14:00:50 <dhellmann> #startmeeting releaseteam
14:00:50 <openstack> Meeting started Fri Jan  8 14:00:50 2016 UTC and is due to finish in 60 minutes.  The chair is dhellmann. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:51 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:53 <openstack> The meeting name has been set to 'releaseteam'
14:01:13 <ttx> added a few things in there
14:01:14 <dhellmann> our agenda is under week R-13 on the mitaka plan etherpad
14:01:17 <dhellmann> #link https://etherpad.openstack.org/p/mitaka-relmgt-plan
14:01:45 <dhellmann> yeah, those are good topics
14:02:00 <dhellmann> I don't think we have any carry-over business from before the break
14:02:15 <dhellmann> #topic priority reviews
14:02:28 <dhellmann> #info reno fix for ordering with merge commits: https://review.openstack.org/#/c/254372/
14:02:30 * ttx reviews
14:02:46 <dhellmann> that needs one more +2, I think, unless there are issues
14:02:55 * ttx should add reno on his dashboards
14:03:15 <dims> i have to star-it for later dhellmann
14:03:30 <dhellmann> ttx: there's a dashboard definition file in the gerrit dashboard creator repo
14:03:31 <dhellmann> dims : ack
14:04:03 <dhellmann> the only thing that dashboard creator is missing for certain is the new repo to create the release schedule from yaml files
14:04:46 <dhellmann> #info https://review.openstack.org/262003 update release liaison duties
14:04:54 <dhellmann> are there any other high-priority reviews?
14:06:03 <dims> +1 to 262003
14:07:01 <ttx> 254372 looks good, I can approve now, unless dims wants to review it as well
14:07:35 <dims> ttx : am ok if you are good, i will try it out later and see if i spot any problems
14:07:47 <ttx> alrighty then
14:07:53 <dhellmann> sounds good, and I'll go ahead and cut a release today
14:08:43 <dhellmann> the liaison duties patch is less urgent, so let's move on
14:08:52 <dhellmann> #topic Review things we should complete before mitaka-2
14:09:02 <dhellmann> #info release-tools wrap up / make sure RC process is not screwed ?
14:09:04 <ttx> I reviewed it already, needs a second ptg-core +2
14:09:23 <dhellmann> ok, I'll poke some of them today
14:09:38 <dhellmann> ttx, what work do you anticipate for the wrap up here?
14:09:41 <ttx> I think that could be done post-m2, but would like to try to get it done before if I can
14:09:58 <ttx> it's mostly me focusing on them for a bit without interruption
14:10:09 <dhellmann> ok
14:10:26 <ttx> also we might want to keep some tooling around until the last kilo stable releases are done
14:10:33 <ttx> i.e. mid-2016
14:10:41 <ttx> we may already have removed tools that they use, actually
14:10:48 <dhellmann> oh, hmm
14:11:14 <dhellmann> launchpad-related scripts?
14:11:45 <ttx> and to push the tags
14:12:06 <dhellmann> ok. I expected that we would be able to use release.sh even for stable tags.
14:12:24 <dhellmann> unless it is leaving out steps we did for past series
14:12:28 <ttx> they use upload_release.py for example
14:12:35 <ttx> see https://wiki.openstack.org/wiki/StableBranchRelease
14:12:37 <dhellmann> ah, right
14:12:49 <dhellmann> ok, so we should probably recover that script
14:12:56 <ttx> I'll look into that
14:13:18 <dhellmann> and set some sort of reminder so we remember to delete it when kilo is done
14:13:29 <dhellmann> although maybe that's not so important
14:13:47 <dhellmann> #info migration to releases.o.o ?
14:14:12 <dhellmann> I haven't started the migration, yet. I need to confer with infra. I'll put that on my list for next week.
14:14:12 <ttx> yeah, was wondering if that would not be something we'd rather complete long before M release
14:14:24 <ttx> so that people get used to the url
14:14:41 <dhellmann> yeah, it would be good to have it set up for the end-of-cycle marketing materials that want to link to release info, too
14:15:11 <dhellmann> did we decide on releases.o.o or release.o.o (singular)?
