00:01:22 <thinrichs> #startmeeting CongressTeamMeeting 00:01:22 <aimeeu> hello! 00:01:23 <openstack> Meeting started Thu Sep 22 00:01:22 2016 UTC and is due to finish in 60 minutes. The chair is thinrichs. Information about MeetBot at http://wiki.debian.org/MeetBot. 00:01:24 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 00:01:26 <openstack> The meeting name has been set to 'congressteammeeting' 00:01:55 <thinrichs> Agenda for today… 00:02:01 <thinrichs> 1. Newton release 00:02:24 <thinrichs> 2. Other status 00:02:31 <thinrichs> 3. Barcelona 00:02:47 <thinrichs> Anything else? 00:03:40 <thinrichs> #topic Newton 00:04:23 <thinrichs> What's our status on Newton today? I saw ekcs's email and was trying to reproduce. Are there other issues that we think are release blockers? 00:05:27 <ekcs> No other ‘release blockers’ iam aware of. 00:05:58 <ekcs> That are not already in progress. 00:06:10 <thinrichs> I was trying to replicate the nova problem right before the meeting and with the latest version of congress am seeing keystone and ceilometer errors when polling. 00:06:40 <ekcs> keystone error is known. Finally made progress on it today: https://bugs.launchpad.net/congress/+bug/1597555 00:06:42 <openstack> Launchpad bug 1597555 in congress "keystone_driver fails to poll user data" [High,New] - Assigned to Aimee Ukasick (aimeeu) 00:06:42 <thinrichs> Could be I have an old version of ceilometer/keystone. Does unstack/restack always pull the latest versions? 00:06:55 <ekcs> yes on latest version. 00:07:33 <ramineni_> ekcs: nova flavor list is failing in the mail you mentioned? 00:07:46 <ekcs> I’ll try to reproduce ceilometer issue. 00:08:07 <ekcs> ramineni_: yes. in fact all nova seems to be failing even when accessed from horizon. 00:08:08 <thinrichs> Running that now then. Hopefully I can report back on ceilometer by end of meeting. 00:08:28 <thinrichs> ramineni_: how about from the CLI? 00:08:47 <ekcs> thinrichs: haven’t tried. can do it right now. 00:09:18 <thinrichs> CLIs were supposed to be frozen weeks ago, so if it's a server regression it should break the CLI too 00:09:39 <ramineni_> hmm, strange .. we have tempest tests for checking that .. but they are passing right 00:09:51 <ramineni_> https://github.com/openstack/congress/blob/master/congress_tempest_tests/tests/scenario/congress_datasources/test_nova.py 00:10:36 <thinrichs> I didn't see a nova error in my setup; could be old. 00:10:41 <thinrichs> Are either of your Nova's old? 00:10:54 <thinrichs> The gate is pulling the latest from master 00:10:57 <ekcs> fails with CLI too. it may be my env or devstack problem. 00:11:06 <ekcs> $ nova flavor-list 00:11:06 <ekcs> ERROR (ConnectFailure): Unable to establish connection to http://192.168.218.145:8774/v2.1 00:11:15 <thinrichs> Other services working? 00:11:22 <ekcs> some kind of nova-devstack mismatch. 00:11:23 <ramineni_> is nova service running without any errrors 00:11:25 <ekcs> yea others working. 00:11:31 <ramineni_> may be its not started properly 00:12:04 <ramineni_> ekcs: you can try restarting the service and check agai 00:12:05 <ekcs> lol. yea nova api failed to start. oops. 00:12:25 <ramineni_> :) 00:12:35 <thinrichs> ekcs: that's an easy fix, thankfully :) 00:12:57 <thinrichs> ramineni_: are you pushing your devstack patch to default to 1 process for Newton? 00:13:12 <ekcs> yea. at least it’s not ours to fix. nova fails to start in standard devstack install. anyway. 00:13:30 <ramineni_> thinrichs: actually, planning to do that .. but somehow jenkins seems to fail 00:13:45 <ramineni_> but devstack passes fine when i use that patch in my env 00:13:52 <ramineni_> still figuring out the issue 00:14:08 <thinrichs> That's a weird error—says no hosts available. Didn't even run the tests. 00:14:24 <ramineni_> ya.. congress itself not starting 00:14:46 <ramineni_> bt i couldn see any specific error 00:15:31 <thinrichs> Seems there may be a couple of other patches we're hoping to land in Newton. Is that right? 00:15:39 <thinrichs> ekcs: your Engine patch 00:16:01 <ekcs> right. 00:16:05 <thinrichs> ramineni_: your congress docs 00:16:11 <ramineni_> yes 00:16:23 <ekcs> and ideally we can also solve the stacktrace issue. 