17:00:58 #startmeeting qa 17:01:02 Meeting started Thu Nov 19 17:00:58 2015 UTC and is due to finish in 60 minutes. The chair is mtreinish. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:03 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:01:06 The meeting name has been set to 'qa' 17:01:19 hi, who's here for today's meeting? 17:01:23 o/ 17:01:28 . 17:01:31 hi 17:01:33 #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Proposed_Agenda_for_November_19th_2015_.281700_UTC.29 17:01:37 ^^^ today's agenda 17:02:00 looks pretty boilerplate today 17:03:04 andreaf, sdague, mkoderer, jordanP, dtroyer, afazekas: around? 17:03:15 mtreinish: what's up? 17:03:26 andreaf: meeting time... 17:03:43 mtreinish: oh, I'm in a different TZ, I got confused, sorry 17:04:31 well let's get start, maybe more people will trickle a bit later 17:04:42 #link Specs Reviews 17:04:50 #topic Specs Reviews 17:05:08 #link https://review.openstack.org/#/q/status:open+project:openstack/qa-specs,n,z 17:05:17 does anyone have any open spec reviews to discuss today? 17:05:32 slowrie: https://review.openstack.org/#/c/247575/ 17:05:54 #link https://review.openstack.org/#/c/247575/ 17:06:09 dwalleck: is that that first of the cli refactor specs we discussed at summit? 17:06:10 We put up one of the first new tempest command line specs. More are on the way 17:06:23 mtreinish: yup 17:07:08 dwalleck: ok, cool I'll take a look. Those specs shouldn't be too involved so it'll hopefully go quick :) 17:07:27 mtreinish: #link https://review.openstack.org/#/c/173334/ 17:07:49 I have to respin it to address oomichi comments, but I'd like to get more eyes on it 17:07:53 mtreinish: Hoping so! I might ping you back about the run one. I'm trying to refresh my memory about what we discussed around testrepository.ui 17:08:43 andreaf: ok, yeah that'll be a good one to get moving soon 17:09:35 dwalleck: sure, np 17:09:39 #link https://review.openstack.org/230183 17:10:03 that one is from SpamapS about some additional instrumentation of gate runs 17:10:30 it LGTM, but we talk about it semi-regularly so I might be too close to fairly evaluate the spec 17:10:30 I like it, interesting data to have 17:11:29 ok, are there any other specs to discuss? 17:12:45 ok, then let's move on 17:12:58 #topic Priority Items 17:13:05 #link https://etherpad.openstack.org/p/mitaka-qa-priorities 17:13:45 so for those who weren't at summit that's the etherpad we're using for tracking at a high level what this cycle's prio items are 17:14:39 we've already closed out one item from the list 17:14:56 \o/ 17:14:57 sdague got multinode grenade voting everywhere 17:15:28 nice :) 17:15:54 does anyone else have anything to discuss on this topic? 17:16:16 we can talk about specific items on the list here, or in the project specific sessions. It doesn't really matter 17:17:21 'Drop extras.d from devstack (sdague) M1' <<-- -we still have 80-tempest.sh in there 17:18:09 I hope we are not going to shoot ouselve in the foot 17:18:17 jordanP: yeah, that's something we have to work to drop. I don't think it's used by any of the gating jobs right now, so I'm not sure what it's for 17:18:47 mtreinish, how could it be not used ? 17:18:55 this is how tempest gets installed no ? 17:19:49 jordanP: sdague has a script that prints very obviously when extras.d is used and is tracking that with logstash 17:20:01 we don't have that on every run, so I assumed it's not used 17:20:33 humm... 17:21:07 mtreinish, "local exceptions="50-ironic.sh 60-ceph.sh 80-tempest.sh" 17:21:10 that's why... 17:21:40 anyway, a fix should be easy and we can react quickly when it's time, i guess 17:21:41 ah ok 17:22:43 jordanP: yeah, I'm not too worried about it 17:23:43 ok, well if there isn't anything else I guess we can move on 17:24:03 #topic Tempest 17:24:12 yeh, so those will be exempted 17:24:26 until we can migrate them out 17:24:26 andreaf, is making good progress on "cred provider" (when he is not breaking the world haha!) 17:25:05 jordanP: oh, you found out about my secret evil plan 17:25:15 jordanP: heh, that happens a lot with the cred provider stuff. 17:25:48 jordanP: I think it took us at least 50 revs to get the accounts providers right :) 17:25:55 jordanP, mtreinish: next step is to get client.py out of the credential providers, and the route I'm taking is to split credentials provisioning and resource provisioning 17:26:00 yeah, I can image 17:26:10 *imagine 17:26:50 andreaf: that's not a small task... 17:27:02 jordanP, mtreinish: but I need to do that in a way that still maintains the link between the creds and their resource 17:27:04 resources 17:27:53 yes I'm trying to find out the best way to do that in small steps 17:28:38 to be fair, we never really thought about external consumption when we wrote all of that 17:29:33 ok, is there anything else on tempest to discuss? 