14:00:43 <mriedem> #startmeeting nova
14:00:44 <openstack> Meeting started Thu Oct 19 14:00:43 2017 UTC and is due to finish in 60 minutes.  The chair is mriedem. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:45 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:48 <openstack> The meeting name has been set to 'nova'
14:00:55 <mriedem> hi
14:00:55 <dansmith> o/
14:00:56 <gibi> o/
14:00:57 <takashin> o/
14:00:57 <efried> &/
14:01:02 <alex_xu> o/
14:01:05 <gmann> o/
14:01:09 <edleafe> \o
14:01:12 <tssurya> o/
14:01:42 <mriedem> ok let's get started
14:01:45 <mriedem> #link agenda https://wiki.openstack.org/wiki/Meetings/Nova#Agenda_for_next_meeting
14:01:45 <bauzas> _o
14:01:48 <bauzas> \o even
14:01:54 <mriedem> #topic release news
14:02:01 <mriedem> #link Queens release schedule: https://wiki.openstack.org/wiki/Nova/Queens_Release_Schedule
14:02:08 <mriedem> #info Oct 19: q-1 milestone, nova spec freeze (today)
14:02:13 <cdent> …..\o
14:02:17 <mriedem> i know people are still going through some spec reviews
14:02:40 <mriedem> #info Blueprints: 63 targeted, 42 approved, 1 complete
14:02:51 <mriedem> questions about the release?
14:03:11 <mriedem> #topic bugs
14:03:20 <mriedem> nothing marked critical
14:03:25 <mriedem> #info 16 new untriaged bugs (up 5 from last week)
14:03:37 <mriedem> gate status appears to be ok
14:03:46 <mriedem> zuulv3 issues seem to be under control
14:03:58 <mriedem> 3rd party ci, no news really
14:04:16 <mriedem> for CI there was one thing i wanted to mention,
14:04:25 <mriedem> the grenade live migration job has been broken since we switched over to queens,
14:04:39 <mriedem> because of some ceph setup stuff with singleconductor mode in the grenade jobs,
14:04:44 <mriedem> i thought it would be a relatively easy fix https://review.openstack.org/#/c/508271
14:04:56 <mriedem> it might still be, but i haven't had the time to poke at this ^
14:05:04 <mriedem> so if someone wants to help me out it'd be appreciated
14:05:21 <mriedem> really all that's left is sourcing the localrc that grenade creates so the scripts can get to the devstack variables
14:05:52 <mriedem> the alternative is not running with ceph-backed ephemeral storage for the grenade live migration job, but that is a pity
14:06:11 <mriedem> right now we're just running block storage
14:06:24 <johnthetubaguy> just saw nfs was still commented out :(
14:06:36 <mriedem> #help mriedem is looking for someone to help poke at fixing the grenade live migration job https://review.openstack.org/#/c/508271/
14:06:44 <mriedem> yeah, NFS
14:06:45 <mriedem> psh
14:06:53 <mriedem> moving on
14:07:03 <mriedem> #topic reminders
14:07:11 <mriedem> Forum topics have been reviewed and the summit schedule is available.  Start working on Forum session discussion etherpads and presentations. https://wiki.openstack.org/wiki/Forum/Sydney2017
14:07:37 <mriedem> some of the forum sessions are under my name since i submitted them but i'll probably look to offload some of that to others...
14:08:02 <mriedem> anything to discuss about the forum
14:08:02 <mriedem> ?
14:08:21 <cdent> do we know if jay’s going to be there? otherwise I can probably moderate the placement one if mriedem doesn’t want to.
