Friday, 2021-03-19

*** vishalmanchanda has quit IRC00:09
*** tosky has quit IRC00:14
*** penick has joined #openstack-tc00:34
*** penick has quit IRC00:59
*** penick has joined #openstack-tc01:00
*** penick has quit IRC01:05
*** lbragstad has quit IRC02:06
*** evrardjp has quit IRC03:33
*** evrardjp has joined #openstack-tc03:33
*** ianychoi has quit IRC03:37
*** vishalmanchanda has joined #openstack-tc04:16
*** penick has joined #openstack-tc04:36
*** jaosorior has quit IRC05:14
*** penick has quit IRC05:24
*** penick has joined #openstack-tc05:28
*** penick has quit IRC05:38
*** Luzi has joined #openstack-tc06:04
*** jaosorior has joined #openstack-tc06:21
*** e0ne has joined #openstack-tc06:32
*** lpetrut has joined #openstack-tc06:42
*** e0ne has quit IRC07:13
*** dklyle has quit IRC07:35
*** rpittau|afk is now known as rpittau08:24
*** andrewbonney has joined #openstack-tc08:51
*** tosky has joined #openstack-tc08:54
*** e0ne has joined #openstack-tc09:47
*** ricolin has quit IRC10:01
*** e0ne has quit IRC10:35
*** e0ne has joined #openstack-tc10:36
*** e0ne has quit IRC11:40
*** e0ne has joined #openstack-tc11:42
*** lbragstad has joined #openstack-tc12:22
*** e0ne has quit IRC12:55
*** lbragstad_ has joined #openstack-tc13:25
*** lbragstad has quit IRC13:33
*** irclogbot_2 has quit IRC13:33
TheJuliaSo, I was pondering/wondering13:34
TheJuliaSince we're no longer trying to move at warp speed, perhaps it is time to reconsider the coordinated release model and encourage a model where projects release at their own cadence but maybe still do a yearly alignment for a formal named release?13:35
*** purplerbot has quit IRC13:36
*** irclogbot_2 has joined #openstack-tc13:36
evrardjpTheJulia: +113:41
evrardjpThere were multiple ideas in the past to promote this!13:41
evrardjpPutting projects in charge of the release isn't a bad idea, certainly for the most mature ones.13:42
evrardjpOf course, the release team should voice here, and maybe it shouldn't be an all or nothing (could be an opt-in)13:43
*** purplerbot has joined #openstack-tc13:44
fungiso switch from two coordinated releases a year to one, basically (with more intermediate releases)13:57
*** e0ne has joined #openstack-tc14:02
*** e0ne has quit IRC14:04
*** jaosorior has quit IRC14:09
*** e0ne has joined #openstack-tc14:16
*** e0ne has quit IRC14:16
TheJuliaYeah, basically14:39
fungiin the past it was asserted that ~nobody consumes/redistributes the intermediate releases, so that would effectively mean new features don't make it into the hands of users for a year instead of half a year, but that may not still be the case today14:44
TheJuliaI guess, in part, I'm worried about burning folks out and setting appropriate cadence.14:46
fungiyep, a valid concern, for sure14:46
TheJuliaI know some *now* do consume some intermediates, like ironic intermediates get consumed and shipped, so that enables new features to get consumed/used. At the same time, are people really shipping a full coordinated release every six months14:47
TheJuliabecause that is also a super heavy lift for any involved party as well14:47
TheJuliaConsuming an intermediate where appropriate may be a lot easier and would help with things like upgrade stories in the grand scheme of things14:48
TheJuliadeprecation and backporting rules would likely need to be revisited, but nothing particularly insane. Just thinking more from a human sanity/end user benefit standpoint.14:53
*** Luzi has quit IRC14:55
fungiyeah, i really have no idea. like, what is rdo's release model? or ubuntu cloud's?14:55
fungii know debian tries to update packages after the coordinated release and then apply stable point releases thereafter, but i don't really know much about any other distributors' models14:56
fungipart of it is also about things like how many versions of oslo.config are you going to package... do ironic's intermediate releases use the oslo libs from the previous coordinated release, or from the upcoming coordinated release?14:58
*** rpittau is now known as rpittau|afk15:08
*** lbragstad_ is now known as lbragstad15:11
*** dklyle has joined #openstack-tc15:35
spotzRDO is a trailing release model15:43
fungispotz: but does rdo also ship intermediate releases, like the ones ironic produces during the cycle leading up to the coordinated release?15:47
spotzfungi: I'd call it more like dailies then a set release15:53
spotzI can grab someone and pull them in here if you need all the specifics15:54
fungigot it, so snapshots of everything, but not the intermediate releases themselves15:54
spotzNot from what I've seen no15:54
spotzIf we need them we can probably discuss it with the team15:55
fungispotz: yeah, the bigger question is what it would mean for, say, rdo if we started focusing more on intermediate releases which happen at random times for different projects throughout the year15:55
spotzTRying to see who from the team isn't already marked as away for the weekend15:57
spotzMight need to be a Monday question15:58
*** ykarel has joined #openstack-tc15:59
fungii doubt it's urgent15:59
spotzOk ykarel is here to ask fungi15:59
ykarelfungi, spotz, what's the question15:59
fungiykarel: proposal from TheJulia starts here:
fungii'm mainly curious what that would mean for major distributors of openstack currently, rdo as an example16:01
spotzWhich is where I started paying attention:)16:01
fungidata points like that would probably weigh into any such changes to openstack's overall release model16:02
ykarelfungi, as part of RDO release, we follow upstream releases, like for wallaby, clients and non-clients already released, RDO too built those tagged versions16:04
ykarelif a new release for a library get's released during release or post release we will get that build too16:04
ykarelwe follow upper-constraints for this16:04
ykareland for project which are not in u-c and follow rc model we pick latest releases16:05
