14:01:22 #startmeeting releaseteam 14:01:23 Meeting started Fri Apr 1 14:01:22 2016 UTC and is due to finish in 60 minutes. The chair is dhellmann. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:24 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:26 The meeting name has been set to 'releaseteam' 14:01:28 o/ 14:01:34 courtesy ping: ttx, dims, lifeless, ajiang 14:01:55 that should have been AJaeger 14:02:08 our agenda is under R-1 in https://etherpad.openstack.org/p/mitaka-release-tracking 14:02:19 #topic dead links on mitaka page 14:02:29 thanks for doing that review, ttx 14:02:31 #link https://etherpad.openstack.org/p/dude-where-are-my-tarballs 14:02:41 I have to admit I found more than I expected 14:02:42 we have some fixes in the queue already, some are easy 14:02:49 #info fix tarball base setting for a few projects: https://review.openstack.org/300457 14:02:59 some are harder 14:03:02 #info drop unpublished murano-apps: https://review.openstack.org/300473 14:03:13 I don't know if we want to do that, or just turn their tarball link setting to "none" 14:03:49 I added that setting for one of the deployment projects (ansible or puppet?) since they don't build tarballs 14:04:05 what do you think? delist or just turn off the links? 14:04:06 checking 300457 14:04:58 +2 14:05:10 looking at murano-apps 14:05:44 We could no-tarball it 14:05:55 as a first step 14:05:58 yeah, that feels better, I'll make that update 14:06:06 then discuss with PTL if it makes sense to even list 14:06:36 what about monasca-thresh 14:06:39 yeah, same for astara 14:07:31 I was going to treat monasca-thresh the same way, but haven't submitted that one 14:08:10 feels like an oversight in the case of monasca-thresh, not even sure how functional monasca is without it 14:08:18 they have a bunch of missing parts to their release build, I need to discuss with rolland when he comes online 14:08:25 yeah 14:08:38 murano-apps is a different animal 14:08:46 it probably has no jobs on purpose 14:09:18 what is that? 14:09:30 hmm 14:09:36 so monasca-thresh is a Java component 14:10:18 hmm, I don't think we even know how to publish that 14:10:23 (cue told you so music about accepting Java bits without toolchain to handle them) 14:11:04 yeah 14:11:11 I see a couple of options: 14:11:15 1. no-tarball 14:11:16 dhellmann: yeah, so no-tarball them and then open another review to just drop it from "mitaka" 14:11:19 2. delete entirely 14:11:30 3. build the "tag but don't show" feature 14:11:41 personally I don't think something without a proper job to build a source code tarball can be "released" 14:11:49 wfm 14:12:06 if we're going to delete it, there's no need to no-tarball them, right? 14:12:12 so I'll +2 the removal from mitaka, but will let the PTL defend themselves before 14:12:19 ah 14:12:27 sure. I just want to remove the 404 14:12:30 immediately 14:12:38 sure 14:12:43 and then decide what is the long-term solution 14:13:04 are we talking removing all of monasca, or just thresh? 14:13:34 murano-apps is a collection of murano packages, which is a bit specific too 14:13:54 just thresh... 14:14:04 I bet people can find it somewhere else like on maven 14:14:20 which is why nobody bothered to create tarball jobs 14:14:51 also source code is kind of a dirty thing in Java land. they release jars 14:15:10 ah 14:15:18 * dhellmann is ignorant of these things 14:15:25 don't even get me started 14:15:30 let's keep moving, then :-) 14:15:33 #info add missing release jobs: https://review.openstack.org/300474 14:15:52 those are the projects where it seemed to make sense to add the missing jobs 14:18:40 looks like there's a validation failure, so I'll keep working on that 14:18:40 for fun, http://ttx.re/the-real-problem-with-java-in-linux-distros.html 14:18:40 #info add validation for release jobs: https://review.openstack.org/300492 and https://review.openstack.org/300493 14:18:40 * dhellmann adds blog post to his reading list 14:18:40 bit old, still valid :) 14:18:40 there's sample of the combined validation output in http://paste.openstack.org/show/492734 14:18:40 I should probably not require the jobs when we have tarball mode set to none, so I'll add that as a third patch 14:18:40 is there anything else on the links topic? 14:18:45 does that cover them all ? 14:19:00 do we have a fix for monasca-log-api 14:19:30 oh, I missed that one at the top 14:19:42 I'll no-tarball that, too 14:20:04 What about monasca-ceilometer 14:20:14 not in https://review.openstack.org/#/c/300474 14:20:22 intentional ? 