*** bauzas7 is now known as bauzas | 00:49 | |
ttx | #startmeeting releaseteam | 14:00 |
---|---|---|
opendevmeet | Meeting started Fri Jun 6 14:00:30 2025 UTC and is due to finish in 60 minutes. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. | 14:00 |
opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 14:00 |
opendevmeet | The meeting name has been set to 'releaseteam' | 14:00 |
elodilles | o/ | 14:00 |
ttx | Ping list: release-team elod | 14:00 |
ttx | Hi! | 14:00 |
ttx | Our agenda today... | 14:01 |
ttx | #link https://etherpad.opendev.org/p/flamingo-relmgt-tracking#L124 | 14:01 |
ttx | #topic Review task completion | 14:01 |
ttx | - Manually push 2023.2-eol tags for Kolla (frickler) | 14:02 |
ttx | elodilles: do you know if that was done? | 14:02 |
elodilles | yepp, as far as i know frickler did it | 14:03 |
elodilles | but let me quickly double check it | 14:03 |
ttx | perfect | 14:03 |
ttx | - Review any remaining milestone-1 exceptions (elod) | 14:03 |
ttx | List is empty so I suspect that is done too | 14:04 |
elodilles | yepp | 14:04 |
elodilles | there was oslo.limit that remained: https://review.opendev.org/c/openstack/releases/+/949457 | 14:05 |
elodilles | but got released meantime | 14:05 |
ttx | - Propose 2025.1 Epoxy cycle trailing final release patches if not yet released (elod) | 14:06 |
frickler | kolla eol tags are done nad braces removed | 14:06 |
elodilles | yepp, i've just double checked and found the same result o:) thanks frickler :) | 14:08 |
elodilles | cycle trailing release for openstack-ansible: https://review.opendev.org/c/openstack/releases/+/951745 | 14:08 |
elodilles | all the other cycle-trailing projects are ready with their 2025.1 Epoxy releases | 14:08 |
ttx | - Remind the Foundation that the next Release Name selection process should be started (ttx) | 14:09 |
ttx | That was done... | 14:09 |
elodilles | it seems that openstack-ansible needs some more time for their final release | 14:09 |
ttx | - Send weekly email "Between Milestone-1 and Milestone-2" (ttx) | 14:09 |
ttx | That will be done in a few minutes | 14:09 |
ttx | elodilles: yeah that's fine, as long as they know it should be out soon | 14:10 |
elodilles | noonedeadpunk: ^^^ | 14:10 |
ttx | #topic Assign R-14 week tasks | 14:10 |
ttx | Looks like they are all aissgned | 14:10 |
ttx | #topic Review weekly countdown email | 14:11 |
ttx | #link https://etherpad.opendev.org/p/relmgmt-weekly-emails | 14:11 |
ttx | Looking good? | 14:13 |
elodilles | yepp, LGTM | 14:13 |
ttx | OK will send after meet | 14:13 |
frickler | is it too early to add the next PTG date? | 14:13 |
ttx | A bit.. But a good add for the coming weeks | 14:13 |
ttx | #topic Open Discussion | 14:13 |
ttx | I was looking up stale patches and was wondering if we should hold on https://review.opendev.org/c/openstack/releases/+/948217 or if it's good to go | 14:14 |
noonedeadpunk | yeah, we're pretty much having just couple of patches to land: https://review.opendev.org/q/topic:%22osa/el10-stable/2025.1%22 | 14:14 |
elodilles | noonedeadpunk: ACK, thanks for the heads up :) | 14:15 |
ttx | the comments in there are not crystal clear | 14:15 |
elodilles | noonedeadpunk: we are looking forward to the OSA Epoxy release :] | 14:15 |
noonedeadpunk | me too... | 14:16 |
noonedeadpunk | this time we scoped small, delivered in time, and ended up as ususal | 14:16 |
elodilles | ttx: i think we should merge this before the above EOL patch: https://review.opendev.org/c/openstack/project-config/+/951842 | 14:16 |
frickler | ah, yes, that's still on my review list | 14:17 |
ttx | ah ok | 14:17 |
ttx | I'll drop a note that we are holding on that | 14:17 |
*** haleyb is now known as haleyb|out | 14:17 | |
elodilles | ttx: so that the tagging should happen and we don't have to do manual tagging like with the Kolla patch (well, the other option is bobcat state revert, then when everything is ready, revert of the revert) | 14:18 |
elodilles | ttx: yepp, that's a good idea | 14:18 |
ttx | done | 14:18 |
elodilles | thanks o/ | 14:18 |
ttx | anything else we need to cover today? | 14:18 |
frickler | let me mention once again the issue with the oslo.utils 2024.2 release https://review.opendev.org/c/openstack/designate/+/951581/comments/e05b4bfc_300d1141 | 14:18 |
ttx | Enjoy the two-week break | 14:18 |
frickler | somehow it looks like the change to oslo_utils as name is breaking some stuff | 14:19 |
ttx | frickler: yeah that is weird | 14:19 |
elodilles | frickler: are those breaking things all using/related to pkg_resources ? | 14:19 |
frickler | probably, yes. but not easy to replace that in stable branches | 14:20 |
elodilles | how many breaking projects are we aware of? | 14:21 |
frickler | I'm only aware of the designate/neutron/sdk job issue. I think octavia was also affected but got fixed | 14:22 |
elodilles | frickler: ACK, at least those are not that many then | 14:23 |
elodilles | oh, i see there is a 'reinstall' issue during devstack runs as well :S | 14:23 |
frickler | yes, I assume that these are two facets of the same underlying issue, though | 14:24 |
elodilles | ("oslo.utils reinstalled every time when it appears as a dependency" as you wrote) | 14:24 |
