Monday, 2020-10-19

*** tetsuro has joined #openstack-release00:00
*** brinzhang has joined #openstack-release00:16
*** e0ne has joined #openstack-release00:30
*** ricolin has quit IRC01:01
*** brinzhang_ has joined #openstack-release01:56
*** brinzhang has quit IRC01:59
*** tetsuro has quit IRC02:55
*** tetsuro has joined #openstack-release02:56
*** tetsuro has quit IRC03:00
*** evrardjp has quit IRC04:33
*** evrardjp has joined #openstack-release04:33
*** brinzhang_ has quit IRC05:13
*** brinzhang_ has joined #openstack-release05:13
*** brinzhang0 has joined #openstack-release06:00
*** ricolin has joined #openstack-release06:01
*** brinzhang_ has quit IRC06:04
*** sboyron has joined #openstack-release06:45
*** iurygregory has joined #openstack-release06:51
*** jtomasek has joined #openstack-release06:51
*** slaweq has joined #openstack-release06:57
*** e0ne has quit IRC07:04
*** e0ne has joined #openstack-release07:05
*** brinzhang_ has joined #openstack-release07:14
*** brinzhang0 has quit IRC07:17
*** rpittau|afk is now known as rpittau07:26
*** tosky has joined #openstack-release07:29
*** jbadiapa has joined #openstack-release08:15
openstackgerritCarlos Goncalves proposed openstack/releases master: Release octavia 6.1.0 and 5.0.3
openstackgerritMerged openstack/releases master: New major release for tripleo-upgrade.
*** brinzhang0 has joined #openstack-release09:05
*** brinzhang_ has quit IRC09:09
*** suryasingh has joined #openstack-release09:11
*** brinzhang_ has joined #openstack-release09:20
*** brinzhang_ has quit IRC09:20
*** brinzhang has joined #openstack-release09:21
*** brinzhang0 has quit IRC09:24
*** dtantsur|afk is now known as dtantsur09:55
ricolinSome weird thing happened to heat release note, appears heat victoria release note adopt all release notes from previous cycle10:37
ricolinany hints?:)10:38
*** ianychoi_ has joined #openstack-release11:23
*** ianychoi__ has quit IRC11:27
*** marios has joined #openstack-release11:30
hberaudricolin: I take a look11:34
*** ricolin has quit IRC11:43
hberaudricolin: I guess we need that
openstackgerritDaniel Bengtsson proposed openstack/releases master: New bugfix release for oslo.messaging.
openstackgerritDaniel Bengtsson proposed openstack/releases master: New bugfix release for pbr.
*** slaweq has quit IRC12:19
*** slaweq has joined #openstack-release12:21
*** slaweq has quit IRC12:25
*** ricolin has joined #openstack-release12:28
*** priteau has joined #openstack-release12:30
*** slaweq has joined #openstack-release12:31
*** slaweq has quit IRC12:54
*** slaweq has joined #openstack-release12:57
openstackgerritMerged openstack/releases master: Release os-win 5.2.0 (Victoria)
openstackgerritMerged openstack/releases master: Add examples of previous highlights
*** sboyron_ has joined #openstack-release13:23
*** sboyron has quit IRC13:26
*** sboyron has joined #openstack-release13:31
*** sboyron_ has quit IRC13:32
openstackgerritHervĂ© Beraud proposed openstack/releases master: Add Wallaby deliverable files
*** jtomasek has quit IRC13:34
-openstackstatus- NOTICE: Open Infrastructure Summit platform issues are being worked on by OSF events and webdev teams, status updates will be available in the conference "lobby" page as well as the #openinfra-summit channel on Freenode (though it is presently not logged)13:35
*** vishalmanchanda has joined #openstack-release13:47
*** diablo_rojo has joined #openstack-release13:57
*** jtomasek has joined #openstack-release14:00
*** jtomasek has quit IRC14:02
cgoncalvesrelease team: Octavia EOL'd Queens and Rocky months ago but their stable branches still exist. could you please remove them? thanks14:04
