Tuesday, 2022-03-29

fungithe other trove backport (835494) is also failing a bunch of jobs01:31
fungii suspect it may need to depends-on the passing commit01:33
fungiTypeError: __init__() got an unexpected keyword argument 'tenant'01:33
fungiyeah01:34
fungii'll update the commit message01:34
fungier, rebase i mean01:34
fungi! [remote rejected]   HEAD -> refs/for/master%topic=stable/yoga (change https://review.opendev.org/c/openstack/trove/+/834373 closed)01:44
opendevmeetfungi: Error: "remote" is not a valid command.01:44
fungithis is very confusing01:44
fungioh01:44
fungithere we go. the branch is still missing its .gitreview change, that was why01:45
fungilooks like rebasing it has gotten the unit tests passing02:17
fungiunfortunately my murano fix is still not coming along as i'd hoped02:17
fungii worry i'm a bit out of my depth with it, and it's getting late02:41
fungithe current iteration is spewing tracebacks like https://zuul.opendev.org/t/openstack/build/8f5d43c35053445985f62dddc2e1745f/log/job-output.txt#1426-144802:41
fungii guess i need to update the db sessions to use project_id as well02:43
fungier, no, it must be that package_upload() is getting passed a nonetype tenant_id02:45
fungii'll ask on the ml for help02:45
fungithe stable/yoga backports for trove have been approved for an hour now but are still not in the gate pipeline03:23
fungiapparently it just wanted a recheck. maybe something about how i originally applied the workflow and code review votes failed to match the comment-added event03:33
*** abhishekk is now known as akekane|home04:54
*** akekane|home is now known as abhishekk04:54
*** marios is now known as marios|ruck05:22
ttxI can live with a broken Murano release if it comes down to it06:46
ttxfungi: Thanks a lot for driving this overnight!06:47
*** jbadiapa|off is now known as jbadiapa07:20
ttxapproved Sahara rc207:40
elodillesyepp, thanks for working on it07:42
ttxpostign trove rc2 now07:42
elodillesttx: ack07:43
elodillesttx: i wanted to make the generated patch to merge, but it seems the recheck did not work ( this https://review.opendev.org/c/openstack/trove/+/832972 and this https://review.opendev.org/c/openstack/trove/+/832974 )07:44
ttxworking off https://review.opendev.org/c/openstack/trove/+/83549207:45
elodillesoh, the recheck worked, but it is still on the gate :S07:45
elodilles(in check queue)07:45
opendevreviewMerged openstack/releases master: Final RC patch for sahara  https://review.opendev.org/c/openstack/releases/+/83503707:47
elodillesbut anyway, they are not that important since we are in lack of time07:47
opendevreviewThierry Carrez proposed openstack/releases master: Release Trove Yoga rc2  https://review.opendev.org/c/openstack/releases/+/83554407:47
ttxThat leaves murano, Andy Botting proposed a new rev on the patch07:48
ttxcurrently in check07:48
ttxIf we get a patch that works we can backport it and fungi can approve them, then we can regenerate our release patch alright07:49
elodillessounds good!07:49
elodillesttx: about the other, setuptools problem: if i understood correctly you also preferred temporary pinning setuptools (with this patch: https://review.opendev.org/c/openstack/releases/+/835423 )07:53
elodilles(though 1 out of 3 in-place workarounds has landed: https://review.opendev.org/q/topic:setuptools-issue-3197 , but those would also need last-minute re-releases)07:54
ttxyes approved07:54
elodillesttx: awesome, thanks!07:55
opendevreviewMerged openstack/releases master: Pin setuptools for Yoga release  https://review.opendev.org/c/openstack/releases/+/83542308:04
ttxFWIW in the case we can't get Murano fixed very soon, I would be fine releasing Murano as-is and fix it in stable release. Like I said, it's not the first time it happened (we usually realize it's broken after the release :))08:17
ttxIdeally we'd post the release patch at least 24 hours before approving it to give people tie to +1 it08:18
ttxSo if we don;t have a new Murano release in a couple of hours, I would move on08:19
elodillesttx: i agree08:25
ttx(it seems to fail tests on more than just that oslo.context thing)08:26
elodillesyes, i just looked the results and saw now we are at the same failures than with the 'lower-constraints' job (which means, with oslo.context<4.0.0)08:28
