14:01:00 <johnthetubaguy> #startmeeting nova 14:01:01 <openstack> Meeting started Thu Aug 6 14:01:00 2015 UTC and is due to finish in 60 minutes. The chair is johnthetubaguy. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:05 <openstack> The meeting name has been set to 'nova' 14:01:06 <ndipanov> hello 14:01:07 <alaski> o/ 14:01:07 <johnthetubaguy> hello everyone 14:01:12 <rgerganov> hi 14:01:13 <alex_xu> \o 14:01:13 <markus_z> o/ 14:01:14 <mriedem> o/ 14:01:14 <raildo> hi 14:01:16 <claudiub> o/ 14:01:16 <abhishekk> O/ 14:01:17 <edleafe> o/ 14:01:17 <rlrossit> o/ 14:01:19 <johnthetubaguy> #link https://wiki.openstack.org/wiki/Meetings/Nova 14:01:20 <ctrath> \o 14:01:22 <lxsli> o/ 14:01:35 <n0ano> o/ 14:01:40 <andreykurilin> o/ 14:01:50 <hshiina> hi 14:01:59 <meghal> hello 14:02:01 <dansmith> o/ 14:02:03 <n0ano> o/ 14:02:06 <atuvenie> o/ 14:02:10 <johnthetubaguy> #topic Release Status 14:02:29 <johnthetubaguy> so we have just passed this: non-priority feature freeze 14:02:32 <johnthetubaguy> #link https://wiki.openstack.org/wiki/Nova/Liberty_Release_Schedule#Non-priority_Feature_Freeze 14:02:46 <johnthetubaguy> #link https://etherpad.openstack.org/p/liberty-nova-non-priority-feature-proposal-freeze 14:02:58 <johnthetubaguy> we reviewed the exception requests 14:03:16 <johnthetubaguy> actually had quite a few blueprints merged over the last week, which was good to see 14:03:23 <johnthetubaguy> but now its time to focus on bug fixes 14:03:36 <johnthetubaguy> and completing the priority efforts 14:04:33 <johnthetubaguy> #link http://lists.openstack.org/pipermail/openstack-dev/2015-August/071487.html 14:04:38 <johnthetubaguy> anyways any questions on all that? 14:05:05 <johnthetubaguy> #topic Bugs 14:05:15 <garyk> is there any chance of getting FFE's or should we just focus on bugs? 14:05:24 <johnthetubaguy> markus_z helped sort out a bug triage day 14:05:39 <markus_z> One critical bug is in progress: https://bugs.launchpad.net/nova/+bug/1481220 14:05:39 <openstack> Launchpad bug 1481220 in OpenStack Compute (nova) "Cannot boot from volume-backed instance snapshot" [Critical,In progress] - Assigned to Feodor Tersin (ftersin) 14:05:41 <uvirtbot> Launchpad bug 1481220 in nova "Cannot boot from volume-backed instance snapshot" [Critical,In progress] https://launchpad.net/bugs/1481220 14:05:42 <uvirtbot> Launchpad bug 1481220 in nova "Cannot boot from volume-backed instance snapshot" [Critical,In progress] 14:05:44 <johnthetubaguy> garyk: there were announced on the ML, see previous link 14:06:32 <garyk> ok 14:06:54 <markus_z> Right, yesterday was "bug triage day": http://lists.openstack.org/pipermail/openstack-dev/2015-August/071338.html 14:07:32 <johnthetubaguy> markus_z: are we lined up for the next bug day? 14:07:49 <johnthetubaguy> August 12: bug review day 14:07:59 <johnthetubaguy> does that work for people? 14:08:00 <markus_z> I'm not quite sure yet, sorry. Will make a query later and post ist to the ML 14:08:13 <johnthetubaguy> markus_z: thats cool, just wanting to advertise here 14:08:27 <johnthetubaguy> #info August 12: bug review day 14:08:48 <johnthetubaguy> #info feature freeze exception results can be seen here: https://etherpad.openstack.org/p/liberty-nova-non-priority-feature-freeze 14:08:58 <johnthetubaguy> markus_z: cool, any more on bugs? 14:09:13 <markus_z> nope, the rest as usual 14:09:17 <johnthetubaguy> mriedem anything about the gate we need to talk about? 