| elodilles | hmm, there are some *new* final RC patches, and a tempest release as well. I thought that maybe we are too late for that, but i'm still hesitant | 07:53 |
|---|---|---|
| elodilles | hberaud ttx ^^^ | 07:53 |
| opendevreview | Slawek Kaplonski proposed openstack/releases master: Final RC patch for neutron https://review.opendev.org/c/openstack/releases/+/877654 | 08:13 |
| hberaud | elodilles: I replied with a comment on the patch | 08:23 |
| hberaud | elodilles: One solution could be to cut the RC2 at the commit before the one who requires the tempest changes, and then release it next week, once the final release is past. | 08:26 |
| hberaud | what do you think about that? | 08:26 |
| elodilles | hberaud: sounds OK, though we have to discuss that with cinder and tempest folks as well, so it takes time too, and then we delay cinder's RC2 as well :/ | 08:43 |
| hberaud | yeah | 08:43 |
| elodilles | I'm tend to accept the tempest release now that I read the story, though it's a bit unfortunate and we should avoid similar situations in the future :S | 08:44 |
| hberaud | whoami-rajat: FYI ^ | 08:45 |
| elodilles | we have clear deadlines, of course exceptions can be made, but those should be well assessed | 08:45 |
| hberaud | yeah | 08:46 |
| elodilles | at least that is my view from release management perspective :) | 08:46 |
| hberaud | Concerning myself I'd prefer to ignore those tempest changes and requirements | 08:46 |
| hberaud | but if we decide to continue with that I won't block | 08:47 |
| elodilles | OK, so you say that we should not release now Tempest 34.1.0, right? | 08:48 |
| hberaud | yes | 08:48 |
| hberaud | and cut cinder's RC2 the commit before the one related to the tempest changes | 08:49 |
| elodilles | hberaud: ++ | 08:49 |
| elodilles | sounds good to me. | 08:49 |
| hberaud | and I'd also ignore the skyline patches | 08:49 |
| elodilles | yepp, that should be our safest bet | 08:51 |
| elodilles | though they were proposed on Friday ~ 3 PM (CET) | 08:52 |
| elodilles | so technically within the deadline we stated (March 17th) | 08:52 |
| elodilles | so I am hesitant with that as well | 08:54 |
| hberaud | indeed | 08:56 |
| hberaud | I don't think the skyline patches can be dangerous (in a widely manner), though | 08:56 |
| hberaud | so lets +2 skyline | 08:57 |
| elodilles | yes, they seems to be like bugfix + packaging patches | 08:57 |
| hberaud | because you are right the deadline was respected | 08:57 |
| elodilles | we should make sure next time that rc2 patches are proposed much earlier :) | 08:58 |
| elodilles | (when I generated the rc2 patches I didn't find patches on skyline afair) | 08:58 |
| hberaud | and indeed they propose lot of fixes | 08:58 |
| elodilles | neutron rc2 was also updated, can you review that, too? https://review.opendev.org/c/openstack/releases/+/877654 | 08:58 |
| whoami-rajat | elodilles, hberaud IIUC, we are going to ignore the cinder and tempest changes for RC2 but they're going to make it into the final 2023.1 GA release, right? | 09:01 |
| elodilles | whoami-rajat: nope | 09:01 |
| elodilles | whoami-rajat: if we ignore them for RC2 then they won't be part of the GA | 09:02 |
| elodilles | (as latest RC patches would be marked as final release) | 09:03 |
| whoami-rajat | hmm, that doesn't sound good, the point of delaying RC2 was to include the cinder change, it is a major issue which came to our notice few weeks ago and was important to fix it in the 2023.1 release | 09:03 |
| whoami-rajat | what is the problem with inclusion in 2023.1? we can't release tempest right now? | 09:04 |
| hberaud | whoami-rajat: we are hesitant to release a new version of tempest so close to the final deadline | 09:05 |
| hberaud | because if it contains an unforseen problem it could broke everything | 09:06 |
| hberaud | that's our main point | 09:06 |
| elodilles | whoami-rajat: yes, cycle-with-intermediary release deadline is way over. we don't say it's impossible to release now, but a release has to be well justified | 09:06 |
| whoami-rajat | ok, gmann and I can guide through the changes if it helps understanding the risk situation of the changes | 09:07 |
| elodilles | whoami-rajat: if you say that it is a major issue, and it is *safe* to release, then we can grant the exception | 09:08 |
| elodilles | hberaud: right? ^^^ | 09:08 |
| hberaud | WFM | 09:08 |
