14:01:24 <gibi> #startmeeting nova 14:01:25 <openstack> Meeting started Thu Jan 11 14:01:24 2018 UTC and is due to finish in 60 minutes. The chair is gibi. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:26 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:29 <openstack> The meeting name has been set to 'nova' 14:01:32 <efried> \o 14:01:34 <takashin> o/ 14:01:39 <cdent> o/ 14:01:40 <bauzas> _o 14:01:48 <bauzas> (I'm lazy) 14:01:50 <gmann> o/ 14:02:00 <gibi> I'm not prepared so I'm sorry in advance :) 14:02:01 * efried hands bauzas a crutch 14:02:09 * gmann while eating dinner 14:02:33 * gibi opening the wiki for the agenda 14:02:40 <edleafe> \o 14:02:57 <gibi> let's start 14:03:00 <gibi> #topic Release News 14:03:12 <gibi> #link Queens release schedule: https://wiki.openstack.org/wiki/Nova/Queens_Release_Schedule 14:03:18 <gibi> #link Release schedule recap email: http://lists.openstack.org/pipermail/openstack-dev/2018-January/125953.html 14:03:32 <gibi> we have some upcoming deadlines 14:03:37 <gibi> #info Jan 18: non-client library (oslo, os-vif, os-brick, os-win, os-traits, etc) release freeze 14:03:42 <gibi> #info Jan 25: q-3 milestone, Feature Freeze, final python-novaclient release, requirements freeze, Soft String Freeze 14:04:38 <gibi> I don't think the number of completed blueprints on the agenda is up to date so I will not copy that here 14:04:55 <gibi> anything else about the release? 14:05:33 <gibi> then, moving on 14:05:43 <gibi> #topic Bugs 14:05:52 <gibi> We don't have critical bugs 14:06:17 <gibi> #info 43 new untriaged bugs [data could be outdated] 14:06:44 <gibi> Gate seems to be really slow 14:07:12 <gmann> ya it was ~300 patches in check pipeline today 14:07:24 <gibi> I personally followed a patch in the gate queue that took more than 24h to bounce back 14:07:36 <efried> I've been hammering on a couple of mine for three days now. 14:07:44 <ildikov> gmann: and almost 12 hours waiting time :( 14:08:05 <gmann> yea 14:08:10 <gibi> does somebody has some info from infra about the possible solution? 14:08:20 <bauzas> I have zero idea on the root cause 14:08:24 <bauzas> of the slowness 14:08:29 <bauzas> that's in my pipe for later 14:08:50 <efried> They've been doing... stuff. Like rebooting servers and patching kernels. Nothing has really seemed to help so far. 14:09:09 <gibi> efried: thanks 14:09:15 <gibi> let's move on 14:09:38 <gibi> after a quick glance at http://ci-watch.tintri.com/project?project=nova&time=7+days the third party CI state look OK 14:10:14 <gibi> anything else about bugs or the gate? 14:10:41 <gibi> #topic Reminders 14:10:48 <gibi> #link queens blueprint status https://etherpad.openstack.org/p/nova-queens-blueprint-status 14:11:07 <gibi> and we have an etherpad for the rocky ptg as well #link https://etherpad.openstack.org/p/nova-ptg-rocky 14:11:45 <gibi> #topic Stable branch status 14:11:53 <gibi> #link stable/pike: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/pike,n,z 14:11:56 <gibi> #link Etherpad for bugs tracked for Pike backports: https://etherpad.openstack.org/p/nova-pike-bug-fix-backports 14:12:11 <bauzas> efried: you imply the slowness could be due to any Meltdown patch ? scary. 14:12:41 <gibi> #link stable/ocata: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/ocata,n,z 14:12:44 <gibi> #info Ocata 15.1.0 released 14:12:49 <efried> bauzas The knee-jerk "fix" for meltdown would be to disable the offending CPU optimization, so it's possible. But I sorta doubt it. 14:13:13 <gibi> #link stable/newton: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/newton,n,z 14:13:16 <gibi> #link Newton 14.1.0 release request: https://review.openstack.org/#/c/529102/ 14:13:19 <gibi> Do we still have newton? 14:13:37 <gibi> at least the review list seems empty 14:13:39 <ildikov> bauzas: efried: I think it's more of an outage somewhere after a quick glance on the chats in the infra channel 14:14:03 <bauzas> gibi: Newton is EOL 14:14:09 <gibi> \o/ 14:14:17 <bauzas> well 14:14:21 <bauzas> I think it is 14:14:36 <bauzas> I haven't seen personnally the EOLness 14:14:58 <gibi> me neither 14:15:04 <gibi> anyhow moving on 14:15:08 <gibi> #topic Subteam Highlights 14:15:18 <gibi> Cells v2 14:15:34 <gibi> do we have somebody to talk about cells 14:15:35 <gibi> ? 14:16:08 <efried> ENODAN 14:16:27 <gibi> I see that there was a meeting #link http://eavesdrop.openstack.org/meetings/nova_cells/2018/nova_cells.2018-01-10-20.59.log.html 14:16:55 <gibi> let's move on to scheduler then 14:17:03 <efried> \o 14:17:05 <efried> Alternate hosts had one open patch to address a bug; the patch is now merged (https://review.openstack.org/#/c/531022/). There's now one patch on the series: https://review.openstack.org/#/c/526436/ 14:17:05 <efried> Consuming limits in /allocation_candidates: dansmith has a WIP: https://review.openstack.org/#/c/531517/ 14:17:06 <efried> Nested + allocation_candidates series is making good progress: https://review.openstack.org/#/c/531443/ 14:17:06 <efried> ComputeDriver.update_provider_tree() series is stalled in the gate for days, but otherwise making good progress. 