elodilles | i think this is not because of our tooling, rather a pip issue, but maybe i'm wrong. i'll look into this when i'll have some time | 14:25 |
frickler | iirc we changed our tooling somehow to generate oslo_utils instead of oslo.utils, something in pbr? | 14:26 |
ttx | I'll add a note to continue duscussing it at next meeting in case we did not make any progress | 14:28 |
frickler | maybe fungi remembers better | 14:28 |
ttx | OK, anything else? | 14:29 |
elodilles | ttx: nothing else from me | 14:30 |
frickler | not from me either | 14:30 |
ttx | alright then | 14:30 |
ttx | #endmeeting | 14:30 |
opendevmeet | Meeting ended Fri Jun 6 14:30:49 2025 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 14:30 |
opendevmeet | Minutes: https://meetings.opendev.org/meetings/releaseteam/2025/releaseteam.2025-06-06-14.00.html | 14:30 |
opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/releaseteam/2025/releaseteam.2025-06-06-14.00.txt | 14:30 |
opendevmeet | Log: https://meetings.opendev.org/meetings/releaseteam/2025/releaseteam.2025-06-06-14.00.log.html | 14:30 |
fungi | oh, checking... | 14:31 |
frickler | fungi: essentially the question is what changed to make the oslo.utils release produce an oslo_utils tarball instead with "." in the name | 14:33 |
fungi | the switch from running `setup.py sdist` to `build` brought in a newer version of setuptools which applies the sdist normalization rules that pypi also now requires | 14:33 |
fungi | sdists are now required to use _ instead of - or . in the name portion of their filenames, even if the sdist name itself has a - or . | 14:34 |
fungi | note that the setuptools version which started forcing that at sdist build time was released almost a year ago now, but we didn't see it for many months because we were ending up running older setuptools before the switch to using pyproject-build as the frontend | 14:36 |
frickler | fungi: right, but where was that change made, in release tooling or in pbr? | 14:36 |
frickler | also iiuc due to the pypi changes there is no way to revert to the old behavior for stable branch releases? | 14:37 |
fungi | frickler: the change took effect for us when https://review.opendev.org/c/zuul/zuul-jobs/+/940273 merged on 2025-01-30 | 14:39 |
fungi | and yes, it's necessary that it apply to stable branch releases too | 14:40 |
fungi | we could specify `legacy_setup: true` on that role to get back the old behavior, but i think at this point pypi is rejecting uploads for non-normalized sdist filenames | 14:41 |
fungi | what's the actual problem it's causing? i'm surprised if the dist filename (not to be confused with the dist name) is referenced in any consuming project's code | 14:43 |
fungi | the dist name should still be able to be oslo.utils, even if the filename for the sdist looks like oslo_utils....tar.gz | 14:43 |
fungi | also that's independent of the import name as well, which could be something else entirely too | 14:44 |
frickler | https://zuul.opendev.org/t/openstack/build/a22badc4461140c9bc7786d5ab29fb3b is a typical failure, with the error seen in the designate-api log | 14:46 |
frickler | https://zuul.opendev.org/t/openstack/build/a22badc4461140c9bc7786d5ab29fb3b/log/controller/logs/screen-designate-api.txt#44 | 14:46 |
fungi | did this start on 2025-01-30 or later? | 14:48 |
frickler | note that we had a similar issue in reqs tooling, needed https://review.opendev.org/c/openstack/requirements/+/948811 before the u-c bump could get merged | 14:48 |
fungi | i guess maybe after the first stable point release for oslo.utils? | 14:48 |
frickler | well this was first seen with the 7.3.1 release which was made well after that date | 14:48 |
frickler | and yes, likely no other similar release before that | 14:49 |
fungi | this seems more like a behavior change in pkg_resources | 14:49 |
frickler | or something is special about oslo.utils compare to other oslo libs | 14:50 |
elodilles | pkg_resources is deprecated as far as i understand and does not handle the new normalized name thing | 14:55 |
fungi | correct, i'm trying to track down the exact reference | 14:55 |
fungi | frickler: i suspect this is part of the cause... https://opendev.org/openstack/oslo.utils/src/branch/stable/2024.2/setup.cfg#L25-L27 | 15:02 |
fungi | ah, that's because oslo_utils is the import name: https://opendev.org/openstack/oslo.utils/src/branch/stable/2024.2/oslo_utils | 15:04 |
fungi | so probably unrelated | 15:05 |
fungi | metadata.name lists it as oslo.utils, which is what should show up in the pkg_resources list of installed dists | 15:05 |
fungi | looking at both the stable branch copy of bin/designate-api-wsgi vendored in the designate repo and the current version of the same in the string literal in pbr.packaging._wsgi_text it's unclear to me how pkg_resources is coming into that traceback in the example failure | 15:13 |
fungi | this might be something stephenfin has more context for, since he's tackled a lot of the wsgi script churn recently | 15:15 |
fungi | i guess this is related to something searching for plugin entrypoints at startup? | 15:19 |
frickler | hmm, we have pbr pinned to 6.1.0 in 2024.2 u-c, maybe we can bump that? | 15:22 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!