15:00:22 <ttx> #startmeeting releaseteam
15:00:22 <amotoki> thanks!
15:00:22 <openstack> Meeting started Fri Dec  7 15:00:22 2018 UTC and is due to finish in 60 minutes.  The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:23 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:25 <openstack> The meeting name has been set to 'releaseteam'
15:00:28 <ttx> ohai!
15:00:29 <Chenjie> guys, thank you!
15:00:39 <dhellmann> o/
15:00:41 <diablo_rojo> o/
15:00:47 <ttx> been a while.... where did smcginnis pack the pinglist
15:00:55 <armstrong> o/
15:01:12 <ttx> fungi, evrardjp maybe?
15:01:32 <fungi> ohai
15:01:34 <dhellmann> yeah
15:01:34 <ttx> Agenda around line 192
15:01:40 <dhellmann> I don't see a list in the etherpad
15:01:50 * fungi was sucked into openstack-discuss
15:02:00 <ttx> sort of a wormhole
15:02:07 <fungi> indeed
15:02:15 <ttx> #topic how to fix karma-subunit-reporter 0.0.4 tag
15:02:40 <ttx> So I've been importing release history for some rarely-released QA deliverables
15:02:43 <dhellmann> in the past we've just ignored malformed tags
15:03:12 <ttx> karma-subunit-reporter has a "v0.0.4"
15:03:34 <ttx> So I should only mention 0.0.3 and be done ? That might result in someone proposing 0.0.4 next and confusion?
15:04:12 * ttx is actually tempted to release a 0,0,5
15:04:23 <fungi> ahh, 2-year-old tag pushed by a human
15:04:33 <ttx> yes, pre-release-repo
15:04:45 <diablo_rojo> how vintage
15:04:52 <ttx> dhellmann: what do you suggest?
15:05:15 <dhellmann> so v0.0.4 is the latest release?
15:05:20 <ttx> yes
15:05:20 <fungi> yes
15:05:31 <fungi> #link https://git.openstack.org/cgit/openstack/karma-subunit-reporter
15:05:35 <dhellmann> I guess we could retag it, then
15:05:58 <fungi> note that our release pipeline regex wouldn't have matched that tag anyway
15:05:59 <dhellmann> we could ignore it, re-tag it, or tag a new version
15:06:15 <fungi> the tag pipeline would have because it fires for any signed tag regardless of content
15:06:19 <dhellmann> maybe using a new version number would be better
15:06:23 <ttx> #link https://review.openstack.org/#/c/622919/
15:06:31 <ttx> dhellmann: yeah, that's what I was thinking
15:06:45 <dhellmann> less confusing than having 2 packages with the "same" version
15:07:34 <ttx> looks like my technique to match tags and commit SHAs is not valid, too
15:08:03 <dhellmann> we have a tool in the releases repo to import existing history
15:08:21 <ttx> I used git show-ref --tag
15:08:55 <fungi> how does that not work?
15:08:56 <ttx> dhellmann: do we? the tools I found did seem to do something else, or not work for _independent
15:08:58 <dhellmann> at least we used to
15:09:03 <dhellmann> tools/build_tag_history.sh looks relevant
15:09:08 <dhellmann> although that's clearly importing older stuff
15:09:15 <dhellmann> ah, tag_history_from_git.py
15:09:53 <ttx> fungi: the validation complains for example that Version 0.1.4 in openstack/devstack-tools is on commit '1384ec992db4ea1692e66427dfa56edfe9a75528' instead of 'f77cebc8dc8994f6a76bb3f46dd0f2ff2d678244'
15:09:59 <dhellmann> you might want to modify that to make it just pull everything into the independent file
15:10:01 <ttx> while show-ref gives:
15:10:08 <ttx> f77cebc8dc8994f6a76bb3f46dd0f2ff2d678244 refs/tags/0.1.4
15:10:15 <fungi> huh, i'm going to dig into that
15:10:23 <fungi> oh!
15:10:27 <dhellmann> a signed tag gets a git object of its own
15:10:28 <fungi> i get it
15:10:36 <dhellmann> unsigned tags dono't
15:10:37 <ttx> dammit git
15:10:52 <dhellmann> this was exciting to learn about when I was writing reno :-)
15:10:54 <ttx> I'm too old for that sh%t :)
15:10:55 <fungi> yes, show-ref <obj> for a tag gives you a hash of the tag, not the commit it wraps
15:11:17 <fungi> i misunderstood what you were looking for there for a moment
15:11:38 <ttx> Anyway, moving on... (still if someone has the magic git command that displays the SHA I want, I'll take it)
15:12:01 <ttx> #topic networking-arista release job errors
15:12:04 <dhellmann> ttx: look in tools/tag_history_from_git.py for the git show command that uses
15:12:21 <ttx> We had a number of job fails this week, I think that's the only one still open though
15:12:43 <dhellmann> arista isn't an official project, and it looks like someone is manually uploading releases and then pushing tags
15:12:44 <ttx> #link http://logs.openstack.org/47/47808fbdc25422e35d3448236bdb4123407f4242/release/release-openstack-python/b3d41d9/
15:12:57 <ttx> dhellmann: so I guess my question is... should we care at all ?
