21:00:11 <mriedem> #startmeeting nova 21:00:12 <openstack> Meeting started Thu Feb 16 21:00:11 2017 UTC and is due to finish in 60 minutes. The chair is mriedem. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:13 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:15 <openstack> The meeting name has been set to 'nova' 21:00:19 <dansmith> o/ 21:00:22 <takashin> o/ 21:00:29 <viks> hi 21:00:36 <dtp> o/ 21:00:39 <bauzas> \o 21:00:40 <melwitt> o/ 21:00:42 <efried> o/ 21:00:43 <raj_singh> o/ 21:00:56 <edleafe> \o 21:01:11 <ChrisU> hi 21:01:25 <mriedem> hey everyone, let's get started 21:01:28 <mriedem> #link agenda https://wiki.openstack.org/wiki/Meetings/Nova 21:01:46 <mriedem> #info there will not be a nova team meeting next tuesday 2/23 due to the PTG 21:01:52 <mriedem> oops 21:01:56 <mriedem> s/tuesday/thursday/ 21:01:59 <mriedem> #undo 21:02:00 <openstack> Removing item from minutes: #info there will not be a nova team meeting next tuesday 2/23 due to the PTG 21:02:07 <mriedem> #info there will not be a nova team meeting next Thursday 2/23 due to the PTG 21:02:13 <mriedem> whew 21:02:18 <mriedem> #topic release news 21:02:24 <mriedem> #link Ocata release schedule: https://wiki.openstack.org/wiki/Nova/Ocata_Release_Schedule 21:02:30 <mriedem> #info RC2 is due today Thursday 2/16 21:02:49 <mriedem> #link nova ocata rc2 tag patch https://review.openstack.org/#/c/435092/ 21:03:05 <mriedem> ^ is dependent on 2 changes in the gate, once those merge i'll update the hash and we'll be good to GA 21:03:18 <dansmith> woot 21:03:23 <mriedem> let's not all golf clap at once 21:03:35 <mriedem> #link Ocata retrospective: https://etherpad.openstack.org/p/nova-ocata-retrospective 21:03:45 <mriedem> so like in newton, 21:03:54 <mriedem> we have an etherpad for a retrospective on ocata 21:04:00 <mriedem> we'll go over that at the PTG at some point 21:04:42 <mriedem> any questions on the release? 21:04:49 <mriedem> #topic bugs 21:04:57 <mriedem> no critical bugs 21:05:18 <mriedem> i've been trying to do some triage this week 21:05:23 <bauzas> as well 21:05:27 <mriedem> our numbers are semi normal 21:05:36 <bauzas> nothing really blocker found 21:05:56 <mriedem> yup 21:06:07 <mriedem> #link check queue gate status http://status.openstack.org/elastic-recheck/index.html 21:06:24 <mriedem> the gate has been good this week except for the whole cells v1 job being broken for an entire day thing the other day 21:06:27 <mriedem> that kind of sucked 21:06:55 <mriedem> there isn't much news on 3rd party CI, xenserver ci hit some issues but it sounds like they are working through them 21:07:11 <mriedem> i've secretly been wondering if vmware ci is running anymore, but haven't done any work to check 21:07:14 <mriedem> it's just been super quiet 21:07:19 <bauzas> hence the depends-on ? 21:07:25 <mriedem> bauzas: ? 21:07:30 <bauzas> nvm 21:07:32 <mriedem> ok 21:07:39 <mriedem> anything for bugs? 21:07:49 <mriedem> #topic reminders 21:07:54 <mriedem> #info Move and/or repropose specs for Pike in the nova-specs repo. 21:08:03 <mriedem> those are starting to filter back in, 21:08:12 <mriedem> but i haven't personally been actively reviewing much yet 21:08:15 <mriedem> a few re-proposals is all 21:08:25 <mriedem> i expect spec reviews to pick up after the PTG 21:08:36 <mriedem> and i'll start tracking the burndown rates on those again like in newton 21:08:49 <mriedem> #topic Stable branch status: https://etherpad.openstack.org/p/stable-tracker 21:09:09 <mriedem> #link patch is up for a newton 14.0.4 release https://review.openstack.org/#/c/434928/ 21:09:13 <mriedem> tonyb: ^ 21:09:27 <tonyb> mriedem: \o/ I'll review it today 21:09:28 <mriedem> that's got some goodies in it for placement and cells stuff 21:09:56 <mriedem> stable/mitaka is basically blocked right now 21:10:05 <mriedem> because of http://status.openstack.org//elastic-recheck/index.html#1664793 21:10:19 <mriedem> #info stable/mitaka changes are blocked due to bug 1664793 21:10:20 <openstack> bug 1664793 in OpenStack Compute (nova) "test_stamp_pattern timing out waiting for attached device to show up in guest" [Undecided,Confirmed] https://launchpad.net/bugs/1664793 21:10:29 <mriedem> that's because the ceph job is voting in mitaka 21:10:36 <mriedem> and the ceph job fails at a decent rate due to that test 21:10:54 <mriedem> given mitaka isn't getting much for changes it's less of a priority atm 21:11:08 <mriedem> and fwiw it's not just failing on the ceph job 21:11:43 * tonyb was plannign to look at it during/post PTG 21:11:43 <mriedem> moving on 21:11:59 <mriedem> #topic subteam highlights 21:12:07 <mriedem> there was no cells v2 meeting this week due to conflicts 21:12:19 <mriedem> dansmith is working on content for the PTG, 21:12:26 <dansmith> I added some things 21:12:27 <mriedem> and cells aware support in the API 21:12:53 <mriedem> edleafe: anything for scheduler subteam this week? 21:12:57 <edleafe> Discussed some changes pending for nested resource providers, and how they might make related patches cleaner. 