Tuesday, 2023-05-30

fricklerwhat has happened to the action item of fixing zuul config errors? is there someone still working on this? if not, I'd like to propose a new approach to finally get rid of at least most of them07:24
fricklero.k., I had to go back to last year's channel logs to rediscover https://etherpad.opendev.org/p/zuul-config-error-openstack08:13
fricklerI updated the current status there, it comes likely as no surprise that the largest offenders there are the same projects that lack in CI maintainance also in other metrics, see e.g. https://review.opendev.org/q/topic:reno-2023.1+status:open08:17
fricklertc-member: so my proposal is: someone (hopefully another volunteer, but possibly me) generates patches to fix these errors (first gate: queue: issue, then project renames), then I volunteer to force-merge them08:18
fricklerthis would work around both issues: lack of PTL activity and lack of CI health08:19
fricklerplease let me know what you think about this and whether a formal resolution would be needed08:20
fricklerit seems elod has already started some patches, the open ones here would be the first one to be handled this way https://review.opendev.org/q/topic:fix-zuul-queue-syntax08:23
frickler(didn't check in detail, but it looks like those are the cases where even the master branch is still broken)08:24
fricklerI guess the highlight should have been tc-members? just making sure to ring your alarm bells properly :D08:25
fricklerfor context, the reason I'm pushing this again now is that I've recently found some other config errors that were almost impossible to find in this huge pile of old crap, e.g. https://review.opendev.org/c/openstack/project-config/+/884563 and https://review.opendev.org/c/openstack/devstack/+/88456208:27
fricklerwhile we're at it, I'd like to also alert you of another issue: the backlog in updating upper-constraints08:40
fricklerif you look at e.g. https://review.opendev.org/c/openstack/requirements/+/881897 you'll see a growing number of libs that cannot be updated because some projects didn't adapt to them yet08:41
fricklerwhich of course is a self-enforcing feedback loop, since those libs are capped by u-c, projects most often don't even see the need to adapt, because all their u-c capped tests keep on working just fine08:42
fricklerin addition, there are some updates that even aren't appearing at all, because projects have capped things, thus blocking even proposed u-c updates08:43
fricklermost prominent example is sphinx, which we still cap at 4.5.0, most recent release is 7.0.108:43
knikollao/13:20
knikollafrickler: related to zuul config errors, i was the one who volunteered to work on that but I deprioritized the work and haven't been able to spare the time yet. 13:21
jrosserfrickler: it seems quite hard to stumble across the list of config errors if you don't know where to look13:47
TheJuliafrickler: I long for a return to the bot proposing new versions, personally :)13:51
fricklerTheJulia: that's happening again since last cycle, see the patch I cited, but that's not enough if nobody follows up on those patches13:53
fricklerjrosser: not sure what you mean to say by that13:54
jrosserwell, i can't obviously navigate to it from zuul.opendev.org13:54
jrosseri tried the other day and failed, and just now used the url from your etherpad13:55
TheJuliafrickler: I was thinking project side, not just g-r13:55
jrossernot saying it can't be done, but there might be some UI improvement to make it easier to find13:55
fricklerTheJulia: ah, but projects don't have local pinned constraints any more. so the best a bot could do would maybe propose dummy patches for each affected project with a depends-on on the u-c bump13:57
fricklermaybe a script that does that for failing tox-cross jobs could be made13:58
TheJulia... which could serve as a launching pad to fix issues. You don't know what to fix if you can't see the failure13:58
TheJuliaAnd if the delta in versions is invisible... well back to that self fulfilling loop.13:59
fricklerwell, people could follow/check the reqs updates. like have that be part of their weekly PTL routine13:59
TheJuliaThat requires a context switch14:00
TheJulia(to the bigger picture)14:00