14:15:33 <dhellmann> plural feels more natural, I think
14:15:46 <dhellmann> that may just be habit at this point though :-)
14:15:49 <ttx> was it tarballs
14:16:07 <ttx> yeah, tarballs.o.o, so releases.o.o I guess
14:16:13 <dhellmann> yeah, makes sense
14:16:19 <dims> +1
14:16:23 <dhellmann> ok, I'll see about starting that next week then
14:16:34 <dhellmann> #info consistency checks on release models ?
14:16:49 <ttx> another thing that makes sense being frozen starting midcycle
14:17:00 <dhellmann> we had a bunch of changes to the governance repo proposed along these lines
14:17:06 <ttx> so we could say that changes are fair game until m-2
14:17:15 <ttx> after that, unlikely to be changed
14:17:28 <ttx> which means I need to complete the consistency checks
14:17:40 <dhellmann> good idea, do you want to send an email to that effect, to follow up on the thread you started about this already?
14:18:08 <ttx> the bunch of changes we have are the result of me directly pinging affected PTLs
14:18:23 <dhellmann> oh, I thought that came from the email thread :-(
14:18:40 <ttx> I'll come up with something
14:19:15 <dhellmann> ok. we should announce the lockdown deadline asap
14:19:44 <dhellmann> is there anything else we need to review?
14:20:05 <ttx> do you think next week release post would be too late ?
14:20:22 <dhellmann> no, early next week should be fine
14:20:34 <dhellmann> I don't want to wait for next week's countdown email, though, since that's thursday
14:20:36 <ttx> we can say must be proposed before m2, not necessarily processed
14:20:41 <dhellmann> right
14:21:06 <ttx> ok, I'll handle that
14:21:13 <dhellmann> great, thanks
14:21:24 <dhellmann> #topic status of automation work
14:22:06 <dhellmann> with the changes that just merged today, we have all of our tools in place and it's possible to release and announce a new dist with a single command manually
14:22:24 <dhellmann> I spoke with the infra team this week at their meeting, and there are 2 things I took away
14:22:45 <dhellmann> first, we're blocked going much further with automation until they have the key management spec implemented
14:22:59 <dims> ok
14:23:02 <dhellmann> that's a priority for them, so I expect progress, but we'll be doing releases by hand for a while longer
14:23:34 <dhellmann> second, because the release tools will run on the secure node, it's likely they're going to ask us to move them into the project-config repo in the jenkins scripts part of the tree
14:23:52 <dhellmann> that will make changing them a little more difficult, since we'll need to get infra to approve
14:24:11 <dhellmann> we'll see if that turns out to be an issue
14:24:54 <dims> interesting
14:25:44 <dhellmann> the release script install the launchpad lib stuff, so I'll have to see how they feel about that
14:25:51 <dims> is this just so they are aware of the changes we are doing?
14:25:54 <dhellmann> I'll also see if we can get +2 on that part of the tree
14:26:14 <ttx> dhellmann: could we publish the tools on Pypi and could they rely on that ?
14:26:16 <dhellmann> well, it came up in the context of security for the code that runs there on the node
14:26:30 <ttx> ah
14:26:40 <dhellmann> ttx: I got the impression they didn't want to have to install things to the nodes, that it might be a case of building the scripts into the image
14:27:02 <dhellmann> we didn't have time in the meeting to get into all of the implications, so I'll raise that with them separately
14:27:20 <ttx> ack
14:27:24 <dhellmann> anyway, just a heads-up on that
14:27:25 <dims> they build images almost daily... so should not be too much
14:27:34 <dims> thanks dhellmann
14:27:36 <dhellmann> true
14:27:49 <dhellmann> ok, moving on
14:27:50 <dhellmann> #topic sorting out constraints issues for stable/liberty
14:28:00 <dhellmann> #link https://etherpad.openstack.org/p/stable-liberty-constraints-sanity
14:28:34 <dhellmann> mriedeman, dansmith, sdague, dims, lifeless, and I talked about the stable library spec this week
14:29:02 <dhellmann> as a consequence, mriedeman, dansmith, and I spent some time yesterday trying to figure out how to roll back the upper-constraints settings for stable/liberty
14:29:08 <ttx> the backward compat one, right ?