21:13:00 <edleafe> Documenting needed functional tests is progressing 21:13:02 <edleafe> Some pre-PTG discussion, but most is on the etherpad 21:13:03 <edleafe> short meeting :) 21:13:23 <mriedem> ok. the nested resource providers spec probably needs to be reproposed 21:13:25 <mriedem> if not already 21:13:30 <edleafe> yeah 21:13:47 <edleafe> I thought jaypipes said that was done, but I could be wrong 21:13:52 <mriedem> jaypipes is also working on patches for aggregates and shared storage 21:14:07 <mriedem> also, 21:14:27 <mriedem> #link etherpad talking about how to design for shared inventory https://etherpad.openstack.org/p/decision-finding-shared-inventory 21:14:47 <mriedem> tdurakov: anything for the live migration subteam? 21:15:07 <mriedem> might not be around 21:15:15 <mriedem> sdague: were you at the api meeting this week? 21:15:34 <ketonne> Veritas is proposing a scheduler filter blueprint for Pike. https://blueprints.launchpad.net/nova/+spec/veritas-hyperscale-filter 21:15:36 <mriedem> i think the api subteam is mostly prepping for ptg 21:15:47 <mriedem> ketonne: please hold that for open discussion 21:16:15 <mriedem> i don't see anyone around for the sriov/pci subteam, 21:16:22 <mriedem> sfinucan has some topics for the PTG related to SRIOV 21:16:41 <sdague> mriedem: I was at the meeting, it was mostly about policy improvements 21:17:04 <mriedem> ok 21:17:14 <mriedem> lbragstad wants to talk about all of that at the ptg 21:17:16 <mriedem> with the keystone people 21:17:36 <mriedem> at some point we have to clone johnthetubaguy because he's got 3 major efforts going on 21:18:00 <mriedem> notes from gibi on the notifications meeting that didn't happen: 21:18:02 <mriedem> No meeting this week as nobody showed up for the Valentine's day editionĀ :/ 21:18:24 <mriedem> he also mentions bug 1665263 which was in the ML 21:18:24 <openstack> bug 1665263 in OpenStack Compute (nova) ocata "instance.delete notification is missing for unscheduled instance" [High,Confirmed] https://launchpad.net/bugs/1665263 21:18:32 <mriedem> basically, 21:18:37 <mriedem> with the move of instance creation from api to conductor, 21:18:47 <mriedem> we broke when instance.delete notifications happen for an unscheduled instance 21:18:55 <mriedem> since that happens in conductor now 21:19:15 <mriedem> #link ML thread on instance.delete notifications bug http://lists.openstack.org/pipermail/openstack-dev/2017-February/112259.html 21:19:32 <mriedem> i suggested some ideas on where we can do that same thing in conductor, but it starts to feel a lot like whack a mole 21:19:41 <mriedem> as is a lot of the instance/build request delete code between api and conductor 21:19:49 <mriedem> dansmith: you might have some ideas there 21:19:57 <dansmith> yeah I have it on my list to think about 21:20:17 <mriedem> excellente 21:20:21 <mriedem> moving on 21:20:24 <mriedem> #topic PTG planning 21:20:30 <mriedem> #link Pike PTG planning: https://etherpad.openstack.org/p/nova-ptg-pike 21:20:40 <mriedem> i started organizing that this week, 21:20:57 <mriedem> as noted in the ML, if you have a topic but aren't going to be around wed-friday, please note that in the etherpad so we can get to your item when you're around 21:21:34 <mriedem> general things i'm taking away from this is there is going to be a lot more cross-project / workgroup stuff happening on monday and tuesday that affect nova items than were probably previously anticipated 21:21:52 <mriedem> which isn't great since some people aren't coming until wednesday 21:22:15 <mriedem> but, i'm assuming we just work through stuff monday and tuesday on what we can and those that were there can feed that back into the nova sessions wed-friday 21:22:26 <mriedem> things like API capabilities, quotas, etc 21:22:46 <bauzas> how can we know when/where ? 