16:00:23 <gibi> #startmeeting nova
16:00:24 <openstack> Meeting started Thu Jan 14 16:00:23 2021 UTC and is due to finish in 60 minutes.  The chair is gibi. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:28 <openstack> The meeting name has been set to 'nova'
16:01:45 <gibi> o/
16:02:00 <elod> o/
16:02:12 <dansmith> o/ (but in another meeting)
16:02:17 <gmann> o/
16:02:40 <bauzas> o/
16:02:53 <gibi> #topic Bugs (stuck/critical)
16:02:58 <gibi> No Critical bugs
16:03:02 <gibi> #link 16 new untriaged bugs (-1 since the last meeting): #link https://bugs.launchpad.net/nova/+bugs?search=Search&field.status=New
16:03:19 <gibi> is there any bug we need to discuss?
16:04:34 <gibi> #topic Gate status
16:04:39 <gibi> gate on master is not blocked
16:05:31 <gibi> I see some glance multistore test timeout
16:05:49 <gibi> but I was not able to create a logstash signature for it
16:06:01 <dansmith> I'm looking at the multistore thing now
16:06:07 <gibi> dansmith: thanks!
16:06:10 <dansmith> other than that,
16:06:17 <dansmith> the gate seems rather sick just in terms of backlog
16:06:23 <dansmith> 24 hours to get a result on a patch
16:06:42 <gibi> yes
16:06:42 <gibi> 5:43 < fungi> there was a cinder host outage in rackspace which took down one of the devices which make up the data store for our afs02.dfw server so lots of very large volumes (including package
16:06:47 <gibi> mirrors) are having to be re-replicated from scratch and it's taking a while
16:06:50 <gibi> this is probalby the reason ^^
16:07:15 <fungi> how would that impact it?
16:07:21 <dansmith> ah
16:07:36 <gibi> fungi: I assume it made our job execution slower than usual
16:07:44 <gibi> hence the backlog
16:07:51 <fungi> our package mirrors are a couple days old, tarballs site content is about a day behind, same for docs site...
16:08:04 <fungi> but none of that should be affecting anyone's jobs
16:08:13 <gibi> fungi: sorry, then my mistake
16:08:50 <dansmith> okay well, it seems rather unhealthy at the moment
16:09:13 <dansmith> I dunno if that's due to fails or just load, but I have things I submitted 20h ago that haven't even started
16:09:14 <gibi> yepp, but then I don't know the reason
16:09:23 <dansmith> ack
16:09:37 <gibi> we also saw https://bugs.launchpad.net/nova/+bug/1882521 reappearing on the gate
16:09:39 <openstack> Launchpad bug 1882521 in OpenStack Compute (nova) ussuri "Failing device detachments on Focal" [Undecided,New]
16:09:52 <bauzas> :(
16:10:03 <gibi> so I started working on the libvirt event based detach solution
16:10:15 <gibi> it will take couple of day before it will be ready for review
16:10:26 <gibi> https://review.opendev.org/c/openstack/nova/+/770246
16:11:14 <gibi> any other gate issue we need to talk about?
16:11:54 <fungi> dansmith: there is a significant backlog, seems like a combination of some projects sending changes into the gate which have unstable testing, and also just lots of additional activity the past week (maybe leading up to milestone 2?)
16:12:21 <fungi> and more generally, some projects just run a massive number of jobs these days
16:12:49 <fungi> (not to pick on neutron, but expand a neutron change in the status page some time and take a look at what it runs)
16:13:00 <dansmith> fungi: yeah, getting the scoop from clark now
16:14:30 <gibi> if nothing else about the gate then moving on
16:14:38 <dansmith> +1
16:14:45 <gibi> #topic Runway status
16:14:54 <gibi> #link https://etherpad.opendev.org/p/nova-runways-wallaby
16:15:16 <gibi> we have couple of thing ready to be reviewed
16:15:25 <gibi> Cyborg shelve/unshelve support  https://review.opendev.org/#/c/729563/
16:15:54 <gibi> https://blueprints.launchpad.net/nova/+spec/compact-db-migrations-wallaby (stephenfin) https://review.opendev.org/q/topic:%22bp%252Fcompact-db-migrations-wallaby%22+status:open
16:16:04 <gibi> this has my +1 all over it
16:16:10 <gibi> but obviously we need more eyes
16:16:43 <gibi> https://blueprints.launchpad.net/nova/+spec/support-interface-attach-with-qos-ports (gibi) https://review.opendev.org/q/topic:%22bp%252Fsupport-interface-attach-with-qos-ports%22+status:open
16:17:03 <gibi> stephen has +2s on the qos series so that also needs one more core to look through
16:17:36 <gibi> from     https://blueprints.launchpad.net/nova/+spec/modernize-os-hypervisors-api (stephenfin) only the client side patch remains but that already has got some review feedback
16:17:51 <stephenfin> yup, I'll respin that shortly
16:18:02 <stephenfin> need to address gmann's comments on the policy follow up too
16:18:12 <gmann> yeah policy one too
16:18:14 <gibi> good point there is a policy follow up there
16:18:17 <bauzas> oh, forgot to review this series
16:19:03 <gibi> last but not least the Nova supports password encrypted VNC (brinzhang) https://review.opendev.org/c/openstack/nova/+/622336 also progressing well. I left some feedback on it
16:19:17 <gmann> is this next possible candidate for 2.89 microversion  ?
16:19:26 <gibi> yes
16:19:32 <gmann> and then remove-tenant one. or we just review and see how it goes?
