| *** bauzas_ is now known as bauzas | 01:15 | |
| ttx | haleyb: better if you propose it, as the one we propose is later so it gives the RC a bit less soaking time | 06:48 |
|---|---|---|
| opendevreview | Boxiang Zhu proposed openstack/releases master: Add Skyline highlights (Dalmatian release) https://review.opendev.org/c/openstack/releases/+/930264 | 07:40 |
| opendevreview | Boxiang Zhu proposed openstack/releases master: Add Skyline highlights (Dalmatian release) https://review.opendev.org/c/openstack/releases/+/930264 | 07:42 |
| opendevreview | Boxiang Zhu proposed openstack/releases master: Add Skyline highlights (Dalmatian release) https://review.opendev.org/c/openstack/releases/+/930264 | 07:50 |
| zigo | hberaud: Hi there! Do you know the situation wrt to Eventlet and Python 3.13 ? | 09:09 |
| zigo | Is the patch at https://github.com/eventlet/eventlet/pull/966 enough to fix all 3.13 issues? | 09:09 |
| zigo | Looks like it doesn't. | 09:15 |
| zigo | Here's what I get when I build greenlet with 3.13 and then run eventlet unit tests: | 09:34 |
| zigo | https://paste.opendev.org/show/bUnyOgZa4DZLJnNsdKG7/ | 09:34 |
| zigo | Looks like we're again stuck with a dnspython issue ... :/ | 09:34 |
| zigo | After rebuilding greenlet + cffi for 3.13, I get these when building eventlet: | 10:51 |
| zigo | https://paste.opendev.org/show/bIlUz1vnxwPegQo7qNue/ | 10:51 |
| zigo | So it's not fully addressed... :/ | 10:51 |
| elodilles | strange that https://review.opendev.org/c/openstack/nova/+/930007 patch was merged, promote job succeeded, and still we don't have reno page: https://docs.openstack.org/releasenotes/nova/2024.2.html | 11:10 |
| elodilles | bauzas: fyi ^^ | 11:10 |
| elodilles | does anyone see how this could happen? | 11:12 |
| elodilles | maybe because the stable/2024.2 job patch finished just after the job on master? https://zuul.opendev.org/t/openstack/builds?job_name=promote-openstack-releasenotes&project=openstack/nova | 11:15 |
| ttx | maybe fungi can help us understand what happened there ^ | 11:22 |
| fungi | usually it's a race condition with two release notes jobs running at the same time for different branches, where the one that was from before the branch existed finishes slower than the one with the branch present. i'll see if i can match up timings | 12:43 |
| bauzas_ | elodilles: looking | 13:10 |
| *** bauzas_ is now known as bauzas | 13:14 | |
| elodilles | bauzas: fungi wrote already about the root cause ^^^ | 13:16 |
| elodilles | fungi: ACK, thanks! | 13:16 |
| bauzas | how can we fix this ? | 13:17 |
| fungi | well, i'm only just now looking | 13:18 |
| fungi | https://zuul.opendev.org/t/openstack/build/3ed83f21d65c4b15b26412a8c11160e0/log/job-output.txt#150-154 says the build with the stable branch was copied into afs between 09:48:02-09:48:06 and then https://zuul.opendev.org/t/openstack/build/cb235df6a8f74e659493202cb07ca52d/log/job-output.txt#147-151 shows the version without the stable branch getting copied between 09:46:53-09:46:56 | 13:20 |
| fungi | so the next thing to check is whether the build from 930007 actually included the stable branch notes | 13:21 |
| fungi | the preview build for it from the gate pipeline does contain the desired page: https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_d07/930007/1/gate/build-openstack-releasenotes/d07f992/docs/2024.2.html | 13:23 |
| fungi | oh, wait 930007 is the master branch change though | 13:26 |
| fungi | okay, so where i got confused is that the stable branch change (930006) does not include the page for its release notes: https://47f28db2c59a0788c3af-c883f70fc8be918d3693e4049b00789d.ssl.cf1.rackcdn.com/930006/1/gate/build-openstack-releasenotes/46b006f/docs/ | 13:28 |
| fungi | and that published after the master branch change whose build did contain release notes for the stable branch | 13:29 |
| fungi | so looks like if we reenqueue 930007,1 (master branch change) into the promote pipeline, that will publish the content with the new stable branch release notes | 13:31 |