elodillesthis has passed meanwhile if you want to +2 it: https://review.opendev.org/c/openstack/releases/+/83554408:29
elodilles(your trove release patch)08:29
ttxdone08:30
elodilles\o/08:31
opendevreviewMerged openstack/releases master: Release Trove Yoga rc2  https://review.opendev.org/c/openstack/releases/+/83554408:44
elodilleshmm, we have this as well: https://review.opendev.org/c/openstack/releases/+/83546109:33
elodillesthough this maybe contain a high important bug fix, but it's not release-critical and does not affect installation09:37
elodillesin my opinion09:38
ttxSince we are late anyway I'd consider fixing it, though09:38
ttxIt seems to be rather simple and not very risky09:39
ttxand very visible bug09:39
ttxso I'd support including it09:40
elodillesttx: hmmm, OK, let's review the patch and release it09:41
ttxI'll look up the patches09:41
ttx(in horizon)09:41
elodillesttx: one more thing: this would need a RFE as well since horizon is in the upper-constraints.txt09:43
ttxah hm09:43
elodillesmaybe better to hold this release :/09:43
elodilles(this is the bug btw: https://bugs.launchpad.net/horizon/+bug/1950659 )09:44
ttxThe trick is... I'm not sure that's a valid stable update. It does change behavior very significantly09:44
ttxso if we don;t release it pre-release, we create another issue09:44
*** marios is now known as marios|ruck09:45
ttxsee https://review.opendev.org/c/openstack/horizon/+/835143/1/releasenotes/notes/keystone-admin-endpoint-no-longer-required-06a8d29dfdb3b1cd.yaml09:45
elodillesthat's true :/09:46
ttxWhat happens if we grant that RFE?09:48
ttxanything else needs to be re-released?09:48
ttxI feel like those days teh impact is limited09:49
elodilleswell, we need upper-constraints.txt to be updated as soon as possible09:50
elodillesin that case09:51
elodillesand an ACK from prometheanfire 09:51
ttxFTR that would be a +1 on https://review.opendev.org/c/openstack/releases/+/83546109:52
ttxI find it normal to release horizon late -- after all it needs to digest changes in everything upstream of it09:53
ttxbut yeah that's about as late as it can get09:53
elodillesyepp, that is the generated patch, we need09:54
zigoFYI, trove-dashboard needs another release, including this patch: https://review.opendev.org/c/openstack/trove-dashboard/+/83349710:29
zigoIt's currently broken with latest Horizon release, since https://review.opendev.org/c/openstack/horizon/+/81630310:29
zigoThough the package seems still broken ...10:30
ttxthat one could be a stable release10:42
ttxonce it's unbroken10:42
elodillesi agree with ttx 10:47
fungiokay, i'm awake and catching up. sounds like murano is a long shot, the second trove fix bounced out of the gate but has been rechecked, anything else to do related to the oslo.context tenant removal?11:22
ttxI think we are good. Murano is a pretty long shot at this point, we are holding on Horizon RFE before regenerating the final release patch11:24
elodillesfungi: trove is released and sahara as well11:24
ttxeverything else is A-OK11:24
fungioh trove got through on the second try? excellent!11:25
elodillesyepp11:26
elodillesso yes, now horizon (22.1.0) is the question -- https://review.opendev.org/c/openstack/releases/+/83546111:26
elodillesthough it's not a cycle-with-rc so we won't need to update anything in yoga-final patch11:28
*** dviroel|out is now known as dviroel11:28
elodillesso far with ttx we discussed this and found that it would be better to release it - despite how late we are with it ( https://meetings.opendev.org/irclogs/%23openstack-release/%23openstack-release.2022-03-29.log.html#t2022-03-29T09:33:16 )11:30
elodillesprometheanfire: ^^^ if you are around and could share your opinion that would be awesome11:30
fungielodilles: you may be right about needing to bump oslo.context in murano's lower-constraints job. hberaud: do you happen to know off the top of your head what the earliest version of oslo.context is which should work with project_id instead of tenant?12:24
elodillesfungi: hberaud is on PTO today and tomorrow12:36
fungii'll try to work it out fro the release notes12:36
fungithanks!12:36
elodillesunfortunately releasenotes are mostly empty :(12:39
elodilleshttps://docs.openstack.org/releasenotes/oslo.context/12:39
fungiyeah. i grepped for tenant and only found the most recent one where it's removed12:41
fungisame for project id12:41