14:29:15 <dhellmann> we've turned off the bot that was proposing changes
14:29:17 <dhellmann> ttx: yes
14:29:30 <dhellmann> we're keeping notes in the etherpad ^^ as we experiment
14:30:00 <dhellmann> at this point, we're still seeing if we can tell which versions of things work reliably so we can prepare a proposal for the rollback
14:30:34 <dims> dhellmann : right. i checked with dansmith that o.vo that's in stable/liberty branch works ok now
14:31:15 <dhellmann> dims : yes, I think that's what I remember from yesterday. He did have some concern that we didn't have sufficient test coverage to make him comfortable with rolling all the way back, though, so there's a list of test jobs to add in that etherpad, too
14:31:43 <dims> dhellmann : y, i added a grenade job for o.vo (multi-node)
14:31:53 <dhellmann> mriedeman is driving the work, as part of the stable team, but I thought I'd mention it all here for cross-communication
14:32:00 <dhellmann> dims : excellent, thanks
14:32:04 <dims> good call dhellmann
14:32:41 <dhellmann> ok, that looks like everything from the formal agenda
14:32:41 <dims> i hope we can come up with a release-constraints.txt for stable/liberty and unblock the bot soon
14:32:44 <dhellmann> #topic open discussion
14:33:17 <dhellmann> dims : yes, we'll need both sets of test jobs first, but I agree
14:34:36 <dhellmann> oh, one other topic that came up was the difficulty in reviewing new requirements because of the changes in the constraints file
14:34:39 <dhellmann> #link https://review.openstack.org/#/c/264907/
14:35:09 <dhellmann> I proposed some alternate instructions, but lifeless objected on the grounds that they result in incomplete constraints, which means we don't understand all of the things we're adding to our dependencies
14:35:26 <dhellmann> this came up in the context of the mimic requirement for ironic's functional tests
14:35:47 <dims> dhellmann : if we saw mimic review, we would not have noticed twisted dragged in
14:35:50 <dims> right
14:35:54 <dhellmann> there was an email thread
14:35:55 <dhellmann> #link http://lists.openstack.org/pipermail/openstack-dev/2016-January/083510.html
14:36:30 <dhellmann> dims : right, and when I did look at the constraints file with a dozen changes or so, I had no idea whether they were all needed or if some were just updates generated by the script because there had been a new release
14:36:38 <dims> dhellmann : we could add a job that calculates the u-c changes so reviewers don't have to do it by hand?
14:36:59 <dhellmann> dims : yeah, that's what lifeless and I came up with, but we need someone to write that part of the validation job
14:37:08 <dhellmann> it could be part of an existing job, because we wouldn't want it to fail, just warn
14:37:43 <dhellmann> that way if a new release of some third-level dependency happens it doesn't invalidate the proposed constraints change
14:38:04 <dhellmann> and we'll still get that update when the bot runs the next time
14:38:54 <dhellmann> ok, I think that's everything for this week, then
14:39:06 <dhellmann> have a good weekend, everyone!
14:39:08 <dims> dhellmann : do we have a TODO list somewhere? :)
14:39:25 <dhellmann> dims : we're tracking all of that in the etherpad https://etherpad.openstack.org/p/mitaka-relmgt-plan
14:39:45 <dhellmann> there are lists at the top organized by theme, and we have date-based lists at the bottom in the schedule section
14:40:15 <dims> dhellmann : cool thanks
14:40:45 <dhellmann> ok, I'll give back the 20 minutes remaining then
14:40:48 <dhellmann> #endmeeting