 21:07 |
| clarkb | but ya I think it is good | 21:07 |
| elodilles | I can see it there, too | 21:08 |
| clarkb | alright I'm off for a bit. Back later | 21:08 |
| fungi | but would definitely be good for the branch delete script to spit out the result of `git show-ref origin/stable/whatever` before deletion (and after a remote update of course) | 21:08 |
| fungi | just so that next time there's no question as to what the last known state of the branch was before it was deleted | 21:09 |
| elodilles | fungi: to be honest, the script what is meant to be used does check if the tag exists and matches with the head ( list_eol_stale_branches.sh ) | 21:11 |
| *** dviroel is now known as dviroel|out | 21:11 | |
| elodilles | fungi: but since these are special cases, as I mentioned in my mail, it needs some manual tagging & deletion | 21:12 |
| fungi | got it | 21:16 |
| elodilles | but definitely i will add some extra checks (with some scripting) before i continue with the stable/newton and older branch cleanup :S | 21:17 |
| elodilles | fortunately in the usual cases I have to run 'list_eol_stale_branches.sh' which is safe in this sense :) with that, this kind of error is not possible | 21:19 |
| elodilles | but now i really need to sleep. thanks again for the help! o/ | 21:21 |
| opendevreview | Ian Wienand proposed openstack/openstack-zuul-jobs master: wheels: Add "wheel" to build venv https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/843222 | 23:33 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!