14:06:46 #startmeeting releaseteam 14:06:47 Meeting started Fri Mar 4 14:06:46 2016 UTC and is due to finish in 60 minutes. The chair is dhellmann. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:06:48 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:06:50 The meeting name has been set to 'releaseteam' 14:07:10 courtesy ping: ttx, dims, lifeless 14:07:16 o/ 14:07:29 #link our agenda is under R-5 on https://etherpad.openstack.org/p/mitaka-relmgt-plan 14:07:45 that etherpad is starting to be so huge it's pretty slow 14:08:09 yeah, I've started seeing letters inserted in the wrong order when I type, too. Next cycle we probably want to split the calendar into a few pages 14:08:17 maybe one per milestone 14:08:20 o/ partially here 14:08:26 anyhow, let's start 14:08:31 #topic m3 deadline 14:08:41 So we have 3 misses 14:08:45 thanks, ttx, for putting together the list of projects that haven't done releases 14:08:48 On managed one (barbican) 14:09:01 and trove-dashboard 14:09:15 In the official list, Astara missed the date again 14:09:17 I wonder if trove-dashboard was merged into horizon? amrith, david-lyle? 14:09:27 I've pinged adam_g about that, but I think he's west coast 14:09:42 dhellmann: about barbican, you noted that they were just waiting for a late patch 14:09:58 I haven't sent the m3 email announcement yet 14:10:01 yeah, I didn't get a patch id 14:10:05 I'll ping redrobot 14:10:12 so we could squeeze it in 14:10:40 adam_g: what about Astara ? 14:10:48 ok 14:11:33 dhellmann: are you fine adding barbican and trove-dashboard if we can get them in the next hours ? 14:11:33 since astara isn't managed, I think it's safe to announce m3 without them, even if we let them slip a tag in later today 14:11:44 sure 14:12:02 I'd like to get both barbican and trove-dashboard in though, assuming amrith actually wants a trove-dashboard release 14:12:07 hello 14:12:10 ack 14:12:23 I need to come up with a better way to track who's keeping up with the schedule next cycle. A checklist of deliverables. 14:12:29 dhellmann, trove-dashboard should be released, yes. 14:12:40 I wasn't sure whether it is treated as a 'client library' 14:12:50 and the understanding we had was that horizon would get it 14:12:52 no, it's treated as an extra thing 14:13:01 same deadlines as trove itself 14:13:10 if follows cycle-with-milestones, so we need a 0b3 tag 14:13:10 well, extra thing with its own release or extra thing that horizon will get? 14:13:23 most projects release the -ui extensions at the same time as the main thing 14:13:23 ok, a b3 tag shall be produced ;) 14:13:25 amrith : according to http://governance.openstack.org/reference/projects/trove.html the trove team manages it 14:13:43 dhellmann, ok, I'll make a b3 tag. 14:13:46 amrith: thx 14:13:48 amrith : great, thanks! 14:13:56 what's the convention re: the version numbering? Same as trove I assume? 14:14:08 amrith: well... 14:14:10 when they are to be release together (same time) 14:14:14 the issue is it's been using it's own 14:14:17 amrith : it doesn't have to be 14:14:25 https://review.openstack.org/#/c/288339/ 14:14:34 I see 6.0.0.0b2 there now 14:14:47 yikes 14:15:02 that's 'interesting'. 14:15:05 amrith: you might want to post a follw-up patch to https://review.openstack.org/#/c/288339/ 14:15:09 though that's probably an old one 14:15:27 dhellmann: no, it's the m2 one 14:15:32 well, that's about 4h old. I think ttx just posted that this AM 14:15:42 ah, yeah, I see that's a new patch now 14:15:52 yes, it's just documenting the existing patch 14:15:55 tag 14:16:03 got it 14:16:23 how badly would the wheels come off if I sent up a 5.0.0.0b3? 14:16:34 very badly, you can't go backwards 14:16:45 seems to make sense that Trove 5.0.0.0b3 goes with trove-dashboard 5.0.0.0b3, no? 14:17:04 * amrith channels my inner Trump 14:17:12 yes, but things would be very unhappy 14:17:22 ok, so going backwards is out. 14:17:25 like they would continue to use 6.0.0.0b2 anyway 14:17:35 your users would come and find you and do unpleasant things to you for breaking their upgrades 14:17:35 so it may be too late to catch this bus 14:17:56 I'd say do a 6.0.0.0b3 14:18:00 right 14:18:01 and maybe align next cycle 14:18:02 and at a later time, maybe trove has to skip release 6 and go straight to 7 in lock step with dashboard. 