16:00:08 #startmeeting tc 16:00:08 Meeting started Wed Feb 15 16:00:08 2023 UTC and is due to finish in 60 minutes. The chair is gmann. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:08 The meeting name has been set to 'tc' 16:00:11 #topic Roll call 16:00:12 o/ 16:00:13 o/ 16:00:14 o/ 16:00:22 o/ 16:00:39 o/ 16:00:44 today Absence: 16:00:52 knikolla[m] have conflict with another meeting 16:00:54 JayF out 2/15 (covid) 16:01:15 o/ 16:01:17 JayF: sad to hear that, please take rest and get well soon 16:01:38 ++ for JayF, -- for covid 16:01:50 #link https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee#Next_Meeting 16:01:53 today agenda ^^ 16:01:53 * bauzas sits in the back 16:02:41 * gmann bauzas you are very welcome to come front :) 16:02:50 let's start 16:02:53 #topic Follow up on past action items 16:03:07 noonedeadpunk to add PyPi access policy in governance documentation 16:03:44 yeah, I didn't have progress on that :( 16:03:50 noonedeadpunk: any update on this? or you want to discuss it in the next topic 16:03:58 k, let's continue this 16:04:03 I do have it in my backlog though 16:04:09 #action noonedeadpunk to add PyPi access policy in governance documentation 16:04:14 noonedeadpunk: thanks 16:04:18 so let's leave it for the next week 16:04:36 sure 16:04:41 gmann to prepare etherpad for grenade skip upgrade job data and send email asking required projects to add job 16:05:05 ah, I missed it. no progress on this. will continue this item 16:05:11 #action gmann to prepare etherpad for grenade skip upgrade job data and send email asking required projects to add job 16:05:35 #topic Gate health check 16:05:48 and this is reason I am stuck on my many other tasks :) 16:06:06 yeah, so, 16:06:07 dansmith: please go ahead to summarize the gate sitatuion now 16:06:10 o/ 16:06:23 things are slightly better now that we merged a couple of things, but it's still pretty rough IMHO 16:06:29 gmann: did the timeout thing merge? 16:06:54 definitely rough IMHO 16:06:54 I've got a patch up that as of this morning looks like it may be dropping mysql memory usage by a considerable amount, but I need to check the impact on performance to see if it's worth it 16:07:04 but that may be an option to help with the OOMs 16:07:10 but there are still a lot of other failures 16:07:18 tons of volume-related failures specifically, 16:07:30 and then lots of other small cuts which together make the situation pretty bad 16:07:54 I fixed the glance locations occasional-failure, and then fixed that fix 16:08:03 ok 16:08:09 we also had some problems with dhcp clients 16:08:12 gmann has a patch up to increase the timeout, which we were going to do temporarily, but ... 16:08:16 (at least in the nova CI) 16:08:27 bauzas: yep, plenty of unable to ssh to test instances for sure 16:08:30 timeout series is merged and I abandon the temporary increasing the timeout https://review.opendev.org/q/topic:bug%252F2004780 16:08:56 oh, so no timeout increase because of the test splitting you mean? 16:09:12 dansmith: let's observe the timeout failure now and if we still get then I will retsore that https://review.opendev.org/c/openstack/tempest/+/873472 16:09:16 yeah 16:09:17 ack 16:09:34 I didn't see any timeouts this morning in the few patches I looked at, so perhaps that helped 16:09:57 k, let's continue monitor those 16:10:04 yeah 16:10:14 but we really need more help working on these things I think 16:10:21 especially the volume ones 16:10:28 maybe after FF people will be able to jump on them more 16:10:37 yeah 16:10:48 Well, I'm not sure if it's timeout that helped or load reduced on providers, as we instead of timeouts now get some jobs just 5mins before timeout 16:11:10 And waaaay less timeouts for the past week overall 16:11:25 ack 16:12:00 gate is very unstable overall during this release especially in last month 16:12:07 or since tox4 time 16:12:08 indeed 16:12:19 pretty much all year so far I think 16:12:26 yeah 16:12:26 first time I'm creating an etherpad to track the most CI failures 16:12:39 humm. but +1 on tracking those 16:12:49 anything else on gate things? 16:13:22 #topic TC 2023.1 tracker status checks 16:13:26 #link https://etherpad.opendev.org/p/tc-2023.1-tracker 16:13:45 please go ahead if anyone has any updates on the items assigned the to them 16:14:29 one is from slaweq on guidelines of using release version/project version in doc/releasenotes 16:14:38 and patch is up for review #linkhttps://review.opendev.org/c/openstack/governance/+/872769 16:14:49 please check and vote/feedback accordingly 16:15:50 rosmaita: anything you would like to update on i18? I think you had meeting? sorry i missed that 16:15:56 was it yesterday or today? 