14:08:27 <mriedem> he will not be there
14:08:44 <dansmith> I'll be there and present for placement
14:09:02 <mriedem> we had talked about doing a call next week or some point before the forum,
14:09:05 <mriedem> to cover some content with jay
14:09:11 <mriedem> will keep you in the look chris
14:09:14 <mriedem> *loop
14:09:16 <cdent> thanks
14:09:17 <bauzas> oh sad for jay
14:09:44 <bauzas> I can help of course
14:09:57 <mriedem> #topic stable branch status
14:09:58 <bauzas> but let's leave cdent or dansmith for that :)
14:10:07 <mriedem> #link stable/pike: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/pike,n,z
14:10:13 <mriedem> #link Etherpad for bugs tracked for Pike backports: https://etherpad.openstack.org/p/nova-pike-bug-fix-backports
14:10:26 <mriedem> there is one more pike thing i wanted to get merged before a release
14:10:39 <mriedem> dansmith: https://review.openstack.org/#/c/513198/
14:10:49 <dansmith> mriedem: was literally just doing it
14:10:53 <mriedem> yay
14:11:02 <mriedem> there is other stuff in pike but it can wait for the 16.0.3
14:11:11 <mriedem> #link stable/ocata: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/ocata,n,z
14:11:23 <mriedem> not much activity there right now as ocata is the middle child of the stable releases
14:11:29 <mriedem> and therefore gets the least amount of attention
14:11:46 <jaypipes> mriedem: happy to do that. still also working on some more articles.
14:11:49 <mriedem> #link stable/newton: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/newton,n,z
14:12:00 <mriedem> newton is flushed
14:12:01 <mriedem> #info Oct 18 : stable/newton branches get tagged EOL: https://review.openstack.org/#/c/512905/
14:12:05 <mriedem> today is newton eol
14:12:09 <mriedem> ^ is the last release request
14:12:09 <bauzas> wow
14:12:21 <bauzas> time flies
14:12:33 <mriedem> #topic subteam highlights
14:12:41 <mriedem> dansmith: no cells meeting but want to mention anything?
14:13:07 <dansmith> not really, just that edleafe's set needs the most attention I think
14:13:15 <dansmith> the buildrequests patch of mine second to that,
14:13:19 <dansmith> I think it just needs one more
14:13:27 <dansmith> but otherwise I think we're in good shape
14:13:48 <mriedem> ok. i said i'd review alternate hosts this week, but stable and spec freeze so...
14:13:52 <mriedem> maybe tomorrow
14:14:10 <mriedem> edleafe: scheduler highlights?
14:14:15 <edleafe> We discussed the lull in the Nested RP series while Jay was on vacation, and whether or not it would be a good thing to have more than one person involved in major efforts like this.
14:14:19 <edleafe> The other patch series all seem to be moving ahead without issue.
14:14:21 <edleafe> We also discussed the possibility of a "recovery mode" for self-healing when the placement data gets severely out of sync with reality.
14:14:24 <edleafe> That's it.
14:14:59 <mriedem> along the lines, i thought the other night about writing something in nova-manage that could give a report of what placement says your allocations are and what we think they should be
14:15:05 <mriedem> to detect if we're leaking anywhere
14:15:10 <mriedem> but...just a thought
14:15:10 <jaypipes> edleafe: have we seen the placement data get severely out of sync with reality?
14:15:30 <edleafe> jaypipes: no, but the use case was that the db gets whacked
14:15:36 <edleafe> or something similar
14:15:50 <mriedem> we said in the meeting if the db gets dropped it's not really on nova
14:16:10 <mriedem> i was more concerned with leaks because of the numerous bugs we've had about cleaning up allocations on failed moves
14:16:24 <cdent> jaypipes: the reason I initially brought it up was in the case that the db got _deliberately_ wacked and state reset was required, using base reality (the compute nodes themselves). It was speculative talk.
14:16:40 <dansmith> jaypipes: we have seen it get out of sync when we have a bug and then need patching up, which I think is likely
14:16:47 <mriedem> oh yeah, which reminds me about a post-script for the ci jobs to dump allocations from placement to see if we're leaking anything, which depends on the osc plugin, which still needs review
14:17:12 <dansmith> jaypipes: I think we do need the ability to fix those holes somehow, but it's going to be pretty intense I think
14:17:28 <mriedem> cdent: and because the RT no longer auto-heals allocations right?