00:16:27 <thinrichs> Maybe a couple of others 00:16:36 <thinrichs> ekcs: right. That would be nice to get in. 00:16:52 <thinrichs> So that's about 6 patches. 00:17:20 <ekcs> not as high priority but I’d also like to add docs in haproxy config 00:17:26 <thinrichs> Let's try to get those merged by next Wed. That'll give us 2 days to do some sanity checking before releasing the week after. 00:17:41 <thinrichs> doc changes make sense—more clarity is always good 00:18:17 <ekcs> on backporting to newton, does it make sense to wait and do it all at once that way we don’t miss anything? 00:18:27 <thinrichs> (Release dates are Oct 3-7) 00:18:46 <ekcs> just look at rc1 tag and backport everything since (wit some exceptions) 00:19:10 <thinrichs> ekcs: we can do it either way 00:19:12 <ekcs> trouble is merging takes time too so waiting to do all at once may not be best. 00:19:46 <thinrichs> If someone wants to take responsibility to do the merging, that's fine. I'd probably just have everyone backport themselves once the patch merges into master. 00:20:13 <thinrichs> As I'm thinking, it's better to have the person who wrote the patch backport it in case there are conflicts or failures. 00:20:42 <thinrichs> Especially if some patches depend on others 00:20:46 <ramineni_> do we backport everything , only 6 patches mentioned now right 00:21:02 <ramineni_> and anythoer critical ones? 00:21:17 <thinrichs> ramineni_: just the ones that need to go in. We're aiming for stability at this point. 00:21:24 <thinrichs> ramineni_: right —the critical ones 00:21:29 <thinrichs> and the docs 00:21:43 <ramineni_> ya, ok 00:21:45 <ekcs> several other important fixes since RC1. 00:22:02 <ekcs> say this one: https://review.openstack.org/#/c/371648/ 00:22:32 <thinrichs> Definitely looks like we should merge that one. 00:22:32 <ekcs> this one makes sense too because things would not work right in python3 is __ne__ isn’t defined. https://review.openstack.org/#/c/372823/ 00:23:41 <ekcs> and adding __ne__ should be zero to low risk. 00:23:56 <thinrichs> That one's less clear. I don't think OpenStack is officially supporting python3 in prod yet. But the release team makes the call. And all cores are on the release team. 00:24:08 <thinrichs> If there's any quesiton, get someone to +1 it before backporting. 00:24:38 <ekcs> ok. 00:25:15 <thinrichs> Here's the Horizon one... 00:25:16 <thinrichs> https://review.openstack.org/#/c/374455/ 00:25:52 <thinrichs> FYI you can do most of these through the gerrit front-end with the Cherry Pick button 00:26:05 <thinrichs> So in summary... 00:26:18 <thinrichs> merge everything into master first 00:26:26 <thinrichs> then if critical backport to newton 00:26:50 <thinrichs> all patches (except maybe docs) in by next Wed (Sept 28) 00:26:54 <thinrichs> Sept 28 00:27:25 <thinrichs> all patches in by next Friday Sept 30 00:27:30 <thinrichs> Any questions/comments? 00:28:16 <ekcs> great. 00:28:27 <thinrichs> #topic Barcelona 00:28:37 <thinrichs> Who is coming to Barcelona? 00:28:44 <thinrichs> (I am) 00:28:44 <ekcs> me 00:28:58 <ramineni_> me too 00:29:13 <thinrichs> aimeeu: you coming to Barcelona? 00:29:57 <aimeeu> Complicated. My current org can't send me, but I move to a new org on 1 Oct and my new boss might. I will know early next week. 00:30:23 <thinrichs> aimeeu: understood. We're hoping to see you there! Good luck! 00:30:43 <aimeeu> I'm hoping to be there! 00:31:13 <thinrichs> We need to decide what topics we are discussing in our meetups. 00:31:19 <thinrichs> We have 3 small rooms and 1 large room 00:32:02 <thinrichs> Do we have topics people feel we should discuss? 00:33:01 <aimeeu> more use/test cases ? 00:34:01 <thinrichs> aimeeu: I like that. Concrete pain points we can solve now that Congress is HA. 00:34:18 <ekcs> that’s a good idea. 00:34:24 <aimeeu> I've gathered some potential test cases from contacts at work 00:34:48 <thinrichs> aimeeu: Awesome! 00:34:49 <ekcs> usability and ease-of-adoption improvemints is something i think worth thinking about depending on other topics. 00:35:00 <aimeeu> +1 00:35:13 <thinrichs> ekcs: Both of those are really key at this point I'd say +1 00:35:25 <ramineni_> agree 00:35:31 <aimeeu> I will also ask bryan_att if he's got any pressing issues 00:36:26 <thinrichs> aimeeu: masahito might be able to something similar 00:36:33 <thinrichs> aimeeu: know if bryan is heading to the summit? 