16:16:00 yesterday 16:16:03 k 16:16:36 quick update ... we figured out some basic stuff that weblate needs to set up the server 16:16:55 thx for all reviews in advance :) 16:16:57 i have a meeting with weblate tomorrow to convey that so they can start setting up the host 16:17:23 ++ 16:17:49 after its setup and openid for openinfra is working, we will move all discussion to the openstack-discuss list 16:18:14 perfect 16:18:17 and start working on recruiting an engineer to handle the weblate-to-gerrit plumbing 16:18:38 hope we get volunteer 16:19:20 yeah, the idea is to simultaneously look for a volunteer while refining the job description 16:19:36 ack 16:19:44 in case we need to ask the foundation for short term contractor support 16:19:53 that's all 16:19:53 rosmaita: feel free to propose it as upstream opportunity also if that can help 16:20:08 ok, that is a good idea 16:20:27 that will help to showcase the volunteer asking in that place too 16:20:38 or even for long term also 16:20:52 yeah, and i just remembered, i am supposed to revise the proposed survey of operators 16:21:10 to get some data but also see if they are interested in sponsoring an engineeer 16:21:20 TC survey questions or i18 SIG specific ? 16:21:57 i18n specific 16:22:00 k 16:22:18 thanks rosmaita for all your hard work on this 16:22:23 we have some on the current user survey, this was going to use a web survey to get more immediate data 16:22:35 ok 16:22:47 that is good idea 16:23:15 this is the survey (not revised yet): https://forms.zohopublic.com/rosmaitafossdev/form/OpenStacktranslationsusageandcontributionsurvey/formperma/9W0PYjoo61tAShU9B1GQj3m52K43uty-KtXDhIlOUe4 16:23:21 (nice url!) 16:23:32 :) 16:23:49 #link https://forms.zohopublic.com/rosmaitafossdev/form/OpenStacktranslationsusageandcontributionsurvey/formperma/9W0PYjoo61tAShU9B1GQj3m52K43uty-KtXDhIlOUe4 16:24:03 i'll get that updated in the next day or so 16:24:13 thanks for the link. anything else rosmaita on this? 16:24:15 cool 16:24:48 main difference is that we know for sure now about the hosted weblate, so focus will be on usage and will you contribute to the engineering "plumbing" work 16:25:13 that's all 16:25:23 sure, getting usage data is always helpful 16:25:32 k. let's move to next topic 16:25:34 #topic Deprecation process for TripleO 16:25:40 #link https://lists.openstack.org/pipermail/openstack-discuss/2023-February/032083.html 16:26:04 you might have seen the email discussion about TripleO shutting down the maintenance 16:26:18 and we discussed it in this week or last week may be 16:26:44 but with more TC members here, let's conclude the way and get agreement and then to communicate the same to TripleO team 16:27:17 should I give some background on what is the situation or everyone knows/remember? 16:27:39 I think we should do business as usual and just follow deprecation path 16:28:14 TripleO has stable/wallaby and then stable/zed. no stable/xena, stable/yoga. and team want to maintain until stable/wallby 16:28:31 I agree with noonedeadpunk - this isn't really different than deprecation of e.g. networking-midonet which we did some time ago 16:28:35 with adding some note to the readme file of stable/zed that it needs some maintenance and it's not realted in any way to coordinated release or smth 16:28:56 not sure about ^ even 16:28:59 difference is stable/zed exist and team does not want to maintain it as last supported stable branch 16:29:19 slaweq: it's different in that a later release will be deprecated before an earlier one right? 16:29:40 so marking deprecated until stable/wallaby which means it will be supported as last maintain branch but keep stable/zed also open 16:29:47 yeah, so maybe we just don't deprecate later release? 16:29:49 and master README.rst explain it 16:30:15 noonedeadpunk: yeah, so not deprecating zed and just putting a comment in the readme that it's not maintained? 16:30:23 yeah, why not deprecate master only and keep stable releases as they are now 16:30:25 it will be deprecated but without support just for anyone want to maintain it from there 16:30:28 gmann: I think master README will say that project is deprecated? 