fungiso i guess i'll have to resort to git log12:41
fungiproject_id seems to be in use as far back as when the files got reshuffled in 201412:43
fungii'm trying to find where tenant was actually logged as a deprecation or documented12:45
fungino luck. seems like it's just been assumed to be deprecated but never really called out anywhere as such that i can find12:49
fungithe commit message for the removal simply asserts "The tenant argument of RequestContext have been deprecated long time ago" but doesn't seem to remove any actual deprecation warnings or logging or documentation notices12:51
elodillesweird12:52
fungii'll try raising the lower constraint to 3.0.0 and see what happens12:52
elodillesi'm now testing locally the l-c job12:52
elodillesOK, 4.0.0 works, 3.4.0 does not :'(12:53
fungihuh, maybe the l-c bump in that trove change was needed after all, then12:53
elodillesso something is fishy here12:54
elodillesyes, it seems that is needed12:54
fungithere were only two non-merge commits between 3.4.0 and 4.0.012:54
fungione is a testing change12:55
fungithe other is the tenant removal change12:55
fungianyway, i don't put much stock in the lower-constraints jobs, if we can work around this by upping the minimum for oslo.context to 4.0.0 there, it's fine by me13:00
elodillesit's OK to me too, as the l-c test clearly shows 4.0.0 is needed13:01
elodilleszhurong decided to drop the l-c job instead13:02
fungioh, that's fine by me then13:05
fungiit would be irrelevant in stable/yoga anyway (or should be) since we previously got consensus that lower-constraints jobs aren't really maintainable on stable branches13:06
elodillesyepp, there are projects who dropped lc completely from their stable branches13:08
elodillesnow this is master branch though :S13:09
fungiit's also a project with very nominal levels of maintenance, so i completely understand if its maintainer(s) think also supporting the lower-constraints job in master is too much for them13:10
*** amoralej|off is now known as amoralej13:23
prometheanfireelodilles:  I'm out today for house move13:56
opendevreviewElod Illes proposed openstack/releases master: Add release note link for heat  https://review.opendev.org/c/openstack/releases/+/83563214:00
elodillesprometheanfire: oh, ack, that sounds fun :] ... or it depends. o:) enjoy the new place! :)14:03
elodillesttx: then the decision about horizon is on us :)14:04
prometheanfireya, it's a good thing14:07
elodillesthat's cool then :]14:14
elodillesttx: I've +2'd the horizon release patch then - https://review.opendev.org/c/openstack/releases/+/83546114:14
ttxhorizon is a go!14:40
ttxelodilles: I propose we regenerate the release patch once that lands14:41
ttxMurano will not be perfect but I can live with that14:42
ttx(I did push for its retirement last August, i stand by that)14:43
ttxor did I miss a Murano fix?14:43
ttxoh, I did14:44
fungibackport is in the gate now14:44
elodillesttx: the murano fix is in the gate queue14:44
fungiother contributors fixed up my patch to be much less terrible ;)14:44
elodillesi'll prepare an rc2 patch for murano, i'll need review on that, and then i'll update the yoga-final patch14:45
elodillesttx: one minor patch to land: https://review.opendev.org/c/openstack/releases/+/83563214:46
elodilles(reno link for heat)14:46
opendevreviewMerged openstack/releases master: Release horizon 22.1.0(Yoga)  https://review.opendev.org/c/openstack/releases/+/83546114:56
ttx+214:57
elodilles+2+W14:57
elodilles~o~14:58
fungiw00t15:06
opendevreviewMerged openstack/releases master: Add release note link for heat  https://review.opendev.org/c/openstack/releases/+/83563215:16
*** dviroel is now known as dviroel|lunch15:17
clarkbfungi: catching up on release status seems we are good with oslo.context but what about the setuptools issue with multiple packages and needing a virtualenv/tox update?15:30
fungiclarkb: if tox gets updated then all will be fine, otherwise we need to put a workaround in place15:31
fungi(i think a workaround was merged to the script already?)15:31
clarkbfungi: release script you mean?15:37
fungiyes15:37
fungier, not tox, virtualenv release i mean15:37
clarkbok definitely ping if I can help debug further or review changes or figure out a workaround15:38
fungiclarkb: https://review.opendev.org/835423 Pin setuptools for Yoga release15:39