14:17:32 <mriedem> ^ is part of the reason
14:17:41 <jaypipes> dansmith: well, that "use case" is around 85% of the code in the resource_tracker.py file currently ;)
14:17:47 <dansmith> I don't think the whole db getting killed is a good example, personally
14:17:51 <jaypipes> dansmith: so it wouldn't be a new use case...
14:17:56 <jaypipes> but sure, cool with me.
14:18:13 <mriedem> alright, moving on
14:18:28 <mriedem> alex_xu: api meeting highlights?
14:18:36 <alex_xu> Talk about the spec https://review.openstack.org/#/c/508101/ - "Spec for API extensions policy removal", decided that the spec is only about deprecating the policy rule which have non-admin default rule.
14:18:42 <alex_xu> that's it
14:19:19 <mriedem> so i haven't gone through that yet, but i assume it's for deprecating policy rules on things like BDMs in a server create request yeah?
14:19:36 <mriedem> like, we don't actually expect anyone ever to disable being able to do BDMs in a POST /servers request at this point
14:19:42 <gmann> mriedem: for attributes in response.
14:20:00 <mriedem> ok
14:20:05 <alex_xu> yea, for those extended server attributes
14:20:22 <gmann> mainly when API add attributes in Show list detail server response
14:20:23 <mriedem> if i can't get my accessIPs in my server response, i'll be so mad
14:20:50 * mriedem still doesn't know what access IPs are used for
14:20:53 * bauzas needs to disappear for 20 mins
14:21:01 <mriedem> ok moving on
14:21:05 <mriedem> now that bauzas is gone...
14:21:14 <mriedem> gibi: notifications?
14:21:20 <gibi> really short meeting
14:21:28 <gibi> no real progress last week
14:21:35 <gibi> and today bp service-create-destroy-notification
14:21:39 <gibi> has been merged
14:22:02 <gibi> no meeting next week as I will be on vacation on Monday Tuesday
14:22:06 <gibi> that is all
14:22:22 <mriedem> gibi: speaking of that spec, https://review.openstack.org/#/c/444731/ - i don't know if the author still plans on working on it, you might want to reach out in case you're interested in picking it up
14:22:25 <mriedem> you might not be though
14:22:39 <mriedem> could be good for anyone new looking to work on something too
14:22:50 <gibi> mriedem: I will try to catch the author
14:22:56 <mriedem> ok last thing,
14:22:58 <mriedem> cinder updates
14:23:11 <mriedem> ildikov needs to update the multiattach spec,
14:23:25 <mriedem> there was a patch split out of the refresh_connection_info changes https://review.openstack.org/#/c/512626/
14:23:28 <mriedem> johnthetubaguy: ^ should be easy
14:23:36 <mriedem> would +2 but i added the test
14:23:45 <johnthetubaguy> mriedem: will have a peak at that topic again
14:23:54 <mriedem> stvnoyes updated the live migration patch, i ran CI again on that last night and it looks happy, just about to confirm and then will +2
14:24:04 <mriedem> otherwise we're waiting on cinder changes starting here: https://review.openstack.org/#/c/510201/
14:24:25 <mriedem> #topic stuck reviews
14:24:29 <mriedem> nothing on the agenda
14:24:32 <mriedem> anyone have something to mention?
14:24:47 <mriedem> #topic open discussion
14:24:54 <jaypipes> mriedem: we have a priority review list for specs last day today?
14:24:54 <mriedem> nothing on the agenda, so i'll open it up
14:25:01 <mriedem> jaypipes: no
14:25:16 <mriedem> jaypipes: check your email, i sent you the ones i've been watching
14:25:17 <mriedem> but that's it
14:25:31 <jaypipes> mriedem: kk
14:25:40 * jaypipes checks 800+ unread emails
14:25:56 <mriedem> per usual, we're already over committed for the release so not everything is going to make it, and i don't expect spec freeze exceptions
14:26:22 <mriedem> ok anything else?
14:26:32 <mriedem> alright thanks everyone
14:26:34 <mriedem> #endmeeting