*** jtomasek has joined #openstack-release14:04
*** jtomasek has quit IRC14:05
*** jtomasek has joined #openstack-release14:06
smcginniscgoncalves: I don't believe we have access to do that. You can ask the infra folks to do that. Probably best to ask on the ML.14:07
cgoncalvessmcginnis, I asked infra first. they redirected me to here :)14:11
hberaud+1 for the ML approach to centralize the debat14:11
clarkbI don't think we need the ML14:12
*** ianychoi_ is now known as ianychoi14:12
clarkbin the past the release team has done branch cleanups. We have had to deputize a release team member to do this as a temporary gerrit admin14:12
clarkbI don't mind doing the deletions as part of the infra team. The concern is more that queens in particular has an eol tag then a bunch of commits after that14:13
clarkbthe release team needs to weigh in on how we want to untangle the EOL'd but not EOL'd branches then we can delete branches for you14:13
clarkbmy concern with bringing this up with the release team is we seem to have a gap somewhere where we can tag an eol tag but then keep committing to a branch14:13
hberaudclarkb: I'm not too familiar with the eol process I suppose teams need to agree too, as they propose to move eol on their own, isn't?14:17
jrosserhow does it relate to the dates on here?
clarkbmaybe? it all changed with the em process and I think that is where the "bug" was introduced. Previously the release team just did it for everoyne on a scheduled day14:17
*** elod has quit IRC14:17
hberaudclarkb: ack I see14:17
cgoncalveshberaud, the octavia team proposed EOL:
hberaudclarkb: then I need to rewind on the em process14:18
clarkbya it seems like we're allowing projects to EOL but then not doing any branch cleanup coordination anymore14:18
clarkbinstead of thinking of this as a specific octavia problem, we should look through which things we've eol'd but not cleaned up. Check if we need new tags. Then once audited and tagged as necessary do the branch removals14:19
*** elod has joined #openstack-release14:19
hberaudI need to take a look to the process first to get the details but maybe one approach could be something like 1. team ask for eol 2. release team validate and ask to infra to delete branches 3. infra trigger the delete14:20
hberaudmaybe we could define some related schedule in our cycle process14:20
hberaudto avoid to forget this part14:21
clarkbanother option is to not delete the branches, but that is probably a bigger discussion14:21
cgoncalvesclarkb, the ask for deleting branches is so that we can remove old zuul jobs. for example, see
hberaudclarkb: possibly some could try to push patches against these branches, isn't? I mean the delete avoid confusing14:22
clarkbyes, there are good reasons to delete the branches, but its not our only option. We could probably do a process similar to repo retirement on EOL'd branches and just commit a "this is EOL Readme" and remove everything else14:23
hberaudclarkb: ah yes I see14:23
noonedeadpunkum, I'm not really sure how deployment project should handle this indeed14:23
noonedeadpunkthen I guess we should drop branches for everything at once?14:24
noonedeadpunkand say that release is eol overall?14:24
hberaudthis is one option14:24
jrosserthis is specifically why i ask about the dates given here for EOL - are they correct and up to date?