14:18:05 ok, sounds good. 14:18:12 i.e. go from 6 to 8 14:18:13 you can fix them next cycle by jumping trove to 7 14:18:16 err 14:18:18 5 to 7 14:18:20 yup 14:18:20 right 14:18:37 ok, I'll send up a trove-dashboard 6.0.0.0.0b3 14:18:40 dhellmann: maybe we can fast-approve https://review.openstack.org/#/c/288339/ 14:18:45 sorry, I'll figure the right number of 0's. 14:18:45 so that amrith posts his on top 14:18:53 done 14:18:58 you want me to dep on 288339? 14:19:09 or just wait for you to merge 14:19:12 amrith: should be in master really soon. 14:19:14 and then submit 14:19:17 ok, thanks 14:19:18 amrith : give it a few minutes to merge, those jobs don't take long 14:19:21 will do it 14:19:27 amrith: thx 14:19:33 amrith : thanks 14:19:33 next 14:19:42 python-brick-cinderclient-ext 14:19:56 missed lib freezae 14:20:10 not used anywhere so smcginnis wanted to do a release next week 14:20:22 At this stage I wonder if it's not simpler to make a first release in newton 14:20:33 there's a 0.1.0 tag, when was that from? 14:20:34 since it won't be used before 14:20:40 hmm 14:20:44 yeah, I just tagged that yesterday 14:20:46 it might have been late 14:20:49 oh 14:21:00 15:12 -0500 14:21:18 I see it now. Ignore me 14:21:35 Next 14:21:41 ok, then cloudkitty 14:21:48 python-cloudkittyclient needed for client lib deadline today 14:21:58 is that project active? 14:22:02 sheeprine told me he would push a tag request now 14:22:17 ok 14:22:38 so it should be there anytime now, just doing a few last-minute tests 14:22:39 how about the server-side projects? 14:22:50 So that's the next topic 14:23:02 Things that do intermediary releases but haven't yet 14:23:24 cloudkitty, cloudkitty-dashboard: sheeprine promised me one in the next 10 days 14:23:36 ironic-ui: do you have news 14:23:53 jroll : what's the plan for ironic-ui? 14:23:58 * dhellmann has no news 14:24:10 Last one I caught this morning was sahara-tests, no idea what plans SergeyLukjanov has for this 14:24:27 we should keep a close eye on those next week 14:24:34 and nag 14:24:42 yeah 14:24:50 I spotted a number of other issues while I went through my lists this morning 14:25:08 I opened reviews for the first 3 14:25:29 * dhellmann waits for gerrit to reload the review dashboard 14:25:31 The 4th is in progress, except angdraug hasn't +1ed my review so it's stuck 14:25:45 The last 5 I want to discuss now 14:25:53 - Missing entry for tacker 0.1.0 (tacker, tacker-horizon, python-tackerclient) 14:26:12 I suspect we want to include that in the mitaka page ? Even if recent ? 14:26:16 I don't see that, did it merge already? 14:26:28 yeah, probably 14:26:45 so we need to file a releases yaml for it 14:26:48 I'll do that 14:27:06 ah, ok, I thought you were saying you'd already done so, sorry I see now 14:27:15 - pbr is release:independent but appears in mitaka deliverables 14:27:29 so this is a mismatch but I have no idea what would be correct 14:27:46 hmm, yeah. 14:27:47 I'll let you sort that one out and file the fix 14:28:03 - tempest is cycle-with-intermediary but hasn't done a "release" 14:28:08 we don't do stable branches, but it seems important to indicate which version of pbr we need 14:28:10 tempest is another odd one 14:28:17 yeah 14:28:37 ttx, dhellmann ... shouldn't dashboard projects also feature in requirements/upper-constraints/global-constraints? How else do we tie a particular dashboard version to a 'release'? 14:28:47 there are some old stable branches, but I don't think mtreinish is using stable branches for tempest any more so maybe we should make it release:independent 14:29:01 amrith: install docs I suspect 14:29:12 ouch 14:29:13 amrith : if nothing actually depends on it, the install guide and the releases.openstack.org page will do it 14:29:17 - Other dashboards should not be considered "libraries" 14:29:42 how then do we manage dependencies? 