14:09:27 <mriedem> yeah 14:09:32 <mriedem> https://bugs.launchpad.net/nova/+bug/1482230 14:09:32 <openstack> Launchpad bug 1482230 in OpenStack Compute (nova) "LibvirtConnTestCase.test_clean_shutdown_first_time causes subunit segfault" [Undecided,New] 14:09:32 <mriedem> is new 14:09:33 <uvirtbot> Launchpad bug 1482230 in nova "LibvirtConnTestCase.test_clean_shutdown_first_time causes subunit segfault" [Undecided,New] 14:09:34 <uvirtbot> Launchpad bug 1482230 in nova "LibvirtConnTestCase.test_clean_shutdown_first_time causes subunit segfault" [Undecided,New] https://launchpad.net/bugs/1482230 14:09:50 <mriedem> 3 hits in check and gate so far, but it's a segfault in libvirt driver tests and i don't see any obvious related changes in the last 24 hours 14:09:58 <edleafe> :30 14:10:03 <edleafe> doh! 14:10:07 <mriedem> so fyi 14:10:27 <johnthetubaguy> yeah, nasty 14:10:36 <mriedem> that's all i know right now 14:10:44 <johnthetubaguy> mriedem: stable branch, anything there worth mentioning? 14:10:53 <mriedem> not for nova 14:11:25 <johnthetubaguy> #topic Regular reminders 14:11:33 <johnthetubaguy> so this is important 14:11:42 <johnthetubaguy> we are trying to scale out our review efforts 14:11:58 <johnthetubaguy> part of that is encourage more folks to review things 14:12:11 <johnthetubaguy> but we are also trying this focus effort using this etherpad: 14:12:19 <johnthetubaguy> #link https://etherpad.openstack.org/p/liberty-nova-priorities-tracking 14:12:25 <johnthetubaguy> lets keep pushing on this 14:12:40 <johnthetubaguy> at the midcycle we spoke about restarting the trivial patch push also 14:12:48 <johnthetubaguy> that is all in the same etherpad 14:13:00 <johnthetubaguy> #link https://etherpad.openstack.org/p/liberty-nova-midcycle 14:13:03 <dansmith> trivial bug fix you mean 14:13:05 <dansmith> right? :) 14:13:11 <johnthetubaguy> dansmith: oops, yes 14:13:41 * johnthetubaguy mutters something about his bad typing skills 14:14:03 <johnthetubaguy> anyways above is the update from the midcycle 14:14:18 <johnthetubaguy> at lot of the time was revisit the actions from the summit: 14:14:20 <johnthetubaguy> #link https://etherpad.openstack.org/p/YVR-nova-liberty-summit-action-items 14:14:29 <johnthetubaguy> anyways, lets move on 14:14:54 <johnthetubaguy> #topic Stuck Reviews 14:15:11 <johnthetubaguy> alaski: I think you are on this review: https://review.openstack.org/#/c/202605/ 14:15:25 <alaski> ahh, yes 14:15:32 <johnthetubaguy> #link https://blueprints.launchpad.net/nova/+spec/inject-nmi 14:15:39 <hshiina> it is mine 14:15:42 <dansmith> that's not stuck 14:15:43 <alaski> I'm not sure it's "stuck", but I did want to discuss it here 14:15:53 <alaski> I added a topic on this to open discussion 14:16:21 <johnthetubaguy> ah, no worries, now is cool 14:16:26 * dims sneaks in late 14:16:51 <dansmith> hshiina: the consensus is use cast() and make sure we have a way to tell if it was not send via the instance actions 14:16:56 <johnthetubaguy> oh, so this is an RPC call, which is surprising to me, but I haven't read about the context 14:16:56 <alaski> at this point I'm against API requests going down to a virt driver before responding 14:17:08 <johnthetubaguy> me too, really 14:17:20 <dansmith> johnthetubaguy: it's a call because I suggested it be a call initially 14:17:27 <alaski> it came up on https://review.openstack.org/#/c/195088/ as well 14:17:31 <johnthetubaguy> we have instance actions to report any failures I guess? 14:17:46 <dansmith> which was a misguided attempt to not punch our users in the face when they're trying to send an NMI to a stuck instance and we just drop it because the virt driver doesn't support it 14:17:50 <dansmith> but instance actions would be much better 14:18:02 <dansmith> as long as we handle the RPC pin case, and the virt-unsupported case 14:18:07 <johnthetubaguy> dansmith: alaski: maybe we should write this down somewhere I guess? 