16:30:29 it's weird because if we have a CVE or something, putting that into wallaby will create a regression in zed 16:30:36 noonedeadpunk: yes 16:30:37 dansmith: yup 16:30:45 Wallaby is EM, so whatever support they will provide, it's still EM 16:31:20 I wonder how they're going to do that as some projects already EOLed W 16:31:24 anyway 16:31:45 yeah, that is ok as they need to do retirement that time anyways 16:31:49 in normal situation also 16:33:06 I think we are ok with this- Deprecate until stable/zed, maintained until stable/wallaby, and mark the project as deprecated on master. 16:33:12 dansmith: I think the main risk in case of CVE if somebody will even backport to Z, they may jsut refuse to merge 16:33:19 is TripleO following the stable policies ? I guess no by what I can read 16:33:19 gmann: I'm confused by your use of "until" 16:33:31 +1 16:33:58 bauzas: nope, they're independant. So stable/zed is basically not coordinated stable/zed I assume... 16:34:13 yeah, it's really confusing 16:34:15 dansmith: I do not think they follow stable policy, at least not in this list #link https://docs.openstack.org/project-team-guide/stable-branches.html#project-teams-which-asserted-they-follow-the-stable-branch-policy 16:34:23 bauzas: ^^ 16:34:32 yeah 16:34:32 eek 16:34:59 I'm not sure what the stable thing really has to do with it 16:35:27 so Project is deprecated, the last stable branch maintained is stable/wallaby, but keep stable/zed content but deprecated 16:35:31 they have a stable/zed branch that looks to match the coordinated release and be stable with everything else, but since they don't claim to support stable officially, they're not on the hook for any of it? 16:36:11 s/it/those procedures/ ? 16:36:27 yeah, we just need to give more clarity on its last supported branch and if where to start if anyone else want to take maintenance 16:37:00 the optics of this for the wider community are just really unfortunate, but okay 16:37:02 as situation is already not good we cannot improve that at this stage 16:37:08 yeah 16:37:34 and something we as TC should work to avoid this situation in future for any project 16:38:50 let's do the voting on the agreement for best possible way forward 16:38:55 is this ok "The project can be deprecated, the last stable branch maintained is stable/wallaby, but keeps stable/zed content and deprecated" 16:39:00 the most obvious thing to avoid such thing is to have more diversity in projects :) 16:39:14 what does mean ` keeps stable/zed content and deprecated` in fact? 16:39:15 slaweq: ++ :) 16:39:16 slaweq: yeah 16:39:27 noonedeadpunk: not delete the branch 16:39:47 yeah and not supported/maintained also 16:39:56 we can write many rules in our docs but if project is maintained by single company and this company changes its goals, docs won't help 16:39:57 yeah, but what `deprecated` basically means in this case 16:40:07 which will be clarified in master README.rst 16:40:24 noonedeadpunk: it reflect that any further work on stable/zed is stopped 16:40:31 as deprecation of master means that all content and CI jobs should be removed 16:40:48 on master only and content removal also 16:41:13 Has anyone from another company stepped up to take over? 16:41:17 i think we need to write this out on an etherpad or something, a one line sentence is not sufficient to describe what we are going to vote on 16:41:21 but here we want to keep stable/zed content in case anyone want to proceed further 16:41:32 spotz[m]: no afaik 16:41:39 gmann: no content removal at all, but you mean delete the master branch I assume? 16:41:59 dansmith: yes, on delete the master means content removal on master 16:42:09 but not branch itself 16:42:20 that is what deprecation step is. 'delete the master content but keep on stable' 16:42:28 gmann: right, what I mean is they claimed they were going to "empty commit" the branch.. you mean just delete the master branch but not do the empty commit anywhere right? 16:42:34 ah yes, not the branch it will have README.rst 16:42:39 will the existence of stable/zed not confuse people? (not sure this is feasible/sensible, but renaming the zed branch may be better) 16:42:39 so basically following regular deprecation process, and adjust readme to reflect support status of stable/zed 16:42:46 dansmith: yes. 16:43:05 arne_wiebalck: it is confusing but renaming it can create more confusion 16:43:07 arne_wiebalck: rename to what? 16:43:24 dansmith: to anything but stable/zed 16:43:27 arne_wiebalck: its already confusing that they're dropping support for newer stuff and only supporting the old stuff, so I think confusion is guaranteed 16:43:35 dansmith: heh 16:43:43 arne_wiebalck: the-artist-formerly-known-as-stable/zed ? 