*** armstrong has joined #openstack-release14:25
clarkbjrosser: sort of, it all chagned with the switch to extended maintenance. That is why you don't see "newer" branches as EOL14:25
hberaudjrosser: I guess yes14:25
noonedeadpunkbecause in eol ing specific project I have no idea how we should handle this tbh....14:26
noonedeadpunkas eventually specific project branch is dropped, our CI just stucks14:26
clarkbI believe the new policy shift was to give more control to projects for when they couldn't keep the branches alive anymore14:26
clarkboctavia doing eol of queens is therefore fine, we just missed the piece of cleaning up branches (and seems we missed that for puppet-* too)14:27
clarkbI think it would be good to take a moment to figure out which things have asserted they are eol, check if tags need to be added, then finish the cleanup fromthere.14:27
noonedeadpunkso this kind of means that in case one single project goes eol before release overall goes to eol - all deployment projects should go to eol as well?14:28
clarkbor the deployment should start deploying from the eol tag14:28
noonedeadpunkah, ok, so instead of branch there will be a tag?14:28
noonedeadpunksorry I missed that part I guess14:28
clarkbyes, I think the eol tagging is still happening, its the branch deletion after that that is not14:29
hberaudclarkb, noonedeadpunk: I added this point to our meeting agenda
noonedeadpunkshould be there any ML to release regarding eol of the projects?14:29
hberaudI thinkit could be worth to discuss this point during our meeting to ensure that everybody with a full history of that point are there14:30
noonedeadpunkor any other means of comunication? I mean that would be really awesome so somehow get this really important info without need to read every single ml topic14:31
noonedeadpunksure, agreed14:31
noonedeadpunkthanks hberaud14:31
hberaudnoonedeadpunk, cgoncalves, clarkb do not hesitate to add bullets that seems important for you (R-25 related meeting agenda) if you think they are missing =>
elodhi, about EM (and EOL) the original discussion covered in this resolution:
hberaudthanks for brough this topic to our attention14:36
hberaudelod: awesome thanks14:36
elodIf I understand correctly the origin of the discussion is some patch that were merged after the EOL. Is that the case?14:41
elodIn that case is that possible to make zuul vote somehow -2 in a case where there's an EOL tag already on the branch of the given project?14:42
hberaudcgoncalves: ^^^14:42
clarkbelod: yes, cgoncalves has requested that we delete a branch, but the eol tag for that branch is no longer current. I don't want to delete a stable branch in that state without coordination from the release team14:42
clarkbelod: not really no. tags and branches aren't directly mappable so that would get dangerous and likely result in false positive -2s14:43
clarkber false negatives?14:43
cgoncalveselod, not quite. we want to remove old job definitions in octavia-tempest-plugin used in queens and rocky but Zuul returns a syntax error that stable/queens is still using that job def:
elodcgoncalves: when we edited the EOL steps ( ) we discussed, that ONLY those jobs should be deleted which are defined somewhere else, not in the given project,14:45
elodbut maybe this is false,14:46
elodas the deletion does not happen instantly,14:46
elodwhich causes similar issue, like yours14:46
clarkbwell deletion hasn't happened at all I think beacuse we've not properly figured out how to make those requests in the new system14:46
clarkbI expect it could be made to be close enough to instant if we sort that out14:47
elodso maybe for a first step we should add another step between 2 and 3 to remove the zuul jobs in the repo, too14:47
elodclarkb: yes, sorry for my wording :) yes, if we keep track of everything correctly, then with the 4th step everything should be fine ( "4. After the branch is tagged with $series-eol, request the infra team to delete the branch." )14:49
elodI understand that these deletion request were not sent,14:49
elodor simply lost in the noise14:50
clarkbI noted on the etherpad that I think it would be good for the delete requests to come from the release team after the release team has confirmed ti is safe/ready to do so14:51
elodanyway, maybe a safety solution would be to remove .zuul.yaml (and similar things) from the repo, before tagging14:51
clarkbI think that is a key detail we're missing because we've gotten a few of these after the fact requests and for most of them we've had to do cleanups first14:51
elodclarkb: ++14:51