14:18:15 <dansmith> alaski already did 14:18:21 <johnthetubaguy> dansmith: yeah, thats cool 14:18:21 <dansmith> in the review, which I thought was clear 14:18:45 <hshiina> sure, 14:18:48 <johnthetubaguy> OK, thats cool, I was meaning in general, but thats probably silly 14:18:58 <johnthetubaguy> hshiina: do you feel unblocked now? 14:19:01 <alaski> I can update the devref if you like 14:19:08 <hshiina> let me confirm i shoud use cast() 14:19:17 <dansmith> hshiina: yes 14:19:18 <alaski> hshiina: yes 14:19:23 <johnthetubaguy> yes, and handle errors using instance actions 14:19:28 <johnthetubaguy> cool 14:19:28 <hshiina> thanks, i feel unlocked 14:19:36 <hshiina> i will revise my patches 14:19:38 <johnthetubaguy> hshiina: awesome, thanks 14:19:48 <johnthetubaguy> #topic Open Discussion 14:19:48 <hshiina> thank you for you time 14:20:10 <johnthetubaguy> #link cinder spec needs review for nova impact: https://review.openstack.org/#/c/149894/ 14:21:17 <mriedem> that was mine 14:21:18 <johnthetubaguy> there are some review requests 14:21:28 <mriedem> nova people should take a gander at that cinder spec, 14:21:42 <mriedem> cinder wants to get to active/active HA support and part of that is removing locks in the volume manager, 14:21:59 <mriedem> however it sounds like that breaks some things in their API, so they want to change the API and start returning errors for requests that used to be OK 14:22:04 <mriedem> for polling status or something 14:22:12 <johnthetubaguy> ah, its for active/active, I guess thats why the distributed lock manager discussion also came up 14:22:13 <mriedem> i've raised up that that's a bad idea... 14:22:16 <mriedem> yes 14:22:19 <mriedem> the ML threads are all about this 14:22:32 <mriedem> trying to steer them down the path of microversions 14:22:41 <mriedem> and mimic the old lock behavior in the v1/v2 cinder api 14:22:51 <mriedem> so if you want HA with cinder you opt in to v2.1 14:22:54 <mriedem> if that becomes a thing 14:23:20 <mriedem> anyway, it'd be good if it's more than just me saying breaking the API contract is bad 14:23:36 <mriedem> FIN 14:23:37 <johnthetubaguy> yeah, breaking the API contract is bad 14:23:42 <johnthetubaguy> sounds good 14:23:53 <mriedem> oh yeah, 14:24:03 <mriedem> they are talking about a cross project design summit session 14:24:08 <mriedem> for this and some other stuff 14:24:18 <johnthetubaguy> you mean a nova + cinder session? 14:24:20 <mriedem> yes 14:24:31 <scottda> mriedem: I think ask of cinder agrees it Is bad. 14:24:37 <johnthetubaguy> yeah, I think that usually means finding a spot in our track, rather than the cross project one 14:24:44 <scottda> Ask/all 14:25:02 <mriedem> scottda: ok, i was hearing some 'well it's broken so let's just fix the current api, make people move' 14:25:14 <mriedem> anyway 14:25:24 <scottda> No, I don't think we're going there... 14:25:26 <mriedem> i'd push to have specs up before summit so it's productive 14:25:36 <johnthetubaguy> mriedem: +1 14:25:43 <mriedem> there was some other thing about callbacks for extended volumes 14:25:46 <scottda> Yes.I'm signed up for the specs. 14:25:49 <johnthetubaguy> cool 14:25:57 <mriedem> scottda: was extended volumes going to come up i nthe cinder meetup? 14:26:13 <mriedem> we can take this elsewhere 14:26:16 <scottda> No, but we can today or tomorrow. 14:26:20 <scottda> Ok 14:26:23 <johnthetubaguy> so there are some review requests in the agenda, feel free to see those in there 14:26:32 <abhishekk> hi 14:26:36 <johnthetubaguy> abhishekk: you are linking an ML thread you want a response to, what is the question there? 14:26:40 <abhishekk> As per discussion, I have moved this spec in the backlog. 14:26:51 <abhishekk> Requesting alaski for having a final look as I have implemented his suggestions. 