fricklerthat's a good thing to do in general, not only look at your single own project14:01
TheJuliaBut if if it all people have time to focus on...14:01
fricklerknikolla: so what was your plan for proceeding? continue trying to work with PTLs? https://review.opendev.org/c/openstack/zaqar/+/861188 seems to have worked out after some time, that would need stable backports as follow-up14:02
fungijrosser: you mean the non-popup version of it? there is a popup you can show by clicking the bell icon along the top of the dashboard14:09
fungiand clicking on any entry in the popup takes you to https://zuul.opendev.org/t/openstack/config-errors14:10
fungimaybe that's not as intuitive as it could be14:10
fricklerand for completeness https://zuul.opendev.org/api/tenant/openstack/config-errors is the json version of it that is much easier to parse, like I showed for the current state in the etherpad14:12
fungiah, yes, i suppose that might eventually be findable through https://zuul.opendev.org/openapi but is currently incomplete14:14
* frickler instead just looked at what the UI is fetching14:15
fricklerbecause doing curl on the UI page isn't really helpful and I got tired of doing copy/paste from the browser14:16
fungimakes sense14:18
jrosserfungi: ahha there it is! i was looking in here https://zuul.opendev.org/tenants, and my expectation was to find a config errors count and click it14:48
knikollafrickler: yes, initially I was planning on working with PTLs. But I'm open to your approach in the cases where that doesn't work out. 14:51
fricklerjrosser: adding a config error column in that table looks like a good idea to me. maybe you can propose it in the zuul matrix channel yourself?15:07
fricklerknikolla: o.k., so let's get this process started again, I can do backports for zaqar and then check the other projects one by one15:08
gmannfrickler: I am ok to force merge those, but if we can get zuul throwing error while those renaming happening can be great but do not know if that is possible. 15:19
gmannand ++ on restarting the work, knikolla I am not sure how it got remove from agenda (I think I moved it to tracker during last cycle ) but I feel we should have it in tc meeting agenda or 2023.2 tacker if you have preapred that15:20
knikollagmann: i have blocked 2 hours on my calendar Thursday to push through finishing the tracker. 15:23
knikollaapologies for the delay. 15:23
fricklerknikolla: oh yeah, this is going to be fun, clean one config error, get the next one for free ... :-( https://review.opendev.org/c/openstack/zaqar/+/88454715:37
knikollatc-member: reminder, meeting in ~1hour. 17:01
knikollafrickler: i'll take a look later today, thanks for proposing the backport. 17:02
knikollatc-members: reminder, meeting in ~1hour.17:02
knikollaI had a missing "s" in the previous ping :)17:02
opendevreviewVladimir Kozhukalov proposed openstack/governance master: Add openstack-helm PTLs missing irc nick name  https://review.opendev.org/c/openstack/governance/+/88475717:34
knikolla#startmeeting tc17:59
opendevmeetMeeting started Tue May 30 17:59:44 2023 UTC and is due to finish in 60 minutes.  The chair is knikolla. Information about MeetBot at http://wiki.debian.org/MeetBot.17:59
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.17:59
opendevmeetThe meeting name has been set to 'tc'17:59
knikolla#topic Roll Call17:59
dansmitho/17:59
noonedeadpunko/17:59
slaweqo/17:59
knikollaHi all, welcome to the weekly meeting of the OpenStack Technical Committee18:00
knikollaA reminder that this meeting is held under the OpenInfra Code of Conduct available at https://openinfra.dev/legal/code-of-conduct 18:00
knikollaToday's meeting agenda can be found at https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee18:00
knikollao/18:00
gmanno/18:00
noonedeadpunkI'd say it's missing important point which is PTG timeslots for Vancouver?18:00
knikollaBefore we quickly move on to the agenda item's I'm going to capture some todo items under my name, to keep me accountable18:00
knikollanoonedeadpunk: yep, that being one of those action items. 18:01
knikolla#action knikolla Book timeslots for PTG in Vancouver18:01
gmannand also what we are going to discuss, some pre captured topics18:01
