21:00:08 <melwitt> #startmeeting nova 21:00:09 <openstack> Meeting started Thu Oct 25 21:00:08 2018 UTC and is due to finish in 60 minutes. The chair is melwitt. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:10 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:12 <openstack> The meeting name has been set to 'nova' 21:00:15 <melwitt> hi everyone, welcome 21:00:21 <takashin> o/ 21:01:01 <mriedem> o/ 21:01:42 <edleafe> \o 21:01:51 <melwitt> let's make a start 21:01:59 <melwitt> #topic Release News 21:02:05 <melwitt> #link Stein release schedule: https://wiki.openstack.org/wiki/Nova/Stein_Release_Schedule 21:02:08 <efried> o/ 21:02:23 <melwitt> today is milestone 1, so we'll be releasing the first nova beta 21:02:34 <melwitt> #info Today is s-1 and we'll be proposing releases for master and stable branches today. os-traits and python-novaclient have already been released this week. os-vif release has been proposed earlier this week. 21:03:02 <melwitt> stable branches don't have to be today, depending on whether we'll want to land specific things before proposing the release 21:03:12 <melwitt> #link Stein runway etherpad: https://etherpad.openstack.org/p/nova-runways-stein 21:03:18 <melwitt> #link runway #1: https://blueprints.launchpad.net/nova/+spec/reshape-provider-tree (bauzas/naichuan) [END: 2018-11-08] https://review.openstack.org/599208 and https://review.openstack.org/#/c/521041 21:03:29 <melwitt> only have one runway occupied at the moment 21:03:58 <efried> eek, /me should really be paying attention to that. /me curses $work 21:04:11 <efried> btw 21:04:21 <efried> that shouldn't be marked against bp reshape-provider-tree 21:04:36 <melwitt> gibi's nested-resource-allocations recently left a runway but review is still needed, so let's keep paying attention to that. and really we should just put it back in the queue since nothing else is in the queue 21:04:37 <efried> I expressed why in one of the reviews. 21:04:48 <melwitt> oh, hm 21:05:10 <efried> We should totally still review it 21:05:10 <melwitt> where should it go then? is it part of the catch-all for NRP? 21:05:17 <efried> It's a vgpu thing. 21:05:33 <efried> It should be in the same bp as the rest of the xen vgpu stuff. 21:05:45 <melwitt> ok, I'll take a look 21:05:45 <mriedem> well, you could argue that for the libvirt change as well then 21:05:49 <efried> Yes 21:05:52 <mriedem> (1) reshape to put vgpu inventory on child provider, 21:05:57 <mriedem> (2) add support for multiple vgpu types 21:06:10 <mriedem> the former is the reshape bp, the latter is the vgpu-stein bp 21:06:11 <efried> And every other blueprint ever in the future that makes use of reshape to satisfy some upgrade scenario related to some feature. 21:06:24 <mriedem> we only need to reshape vgpu right now 21:06:31 <mriedem> i think that was the target for that bp 21:06:54 <mriedem> stuff like pcpu/vcpu would be part of a different spec imo 21:07:09 <efried> we're not going to pick out some patch in the middle of a bp series in the X release and retroactively mark it against the reshape bp. So we shouldn't do that here either. 21:07:13 <efried> was my reasoning. 21:07:32 <mriedem> the reshape bp is targeted to stein you know 21:07:40 <mriedem> do you consider it done? 21:07:46 <mriedem> if not, what gets to done? 21:07:59 <efried> FFU script, I think is the only remaining item. 21:08:23 <efried> which I suppose in the scope of the bp is just framework 21:08:38 <efried> and then the feature bps add "plugins" or whatever you want to call them 21:08:40 <mriedem> so FFU before any driver actually does a reshape? 21:08:52 <efried> no, I'm not saying that. 21:09:12 <efried> I'm not saying closing the reshaper bp is a prereq to any other bp that's dependent on it. 21:09:27 <efried> This can't be the first time a bp depends on a subset of another bp 21:09:40 <efried> but actually 21:10:03 <efried> yes, come to think of it, we need to have the ffu script before we can declare the deps complete in this case. 21:10:22 <efried> because otherwise we can't ffu to stein with those features in place. 21:10:24 <efried> right? 