| whoami-rajat | I agree it is a backward incompatible change but it's for a specific cinder feature, "multiattach" and we want to intentionally break the backward compatibility for those who are using it the wrong way (since it could lead to a data corruption issue) | 09:09 |
| whoami-rajat | not all deployments enable multiattach so the tests won't run for them | 09:09 |
| hberaud | whoami-rajat: so you say the tempest changes are really isolated? | 09:10 |
| whoami-rajat | can we wait for gmann's response ? I think he can better explain the deployment scenarios of tempest and where it can/cannot break | 09:10 |
| hberaud | and won't impact the other deliverables? | 09:10 |
| whoami-rajat | hberaud, it's isolated for a specific cinder feature | 09:10 |
| whoami-rajat | hberaud, with other deliverables, do we mean other projects like nova, neutron etc? | 09:11 |
| hberaud | yeah | 09:11 |
| elodilles | and all the rest :) | 09:11 |
| hberaud | I meant the other deliverables that also relies on tempest | 09:12 |
| whoami-rajat | since nova attaches the volume, the tests include nova and cinder but IIRC not other projects, glance for creating bootable volumes but yeah mostly they're nova cinder related changes | 09:12 |
| whoami-rajat | apart from nova and cinder, I don't think it should affect anything else | 09:13 |
| hberaud | can we do tests against nova with this version to ensure that its gates are not broken? | 09:14 |
| hberaud | (this version of tempest) | 09:14 |
| whoami-rajat | sure, we can | 09:16 |
| whoami-rajat | also the testing now relies on devstack so the deployment needs to happen with devstack to run these tests successfully | 09:17 |
| whoami-rajat | else there are some manual config changes that need to be done in tempest (if deployment tool is not devstack) | 09:17 |
| whoami-rajat | I think tempest gate that approved the changes already runs all the relevant tests in tempest-full-py3, tempest-slow-py3 | 09:18 |
| whoami-rajat | also it runs tests against Zed and Xena branches | 09:18 |
| ttx | Hi hberaud and elodilles , I'm back | 09:20 |
| hberaud | ttx: o/ | 09:20 |
| ttx | let me know if there is anything that needs my urgent attention, otherwise I'll continue to process email | 09:20 |
| hberaud | ttx: yeah we need your feedback about our current discussion | 09:22 |
| ttx | ok reading immediate backlog | 09:22 |
| hberaud | ttx: thanks | 09:22 |
| hberaud | we are discussing about a last minute release of tempest, required by cinder's RC2. elodilles and I thinks we are a bit late to release a new version of tempest, however, cinder require this new version because it fix cinder issues. | 09:24 |
| ttx | If the tempest changes are very limited in scope I'd be tempted to grant it... We still have two days | 09:25 |
| ttx | Let me have a look | 09:25 |
| hberaud | that's me or zuul is running nothing? https://zuul.opendev.org/t/opendev/status | 09:27 |
| ttx | It's you https://zuul.opendev.org/t/openstack/status | 09:28 |
| hberaud | I 'rechecked' a patch ~1h ago so I'd expect to see it enqueued | 09:28 |
| ttx | wrong url | 09:28 |
| hberaud | ah yes, good catch | 09:29 |
| hberaud | thanks | 09:29 |
| ttx | That tempest change is slightly bigger than expected, but still mostly focused on cinder | 09:29 |
| ttx | If that's the only thing we have in terms of exceptions, I'd support accepting it asap | 09:29 |
| hberaud | yeah that's our main exception | 09:30 |
| ttx | same for neutron and skyline... get all the things in this morning and freeze there | 09:31 |
| hberaud | I +2'd the tempest patch | 09:32 |
| ttx | It's better to get that cinder change in now than later | 09:32 |
| hberaud | yeah | 09:32 |
| ttx | since it's breaking compat | 09:32 |
| ttx | ok I'll +2a it | 09:33 |
| hberaud | I also +2d the skyline and neutron patches | 09:34 |
| hberaud | ttx: FYI I decided to wait this morning to generate the final rc patch, I'll generate it soon | 09:35 |
| ttx | +1 | 09:36 |
| hberaud | (to wait those last RC2) | 09:36 |
| ttx | Do we need to hold on the Cinder release for Tempest to be released? | 09:36 |
| hberaud | good question | 09:36 |
| hberaud | can't hurt | 09:36 |
| ttx | i also like to test that we can still merge things after that lands :) | 09:37 |
| ttx | rather than discovering it Wednesday | 09:37 |
| hberaud | yeah good idea | 09:38 |
| * ttx goes back to purging inbox | 09:38 | |
| elodilles | note that this skyline-apiserver rc2 is exactly the *same* as rc1, just realised it: https://review.opendev.org/c/openstack/releases/+/877791 | 09:38 |
| hberaud | good catch I missed that point | 09:40 |