21:22:53 <mriedem> we're also going to have some cross-project slots of time with cinder, keystone, ironic and probably neutron on wed, thurs and/or friday 21:22:57 <mriedem> bauzas: well, 21:23:15 <mriedem> i guess the #openstack-ptg channel is one 21:23:30 <bauzas> um 21:23:34 <mriedem> so if you have your laptop open and are watching it at just the right time, you might see someone broadcast that they are talking about something 21:23:41 <bauzas> okay, that's an answer :) 21:23:46 <mriedem> it's a shitty answer 21:23:47 <mriedem> but yeah i know 21:24:06 <bauzas> at least, we don't need Slack :p 21:24:20 <mriedem> for the cross project sessions with other teams we'll have on wed/thurs/fri i'm working with the other PTLs to try and time box some of those into a schedule 21:24:28 <mriedem> @here yeah 21:24:52 <mriedem> also, 21:25:02 <mriedem> cburgess was crowned the PTG nova dinner party czar last night 21:25:15 <mriedem> #info cburgess is the official Pike PTG nova dinner party czar, all hail his glory 21:25:34 <mriedem> he's in charge of finding a bbq place that seats 20 on thursday night 21:25:41 <bauzas> I did heard of french cut 21:25:50 <bauzas> which amazed me 21:25:50 <mriedem> that's the $100/plate steak place 21:26:01 <mriedem> which is not to be confused with the group place 21:26:03 <mriedem> anyway 21:26:14 <mriedem> anything else for ptg? 21:26:24 <mriedem> #topic stuck reviews 21:26:28 <mriedem> there were none in the agenda 21:26:42 <mriedem> closest thing to stuck anything is probably https://etherpad.openstack.org/p/decision-finding-shared-inventory 21:26:52 <mriedem> #topic open discussion 21:26:58 <mriedem> ok, who has something? 21:27:02 <efried> PowerVM driver update 21:27:13 <mriedem> efried: go 21:27:14 <efried> Pike blueprint is approved, and we're still moving forward apace with the code. 21:27:16 <ketonne> https://blueprints.launchpad.net/nova/+spec/veritas-hyperscale-filter Veritas HyperScale filter blueprint 21:27:23 <efried> Turns out I won't be at the PTG, but thorst_afk and adreznec will be there to pester y'all about getting started on those reviews. We expect to have at least the first (big) change set fully ready and +1ed by the nova-powervm core team in time for the PTG. 21:27:30 <efried> We would like to be aggressive about those reviews early in the cycle, because the #LOC are front-loaded in the earlier change sets. The later change sets will be smaller, thus easier to work as pike velocity increases. 21:27:37 <efried> Our CI is coming along nicely, though still experiencing some growing pains. We're mostly passing when we expect to. Still working towards being perfect like all the other 3rd-party CIs ;-) 21:27:52 <efried> Questions, comments? Else that's it from my end. 21:28:24 <mriedem> efried: the reviews are going to be tough to expect w/o the passing CI 21:28:33 <mriedem> one a 1K+ LOC patch 21:28:35 <mriedem> just fyi 21:28:44 <efried> Yuh, we're (mostly) passing now. 21:28:51 <mriedem> ok 21:29:02 <mriedem> ketonne: ok i see there is a spec https://review.openstack.org/#/c/433855/ 21:29:13 <ketonne> Questions on Veritas HyperScale filter? There are no vendor specifc filters AFAIK 21:29:14 <mriedem> ketonne: as noted earlier, we're mostly not reviewing specs yet, but people are free to look 21:29:24 <bauzas> I cc'd it 21:29:41 <mriedem> ketonne: not typically 21:29:42 <bauzas> I briefly overlooked at the spec and I have concerns but I'll detail them in the change 21:29:47 <mriedem> i don't have the 40 filters memorized though 21:30:16 <mriedem> seems like something that would be better served with placement 21:30:19 <ketonne> Also we have https://blueprints.launchpad.net/nova/+spec/veritas-hyperscale-libvirt-volume-driver Veritas HyperScale volume driver 21:30:20 <mriedem> and traits 21:30:52 <bauzas> mriedem: ketonne: at least my two major concerns is that it requires a out-of-tree package, and the second one is that it verifies storage which is a bit difficult atm 21:30:52 <mriedem> ketonne: on the volume driver, that goes into os-brick 21:31:07 <mriedem> ketonne: do you have veritas support in cinder and os-brick yet? 21:31:58 <ketonne> True. There is no precedence for such filter so is there a point to propose this one 21:32:19 <mriedem> ketonne: it will be reviewed, i'd like jaypipes to look at it from a placement service perspective 21:32:33 <mriedem> because it seems this could be served with aggregates and traits in the placement api 21:32:51 <ketonne> Cinder driver is under review. Re-proposed for Pike 21:33:19 <mriedem> ok, the nova side won't go in without a released connector in os-brick 21:33:33 <mriedem> i'll note the same in the bp in launchpad 21:33:40 <mriedem> anyone for anything else in open discussion? 21:33:45 <ketonne> os-brick connector is also proposed for pike 21:34:01 <mriedem> ok link those into the blueprint whiteboard or description 21:34:21 <ketonne> Details are in the same blueprint. 21:34:28 <mriedem> ok i'm not hearing anything for open discussion 21:34:42 <mriedem> i'll see some of you next week, safe travels 21:34:47 <mriedem> #endmeeting