*** jtomasek has quit IRC14:51
*** jtomasek has joined #openstack-release14:52
smcginnisclarkb: The release team does not have permissions to delete branches. With the EOL patches that have come through since EM, I've usually reminded the teams requesting those that they need to follow through with someone in infra (usually it's been fungi) to have the branch deleted after we've tagged eol.14:54
fungiyeah, in the past we've authorised a release manager or stable branch manager to tag and delete eol branches in bulk14:55
clarkbsmcginnis: ya, I think the problem is the responsibility is falling on the project and they aren't doing the checks first14:55
clarkbthey come to us we say x y z needs to be done then the deletion process ends there14:55
smcginnisIt would be great if we can automate it with the tagging. That does seem a little risky though.14:55
clarkbI'm suggesting the release team should ensure everything is prepped then make the request directly14:55
fungiwhen we stopped coordinating eol, the responsibility has ended up falling on the gerrit admins to do a bunch of one-off cleanups rather than having everyone's old branches deleted in bulk14:56
clarkbonce we've got the gerrit upgrade sorted, figuring out how to delegate branch deletion more safely is likely a good thing to work upstream on14:59
clarkbin particular you currently need force push perms or project ownership to do it14:59
clarkbboth of which imply significanlty more access than removign a single branch14:59
elodsmcginnis: automation would be good, but if we consider it risky, maybe just some kind of an automated 'notification' could be sent? (but I don't know how o:))15:02
smcginnisMaybe a periodic job that checks for foo-eol tags and the presence of stable/foo branches...15:03
hberaudgood idea15:03
otherwiseguysmcginnis: just noticed that it looks like everything ran for but it still says Needs Verified Label?15:06
smcginnisotherwiseguy: It's not required. Sometime the job has issues, for whatever reason. It's really just to help the release team know when a PTL has ack'd the release, but we can also look at the list-changes job logs to see who the PTL and liaison(s) are and can manually check that they've ack'd.15:07
smcginnisOh, I see that was approved on the 15th.15:07
smcginnisI wonder if that got lost in a zuul restart or something.15:07
smcginnisAnd it's a requirements patch, not a release.15:08
smcginnisWe'll see if a recheck gets it going.15:08
otherwiseguyyeah, looks like it just kinda fell off the map in the middle of gate run.15:08
otherwiseguysmcginnis: ok, thanks!15:08
smcginnisWe'll see if this works. If not, we can dig into logs and see what's going on.15:09
openstackgerritSean McGinnis proposed openstack/releases master: init-series: Skip unrelease comment for trailing
*** slaweq has quit IRC15:55
*** slaweq has joined #openstack-release15:59
*** marios has quit IRC16:01
*** prometheanfire has quit IRC16:05
*** tosky has quit IRC16:12
*** prometheanfire has joined #openstack-release16:14
openstackgerritMerged openstack/releases master: New bugfix release for pbr.
openstackgerritMerged openstack/releases master: Release octavia 6.1.0 and 5.0.3
openstackgerritMerged openstack/releases master: New bugfix release for oslo.messaging.
openstackgerritMerged openstack/releases master: Adding a new tools to search topic on the ML
openstackgerritMerged openstack/releases master: Allow to consume projects without the openstack prefix
openstackgerritMerged openstack/releases master: reminder lighten up for QA releases
openstackgerritMerged openstack/releases master: R-3 Tagging cwi branchless deliverables (tempest-plugin)
*** rpittau is now known as rpittau|afk16:43
*** e0ne has quit IRC17:54
*** dtantsur is now known as dtantsur|afk18:03
*** sboyron has quit IRC18:12
*** priteau has quit IRC18:22
*** e0ne has joined #openstack-release18:24
*** e0ne has quit IRC18:35
*** e0ne has joined #openstack-release18:39
*** tosky has joined #openstack-release19:35
*** suryasingh has quit IRC19:50
*** otherwiseguy has quit IRC20:07
*** weshay has quit IRC20:07
*** slaweq has quit IRC20:12
*** slaweq has joined #openstack-release20:14
openstackgerritJose Luis Franco proposed openstack/releases master: Update release and branch to Victoria for tripleo-upgrade.
*** ianw has quit IRC21:13
*** ianw has joined #openstack-release21:15
*** e0ne has quit IRC21:16
openstackgerritMerged openstack/releases master: sahara: tag ocata as EOL
openstackgerritMerged openstack/releases master: sahara: tag pike as EOL
*** slaweq has quit IRC21:35
*** armstrong has quit IRC21:35
*** slaweq has joined #openstack-release21:35
*** slaweq has quit IRC21:41
*** vishalmanchanda has quit IRC21:55
*** otherwiseguy has joined #openstack-release22:05
*** tosky has quit IRC23:16

Generated by 2.17.2 by Marius Gedminas - find it at!