16:19:43 <gibi> brinzhang queued the tenant series after the vnc patch
16:19:51 <gmann> ah i see depends now
16:19:56 <gmann> that is nice.
16:20:06 <gibi> yepp it helps avoiding confusion
16:20:33 <gmann> yeah
16:21:02 <gibi> if you have code ready to review please plug the link into the runway etherpad https://etherpad.opendev.org/p/nova-runways-wallaby
16:21:40 <gibi> moving on
16:21:44 <gibi> #topic Release Planning
16:21:49 <gibi> Milestone 2 is next week, which is spec freeze
16:22:38 <gibi> so if you still have open specs you want to get implemented in W then it is the last chance
16:23:38 <gibi> we spent sizeable effort on https://review.opendev.org/c/openstack/nova-specs/+/750116 so I guess there will be a push there next week before the freeze
16:24:06 <gibi> (Libvirt supports composing cyborg owned vGPU accelerator into domain XML)
16:24:27 <bauzas> my open concerns are still there
16:24:28 <sean-k-mooney> i have 2 specs up for review and artom has 1
16:25:04 <sean-k-mooney> i skimed the vgpu one a few times i can take a look again
16:25:19 <gibi> sean-k-mooney: ack, I will try to spend some time on those new specs tomorrow morning
16:25:20 <sean-k-mooney> i didnt leave review comments before
16:26:14 <gibi> anything else regarding M2 and the incoming release?
16:26:25 <bauzas> I'll loudly explain that we shouldn't think about vGPU features between Cyborg and Nova are competitive
16:26:43 <bauzas> *being* competitve
16:26:58 <sean-k-mooney> bauzas: you mean we should not be deprecating one in favor of the other
16:27:06 <sean-k-mooney> we shoudl offer both
16:27:17 <sean-k-mooney> at least for the next few releases
16:27:21 <gibi> bauzas: and I'm on the side that there is a duplication which means maintenance overhead
16:27:22 <bauzas> we shouldn't even explain in the spec the differences
16:28:21 <sean-k-mooney> gibi: sure but cyborg is not productied so with my downstream had on it will be at least 2 years before it can be used in our poduct
16:28:24 <bauzas> gibi: not if the methods are the same in the libvirt driver
16:28:59 <bauzas> but anyway, will continue to provide my comments
16:29:03 <sean-k-mooney> gibi: currently we dont plan to support it in our next release and its not on the road map for the one after
16:29:18 <bauzas> honestly, how many operators are using Cyborg atm ?
16:29:25 <gibi> sean-k-mooney: I know that
16:29:29 <bauzas> but maybe not asking it there
16:30:00 <bauzas> I'm already loud and that's not the moment we should be discussing about it
16:30:08 <gibi> OK, moving on then
16:30:09 <bauzas> moving on ?
16:30:14 <bauzas> please, yeah
16:30:15 <gibi> #topic Stable Branches
16:30:26 <gibi> copying elod's summary:
16:30:30 <gibi> only stable/rocky remained blocked but fix is on gate - https://review.opendev.org/766492
16:30:35 <gibi> lower-constraints jobs were set to non-voting for Stein and Rocky
16:30:40 <gibi> ~150 open stable patches
16:30:42 <gibi> EOM
16:32:21 <gibi> anything else about stable?
16:32:47 <gibi> #topic Sub/related team Highlights
16:32:51 <gibi> Libvirt (bauzas)
16:33:08 <bauzas> nothing to report, sir.
16:33:15 <gibi> thanks
16:33:20 <gibi> #topic Open discussion
16:33:25 <gibi> we have on item on the agenda
16:33:30 <gibi> (stephenfin) Get consensus on https://review.opendev.org/c/openstack/nova/+/764482
16:33:33 <gibi> I have reworked to replace periods with hyphens, rather than truncate after the first period. Does anyone have concerns with this approach?
16:33:48 <stephenfin> per $subject
16:33:59 <stephenfin> artom and sean-k-mooney have already looked at it and are +1
16:34:43 <stephenfin> I'm just raising this because we discussed it before Xmas and the solution is different to what we agreed, per above
16:34:44 <gibi> I have no objectsion
16:35:38 <bauzas> will review it
16:35:49 <sean-k-mooney> well we agree to ask operators on the ml and they did not have a preference so we just chose one
16:35:57 <sean-k-mooney> and you invited them to comment if they had issues
16:36:07 <stephenfin> yup, exactly
16:36:08 <gibi> I interpret the silence that other are also OK with this as is. But if not the review is still open
16:36:08 <sean-k-mooney> so i think we can just proceed with the review
16:36:15 <bauzas> surely
16:36:18 <stephenfin> gibi: makes sense
16:36:30 <gibi> then it is decided
16:36:31 <gibi> :)
16:36:33 <bauzas> (and stephen removes after the first period, so definitely +1 :D )
16:37:06 <gibi> anything else for today?
16:37:21 <stephenfin> nothing from me
16:37:48 <sean-k-mooney> do we want to have a review day before m2
16:37:55 <sean-k-mooney> or just review as normal for specs
16:38:05 <gibi> I think review as normal, we had two spec review days aroun m1
16:38:13 <gibi> and we have plenty of approved bps
16:38:15 <sean-k-mooney> cool with me
16:39:27 <gibi> then this was all for today
16:39:33 <gibi> thanks for joining
16:39:36 <sean-k-mooney> o/
16:39:37 <gibi> have a nice day
16:39:38 <gibi> o/
16:39:41 <gibi> #endmeeting