knikolla#action knikolla Prepare PTG agenda for Vancouver18:01
knikolla#action Finish TC Tracker 2023.218:01
knikollaI have blocked off time in my calendar for all of those topics this week18:02
* rosmaita sneaks in18:03
knikollaSorry for the delay18:03
knikollaWe can circle back at the end of the meeting if there's further to discuss about that18:03
knikolla#topic Follow up on past action items18:03
knikollaThere is 1 action item from the previous meeting18:03
knikolladansmith: backport memory reduction patch for devstack to yoga and look useful doing it18:04
dansmithfirst part is done, haven't figured out how to do the second part yet18:04
knikollaDepends on the audience, haha. You look pretty useful to us. I don't know if that applies to your supervisor too. 18:04
knikolla#topic Broken docs due to inconsistent release naming18:06
knikollaThere are no updates on this and I failed to capture the to do item that arose from the previous discussion18:06
knikolla#action knikolla To fix link redirect to release from docs18:06
knikolla#topic Gate health check18:07
knikollaAny updates on the state of the gate?18:07
JayFo/18:07
dansmithstill having guest kernel crashes on cirros 0.6.1 but they look different18:08
dansmithstill waiting for enough data to say if it's better or not18:08
dansmithotherwise I don't know of any major gate issues myself, other than the combined effect of the usual bunch of small issues18:08
knikollaThanks dansmith, anything else on this or gate status in general? 18:09
gmannyeah, I have also not seen very visible difference with new cirros18:09
gmannbut overall there are less timeout because of multiple fixes happened in this cycle 18:09
noonedeadpunkYeah, I would say that gates are really goodlately18:09
dansmithmaybe worth mentioning that I did backport the flag, but it has not yet merged: https://review.opendev.org/c/openstack/devstack/+/88397018:09
dansmithbut I just pinged some people to remind them18:10
noonedeadpunkexcept nasty CentOS bumping libvirt version to 9.3...18:10
noonedeadpunkBut yeah, nothing from our side at least18:10
noonedeadpunkhaven't seen any timeouts last couple of weeks18:11
knikolladansmith: I +1ed that to look useful too18:12
dansmithis that all it takes?18:12
fungialso a reminder that opendev is working on dropping package mirrors for fedora (and possibly fedora images after that)18:12
dansmithanyway, nothing else from me18:12
spotzo/18:12
knikollafungi: are there actions to do on our part or things that we need to monitor?18:13
fungiother than speak up if you're heavily reliant on fedora testing, no18:13
fungiso far the arguments for keeping fedora around have been fairly minor18:13
fungiand what we have in opendev is a couple releases behind and ~nobody noticed/cared18:14
noonedeadpunkI think we're missing rocky mirrors as of today?18:14
* noonedeadpunk thinks of utilizing freed space :D18:14
knikollaroger18:15
fungiyes, we've been discussing what would be needed to add rocky linux mirroring18:15
fungibut also we're trying to free space for the upcoming debian release18:15
fungianyway, join the discussion in opendev if you're interested in those topics18:15
knikollaanything else on gate?18:16
noonedeadpunkWe've switched most of testing from CentOS to Rocky nowadays, so we'd use these mirrors in OSA at very least...18:16
knikollathat seems useful and as fungi mentioned, worth discussing further in the opendev channel18:17
noonedeadpunk++18:17
knikolla#topic Keeping Python 3.8 support and clarifying Python version support18:17
knikolla#link https://review.opendev.org/c/openstack/governance/+/882165?usp=search18:18
knikolla#link https://review.opendev.org/c/openstack/governance/+/882154?usp=search18:18
gmannthanks noonedeadpunk for updating those18:18
knikollaThanks noonedeadpunk for updating the proposals. 18:18
dansmiththanks noonedeadpunk for updating those18:18
* dansmith jumps on the bandwagon18:19
noonedeadpunkI've jsut noticed that commit msg for 882165 is wrong....18:19
knikollalol18:19
gmannah18:19
gmannit still says library 18:19
noonedeadpunkI wonder if I should update it....18:19
noonedeadpunklikely good to do...18:19