clarkbhttps://github.com/pypa/virtualenv/pull/2324 is still open with no reviews15:39
fungilooks like it worked there because that doesn't rely on constraints15:39
fungiand yeah, https://github.com/pypa/virtualenv/pull/2324 doesn't seem to be getting fast-tracked so i wouldn't bank on it merging today15:39
clarkboh intereting if you put virtualenv in requires for tox it figure out how to use tha virtualenv for the test envs?15:40
fungiyeah15:41
elodillessorry, yes, the setuptools WA has been merged today morning. it solves our problem15:42
elodillesi've tested it in some early patch set ( https://review.opendev.org/c/openstack/releases/+/835423/2/ )15:43
elodillesafter the release we can revert this actually, and the packages can fix the 'multiple top level' issue by themselves15:44
elodilles(which technically means they need to merge these patches: https://review.opendev.org/q/topic:setuptools-issue-3197 )15:45
clarkbcool thanks for confirming that was addressed15:46
*** amoralej is now known as amoralej|off15:47
fungiin theory, pbr could grow an option to make that unnecessary too16:07
fungihaving pbr build a manifest from everything tracked by git makes the autodiscovery of python files in setuptools quite redundant, so would probably just disable it and maybe throw in a toggle to allow it to still run in case it's needed for some unusual circumstances16:09
ttxStill in gate16:10
ttxelodilles: will you be around to push the final release patch before EOD?16:11
elodillesttx: sure, i'll be here16:14
elodillesi'm waiting for the murano patch to merge, then propose the release patch16:15
elodillesit would be good if you could review that16:15
elodillesand after that i'll update the yoga-final patch16:15
opendevreviewElod Illes proposed openstack/releases master: [murano] Release RC2 for Yoga  https://review.opendev.org/c/openstack/releases/+/83566316:19
elodillesit has merged, and here is the release patch ^^^16:19
elodillesplease review it16:19
elodillesfungi: hmmm, that pbr feature sounds fancy!16:20
elodillesi mean, nice :)16:21
*** dviroel|lunch is now known as dviroel16:21
fungielodilles: the fancy part is what pbr already does (generates the MANIFEST.in file based on git ls-files, more or less), it's just that setuptools also separately tries to find your python files and that's what's started breaking with the ultiple top-level packages error16:26
elodillesi see :-o16:28
*** sfinucan is now known as stephenfin16:31
elodillesttx: btw, this horizon constraints lift patch needs review too (when the gate is ready with the testing): https://review.opendev.org/c/openstack/requirements/+/83564716:31
ttxelodilles: I posted my +2 ahead of the check result16:31
fungielodilles: the murano stable/yoga change just merged16:32
elodillesttx: thanks \o/16:32
fungiahh, i see i'm late to the party, you already said that16:32
ttxFeel free to +2a later. Also you should rebase https://review.opendev.org/c/openstack/releases/+/835322 on top of the murano patch16:33
elodillesfungi: :) we just need the gate to finish with the testing there (with the release patch)16:33
elodillesttx: ack, will do that!16:33
*** marios|ruck is now known as marios|out16:34
ttxelodilles: murano release ready for +2a16:59
ttxame for requirements patch17:00
ttxupdated https://etherpad.opendev.org/p/tenant-projectid-last-minute-fixes17:01
ttxI did +2s murano rc217:02
ttx+2a17:02
elodillesthanks (i was slow, but did the same :D)17:06
elodillesby requirements patch you mean the horizon one? or is there something else? :-o17:07
opendevreviewMerged openstack/releases master: [murano] Release RC2 for Yoga  https://review.opendev.org/c/openstack/releases/+/83566317:20
opendevreviewElod Illes proposed openstack/releases master: Yoga final releases for cycle-with-rc projects  https://review.opendev.org/c/openstack/releases/+/83532217:28
opendevreviewElod Illes proposed openstack/releases master: Mark Yoga as released  https://review.opendev.org/c/openstack/releases/+/83532317:28
elodillesthere we go ^^^17:28
fungithe anticipation is unbearable ;)17:42
elodillesand the suspense is terrible :S :)17:46
elodillesand it will last for a while as yoga-final patch is quite heavy :-o17:49
fungiyup17:50
*** dviroel is now known as dviroel|out20:50

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!