14:20:35 ah, andreas pointed out they didn't have the venv setting in tox, so I'll treat it the same as log-api 14:20:50 ok 14:21:01 They also don't have the zuul layout thing I suspect 14:21:06 o/ fashionably late. sorry 14:21:08 probably 14:21:12 hi, dims 14:21:52 unfortunately I often don't see rolland on irc, so I'll try email to contact him 14:22:24 ttx, should I copy you or openstack-dev on that email since I won't be around a lot of next week? 14:22:51 I'll send it to the dev list, no reason not to do this publicly 14:22:55 ack 14:23:06 ok, moving on then 14:23:09 #topic cinder 14:23:24 I asked yesterday on irc and the cinder cores I could find seemed surprised that they might need an rc3 14:23:29 smcginnis wasn't online at the time 14:23:47 well, they don't need it that much 14:23:50 I expect they'll want a patch release after the final for those translations 14:23:52 yeah 14:23:59 could have been useful to refresh translations to pick up Wed/Thu 14:24:16 but not critical if they missed 14:24:22 ++ 14:24:25 a bit worrying that they would not be available/around though 14:24:31 the thing with master is concerning, but can be dealt with 14:24:43 yeah, I just don't want to forget about it 14:24:50 yeah, I'm not sure about the story there 14:25:02 ok, it's noted on the dashboard, too 14:25:18 do you think we need to push them to deal with merging the patch? 14:25:48 it's like every time you think you can propose both master and stable/mitaka (saying master will merge anytime now) it takes ages to merge in master and merges straight away in stable 14:26:08 maybe we should encourage them to use depends-on 14:26:14 which is why I try to strictly enforce the "must merge in master first" rule 14:26:21 I wonder if that works across branches of the same repo 14:26:22 dhellmann: well no 14:26:34 dhellmann: we use teh same changeID 14:26:39 oh, right 14:26:41 right, won't help 14:26:57 and wouldn't prevent them from merging anyway 14:27:15 last cycle we met with ptls regularly, this cycle not at all 14:27:24 maybe next cycle we need to do a pre-release meeting near the end of the cycle? 14:27:38 "we won't process your release until you come around and have this chat" 14:27:38 what we need is some kind of magic check that would look into a stable change and check that it's a straight backport (or has a #different_but_this_is_normal warning tag in the commit message) 14:28:08 that way you wouldn't unintentionaly shoot yourself in the foot 14:28:43 what about just a check that looks for the cherry-pick in the commit message, and requires a change with that id in master? 14:28:48 not sure what the solution is, but that's something we should discuss 14:28:48 it wouldn't have to look at the contents 14:29:03 that would be great already 14:29:15 ttx : dhellmann : how about language in the bot adding a snippet saying "This change should be merged in master first before you approve in stable"? 14:29:21 you need a way to disable it though, some patches just don't get to master (like different translations) 14:29:30 so at least they have a heads up 14:29:32 y 14:29:49 let's brainstorm that in postmortem / in Austin 14:29:51 * dhellmann adds a note to https://etherpad.openstack.org/p/newton-relmgt-plan 14:29:54 ++ 14:30:00 nothing we can change now 14:30:07 ttx: if there's no cherry-pick reference in the commit message, the check would pass 14:30:07 y too late 14:30:20 although watching stable/mitaka now that we are in pre-reelase freeze for "accidental" merges would be good too 14:30:40 yeah, we should -2 everything in the branch I guess 14:30:50 https://review.openstack.org/#/q/branch:stable/mitaka+is:merged 14:31:16 looking good so far, only a few translations merged in cinder 14:31:21 ttx, if I -2 the pending patches I won't be around to lift the ban 14:31:29 do you want to do it? 14:31:49 or we can revert spurious things 14:31:50 or do you want to leave it open, and revert anything if we end up needing a release? 14:31:51 yeah 14:31:56 let's do that, less work up front 14:32:01 -2 is a bit task-intensive and brittle, since that relies in regular reviews 14:32:02 more for the team that doesn't follow the rules 14:32:10 right 14:32:23 ok, moving on... 14:32:28 #topic other fires 14:32:41 do we know of anything else, or was this just in case something came up? :-) 14:32:55 just in case something came up 14:32:58 ok 14:33:02 or I forgot something important 14:33:06 #topic step-by-step release week actions 14:33:11 that were just the fires I could see 14:33:25 how about if we work these steps out in the etherpad so I don't have to try to copy them over in the right order? 