| stephenfin | frickler: elodilles_pto: hberaud: Sorry, wasn't checking IRC Friday. Rather than reverting the many patches, could we cut a new 12.2.0 release and point the branch there? | 08:40 |
|---|---|---|
| stephenfin | You're going to break a lot of gates if you cut a new 14.1 or 15.0 release since we've been adding jobs to test against SQLAlchemy main and oslo.db master branches, and removing those changes from oslo.db will re-break compatibility with SQLAlchemy 2.x | 08:41 |
| hberaud | stephenfin: I'm not sure our tooling will allow us to create a 12.2.0 from the bobcat deliverable file | 08:44 |
| hberaud | however, I agree that it would be really more easy to do that | 08:44 |
| hberaud | also the problem with a 12.2.0 would be that it wouldn't solve the issue related to the version proposed with bobcat https://releases.openstack.org/bobcat/index.html#bobcat-oslo-db | 08:47 |
| hberaud | people would continue to see 13 and 14 as available for bobcat | 08:47 |
| hberaud | so a 15.0.0 would be required at least | 08:47 |
| hberaud | however 15.0.0 could refer to the sha to the right point on the branch | 08:48 |
| hberaud | as you suggested in your previous comment | 08:48 |
| hberaud | thoughts? | 08:49 |
| stephenfin | that would work | 08:49 |
| stephenfin | have you a full list of projects that are broken with oslo.db master right now? | 08:49 |
| hberaud | AFAIK heat, aodh, masakari, manila | 08:51 |
| frickler | + cinder and maybe more, see https://review.opendev.org/c/openstack/requirements/+/879743 | 09:04 |
| frickler | I have prepared a series of partial reverts already, I'm going to push that anyway, then we can see which path seems easier | 09:04 |
| frickler | a series reverting some of the patches in 2023.2, that is, not all of them | 09:05 |
| frickler | we cannot reasonably change the root of the 2023.2 branch and releasing something for bobcat not from the head of that branch would seem very weird to me | 09:06 |
| frickler | stephenfin: hberaud: ^^ | 09:06 |
| frickler | I did some local testing and my reverted branch at least passes with cinder and does not have additional failures with heat unit tests (there seem to be some unrelated issues there) | 09:08 |
| stephenfin | frickler: that's not accurate since that bumps SQLAlchemy and Alembic also | 09:08 |
| stephenfin | the autogenerated oslo.db-only patches are probably more relevant | 09:09 |
| stephenfin | i.e. https://review.opendev.org/c/openstack/requirements/+/887261 and https://review.opendev.org/c/openstack/requirements/+/893043 | 09:09 |
| frickler | stephenfin: yes, but from the comment in https://review.opendev.org/c/openstack/oslo.db/+/878146 oslo.db requires sqla 2.0.5 at least? | 09:11 |
| frickler | also from the above I see the same projects mostly, just ceilometer instead of cinder | 09:13 |
| stephenfin | I'm re-examining that this morning. I think I was wrong. Looking at it further last week, we do have some compat code to handle sqlalchemy 1.x vs. 2.x differences | 09:13 |
| stephenfin | I'm now thinking that the big difference is probably the removal of autocommit | 09:13 |
| stephenfin | (which isn't supported in SQLALchemy 2.x, but Heat, Manila etc. are still relying on it | 09:13 |
| stephenfin | ) | 09:13 |
| frickler | stephenfin: there is also the removal of oslo_db.sqlalchemy.test_migrations , see e.g. the manila failure https://zuul.opendev.org/t/openstack/build/cf97e9e6b47d43df84f5e19f64a7e21e | 09:17 |
| stephenfin | also re-added that | 10:03 |
| opendevreview | Jiri Podivin proposed openstack/releases master: Version of the role was recently bumped in metadata https://review.opendev.org/c/openstack/releases/+/895654 | 12:46 |
| frickler | ttx: elodilles_pto: copying this from the oslo channel talking about whether to attach a bug report to the reverts (and their renos in particular). is there a release team bugtracker at all? would it make sense to cover both castellan and oslo.db together there? | 12:55 |
| *** ykarel is now known as ykarel|away | 15:04 | |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL ironic-prometheus-exporter stable/train https://review.opendev.org/c/openstack/releases/+/895702 | 16:50 |
| JayF | When EOL'ing multiple series for a project, do you want separate PRs for each is or one large one OK? | 16:52 |
| JayF | I am going to retire pike-xena of networking-generic-switch; and that can be one or 6 PRs | 16:52 |
| * JayF is going to do separate PRs in lack of further guideance | 16:57 | |
| JayF | So I'm confused about the state of this repo: stable/pike branch exists, but it is unreferenced in deliverables. My hunch: it never got a pike release. I'm going to PR to create the yaml file and immediately EOL pike release, but I think it may end up needing manual cleanup | 17:11 |
| opendevreview | Jay Faulkner proposed openstack/releases master: Fix/retire networking-generic-switch pike https://review.opendev.org/c/openstack/releases/+/895705 | 17:13 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL networking-generic-switch stable/train https://review.opendev.org/c/openstack/releases/+/895707 | 17:16 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL networking-generic-switch stable/ussuri https://review.opendev.org/c/openstack/releases/+/895709 | 17:19 |
| opendevreview | Jay Faulkner proposed openstack/releases master: Fix/EOL networking-generic-switch pike https://review.opendev.org/c/openstack/releases/+/895705 | 17:20 |
| frickler | JayF: likely that release was made before release automation was set up for that project. tc members should be able to also eol those manually, but best check with elodilles_pto when they're back | 17:21 |
| JayF | frickler: looking at the ocata/networking-generic-swtich.yaml file, it looks similar to the one I proposed for pike | 17:22 |
| JayF | so I *think* merging that should do what we want and bring the yaml in line with the repo | 17:22 |
| JayF | (it had no automation-created releases for ocata, either) | 17:22 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL networking-generic-switch stable/victoria https://review.opendev.org/c/openstack/releases/+/895711 | 17:23 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL networking-generic-switch stable/wallaby https://review.opendev.org/c/openstack/releases/+/895712 | 17:25 |
| frickler | JayF: oh, there's precedence for that, nice, let me check | 17:27 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL networking-generic-switch stable/xena https://review.opendev.org/c/openstack/releases/+/895713 | 17:30 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL python-ironicclient stable/train https://review.opendev.org/c/openstack/releases/+/895714 | 17:34 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL python-ironicclient stable/ussuri https://review.opendev.org/c/openstack/releases/+/895717 | 17:36 |
| JayF | frickler: btw; IDK if you picked up on it, but this is clearing zuul-config-errors for ironic | 17:36 |
| JayF | Happy to be part of the solution instead of the precipitate now :P | 17:37 |
| frickler | JayF: oh, indeed I didn't remember which branches were affected, but this is nice, thx | 17:38 |
| frickler | interesting, that ocata patch was even proposed by elodilles_pto , so I don't understand why we cannot use this for other pending cleanups | 17:39 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL python-ironicclient stable/victoria https://review.opendev.org/c/openstack/releases/+/895719 | 17:46 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL python-ironicclient stable/wallaby https://review.opendev.org/c/openstack/releases/+/895720 | 17:48 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL python-ironicclient stable/xena https://review.opendev.org/c/openstack/releases/+/895721 | 17:50 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL python-ironicclient stable/xena https://review.opendev.org/c/openstack/releases/+/895721 | 18:08 |
| opendevreview | Jay Faulkner proposed openstack/releases master: EOL networking-generic-switch stable/xena https://review.opendev.org/c/openstack/releases/+/895713 | 18:10 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!