gmannyeah, I think better to update to avoid confusion 18:19
knikollaI think it should be updated, yes. 18:19
slaweq++18:20
dansmithdo it now and we can all re-vote real quick18:20
gmann+118:20
slaweqI was just reading it now and confused by that commit message :)18:20
opendevreviewDmitriy Rabotyagov proposed openstack/governance master: Add py38 as a PTI requirement for Python projects  https://review.opendev.org/c/openstack/governance/+/88216518:20
knikolla++18:20
noonedeadpunkoh, wow, roll-call vote is not reseted on commit msg update o_O18:21
dansmithwell, good this time :)18:21
knikollaYay. 18:21
gmannhumm but why18:21
gmannthis is not good right18:22
noonedeadpunkyeah, but I wonder if we should have `changekind:NO_CODE_CHANGE`18:22
noonedeadpunkfor copying votes18:22
noonedeadpunkit's not reflecting content.... but commit message could be completely changed and nobody will notice that18:22
knikollaI think NO_CODE_CHANGE will still have this same behavior18:23
fungiwill need to take a closer look at the current acl18:23
knikollawe might want NO_CHANGE18:23
gmannand depends-on can be removed which is actually kind of change 18:23
noonedeadpunkyeah18:23
knikolla#link https://gerrit-review.googlesource.com/Documentation/config-labels.html#label_copyCondition18:23
gmannnot sure it keep Code-Reviewor its just RC voting only18:24
noonedeadpunkit's about RC only18:24
gmannk18:24
* noonedeadpunk trying to find that in project-config now18:24
fungi#link https://opendev.org/openstack/project-config/src/branch/master/gerrit/acls/openstack/governance.config18:24
fungii was already looking at it18:24
knikollaThanks fungi18:24
noonedeadpunk#link https://opendev.org/openstack/project-config/src/branch/master/gerrit/acls/openstack/governance.config#L2918:25
noonedeadpunkyeah18:25
gmannit is same for Code-Review18:25
noonedeadpunkhuh, yes, we override default behaviour of CR18:25
fungiclarkb: ^ do you recall seeing this happen with other less-common review labels?18:25
clarkbfungi: no, but that comes from gerrit's migration default iirc. TRIVIAL_REBASE is a superset of NO_CODE_CHANGE I want to say18:26
clarkbor at least is documented to be, but maybe they are wrong18:26
clarkbits all configurable though so just set it to what you want18:26
fungifor the openstack/releases PTL-Approved label we used "changekind:NO_CODE_CHANGE OR changekind:TRIVIAL_REBASE"18:27
gmannso it is changed in all repo config files ?18:27
clarkbgmann: no18:27
knikollaCommit messages are important for governance patches, so I think we should change the current behavior for Roll-Call Votes for governance.18:27
fungimost projects inherit the default Code-Review label definition and don't override it18:27
gmannhttps://opendev.org/openstack/project-config/src/branch/master/gerrit/acls/openstack/tempest.config#L1218:28
fungigit history for that file will probably provide some rationale18:28
clarkbfungi: it was the conversion from the old function which was deprecated/removed18:28
clarkband it is based on what gerrit itself converts it as iirc18:28
clarkbbasically its the 1:1 per gerrit to what you had with afunction18:28
fungii mean reason for why it's got an override for code-review18:28
clarkbits beacuse the have custom labels18:29
clarkbI don't know why the governance repo needs custom labels though18:29
fungilooks like https://review.openstack.org/187326 from 201518:29
JayFIs this something we need to fix live in the meeting? Or can we action item a fix to check up on next week18:30
fungier, no, that mainly just renamed the vote, nevermind18:30
fungiJayF: it is not, no18:30
fungiwe can pick this up in #openstack-infra18:30
knikollaalright, anything else to discuss on Python 3.8? 18:30
knikollaboth patches got a sufficient number of votes. thank you all18:31
knikolla#topic Open Discussion and Reviews18:31
knikollaFloor is open for any other matters to discuss18:31
fungioh, well i continued researching and posted in #openstack-infra, but short answer is that the reason code-review is overridden stems from a change in 2015 to intentionally make it not required for merging18:32
fungi#link https://review.openstack.org/185785 "Alter governance repo voting rules"18:33
