17:00:16 #startmeeting nova_cells 17:00:17 Meeting started Wed Aug 17 17:00:16 2016 UTC and is due to finish in 60 minutes. The chair is alaski. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:19 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:21 The meeting name has been set to 'nova_cells' 17:00:25 o7 17:00:33 o/ 17:00:53 I'm planning for this to be just a quick update meeting 17:00:54 o/ 17:00:58 #topic Testing 17:01:10 dansmith: what's the latest on grenade testing? 17:01:25 alaski: the latest is, we were testing really well, 17:01:29 until we realized it broke everything 17:01:33 \o 17:01:38 well, actually, the grenade test is still in there 17:01:45 because grenade only runs against a few things 17:01:54 the devstack one is ready to circle back 17:02:07 so yeah, for those unaware dansmith got grenade testing cellsv2 migrations which broke cellsv1 and ironic 17:02:20 well, no 17:02:25 my grenade support broke ironic 17:02:33 your devstack patch broke cells and ironic :) 17:02:33 oh fun 17:02:52 dansmith: fair :) it just wasn't exposed until the devstack work 17:03:13 handy links for people wanting to get pop-corn ? 17:03:15 alaski: I mean your devstack patch did it.. I don't actually use your devstack stuff in the grenade stuff 17:03:16 for reasons 17:03:23 ahh, that's right 17:04:01 this is the replacement devstack patch: https://review.openstack.org/#/c/355602/ 17:04:12 https://review.openstack.org/#/c/356097/ 17:04:30 there's like 6-7 patches now around this, I'd have to go searching for the links 17:04:41 that one doesn't fix anything broken, 17:04:50 it just makes sure that users get a warning if they're on cellsv1 17:05:03 but my devstack patch works around that itself (kinda has to) 17:05:09 yeah, I've just got it all grouped together in my head 17:05:15 yeah 17:05:18 highly interdependent 17:05:23 cool for links 17:05:26 is there anything outstanding at this point? 17:05:51 no, just need to land what we have and get the gate to let us do that :) 17:05:51 so tbc, the devstack patch was a canary one ? 17:06:01 np 17:06:02 no 17:06:24 355602 is the devstack support for cellsv2 setup, 17:06:32 which replaces an earlier one that we have since reverted 17:06:48 the devstack patch broke cellsv1 because the migration assumes that there are computes, which do not exist in a v1 parent cell 17:07:04 oh I see 17:07:07 yeah and same reason for ironic really 17:07:21 then there was another failure, 17:07:29 yeah. that was a timing issue because the computes weren't up yet 17:07:33 which was because of the way that command worked with multiple nodes 17:07:41 https://review.openstack.org/#/c/355659/ 17:07:45 which landed this morning, 17:07:50 and should be the last thing to unblock ironic 17:07:51 so the solution is to skip the single command in case of cellsv1 ? 17:07:57 bauzas: yes 17:08:01 that's the cells solution, yeah 17:08:01 k 17:08:34 ooooh https://review.openstack.org/#/c/355659/1/nova/cmd/manage.py is terrible 17:08:41 s/terrible/awesome 17:08:47 thanks for handling all of that dansmith 17:09:02 ('terrible' in the french meaning) 17:09:08 #topic Open reviews 17:09:13 because that 1:N relationship is so much fun 17:09:22 there's https://review.openstack.org/#/c/330995/9 for instance groups 17:09:33 and https://review.openstack.org/#/c/263927/62 for build request changes 17:09:43 those two series need reviews 17:09:47 any others? 17:10:12 https://review.openstack.org/#/c/326906/ Move cell message queue switching and add caching 17:10:45 great 17:11:22 #topic Open discussion 17:12:06 at this point in the cycle I'm not sure how much value this meeting has except for pointing out reviews 17:12:14 which can be done in the -nova channel 17:12:30 woot 17:12:34 alaski: are you waiting for more review before updating https://review.openstack.org/#/c/263927/62 ? 17:12:37 is everyone good with going to bi-weekly for now? 17:12:43 +1 to bi-weekly 17:12:47 +1 17:13:01 +1 17:13:02 I'd be fine with octa-weekly, but sure 17:13:18 mriedem: that one's ready IMO. I haven't pinged you because of the novaclient stuff you're dealing with 17:13:34 dansmith: we'll start here and see :) 17:13:41 alaski: ah ok 17:13:41 fine, fine ;D 17:13:58 #info meeting are bi-weekly until it's time to prep for O 17:14:08 if it makes you feel any better, we still have weekly nova/cinder meetings talking about a feature that's dead for newton :) 17:14:13 so there are bigger wastes of time 17:14:22 heh 17:14:33 that's all I've got today 17:14:37 anyone have a thing? 17:14:44 find sdague 17:14:46 he's gone missing 17:14:51 fin 17:15:04 mriedem: reportedly lunching presently 17:15:16 dagnabbit 17:15:35 case closed 17:15:39 thanks all 17:15:41 #endmeeting