00:36:41 <aimeeu> yes bryan will be there 00:37:02 <thinrichs> aimeeu: I'd be interested in hearing what use cases they are relying on Congress for 00:37:29 <aimeeu> yes exactly. I'm meeting with Bryan on Friday to discuss Congress. 00:37:49 <thinrichs> What would we think of having a session where we invite Bryan/Masahito to give us a boots-on-the-ground report on Congress 00:38:52 <aimeeu> I think that would be excellent. Congress is important to AT&T's ECOMP - I am trying to get more info on that 00:39:00 <ekcs> that’d be great. 00:39:39 <thinrichs> So it sounds like we have 3 topics… 00:39:43 <thinrichs> 1) use case review 00:39:54 <thinrichs> 2) Bryan/Masahito boots-on-the-ground report 00:40:01 <thinrichs> 3) usability and adoption 00:40:17 <thinrichs> We also have 1 large room to have broader, ideally cross-cutting discussions 00:40:20 <thinrichs> with other teams. 00:40:28 <thinrichs> Thoughts on which other teams we should engage? 00:43:11 <thinrichs> Last cycle we talked with Monasca, Watcher, Vitrage 00:43:32 <thinrichs> We can reach out to them again. We've also had someone interested in Quotas reach out. 00:43:37 <thinrichs> Anyone else? 00:44:04 <ekcs> i’ll think more about it. 00:44:37 <thinrichs> It's about a month away, so we have some time, but not a ton since people's schedules get filled up. 00:44:49 <thinrichs> Moving on... 00:44:53 <thinrichs> #topic PTL 00:45:17 <thinrichs> As I'm sure you all noticed, ekcs is taking over as PTL for Ocata. 00:45:33 <thinrichs> So he'll be running the sessions in Barcelona. 00:45:58 <thinrichs> Obviously we'll be working closely together to make the transition a smooth one. 00:46:52 <thinrichs> I'm psyched because I know he's going to do a great job! 00:47:05 <thinrichs> And it's great for the project as a whole to get new folks taking on that role. 00:47:16 <ekcs> thanks thinrichs. I will need plenty of help from everyone :-) 00:47:22 <aimeeu> Thanks for all your hard work thinrichs! 00:48:09 <thinrichs> I've been happy to do it and work with such a stellar team! 00:49:15 <thinrichs> Almost forgot one thing about Barcelona.. 00:49:18 <thinrichs> #topic Barcelona 00:49:31 <thinrichs> We have a tradition of having a Congress Team dinner one night. 00:49:58 <thinrichs> Who wants to try organizing this time? 00:50:39 <thinrichs> ekcs did it last time and will have enough on his plate. 00:50:55 <ekcs> aimeeu may or may not be there. 00:51:05 <thinrichs> ramineni_: would you want to pick a place and get a reservation? 00:52:01 <ramineni_> actually Ill try .. ill get help from you guys ..if im not able to figure out 00:52:48 <thinrichs> ramineni_: I'm happy to help if you need it. ekcs had a good idea of finding 3 different restaurants and having people vote/pick 00:53:21 <thinrichs> Probably someplace that takes reservations. 00:53:30 <ramineni_> thinrichs: ya , will check that 00:53:46 <thinrichs> ramineni_: thanks! 00:53:54 <thinrichs> Just a few minutes left. 00:53:58 <thinrichs> #topic open discussion 00:54:09 <thinrichs> Any status updates we didn't get to? Or other topics to discuss? 00:54:20 <thinrichs> I've had no luck getting multiple devstacks running. 00:54:27 <thinrichs> Crashed my laptop multiple times. 00:54:35 <thinrichs> aimeeu: those instructions you sent out worked great 00:54:48 <aimeeu> glad they helped! 00:54:55 <ekcs> Haven't made more progress on HA testing and tox because I haven't been able to code very much with a big flareup in my nerve issue. I still hope to get it done soon. But if someone wants to add tests, please go for it. The framework is there, not too hard to add additional tests. 00:54:58 <thinrichs> But then things didn't go so well when I tried adding Congress into the 2nd one. 00:55:20 <ekcs> #link https://review.openstack.org/#/c/358927/ 00:56:03 <thinrichs> ekcs: Will try to take a look over the next week. 00:56:16 <ekcs> thinrichs: oh interesting so adding congress crashes your laptop? 00:56:39 <ekcs> thinrichs: otherwise it’s fine? 00:56:50 <ekcs> thinrichs: thanks! 00:59:10 <thinrichs> Out of time for this week. Talk to you all next week! 00:59:16 <thinrichs> #endmeeting