clarkbI think the in person board meeting attendance for yvr needs to be requested real soon now18:33
fungihopefully that at least provides some background on wht it's different from other acls18:33
gmannbut copyAllScoresIfNoChange  was not there 18:33
clarkbfor anyone intending to go that hasn't rsvpd18:33
gmannit says it was default true for everyone but we did not see this behavior earlier https://review.opendev.org/c/openstack/project-config/+/88011518:34
gmanneither old flag was not working or new one added it18:34
noonedeadpunkclarkb: I've asked Alison to count me in, but I'm not aware of details (in terms of where/when exactly)18:35
rosmaitawhich day is the board meeting?18:35
noonedeadpunkMonday18:35
clarkbnoonedeadpunk: http://board.openinfra.dev/meetings/2023-06-1218:35
gmannother repo where new flag NO_CODE_CHANGE  is added had copyAllScoresIfNoCodeChange  to true explicitly 18:35
spotzBoard meeting is Monday at Microsoft18:35
clarkbthat has all the details18:35
knikollaThanks clarkb, spotz18:36
slaweqspotz (@_oftc_spotz:matrix.org) in Vancouver, right?18:36
knikollaslaweq: yes18:36
spotzslaweq: Yeah18:36
gmannyes18:36
spotzIt's in Gastown section so not far18:37
slaweqok, I will be there :)18:37
spotzBut as mentioned you need to let Allison know so you can be on the list for entrance18:37
slaweqshould I get any email with details or something like that? I already asked aprice (@_oftc_aprice:matrix.org) that I would like to be there18:37
spotzI would assume you'd get added to the invite but not sure18:37
gmannthere is space constraints I think good to get confirmation from aprice 18:38
slaweqok, I will confirm with aprice (@_oftc_aprice:matrix.org) to be sure that I'm on the list18:38
slaweqthx18:38
* noonedeadpunk needs to do the same18:38
knikollaIf there's nothing else, we can all get back 20 minutes18:39
noonedeadpunkbtw, talking about the event - does anybody knows if gold memebers rsvp dinner is only through org list or there's a path for tc?18:39
knikollaoh, a dinner18:40
knikollai don't know anything about it. 18:40
gmannmelwitt: it needs to be from company list i think aprice mentioned she will contact company representative to give passes 18:40
gmannbut not sure if TC can get too :)18:40
noonedeadpunkIt's on Tuesday 13/618:41
gmannit was for all members company getting some passes to participate in that18:41
* noonedeadpunk was thinking if to occupy place in company list or not :D18:41
knikollaenjoy it for the rest of us :D18:42
spotzI don't think TC is included. Board and sponsors18:42
gmannspotz: board also? I think it was member company and foundation staff ?18:42
noonedeadpunkBoard as well, yes18:42
dansmithif I'm not going to be in vancouver, can I be excused now? :)18:42
spotzThere's something on Tuesday, I wouldn't have called it a dinner though:)18:43
gmannyeah that one might be different18:43
gmannbut anyways18:43
knikolla#endmeeting 18:43
opendevmeetMeeting ended Tue May 30 18:43:18 2023 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)18:43
opendevmeetMinutes:        https://meetings.opendev.org/meetings/tc/2023/tc.2023-05-30-17.59.html18:43
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/tc/2023/tc.2023-05-30-17.59.txt18:43
opendevmeetLog:            https://meetings.opendev.org/meetings/tc/2023/tc.2023-05-30-17.59.log.html18:43
knikollaEnjoy dansmith 18:43
slaweqo/18:43
dansmiththanks18:43
knikollaWe still have the room, no need to stop the discussion about events in Vancouver18:44
noonedeadpunkthanks and sorry, I was sponsor of offtopic discussions today...18:45
noonedeadpunkcan not stop myself18:45
spotzheheh18:45
knikollaNo that's great, haha18:45
noonedeadpunkI'm slightly confused though by amount of PTG teams who has signed up vs booked slots18:46
spotzWell they need to PTGbot:)18:46
fungialso the usual catch-22 of everyone waiting to see what slots everyone else schedules so they can avoid confilcts18:47
fungiconflicts18:47
noonedeadpunkas it's literally nova, designate, publiccloud-sig, security and osa as of today. 18:47
diablo_rojo_phonenoonedeadpunk: there definitely hasn't been as much activity in signing up for rooms as I thought there would be.18:47
spotzoh and if in Vancouver... https://eventyay.com/e/231a963f18:47