14:27:02 <johnthetubaguy> abhishekk: the patch appears to be abandoned, the snapshot one 14:27:14 <abhishekk> https://review.openstack.org/#/c/206160/ ^^ 14:27:23 <abhishekk> yes, that is because of issue in nova 14:27:39 <abhishekk> Eric Harney has implemented it in cinder, but currently marked his patches as WIP due to issue in nova, 14:27:39 <abhishekk> https://review.openstack.org/#/c/149037/ 14:27:48 <abhishekk> This nova changes solves the problem, but as per comment on the patch it's not right way to address this issue. 14:27:49 <alaski> abhishekk: I can take a look, but this should be targetted to M right? 14:28:01 <abhishekk> alaski: yes 14:28:37 <johnthetubaguy> abhishekk: I think you need to talk to the patch author about their patch, and if they plan on working on it again, about this one: https://review.openstack.org/#/c/149037/ 14:28:45 <abhishekk> I guess nikola was about to working for the nova issue 14:28:50 <alaski> abhishekk: okay, you mentioned the backlog, which isn't where it should go now 14:28:54 <johnthetubaguy> ah, OK 14:29:28 <johnthetubaguy> I would reach out to ndipanov about the details I guess 14:29:35 <jordanP> johnthetubaguy, the NFS driver, and glusterfs and quobytes drivers are effected by https://review.openstack.org/#/c/192736/ 14:29:58 <abhishekk> I had discussion with Eric for this and he said due to nova issue its not good to merge it in cinder 14:29:58 <dguryanov> Also vzstorage driver is affected 14:30:00 <jordanP> johnthetubaguy, the patch closes 2 bug at nova's side. I've been rebasing it for a couple of weeks 14:30:17 <johnthetubaguy> jordanP: those bugs are both low priority, is that correct? 14:30:42 <jordanP> johnthetubaguy, one is low the other "no priority" 14:30:57 <jordanP> "undecided" 14:31:25 <johnthetubaguy> jordanP: I guess I mean, what is the correct priority for those, anyways, we can take that offline 14:32:03 <jordanP> okay, but I'd love this have this merged :) 14:32:04 <mriedem> christ, _volume_snapshot_delete is 250 LOC 14:32:16 <jordanP> mriedem, yeah... :) 14:32:58 <johnthetubaguy> oh wow, its time for some refactoring in there... 14:33:03 <johnthetubaguy> anyways, I think we are about done 14:33:09 <johnthetubaguy> any more for any more? 14:33:19 <meghal> johnthetubaguy alaski needed some input on dynamic quota resources spec targeted for M , it helps achieve quota by flavor, AZ or a combination https://review.openstack.org/#/c/206160/ 14:33:33 <meghal> am working on an implementation at my company and wanted to be sure that I am as close to upstream direction…lxsli, melwitt and harlowja have been providing great inputs…but some initial feedback from you all would be greatly appreciated 14:34:05 <johnthetubaguy> meghal: seems like there are comments form alexis that need addressing in there 14:34:12 <alaski> meghal: okay, I will take a look. but it won't be immediately 14:34:19 <meghal> johnthetubaguy: yes…working on those 14:34:37 <ndipanov> abhishekk, ah yeah 14:34:40 <ndipanov> that thing 14:34:40 <johnthetubaguy> meghal: I am not reviewing specs much right now, focusing on code reviews, while we try push hard on getting more bugs merged 14:34:43 <meghal> alaski: sure…its not time critical…but just some feedback on the direction and the use cases it is addressing 14:34:49 <ndipanov> let's talk it through in -nova 14:34:57 <abhishekk> ndipanov: ok 14:35:01 <johnthetubaguy> OK, so I feel like we are done now? 14:35:07 <meghal> johnthetubaguy: sure…I understand 14:35:07 <johnthetubaguy> any more for any more? 14:35:24 <johnthetubaguy> meghal: I hope to get there eventually, just not right now I am afraid 14:35:34 <johnthetubaguy> OK... thanks all 14:35:39 <abhishekk> alaski: thank you 14:35:44 <johnthetubaguy> #endmeeting