14:29:58 So we have a bit of inconsistency where some -ui or -dashboard or -horizon project are type:library, some are packed with deliverables, and some don't have type 14:30:18 ttx: that may indicate that we need another type: tag? 14:30:25 My current take on them is that they should either be packed with deliverables or not use type: for the moment 14:30:41 dhellmann: maybe, in the mean time I'd like to move those from library to other 14:30:51 ttx: ++ 14:30:55 murano-dashboard, monasca-ui 14:31:02 astara-horizon 14:31:18 I think that is all 14:31:29 the others are already there 14:31:39 That will make it clearer that the deadline don't apply to them 14:31:39 ok, I guess that's an independent conversation. I see your change merged, I'll push up 6.0.0.0b3 in a couple of minutes 14:32:01 dhellmann: ok, will propose the fix for that too 14:33:14 finall last one 14:33:17 training-labs is release:cycle-with-milestones 14:33:20 amrith proposed openstack/releases: Release 6.0.0.0b3 of trove-dashboard https://review.openstack.org/288481 14:33:24 which is kinda very weird 14:33:38 given it didn't tag anything yet 14:33:53 I don't even know what training-labs is, I'll check with loquacities 14:34:18 Alright. All covered 14:34:55 I really thought we'd straightened out the release model tags already. This is more mismatches than I expected. 14:35:19 dhellmann: people keep doing weird things 14:35:19 ready to move on to requirements? 14:35:20 tried to read throught the conversation, but couldnt find the answer — is there anything wrong with murano-dashboard currently? 14:35:42 kzaitsev_mb: not really. I'll remove the type:library tag from it since it's a horizon plugin 14:35:48 kzaitsev_mb : we may want to change some tags, but that's it 14:36:06 oh, got it =) you were talking about different type of tags =) 14:36:18 yeah, governance tags :-) 14:36:29 * dims back on now 14:36:32 #topic freezing requirements 14:36:37 dims: just in time 14:36:40 :) 14:36:53 I have a list there in the etherpad of a bunch of the patches I think we want to land for requirements. 14:37:06 I think we *don't* want to land *any* others, including any future auto-generated constraints updates 14:37:12 does that match what your expectations are? 14:38:12 yes, no more merges either to g-r or constraints. 14:38:16 It looks sane but I'll admit I haven't looked at the whole list 14:38:32 anything we release today should be updated in upper constraints though 14:39:41 dhellmann : agree? 14:40:03 dims : good point, we had one or two more client releases on that list from earlier 14:40:11 I'll add those to the etherpad as they come in 14:40:17 dhellmann : cool 14:41:36 dhellmann : i am debating a oslo.cache update to match the dogpile.cache bump. as projects don't define dogpile directly and rely on oslo.cache to pull that in 14:41:39 ok, we can make adjustments if we find the need 14:41:47 dims : is that a new minimum? 14:41:59 dogpile.cache is a new minimum. yes 14:42:05 yeah, I agree, it would be good to have that set with oslo.cache 14:42:33 do we want to raise the minimum for oslo.cache, then? 14:42:37 ok. i'll spend some time today poking at list_oslo_unreleased_changes 14:42:45 dhellmann : yes, otherwise it would not work 14:43:24 makes sense 14:44:00 the etherpad is starting to slow down a lot, I may go ahead and create a new one for finalizing the release 14:44:03 that's the only one i had, will raise red flags if i see anything else 14:44:08 if so, I'll stick a link in this one 14:44:13 thanks 14:44:16 thanks 14:44:21 #topic lib stable branches 14:44:30 we've scheduled this for early next week 14:44:45 dims, you wanted to try to do some oslo releases earlier? 14:44:49 sorry, branches 14:45:09 I think you'd want to leave oslo.cache out, since we're planning another requirements update there 14:45:28 ttx: do you see any reason to wait until monday? 14:45:29 so can we just create the branches so folks can work on master say middle of next week 14:45:52 so do you want to wait? 14:45:57 dhellmann: breaking things on Friday is a bad idea ? 