diablo_rojo_phoneI plan on sending a reminder out to moderators today.18:47
spotzYeah I just got a conflict for during the TC and leadership session so said I'd get over there if I could18:48
fungialso if you're working on booking ptg sessions, don't forget to bring up both the summit/forum schedule and the ptg schedule side by side when you try to work around potential conflicts18:48
spotzI grabbed the spot right after breakfast for D&I:)18:48
aprice[m]slaweq: you are on the list18:48
* noonedeadpunk only worried about TC PTG slot18:48
spotzWell right after Speed Mentoring:)18:48
aprice[m]for the vancouver board meeting 18:48
fungii booked the two slots prior to thursday lunch for the security sig so we have the option to move our discussions to a restaurant if we want18:49
spotznoonedeadpunk: Yeah except the D&I slot I'm just paying attention to the TC slots18:49
spotzfungi: Good plan!18:49
slaweqThx for confirmation aprice18:49
noonedeadpunkaprice[m]: can you kindly check if I am as well ? :)18:50
knikollaHow many slots do we think we need for the TC PTG?18:50
aprice[m]noonedeadpunk: yes, you are as well 18:50
knikollaI was thinking of them being more office-hours sorts of events considering not the entire TC will be there18:50
spotzThere's also the 2 remote capable rooms if we need as well but we'd need to book a slot here and there vs claim one18:51
gmannknikolla: depends on topics we want to cover there18:51
aprice[m]noonedeadpunk: for the member event, i can speak to that as well. that is for silver, gold and platinum members of the foundation as well as the board and staff. We have limited space so couldnt do the technical leadership for each project, but you should check with your company as each company has a set number of passes. 18:52
gmannaprice[m]: so board members are included in that right ?18:53
aprice[m]gmann: yep! 18:53
fungispotz: knikolla: the remote capable rooms are, as i understand, already spoken for wednesday and half of thursday for forum sessions, but diablo_rojo_phone can probably confirm18:53
gmannaprice[m]: thanks for confirmation18:53
aprice[m]we will be sending more information on the member event and the board meeting this week 18:54
gmann++, great18:54
spotzIt's getting so close!18:54
spotzAnd come eat ice cream(or an alternative for dietary restrictions) after the Marketplace Mixer, link posted above:)18:55
fungispotz: knikolla: also i believe it's not really that the rooms are "remote capable" but rather that they're dedicated rooms, so less background noise that the giant ptg room if you want to fire up a teleconference on someone's computer18:55
spotzAhh, that makes sense:)18:55
noonedeadpunkYeah, as I'd love to do so on one of timeslots18:56
knikollathanks fungi 18:56
fungibut very limited availability, and maybe just for thursday afternoon at this point? i'll defer to the organizers on that part though18:57
fungibasically it's that the forum rooms are available for select ptg sessions on request when there's not a forum session booked in them18:57
aprice[m]is this about reservable PTG space?18:57
aprice[m]sorry, hopping between a few thigns 18:57
fungiyes18:57
aprice[m]ah yes18:57
spotzyeh18:57
diablo_rojo_phoneThere will be separate rooms from the main PTG room that can be reserved.18:58
diablo_rojo_phoneThey do not have hardware set up for conferencing. We can however provide zoom rooms if necessary for use.18:58
aprice[m]yeah, there are two rooms on Thursday afternoon (maybe a little Thursday morning)18:58
diablo_rojo_phoneI will be contacting moderators today for that.18:58
fungidiablo_rojo_phone: are those the same rooms that are the forum session rooms?18:58
aprice[m]+ what diablo_rojo_phone said18:58
aprice[m]fungi: yes 18:58
fungithanks, that's what i thought i had heard18:58
diablo_rojo_phoneLol18:58
noonedeadpunkah, damn, one time I decided not to offtop in chat and missed answer to the question I've jsut emailed....19:00
diablo_rojo_phoneLol no worries. I will email everyone later today. Just out trying to finish a run right now. I've got really excellent timing for scheduling these when there's a flood of questions on IRC lol. 19:04

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