16:43:45 noonedeadpunk: basically yes. the exception is to keep stable/zed content but not maintained 16:43:49 dansmith: perfect 16:44:16 dansmith: I bet there will be people seeing this branch as the latest with content, use it, and read the README much later 16:44:17 not maintained until someone step in :p 16:44:19 but yeah 16:44:33 yes 16:44:37 arne_wiebalck: for sure, which is why it's very strange to be doing this at all 16:44:47 dansmith: yes 16:44:51 but renaming it to something else doesn't make it less confusing, just ... different confusing 16:44:53 I was just confused with haping `deprecated` and `keeping content` as these 2 are contraversary 16:45:05 *having 16:45:10 noonedeadpunk: why? 16:45:15 see, we have confused noonedeadpunk already :-D 16:45:33 that is what we can explain it in README.rst on stable/zed also 16:45:34 Hehe 16:45:41 not just noonedeadpunk, i am completely confused 16:46:02 dansmith: basically if content is there we should support otherwise it is retirement means remove content 16:46:17 Vishal Manchanda proposed openstack/governance master: Step 3: Retire repo from the Governance repo - Retire xstatic-font-awesome https://review.opendev.org/c/openstack/governance/+/872837 16:46:27 I'm just never in favor of *deleting* content that already exists. 16:46:27 dansmith: well, it's part of our "deprecation" definition kind of https://docs.openstack.org/project-team-guide/repository.html#deprecating-a-repository 16:46:33 honestly, I'm quite disappointed by the idea of empty committing stable/zed 16:46:38 or EOLing 16:46:39 here we are saying stable/zed is deprecated so content is there but it is not maintained as deprecated things are maintained until they are removed 16:46:41 just to be clear 16:46:58 as some people may take bits of TripleO without fully subscribing to their delivery model 16:47:02 that is why we want to keep stable/zed content 16:47:33 and we can imagine people contributing to the puppet modules on stable/zed without really contributing to the core of tripleo 16:47:38 other option is just delete the stable/zed and anyone want to maintain it further can take master ^HEAD ? 16:47:50 noonedeadpunk: ack, for full project deletion I see (not sure I really like that either) but that clearly doesn't apply here directly 16:47:51 let's please avoid saying that zed is deprecated then ? 16:47:55 so, why don't we just -em all existing branches of triple-o, and then it's up to the community to decide what branch they want to maintain? 16:48:07 rosmaita: even wallaby? 16:48:17 well, wallaby is em already, right? 16:48:21 in tripleo? 16:48:32 let's say it's unmaintained, unsupported or whatever, but not retired or deprecated :) 16:48:44 marking stable/zed as -em also good option and do not mark it deprecdated or so 16:48:52 it 16:49:09 is train -em in triple-o? 16:49:15 point to keep in mind, "tripleo" consists of lots of git repositories, some of which don't have stable branches, so whatever plan you arrive at needs to take that into account as well 16:49:15 no 16:49:19 but -em is a tag - not a branch 16:49:24 noonedeadpunk: right 16:49:28 in regular release process 16:49:34 yes 16:49:43 rosmaita: train, wallaby, zed are all top-level non-em branches in THT for example 16:50:02 https://github.com/openstack/tripleo-heat-templates/branches 16:50:02 yes, but once it's tagged -em, the Extended Maintenance rules apply 16:50:12 Vishal Manchanda proposed openstack/governance master: Add xstatic-angular-fileupload as Horizon team deliverables https://review.opendev.org/c/openstack/governance/+/873845 16:50:38 but marking zed as -em ahead will serve the purpose right? it is there bt not maintained by core team so anyone else want they can take it up 16:50:49 remember that you can't un-tag, so once em always em even if some new team offers to pick up maintenance of those newer branches 16:50:55 okay I thought -em was after the tag has *replaced* the branch 16:51:09 I see they have a wallaby-em tag as well as a stable/wallaby branch 16:51:11 -eol tags replace deleted branches 16:51:14 Well, the only way we can do branch renaming - is saying that it should not have been created as stable/zed at the first place as project has independant release cadence 16:51:24 -em will use the last rleased tag/hash 16:51:53 But then we're inventoing smth new and it will be waaaay harder to communicate 16:52:01 so how about this - "Following regular deprecation process, and mark stable/zed as EM" ? 