14:33:27 ok, so for here I'd like to know what we'll have to do next week 14:33:33 ack 14:33:45 there's already a "release tasks" section under R-0 14:34:17 dhellmann : "unfreeze requirements in master" 14:34:52 LOL i wish 14:35:05 (password typing repeatedly) 14:35:13 * dhellmann might temporarily turn off the passphrase on his gpg key for that step 14:35:53 ttx, there shouldn't be any races because it will be one patch to the releases repo 14:36:08 ah. 14:38:05 I may need to send that announcement a bit earlier than usual 14:38:29 what time would you send it ? 14:39:02 dhellmann: the old process took about 7 hours to run with all the Launchpad stuff 14:39:04 well you said "shortly before 10" and I might want to do it closer to 9? 14:39:18 which is why we'd wait around 10am CT anyway to do it 14:39:28 but we can certainly have a more precise hour 14:39:39 ok, let's target 9am CT 14:39:49 yeah, if I'm up early CT to do the first couple of steps, I'd like to just announce right away so I can go back to my holiday :-) 14:40:03 9 seems like a safe upper bound, sure 14:40:07 :) 14:40:31 we can prep the patch to switch to released today and then approve it next week 14:40:41 the series status, that is 14:41:06 dhellmann: "update acls for stable/mitaka service projects to allow changes from the team stable maintainers (revert previous change)" 14:41:17 that's https://review.openstack.org/#/c/298866/ right 14:41:37 do we have to do anything on launchpad? 14:42:11 ttx: I thought it was the +2 rights? 14:42:48 ttx: https://review.openstack.org/#/c/292781/ 14:42:54 oh, the tagging rights 14:43:13 right so that's a different thing 14:43:14 we want to give them +2 but keep tagging 14:43:29 yeah, it may not be a straight revert, we might want to prepare those patches today, too 14:44:42 dhellmann: I fear those will need rebasing by next week 14:44:56 probably 14:45:07 rebasing will be faster than figuring it out from scratch 14:46:43 Ack ,added prep to the "before Thursday" section 14:47:01 * ttx checks out old release cheat sheets for things we are forgetting 14:47:02 yeah, I'll do that today and prepare them as a patch series so the rebasing will go more smoothly next week 14:47:25 ttx: we should open the g-r repo, I'll remove my -2s there today 14:47:47 oh haven't done that yet ? 14:48:09 no, I've been letting it slide because of late releases and folks weren't pushing for it 14:49:45 adding everything I can think of 14:49:58 yeah 14:52:18 * dhellmann makes a note to pack a candle 14:52:39 this feels like a complete list, ttx 14:52:42 dhellmann: could you be more explicit as to when you will be fully unavailable next week ? 14:52:51 let me look at the calendar, just a sec 14:53:07 so taht I know when to wait for you and when to just make a call 14:53:13 we travel monday night, so I'll have some email Monday 14:53:25 dims: what is your availability looking like next week ? 14:53:47 tuesday I'll have email on my phone and a laptop that night (european time) 14:53:52 if we get broken by library updates it will more likely be over the weekend 14:53:53 ttx : i'll be around all week 14:54:18 I'll have access again at some point wednesday evening 14:54:31 I'm planning to work thursday morning until 9-10 14:54:40 I'll be up early to get an early start 14:54:53 could you add that to the etherpad ? 14:54:56 sure 14:55:47 oh, and.. no library release on our side 14:56:00 will have plenty enough risk with the things beyond our control 14:56:22 y 14:56:46 dhellmann: it doesn't look too bad, definitely doable if there is no unexpecetd failure somewhere 14:56:51 i have znc setup, so now i can check for my nick when i get back online too 14:57:06 ttx: direct email to me will always buzz my phone, irc is going to be trickier 14:57:27 dims: thx a lot for your availability btw, 3 people is not too much :) 14:57:33 my travel access to irc right now is not great because of the new laptop 14:57:56 yeah, dims, thank you for helping so much! 14:57:58 dhellmann: you can pm me you phone number and I can text you in case of horrible urgency 14:58:05 ttx: absolutely 14:58:17 my pleasure 14:59:08 I have a call starting in a minute, so I should close us out 14:59:18 see you in #openstack-release 14:59:19 #endmeeting