15:13:06 <ttx> And if yes, is there a quick way to tell that we should not care
15:13:11 <dhellmann> we could let them know
15:13:18 <dhellmann> but otherwise I don't think we need to care
15:13:31 <dhellmann> unfortunately we get email about all failures, not just the official repo ones
15:13:56 <dhellmann> so I usually check the officialness first if I don't know off the top of my head
15:14:16 <ttx> agree on the "not caring", I'm just wondering how to quickly tell. I guess it's just a grep away
15:14:44 <dhellmann> yeah
15:15:08 <fungi> other than client-side procmail automation i'm not sure how to go about sorting the wheat from the chaff there
15:15:16 <ttx> ok, that is all I had
15:15:20 <ttx> #topic open discussion
15:15:25 <fungi> or having a separate release pipeline for official openstack projects
15:15:26 <ttx> Anything else ?
15:16:04 <ttx> fungi: I suspect that could be resolved by opendevving
15:16:31 <dhellmann> I've noticed diablo_rojo lurking about the meetings ;-) -- should we be doing some mentoring to train her up as a release manager?
15:16:32 <ttx> if you use a zuul opendev tenant for all things not official?
15:16:45 <ttx> dhellmann: acute eyes
15:17:14 <ttx> dhellmann: we should! Also she might take over the release highlights collection process
15:17:29 <dhellmann> ok, I can sign up for that
15:17:42 <ttx> Although she is a bit sleepy at release meeting time
15:17:59 <ttx> /dcc send diablo_rojo caffeine
15:17:59 <diablo_rojo> :) Yes please :)
15:18:00 <dhellmann> I'm sure we can find a better time to pair up on a few reviews
15:18:12 <fungi> ttx: yeah, separate tenants are likely to happen in the future
15:18:12 <diablo_rojo> ttx, yes please
15:18:30 <dhellmann> diablo_rojo : let's talk offline about your monday/tuesday schedule, since those are my release review days
15:18:38 <diablo_rojo> dhellmann, sounds good
15:19:03 <ttx> I'm on point on Wednesdays, fwiw
15:19:34 <ttx> We use review days so that we can more easily ignore relmgt work on other days
15:19:45 <diablo_rojo> Makes sense to me.
15:19:47 <dhellmann> it'll be good to have someone doing reivews in a later tz, too
15:20:06 <diablo_rojo> Monday/Tuesday are usually lighter days for me so that works.
15:20:12 * ttx looks up base docs to send as preparatory pre-mentoring reading
15:20:14 <diablo_rojo> West coast represent!
15:20:45 <diablo_rojo> ttx, anne may have sent me all that already ;) But send over what you've got in case I missed anything
15:20:59 <ttx> https://releases.openstack.org/#references is probably good reading
15:21:28 <ttx> don't hesitate to flag things that look incoherent, sometimes process moves and docs stay still
15:21:36 <dhellmann> ++
15:22:26 <ttx> Like the process doc is not even mentioning release highlights
15:22:43 <diablo_rojo> Will do!
15:22:50 <ttx> "oops"
15:23:05 <dhellmann> this team is run by a bunch of slackers
15:23:36 <ttx> Yes, can't wait until they are all replaced with capable hands
15:23:50 <ttx> dhellmann: they might see through that ruse though
15:24:15 * diablo_rojo will never forget that the release team was late for their first team photo
15:24:17 <dhellmann> ttx: sssh
15:24:21 <ttx> anyway... anything else to discuss in meeting?
15:24:21 <diablo_rojo> by like...20 min
15:24:35 <dhellmann> wow, gauntlet thrown
15:24:38 <ttx> By definition, the release is never late
15:24:43 <fungi> diablo_rojo: i wonder if you could help get the release team better integrated into the contributor guide?
15:24:56 <ttx> It's everyone else that is too early or too late
15:25:26 <fungi> you'd be well-placed to gauge what's missing from their contributor onboarding as you're going through it, after all
15:25:39 <diablo_rojo> fungi, yeah something I can look into once I know more of the things :)
15:25:50 <diablo_rojo> Definitely can do that.
15:26:18 * fungi apologizes if he seems to be volunteering you for things, given how busy you already are
15:26:24 <diablo_rojo> Though hopefully smcginnis and tonyb have already done most of that being involved with OUI
15:26:33 <fungi> great point
15:26:35 <diablo_rojo> fungi, I do the same to you lol
15:26:46 <fungi> touché!
15:27:10 <ttx> Alright I think we can close then
15:27:21 <fungi> thanks for chairing, ttx!
15:27:28 <diablo_rojo> thanks ttx!
15:27:29 <ttx> Have a great end of week!
15:27:33 <ttx> #endmeeting