21:10:51 <mriedem> well, if you ran it before the drivers did their thing, it wouldn't do anything 21:11:07 <mriedem> i guess in my mind the FFU change would come after b/c that's what happened with the ironic flavors stuff, 21:11:12 <efried> Yup, the placement API endpoint also doesn't do anything until you use it. 21:11:12 <mriedem> but that might not be the best example to follow 21:11:25 <mriedem> anyway, this probably doesn't really matter besides paperwork 21:11:36 <efried> was gonna say, we don't need to take up everyone's time. 21:12:03 <melwitt> so, is the current targeted bp fine then? or are we gonna talk about this more later and let someone know to change the target? 21:12:23 <efried> I will bully mriedem into changing the topic back later. 21:12:36 <melwitt> ok, cool. 21:12:45 <efried> All of it is still targeted for stein until/unless something changes. 21:12:58 <efried> so as matt says, just paperwork 21:13:31 <melwitt> I just put nested-allocation-candidates back in a runway. last patch from the previous runway is -W so I strikethrough'd it as a note 21:13:34 <melwitt> ok 21:13:53 <melwitt> anything else for release news or runways? 21:14:14 <melwitt> #topic Bugs (stuck/critical) 21:14:20 <melwitt> no critical bugs in the link 21:14:27 <melwitt> #link 51 new untriaged bugs (up 1 since the last meeting): https://bugs.launchpad.net/nova/+bugs?search=Search&field.status=New 21:14:34 <melwitt> #link 9 untagged untriaged bugs (up 1 since the last meeting): https://bugs.launchpad.net/nova/+bugs?field.tag=-*&field.status%3Alist=NEW 21:14:52 <melwitt> #link bug triage how-to: https://wiki.openstack.org/wiki/Nova/BugTriage#Tags 21:14:59 <melwitt> #help need help with bug triage 21:15:16 <melwitt> Gate status 21:15:34 <melwitt> #link check queue gate status http://status.openstack.org/elastic-recheck/index.html 21:15:49 <melwitt> I think there was a zuul restart earlier in the morning, so things might need rechecking from that 21:15:54 <melwitt> 3rd party CI 21:16:00 <melwitt> #link 3rd party CI status http://ci-watch.tintri.com/project?project=nova&time=7+days 21:16:19 <melwitt> anything else for bugs, gate status, or third party CI? 21:16:34 <melwitt> #topic Reminders 21:16:45 <melwitt> #link Stein Subteam Patches n Bugs: https://etherpad.openstack.org/p/stein-nova-subteam-tracking 21:17:18 <melwitt> I started using the trivial bug section of this etherpad again ^ 21:18:09 <melwitt> #link Create etherpads for Forum sessions: https://wiki.openstack.org/wiki/Forum/Berlin2018 21:18:25 <melwitt> create your etherpads if you haven't already 21:18:35 <melwitt> anything else for reminders? 21:18:55 <melwitt> #topic Stable branch status 21:18:57 * efried ist noch nicht approbiert 21:19:38 <melwitt> #link stable/rocky: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/rocky,n,z 21:19:59 <melwitt> mriedem has a note here that we need to do a stable/rocky release 21:20:02 <melwitt> I agree 21:20:44 <melwitt> as for whether we should wait for https://review.openstack.org/#/q/topic:bug/1798188 I'm biased but I think we should. it would be nice to get the status check and release note out 21:21:16 <melwitt> that one is about extra instruction needed to handle nova-consoleauth during a rolling upgrade from queens => rocky 21:21:17 <mriedem> probably 21:21:23 <mriedem> need another core 21:21:24 <mriedem> to sign up 21:21:26 <melwitt> and a status check to help with communication 21:22:05 <melwitt> any other opinions? 21:22:18 <melwitt> I know there's only a few of us here 21:22:46 <melwitt> I'll try to get another core tomorrow morning when more people are around 21:23:15 <melwitt> ok, moving on I guess 21:23:17 <melwitt> #link stable/queens: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/queens,n,z 21:23:33 <melwitt> lots o backports 21:23:43 <melwitt> #link stable/pike: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/pike,n,z 21:23:55 <melwitt> #link stable/ocata: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/ocata,n,z 21:24:02 <melwitt> oh 21:24:10 <melwitt> I was about to say, I wasn't sure if I should keep including ocata 21:24:23 <mriedem> probably don't need to 21:24:32 <melwitt> ok 21:25:13 <melwitt> #info let me know if there are any other stable branch backports you want to get in before we cut stable releases for s-1 21:25:39 <melwitt> anything else for stable branch status? 21:25:57 <melwitt> #topic Subteam Highlights 21:26:05 <melwitt> efried: scheduler? 