14:46:08 ttx: very true 14:46:12 dhellmann : we don't have to do it today. just figuring out if we could do it and if there are any obstacles 14:46:19 dims : ah, I see 14:46:42 so i can tell the team on monday in the oslo team meeting 14:47:05 ttx : +1 :) 14:47:20 dhellmann: I'm told ironic-ui should have something this week, I have yet to find out how true this is 14:47:22 we always cut a branch from a release, so you can go ahead and let them work with the understanding that if they commit features we'll have to set up the branch and backport bug fixes 14:47:31 jroll : today's the day :-) 14:47:43 mhm 14:47:43 jroll : if there's someone else I should talk to, let me know? 14:47:50 dhellmann : cool. will pass that along 14:47:55 dhellmann: betherly is leading that 14:48:03 jroll : ack, I'll ping her 14:48:04 I just poked her in ironic channel 14:48:25 k 14:48:33 hi jroll dhellmann 14:49:08 hi, betherly, we were hoping to get an idea of when ironic-ui would be ready for a release or milestone tag? 14:49:28 trying to get through review process. made a whole ton of changes this am to meet reviewer requests so waiting on the +2s before it can merge 14:49:57 have pinged krotscheck to see if he can +2 things today 14:50:32 dhellmann: looking at the tacker 0.1.0 tag it seems to be pretty empty and we shouldn't probably consider it a mitaka release. 14:50:49 so I'll move tacker to the "ping for an intermediary release" section 14:50:50 betherly : ok, it's marked as using the cycle-with-intermediary model, which means we'd like to have a release soon that we can use to fall back on if there are issues with any point releases between now and final 14:50:56 ttx: sounds good 14:51:26 betherly : basically, something *close* to final, with the expectation that you might have bug fixes but not features to land before actual final 14:51:40 or there may be no changes needed, and we could use that version as final 14:52:08 dhellmann: i will do my best to get it up there today for sure i will keep both you and jroll in the loop 14:52:23 betherly : thanks! drop by #openstack-release if you have questions or updates 14:52:39 will do thanks dhellmann! 14:52:47 #topic open discussion 14:52:48 dhellmann: means we shoudl probably ask for a python-tackerclient release ~todayish 14:53:12 ttx: noted 14:54:17 I've started filling in https://etherpad.openstack.org/p/newton-relmgt-plan by adding some of the incomplete action items from this cycle 14:54:49 we can review those when we get closer to the end of the cycle and are ready to start doing planning 14:55:22 dhellmann: would be good to get an idea of what lifeless managed to implement of his mitaka plan for constraints, so that we know what to defer to newton 14:55:31 ttx: do you have a draft of the announcement email ready? 14:55:37 dhellmann: not yet 14:55:46 ttx; yeah, I need to confer with him 14:56:35 ttx: did you see my note about the FFE for ironic-lib under R-4? 14:57:09 they have a FFE, but it's a lib, so they'll need to update the minimum requirement next week 14:57:32 should we make the deadline tuesday? 14:57:48 I think only ironic-python-agent will actually need to have that min updated 14:58:12 if only ironic-python-agent is affected... they can have the whole week 14:58:24 we'll have others as RC bugs are discovered in libs 14:58:34 good point 14:58:50 ok, anything else? we're about out of time 14:58:56 nope 14:59:12 FOr newton we should work on automating the verification I did manually this morning 14:59:21 yeah, definitely 14:59:28 i.e. governance/releases mismatches, and missing releases 14:59:31 it would be nice to have some sort of dashboard 14:59:37 ++ 14:59:43 something to think about 14:59:54 also releases/git mismatches, like missing tags 15:00:09 yes, good point 15:00:24 we have some scripts for importing git tags, but they tend to produce messy results 15:00:32 ok, we're out of time and should give up the channel 15:00:33 thanks! 15:00:36 #endmeeting