16:52:17 How marking it as EM will help us? 16:52:36 yeah, I'm confused on that 16:52:39 no more releases, only best effort CI, community maintenance 16:52:47 it will say core team does not maintain it so it is like unmaintained until anyone else come and maintain it 16:53:00 so new maintainer can fix things there 16:53:17 rosmaita: no 'community maintenance' also 16:53:35 as tripleo team will not maintain it 16:54:09 so wallaby is em in their tree and has no notice in the readme, so users just have to know that it also has the tag so it must be EM? 16:54:13 if you really want to be pragmatic, just mention that stable/zed becomes untested 16:54:26 either we can reflecty the statement of 'not maintained ' in stable/zed README or mark it as -em 16:54:32 point being, I normally look at the release cycles stuff to determine what is supported and not, and don't look at the fact that there's this em tag 16:54:51 k 16:54:53 imo if we want to make an exception in our policies now - then we can go straight to EOLing Zed 16:54:56 so I think just tagging zed as em is something we'll tell ourselves matters, but I don't think it will actually communicate much to the average person 16:55:13 =1 ^ 16:55:18 *+1 16:55:21 https://docs.openstack.org/project-team-guide/stable-branches.html#extended-maintenance 16:55:46 +1 16:55:50 EOLing zed means branch will be deleted so anyone using it form there cannot fix bug 16:56:09 which is not ok but we do not have better option I think 16:56:40 why can't we just mention that the zuul coverage will be minimal ? 16:56:46 could we ask the tripleo team if they would consider just supporting the zed branch like normal? 16:56:52 that ^ 16:56:59 I suspect there may be not much work to be done there 16:57:08 but they said no in email right. they want to support only stable/wallby 16:57:09 ++ 16:57:10 so maybe they can just agree to that as a compromise 16:57:18 gmann: I know what they said 16:57:33 they also said they gonna reduce the number of jobs they run 16:57:39 (iirc) 16:57:58 IMHO we should at least try to ask them - maybe they will agree and it will be much easier :) 16:58:03 slaweq: ++ 16:58:06 even if they don't you could stick a note in the readme and add big warnings to documentation that the branch lives to enable others to help but the original team isn't around anymore. Then send it through the nromal eol process when zed reaches that point 16:58:11 sure, we can ask. 16:58:29 +1 for clarkb's proposal 16:58:29 clarkb: agree, that'd be my preferred backup plan 16:58:32 not sure we have them here but let me ask on email thread 16:58:45 and we will continue the discussion in next meeting to conclude 16:58:45 given they indicated an interest in only maintaining one (old, already em) branch, it would make more sense to just officially retire the project and let them maintain the old version downstream 16:58:49 basically not break our process, but if we don't have anyone to support it, just be upfront about it and let it age out naturally 16:59:18 yeah if they can support that will be best way 16:59:20 fungi: oh, radical approach, I like that :) 16:59:22 we are out of time, let's move to next topic 16:59:47 fungi: ++ 16:59:53 we might need to extend the meeting if everyone is ok for that ? 17:00:13 I will need to drop (wife's waiting for me) - sorry 17:00:57 k, let me go to one last topic then and others we can discuss in next meeting 17:01:04 #topic Select time to discuss the 'Less Diversity' discussion with foundation staff (Jimmy) 17:01:35 as we discussed in board sync up call, as next step we can have discussion with Jimmy (foundation staff) on diversity things 17:01:48 when we want to schedule it? in PTG or before PTG ? 17:01:51 maybe we should call this "Need More Diversity" so people don't get the wrong idea 17:02:07 :) 17:02:11 sure 17:02:30 if on PTG then we can schdule that time otherwise I will put the poll to select the time 17:02:31 I need to run but yeah don’t confuse folks;) 17:02:54 gmann: ptg sounds good to me 17:03:14 for me too 17:03:29 +1 17:03:38 ptg sounds good 17:03:57 cool, I will put this in PTG etherpad "Need More Diversity" 17:04:13 ok. thanks all for joining 17:04:22 #endmeeting