| fungi | i'll do that momentarily | 13:31 |
| bauzas | thanks | 13:32 |
| *** bauzas_ is now known as bauzas | 13:41 | |
| haleyb | ttx: thanks, will do shortly | 13:50 |
| fungi | https://docs.openstack.org/releasenotes/nova/2024.2.html works now, so yeah it was a matter of the two promote jobs being triggered so close together that the one which merged first got delayed (probably waiting for its node assignment) and ended up running later than the change which had merged second | 14:02 |
| fungi | since the promote pipeline is not a dependent pipeline, there is no sequencing guarantee for changes on different projects or on different branches of the same project | 14:03 |
| elodilles | fungi: ACK, thanks for fixing it o/ | 14:09 |
| bauzas | fungi: thanks a lot ! | 14:10 |
| fungi | if those two changes had been merged a little farther apart in time, it would have worked as desired | 14:10 |
| fungi | in zuul development we've talked about things that might fix this scenario, e.g. dynamic per-project semaphores or a branch-agnostic supersedent pipeline manager | 14:13 |
| fungi | it's been a while though... clarkb ^ are you aware of any more recent feature additions that might be able to address the problem of builds for different branches of a project happening in parallel in a supercedent pipeline? | 14:31 |
| clarkb | no I am not aware of anything | 14:53 |
| clarkb | I think if it were me I would remove the semaphore then use a failed_when and try to inspect the failure and determine if the failure is ignorable | 14:54 |
| clarkb | maybe the rc code from rsync is different or you can check for the tmp file rename error message in stdout/stderr | 14:56 |
| clarkb | oh wait you're talking about the sequencing issue above not the every release we remove the semaphore issue | 14:56 |
| clarkb | I think for that a semaphore woudl solve it but then you're back to the other problem | 14:57 |
| clarkb | you could also ensure that publication occurs into distinct locations | 14:57 |
| opendevreview | Brian Haley proposed openstack/releases master: RC2 patch for neutron https://review.opendev.org/c/openstack/releases/+/930345 | 15:01 |
| frickler | ttx: elodilles: are we ready for branching qa + reqs? | 15:03 |
| elodilles | frickler: we still need horizon to branch: https://review.opendev.org/c/openstack/releases/+/930226 | 15:04 |
| elodilles | fortunately we just got PTL-Approved+1 on it | 15:05 |
| elodilles | so it's waiting for a 2nd core review | 15:05 |
| frickler | +3 | 15:07 |
| elodilles | ~o~ | 15:07 |
| opendevreview | Merged openstack/releases master: Cut stable/2024.2 branch for horizon https://review.opendev.org/c/openstack/releases/+/930226 | 15:17 |
| elodilles | gmann kopecmartin: horizon stable cut patch has merged, please go on with the QA release activities ^^^ | 15:18 |
| elodilles | release-team: we still have these to decide what we should do with: https://review.opendev.org/q/topic:dalmatian-rc1-deadline+is:open | 15:22 |
| elodilles | as officially these are needed for requirements branching | 15:22 |
| elodilles | so we should decide ASAP what to do with them. | 15:22 |
| elodilles | either delete them from deliverables (we didn't allow this in the past after milestone-2) or release them broken (as we had to do in the past a couple of occasions) | 15:23 |
| opendevreview | Merged openstack/releases master: [ironic] Add automatic_lessee to cycle highlights https://review.opendev.org/c/openstack/releases/+/928305 | 15:24 |
| frickler | I think https://review.opendev.org/c/openstack/requirements/+/929094 should also get merged before branching reqs | 15:24 |
| frickler | elodilles: I think it would be good to discuss in the TC meeting later today. cc gouthamr | 15:25 |
| frickler | (also sidenote: as you can see I'm back this week, but will be out for the next three weeks. so yes, I'll miss the big fun, too) | 15:26 |