17:30:54 ok, then lets move on 17:30:59 #topic DevStack 17:31:11 does anyone have anything to discuss about devstack this week? 17:31:22 dtroyer, sdague: ^^^ 17:31:26 let's not drop support for ubuntu precise :) 17:31:38 yeh, that 17:31:41 let's drop it 17:31:50 we just merged nova code that won't work there 17:31:58 let's wait for that liberasure package to get into the repo 17:32:34 wait, devstack still supports precise? 17:32:34 sdague, because of libvirt requirement ? 17:32:42 mtreinish: right, exactly 17:32:53 jordanP: because of unconditional flag use on ebtables 17:33:28 sdague, I use devstack to deploy swift and keystone, I don"t need nova 17:33:29 mtreinish: this is the patch in question - https://review.openstack.org/#/c/246973/ 17:33:53 jordanP: then you can force it on there with the override 17:34:03 saying devstack supports precise means it all works there 17:34:40 and means people will file bugs and think we are going to fix it. You just have to run with FORCE=yes 17:35:11 yeah.... just for this ebtables hack :( 17:35:29 no, it's actually not for the hack 17:35:32 it's for the real fix 17:35:43 the hack doesn't care about your ebtables version 17:36:04 jordanP: the hack just calls ebtables with flock, that works on precise :) 17:36:28 anyway, I don't think that upstream devstack should state that it supports precise 17:36:40 because it doesn't in the general case 17:36:41 yeah, ok, I call hack my way around this 17:36:45 *can 17:37:15 jordanP: did you sort out with the swift and ubuntu folks if that package is getting backported? 17:38:13 yeah, there is some good progress, imo. theres a bug report for this which tracks the progress: 17:38:41 ok, I don't want to hold up swift ci if they need this sooner rather than later 17:39:15 so you're on a clock :) 17:39:23 I know 17:40:21 I think other than that, the ebtables hack seems to have helped on the fail rate 17:41:03 though logstash fell over so it's a little hard to tell 17:41:07 sdague: oh, cool. It's good that helped 17:41:09 heh, ok 17:41:18 and now ES is 17 hours behind 17:41:38 sdague: the fail rate graph still looks a little fuzzy: http://status.openstack.org/openstack-health/#/ 17:41:44 we should also land - https://review.openstack.org/#/c/245881/ - to never speak of partail upgrade again 17:41:56 * mtreinish notes we really need to change the graph type to be something more suited to the data 17:42:20 mtreinish: yeh, 25% doesn't mean anything without error bars 17:42:29 https://bugs.launchpad.net/precise-backports/+bug/1517713 (Please backport liberasurecode 1.1.0-2 (main) from xenial) 17:42:29 Launchpad bug 1517713 in trusty-backports "Please backport liberasurecode 1.1.0-2 (main) from xenial" [Undecided,New] 17:43:30 ok is there anything else on devstack? (and grenade) 17:43:53 mtreinish: I don't think so 17:44:02 ok, then let's move on 17:44:06 jordanP: are there any ubuntu people on that bug? 17:44:46 sdague, yes; this bug report was opened as a request by Micah on https://bugs.launchpad.net/precise-backports/+bug/1515710 17:44:46 Launchpad bug 1517713 in trusty-backports "duplicate for #1515710 Please backport liberasurecode 1.1.0-2 (main) from xenial" [Undecided,New] 17:44:51 22 hours ago 17:45:26 #topic OpenStack Health 17:45:38 #link https://etherpad.openstack.org/p/openstack-health-tracking 17:45:55 so the big news on openstack health is: 17:45:57 #link https://review.openstack.org/235009 17:46:23 which is the db migration SpamapS and I were working on that should fix most of the subunit2sql perf issues 17:46:34 woo! 17:46:55 \o/ 17:46:56 we're prepping things to run it next week (and include that in the 1.0 subunit2sql release) 17:47:13 the run time estimates don't look so great, at taking multiple days to finish :( 17:48:16 mtreinish, we could also drop the data and start from scratch 17:48:21 other than that, glauco has been working on some improvements to make the first 2 pages grouping a bit more generic 17:48:49 jordanP: we thought about doing that, but then the dashboard or other tooling isn't really useful for at least a week, while we rebuild the trends 17:49:00 jordanP: that'll be the fallback option if something goes wrong :) 17:49:40 k 17:50:24 ok, is there anything else on openstack health this week? 17:51:42 ok, then lets move on 17:51:47 #topic Critical Reviews 17:52:03 does anyone have a review they'd like to get extra eyes on? 17:53:29 really no reviews :) 17:54:44 ok, if there aren't any reviews I guess let's end here today 17:54:57 thanks everyone 17:55:05 #endmeeting