21:26:16 <efried> #link n-sch meeting minutes http://eavesdrop.openstack.org/meetings/nova_scheduler/2018/nova_scheduler.2018-10-22-14.00.html 21:26:24 <efried> I promised a spec extracting the file format from 21:26:24 <efried> #link kosamara's device-placement-passthrough spec: https://review.openstack.org/#/c/591037/ 21:26:24 <efried> while also incorporating some of the tenets of 21:26:24 <efried> #link jaypipes's Rocky provider-config-file proposal: https://review.openstack.org/#/c/550244/ 21:26:29 <efried> That has since happened: 21:26:29 <efried> #link Spec: Provider config YAML file https://review.openstack.org/#/c/612497/ 21:26:35 <efried> Discussed nrp series which was until recently based at 21:26:35 <efried> #link previous bottom of nrp-in-nova series https://review.openstack.org/#/c/606050/ 21:26:42 <efried> mriedem had asked for that to be rebased on 21:26:42 <efried> #link removal of caching scheduler https://review.openstack.org/#/c/611723/ 21:26:42 <efried> That has since happened, and both have merged, along with some other cleanup. 21:26:50 <efried> Extraction-wise, cdent proposed the possibility of merging 21:26:51 <efried> #link stub table creator https://review.openstack.org/#/c/600161/ 21:26:51 <efried> temporarily, but we decided to give alembic stuff a chance to mature for a bit. 21:27:00 <efried> Also extraction-related, we talked about putting some focus on docs in the placement repo. We pressured^wconvinced tetsuro to take on some of the editing work while we figure out what's needed to get the docs actually building. He has since proposed a couple of patches in that vein: 21:27:07 <efried> #link De-nova-ify doc/README.rst (merged) https://review.openstack.org/612665 21:27:07 <efried> #link De-nova-ify doc/source/index.rst https://review.openstack.org/613200 21:27:07 <efried> And cdent is getting us moving toward 21:27:07 <efried> #link Making tox -ereleasenotes work https://review.openstack.org/613118 21:27:14 <efried> We agreed to start 21:27:14 <efried> #link a new placement-extract etherpad https://etherpad.openstack.org/p/placement-extract-stein-4 21:27:20 <efried> Concern was expressed about backporting 21:27:21 <efried> #link update less in rt https://bugs.launchpad.net/nova/+bug/1729621 21:27:21 <efried> given that the related 0.0 allocation ratio fix came later. 21:27:21 <openstack> Launchpad bug 1729621 in OpenStack Compute (nova) pike "Inconsistent value for vcpu_used" [Undecided,In progress] - Assigned to Radoslav Gerganov (rgerganov) 21:27:25 <efried> END 21:28:06 <mriedem> i've also sent an email to the tripleo and osa teams today, 21:28:14 <mriedem> needing someone to start working on the upgrade work in those projects 21:28:23 <mriedem> the grenade patch is ready to go 21:28:43 <efried> #link mriedem email soliciting osa/tripleo help for extraction http://lists.openstack.org/pipermail/openstack-dev/2018-October/136075.html 21:28:55 <melwitt> thanks, was looking for the link and failing 21:29:18 <melwitt> on that bug, https://bugs.launchpad.net/nova/+bug/1729621 I don't quite understand the backport question? 21:29:18 <openstack> Launchpad bug 1729621 in OpenStack Compute (nova) pike "Inconsistent value for vcpu_used" [Undecided,In progress] - Assigned to Radoslav Gerganov (rgerganov) 21:30:07 <efried> mriedem: that's you. I have only a vague sense of what's going on here, would need to go refresh my memory on all that mess. 21:30:27 <mriedem> remember when the xenapi CI started posting 0 allocation_ratios 21:30:28 <mriedem> ? 21:30:33 <melwitt> yes 21:30:37 <mriedem> it's related to that 21:30:53 <mriedem> we think that regression was somehow caused by the thing that rgerganov is trying to backport 21:31:03 <melwitt> ok, I see. so for queens and earlier we'd need that fix first 21:31:05 <efried> for reference, here's what mriedem said at the time, including a link to code: http://eavesdrop.openstack.org/meetings/nova_scheduler/2018/nova_scheduler.2018-10-22-14.00.log.html#l-125 21:31:06 <mriedem> so i said i'm nervous about that, and we should at least make sure those changes go together 21:31:34 <melwitt> ok, got it 21:32:13 <melwitt> alright, so we know how to move forward there. cool 21:32:29 <melwitt> gmann left a note for the api subteam "No API office hour this week, Not much to share." 