| elodilles | frickler: oh, a new-release patch, indeed | 15:27 |
| elodilles | frickler: i'm sad to hear that :( and looking forward to your future reviews and help o:) | 15:42 |
| ttx | frickler: would be good to know if we can force-merge https://review.opendev.org/c/openstack/ansible-role-thales-hsm/+/930181 | 16:09 |
| ttx | and unblock at least this rc1 | 16:09 |
| ttx | Grzegorz Grasza: any chance you could have a look into https://review.opendev.org/c/openstack/ansible-role-thales-hsm/+/930181 ? We'd like to unblock this Barbican deliverable | 16:13 |
| ttx | elodilles: did you get any answer to your Zun email? | 16:14 |
| frickler | ttx: well technically we (infra-root) can do that, but I'm not sure how helpful that would be in the global view of things. also maybe asking for TC approval might be better? | 16:15 |
| elodilles | ttx: no answer yet | 16:16 |
| ttx | frickler: yeah, definitely TC | 16:16 |
| ttx | frickler: any chance you could relay that? | 16:17 |
| ttx | elodilles: ok will ping again | 16:17 |
| frickler | ttx: I'll try to, though I may end up ending early today. I'm hoping my ping for gouthamr will take care of this, though | 16:19 |
| fungi | yes, i'm happy to bypass inactive reviewers and approve important changes at the direction of the tc | 16:20 |
| gouthamr | +1; adding this to the TC meeting agenda… frickler: would you still be around to summarize this/seed the discussion? | 16:21 |
| fungi | ultimately it's the tc who has the decision making power over openstack's software. they delegate it to project leaders who in turn delegate it to reviewers, but the tc can also delegate that permission to me or another gerrit admin if desired | 16:22 |
| frickler | gouthamr: I'll try to, but I'm having a bit of a headache and I might have to close down early today | 16:23 |
| gouthamr | if not, let me attempt to capture the problem: we have a few unreleased components delaying our coordinated release; we need to move these forward without their PTL/release liaison approvals - because we didn’t have them in time.? | 16:23 |
| frickler | gouthamr: it is even worth, we did proceed already with some other projects like you described, but didn't for these remaining 4 because they also have a failing CI | 16:24 |
| frickler | gouthamr: the patch that ttx mentioned above would fix the CI for one project at least | 16:24 |
| gouthamr | ah ++ | 16:24 |
| frickler | s/worth/worse/ | 16:25 |
| gouthamr | elodilles: was the zun email sent only to the maintainers? | 16:38 |
| gouthamr | (checking so I gather links/context) | 16:40 |
| elodilles | gouthamr: yes, we discussed to send a direct mail to PTL and not to mailing list | 16:40 |
| gouthamr | elodilles: ++ ack | 16:41 |
| * gouthamr pings some Barbican maintainers downstream with https://review.opendev.org/c/openstack/ansible-role-thales-hsm/+/930181 | 16:42 | |
| gouthamr | awesome! That patch should be on its way to merge now :) | 16:51 |
| frickler | so down to three then, thx | 16:51 |
| fungi | it takes a village | 16:52 |
| opendevreview | Ghanshyam proposed openstack/releases master: Release Tempest 41.0.0 as final tag for 2024.2 https://review.opendev.org/c/openstack/releases/+/930093 | 17:09 |
| gmann | elodilles: thanks, ack. I will start that today | 17:17 |
| gmann | elodilles: ^^ tempest final release is ready | 17:17 |
| elodilles | ACK +2'd | 17:26 |
| opendevreview | Goutham Pacha Ravi proposed openstack/releases master: Release ansible-role-thales-hsm RC1 for 2024.2 Dalmatian https://review.opendev.org/c/openstack/releases/+/928512 | 18:47 |
| opendevreview | Ghanshyam proposed openstack/releases master: Create stable/2024.2 for devstack https://review.opendev.org/c/openstack/releases/+/930381 | 20:56 |
| opendevreview | Ghanshyam proposed openstack/releases master: Create stable/2024.2 for grenade https://review.opendev.org/c/openstack/releases/+/930382 | 20:58 |
| gmann | elodilles: frickler ttx ^^ | 20:59 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!