14:17:08 <efried> Granular resource requests is waiting on nested + allocation_candidates (though considering starting to write tests now). 14:17:18 <efried> As of Monday, we're still targeting NRP for Queens, including Xen as the first real consumer; but we do not expect to get to e.g. NUMA modeling or generic PCI stuff yet. 14:17:18 <efried> Placement bug queue is getting a bit deep; a scrub would be helpful. 14:17:26 <efried> Design discussion around conflict detection/resolution revealed that we want report client to raise up conflict exceptions rather than driving any retries itself. The caller (resource tracker) should drive the retries. This will eventually prompt a refactor of how inventory update is handled. 14:17:26 <efried> It was also revealed that we're not tracking generations against RP aggregate associations. This needs to be fixed; cdent volunteered. 14:17:26 <efried> As a matter of priorities, since this is (mostly) a shared RP thing, we could do it in R. (But note that yesterday sean-k-mooney came up with a legit use case for aggregates that's not about shared RPs. http://eavesdrop.openstack.org/irclogs/%23openstack-nova/%23openstack-nova.2018-01-10.log.html#t2018-01-10T14:22:48) 14:17:37 <efried> . 14:17:42 <efried> Questions, comments, concerns? 14:18:08 * edleafe admires efried's status dump 14:18:17 <efried> I learned it from watching you, okay?? 14:18:23 <edleafe> heh 14:18:31 <gibi> efried: thanks, awesome 14:19:04 <gibi> if nothing else on scheduler then moving on to API 14:19:26 <efried> For those of you who didn't get that reference: https://www.youtube.com/watch?v=KUXb7do9C-w 14:19:48 <gibi> it seems there was no API meeting this week 14:19:55 <gmann> yea 14:20:05 <gibi> gmann: anything you want to mention? 14:20:07 <gmann> not mcuh on API side other than ongoing reviews. 14:20:40 <gibi> OK, then Notification 14:20:57 * efried looks around for the notification guy 14:21:02 <gibi> :) 14:21:10 <gibi> we had a short meeting 14:21:22 <gibi> have couple of ongoing bug fixes 14:21:44 <gibi> and we already have notification ehancement bps proposed for Rocky 14:21:57 <gibi> that is all 14:22:05 <gibi> OK moving on to Cinder 14:22:14 <ildikov> o/ 14:22:29 <ildikov> mriedem wrote up a nice summary mail 14:22:32 <ildikov> #link http://lists.openstack.org/pipermail/openstack-dev/2018-January/126127.html 14:22:46 <ildikov> we have the chain in Nova ready for reviews 14:22:57 <ildikov> #link https://review.openstack.org/#/c/271047/ 14:23:19 <ildikov> #link https://review.openstack.org/#/q/topic:bp/multi-attach-volume+(status:open+OR+status:merged) 14:23:43 * gibi already has the multi attach patch open in his browser 14:23:52 <ildikov> and the above link is to see a broader range of patches with Tempest tests and a multi-attach CI job that's experimental 14:24:16 <ildikov> and of course the gate is not in good shape right now, but so far local tests, etc look good 14:24:36 <ildikov> there are some policy changes in flight in Cinder as well to turn the feature on/off 14:24:45 <ildikov> gibi: sounds awesome, thanks! 14:25:04 <gibi> ildikov: anything else you want to mention? 14:25:08 <ildikov> so I'm basically begging for reviews so we can have the first version of the feature supported in Queens 14:25:38 <ildikov> improvement ideas are welcomed in the Rocky PTG etherpad, we already have a section opened there 14:25:47 <ildikov> gibi: and that would be it, thanks :) 14:25:57 <gibi> ildikov: thank you 14:26:03 <gibi> #topic Stuck Reviews 14:26:45 <gibi> we have one on the agenda 14:26:46 <gibi> (tetsuro) Do we need microversion for https://review.openstack.org/#/c/531328/ ? 14:26:49 <tetsuro_> o/ 14:27:15 <tetsuro_> Today's nova /os-hypervisors/details API returns hypervisor_type 'QEMU' for both qemu and kvm backend. This patch changes that value to 'KVM' if kvm is the backend. 14:27:18 <cdent> last I checked the consensus had come round to "yes" 14:27:34 <tetsuro_> Ah, okay. 14:27:55 <gmann> ll check it tomorrow for microversion 14:28:10 <tetsuro_> thanks. 14:28:35 <gibi> cool 14:28:50 <gibi> anything other stuck review we need to discuss here? 14:29:45 <tetsuro_> not anymore from me. 14:29:58 <gibi> #topic Open discussion 14:30:18 <gibi> what is on the agenda was already discussed last week so I guess here the agenda is outdated 14:30:27 <gibi> is there anything to discuss? 14:32:03 <efried> Call it. Thanks for chairing, gibi 14:32:10 <gibi> then, let's close it 14:32:11 <gibi> sorry again for not being an up-to-date chair for this meeting 14:32:17 <gibi> #endmeeting