21:32:38 <melwitt> #topic Stuck Reviews 21:32:51 <melwitt> nothing on the agenda. anyone in the room have a stuck review to bring up? 21:33:05 <melwitt> ok 21:33:11 <melwitt> #topic Open discussion 21:33:19 <melwitt> few items on the agenda 21:33:32 <melwitt> first one 21:33:33 <melwitt> (mriedem): Do we need a specless blueprint to add a config option to run nova-metadata-api per-cell? There was agreement to add this at the PTG (L412 https://etherpad.openstack.org/p/nova-ptg-stein ) 21:33:52 <mriedem> i just don't know how anyone wants that tracked, 21:34:14 <mriedem> i think the change is simply add a config option which if true means the meta-api won't look at the API DB for anything, like instance mappings, and assume everything is local to the cell db 21:34:41 <mriedem> you can, and cern already does, run meta-api per-cell but configured to the api db 21:34:53 <mriedem> even though they are just needlessly querying up to the api to find out the instance is in the cell the yare in 21:34:58 <melwitt> ah, ok 21:35:22 <melwitt> I think I'd err on the side of specless bp since it's a new config option 21:35:33 <mriedem> ok 21:35:59 <melwitt> any other opinions? 21:36:00 <efried> surely should have a bp; is the question whether to have a spec? 21:36:12 <mriedem> i don't think it needs a spec 21:36:18 <mriedem> and no that wasn't the questoin 21:36:23 <mriedem> question was is a specless bp needed 21:36:28 <efried> Okay. 21:36:33 <efried> Yes, IMO. 21:36:36 <mriedem> wfm 21:36:37 <efried> For a new config option. 21:37:01 <mriedem> i just forgot about this for awhile b/c it wasn't being tracked, 21:37:06 <mriedem> so i wanted to know how to track it 21:37:18 <efried> looks like you answered your own question. 21:37:26 <mriedem> just looking for validation 21:37:38 <efried> "Gee, this thing isn't tracked, how should we track it? By tracking it? Cool." 21:37:44 <efried> :P 21:37:53 <melwitt> fair enough to ask if it needs a spec though. I think since this is very straightforward, specless is good. and better than not tracking it 21:38:20 <melwitt> this is just saying "don't waste api db calls on something configured to run local that doesn't need api db calls" 21:38:22 <efried> yuh, was going to say: if we can fit all the words in the bp template, and the spec template sections would be largely "None", no spec. 21:38:33 <melwitt> right 21:38:51 <melwitt> ok, next item 21:38:58 <melwitt> (mriedem): Has anyone made a list of the various cross-project Forum sessions that should include nova participation and determined that we'll have a representative? 21:39:12 <melwitt> this is a good question. I think the answer is no 21:39:33 <melwitt> I know that for the change instance ownership session, that's during the nova project update so I can't go to that 21:39:44 <mriedem> i expect dansmith will be at that 21:39:49 <mriedem> since he wrote os-chown 21:39:52 <mriedem> but, 21:39:53 <melwitt> right 21:39:53 <mriedem> idk 21:40:00 <dansmith> I've said I would already yeah 21:40:11 <mriedem> so it would be good to have an etherpad of the obvious xp sessions and make sure we have reps 21:40:23 <melwitt> yeah, so I think last time I made a list of forum sessions of interest to send to the ML. and I should do that again 21:40:45 <melwitt> and yeah, etherpad, I can do that. can probably use our existing forum etherpad https://etherpad.openstack.org/p/nova-forum-stein 21:41:03 <melwitt> and give a heads up to the ML about it after adding cross-project sessions of interest 21:41:11 <efried> ++ 21:41:22 <melwitt> thanks for bringing that up mriedem 21:41:33 <melwitt> ok, last item is from me, a shout out for sundar 21:41:39 <melwitt> (melwitt): The cyborg-nova interaction spec has been updated in response to review comments and is ready for review again: https://review.openstack.org/603955 21:41:58 * efried <== still in the middle of it 21:42:38 <melwitt> yeah, I figured. just an fyi that he's eagerly awaiting everyone's review :) 21:42:59 <melwitt> ok, that's all in the agenda for open discussion. anyone have anything else they'd like to discuss? 21:43:05 <efried> This guy writes good docs. But has a tendency to be... garrulous 21:43:17 * melwitt googles 21:43:30 <melwitt> verbose, yes 21:43:40 <melwitt> :) 21:43:50 <melwitt> ok, if no one has anything else, we can call it a wrap 21:43:55 <melwitt> thanks everyone 21:43:57 <melwitt> #endmeeting