fungiykarel: got it, so for example ironic 16.2.0 as an intermediate release is shipped as part of a "wallaby preview" then? not as part of victoria because that's already branched and gets its own separate stable backports16:07
ykareland the projects which are independent or not in u-c are picked latest and then updated only when maintainer requests16:07
fungibasically ironic has multiple releases in master in a given cycle, the last of which is included in the coordinated openstack release. just curious what the distribution of the earlier releases of ironic during the cycle is like in rdo16:08
fungido end users of rdo ever see them, or do they only wind up with the coordinated release of ironic?16:08
ykarelfungi, for RDO release we will pick latest whenever we build it for wallay release16:08
ykarelwe will do it during RC phase16:09
spotzykarel: So those wouldn't be included in the 'daily' builds?16:10
fungiso if we went to yearly coordinated releases with a lot more projects doing intermediate releases from master at random times during the year prior to the coordinated release, rdo would probably still only ship the last of each of them which appear in the coordinated release?16:10
spotzI'm glad I invited you over!16:10
ykarelbut NOTE ^ we are talking about CentOS CloudSIG release where we follow tag releases, there is RDO trunk where we follow commits from branch and releases don't matter in those case16:10
spotzfungi: I think that would be a team/project discussion that we'd need to have if upstream release cycles changed16:11
ykarel vs
ykarelspotz, in daily builds we don't consider tag releases for projects not in upper-constraints16:11
fungispotz: yep, totally. mostly trying to figure out if users of major distributions are seeing the intermediate releases today, or only the final releases, and what that might mean if we extended the cycle and focused more on non-coordinated intermediate releases over the course of the year16:12
ykarelfungi, yes right before final releases we will not pick intermediate release as we start building for Centos only after upstream start library releases16:12
ykareland if the releases happen after final releases, those will also be picked up16:13
*** ykarel has quit IRC16:16
*** ykarel has joined #openstack-tc16:17
* ykarel got disconnected16:18
spotzYou typed last ykarel:)16:18
fungiyeah, we waited ;)16:18
ykarelok good :)16:18
* ykarel checks if some of my messages missed16:18
spotzAnd if you need to go we can take it up on Monday or another day. But at least you can have talks with folks based on this16:18
fungiykarel: and the collection containing those intermediate releases is treated as a sort of preview of the upcoming coordinated release?16:18
fungibut yeah, this is not urgent, i'm just trying to get a feel for where users end up encountering intermediate releases from master today16:19
ykarelfungi, you mean for the trunk builds which are based on commits?16:20
ykarelsorry didn't got last question16:20
fungiykarel: ironic 16.2.0 as an example16:21
ykarelfungi, so suppose during RC phase, if ironic-16.2.0 is the latest release, we will get that build16:22
fungior 16.1.0 before it16:22
ykareland when next release i.e 16.3.0 happens it will be auto build if it's before wallaby GA or after wallaby GA16:22
ykarel16.1.0 will not get build,16:23
fungiright, i get that you'll ship the last one as part of wallaby, but will any of your users see the other intermediate releases? sounds like no in the case of rdo16:23
ykarelonly releases post 16.2.0 will be picked16:23
ykarelyes it will not be seen16:23
fungiso under your current release model for rdo, if we had coordinated releases less often and made a lot more intermediate releases, the net result for rdo consumers would be that there are fewer releases (yearly instead of twice a year)16:24
fungithey won't see the intermediate releases made in the interim period16:25
ykarelyes right16:25
fungithanks ykarel! that's what i was trying to figure out16:26
fungiobviously rdo might decide to change their release model to compensate if we shift the focus of our releases, but at least knowing that helps inform discussion about those sorts of decisions16:26
ykarelfungi, yes we may need changes based on upstream, we closely follow it, so keep us informed whenvever there are such changes16:27
spotzAnd if there are any meetings let me know and I'll make sure someone can attend who can answer questions, and I'll try to attend as well16:28
spotzThe more we understand the better decisions can be made within RDO16:28
*** ykarel is now known as ykarel|away16:29
ykarel|awayThanks spotz fungi16:29
fungithanks to you! have a great weekend16:29
spotzThank you! Enjoy your weekend!16:29
*** lpetrut has quit IRC16:47
*** ykarel|away has quit IRC17:31
*** timburke_ has joined #openstack-tc17:47
*** timburke has quit IRC17:50
*** andrewbonney has quit IRC17:50
*** vishalmanchanda has quit IRC19:35
*** timburke_ has quit IRC19:50
*** timburke_ has joined #openstack-tc19:50
*** openstackgerrit has joined #openstack-tc19:53
openstackgerritGhanshyam proposed openstack/project-team-guide master: Clarify optional lower bounds testing
openstackgerritGhanshyam proposed openstack/project-team-guide master: Clarify optional lower bounds testing
*** e0ne has joined #openstack-tc20:02
openstackgerritGhanshyam proposed openstack/project-team-guide master: Clarify optional lower bounds testing
gmanntc-members, fungi updated ^^20:08
gmannalso sent updates on ML,
gmannpinged Keystone team for leaderless things20:12
gmanngot response form Mistral team on trying DPL model and we should see patch soon20:13
gmannthat complete my AI from yesterday meeting.20:13
*** openstackgerrit has quit IRC21:22
*** e0ne has quit IRC21:44
*** lbragstad has quit IRC22:11
*** akahat has quit IRC22:11
*** lbragstad_ has joined #openstack-tc22:12
*** akahat has joined #openstack-tc22:19
*** tkajinam has quit IRC23:27

Generated by 2.17.2 by Marius Gedminas - find it at!