| elodilles | also, we shouldn't forget about oslo.log :/ | 09:44 |
| elodilles | as I see one of the revert patch had a timeout O.o | 09:46 |
| elodilles | I've rechecked it now | 09:46 |
| elodilles | +1: when one of you have a bit of a time for a quick review for a reno link patch: https://review.opendev.org/c/openstack/releases/+/877288 | 09:47 |
| hberaud | thanks, I requested Stephen's review | 09:48 |
| elodilles | hberaud: cool, thanks! | 09:48 |
| opendevreview | Merged openstack/releases master: Release skyline-console RC2 for Antelope https://review.opendev.org/c/openstack/releases/+/877793 | 09:49 |
| opendevreview | Merged openstack/releases master: Final RC patch for neutron https://review.opendev.org/c/openstack/releases/+/877654 | 09:51 |
| opendevreview | Merged openstack/releases master: Release tempest 34.1.0 for 2023.1 https://review.opendev.org/c/openstack/releases/+/877840 | 09:51 |
| hberaud | ttx: tempest is now merged, I think we can go with cinder | 09:56 |
| ttx | wfm | 10:01 |
| opendevreview | Merged openstack/releases master: Add release notes links for 2023.1 Antelope https://review.opendev.org/c/openstack/releases/+/877288 | 10:04 |
| opendevreview | Hervé Beraud proposed openstack/releases master: Fix propose-final-releases command to use release-id https://review.opendev.org/c/openstack/releases/+/877819 | 10:06 |
| opendevreview | Merged openstack/releases master: Final RC patch for cinder https://review.opendev.org/c/openstack/releases/+/877652 | 10:20 |
| hberaud | Now I'm gonna go to generate the final rc patch. | 10:45 |
| hberaud | ttx: how do you want to test that things can be merged now that the tempest patch is merged? | 10:47 |
| ttx | The Cinder patch merged after it, so we should be good | 10:49 |
| hberaud | yeah but cinder was tailored to work with this release so I'm not sure this is a significant test... | 10:50 |
| elodilles | meanwhile I got another question, namely magnum wants an rc2, but I suggested there to not do that but rather release gate fixes after the official 2023.1 Antelope release: https://review.opendev.org/c/openstack/magnum/+/876866 | 11:00 |
| opendevreview | Hervé Beraud proposed openstack/releases master: antelope final rc https://review.opendev.org/c/openstack/releases/+/877942 | 11:01 |
| elodilles | probably it could just fit in before the release, but even the backports are not yet proposed, so it wouldn't be that convenient to push through everything... | 11:02 |
| hberaud | elodilles: yeah I agree with you | 11:03 |
| hberaud | maybe we could use release-test instead to do this test | 11:03 |
| elodilles | release-test does not have tests as far as I know. could it be used to test this? | 11:13 |
| hberaud | no idea | 12:32 |
| hberaud | if it doesn't have tempest tests then that's ot really useful | 12:32 |
| hberaud | *not | 12:33 |
| elodilles | i think release-test is more like testing the release process itself | 12:54 |
| elodilles | hberaud: oslo.log patches got +2s from stephen: https://review.opendev.org/q/topic:revert-oslo-5.1.0-functional-content | 13:39 |
| ralonsoh | elodilles, hi, sorry, I was in PTO. slaweq fixed the patch. Thanks! | 16:16 |
| elodilles | ralonsoh: oh, i see, yes, fortunately slaweq helped us out so rc2 is now released and final release patch is uploaded as well \o/ | 16:17 |
| ralonsoh | thanks! | 16:18 |
| elodilles | np :) | 16:18 |
| gmann | elodilles: hberaud ttx: thanks for understanding and grant the exception for tempest change. Change is mostly for cinder(for whoami-rajat cinder fix for multiattach) and Tempest change is tested in gate (for stable/xena till current master) since Mar 17. | 17:34 |
| opendevreview | Elod Illes proposed openstack/releases master: Release oslo.log for 2023.1 Antelope https://review.opendev.org/c/openstack/releases/+/877999 | 18:42 |
| elodilles | gmann: ack, no worries. (it was a bit unfortunate, we should try to avoid such late releases next time) | 18:44 |
| gmann | elodilles: completely agree. | 18:59 |
| elodilles | hberaud ttx : the oslo.log release patch: https://review.opendev.org/c/openstack/releases/+/877999 | 19:07 |
| opendevreview | Ghanshyam proposed openstack/releases master: Add releasenotes link for Tempest antelope release https://review.opendev.org/c/openstack/releases/+/878003 | 19:08 |
| opendevreview | Elod Illes proposed openstack/releases master: Add release notes links for 2023.1 Antelope https://review.opendev.org/c/openstack/releases/+/878004 | 19:28 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!