17:59:55 #startmeeting trove 17:59:55 Meeting started Wed Nov 16 17:59:55 2016 UTC and is due to finish in 60 minutes. The chair is amrith. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:59:56 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:59:58 The meeting name has been set to 'trove' 18:00:00 o/ 18:00:06 hola 18:00:11 o/ 18:00:15 o/ 18:00:33 SlickNik sends his apologies, won't be able to make it today 18:00:43 hello amrith 18:00:52 courtesy ping for dougshelley66 pmalik vgnbkr mvandijk trevormc aliadil spilla songjian trevormc apsarshaik 18:01:00 hi mariam, peter, matt, ali 18:01:32 not a lot on my agenda for today 18:02:27 #agenda https://wiki.openstack.org/wiki/Meetings/TroveMeeting 18:02:33 let's get started 18:02:46 not sure if anyone else will be here, and this may be a short meeting 18:02:58 #topic Review status 18:03:06 #link https://docs.google.com/spreadsheets/d/1vJxNaoR3VVNS1Cpiz7U--1zyJRZ6ybMxzJoayrjdduo/edit?usp=sharing 18:03:14 o/ 18:03:26 so a small uptick in reviews but as always need to pick up the pace 18:03:27 o/ 18:03:36 the gate has of course been a bit of a drag 18:03:47 but that is now unblocked (for the present) 18:03:52 and we should be able to move forward 18:04:08 as I said at last meeting, there is an ocata deadline tomorrow; more about that later. 18:04:17 anything else about reviews? velocity? status? 18:04:28 Right now, pretty much everything is +1'd or -1'd - nothing to review :-) 18:05:11 vgnbkr, that's good. waiting for some of them to get +2'ed and then merged. 18:05:23 hearing nothing else, let's switch the agenda a bit to 18:05:25 #topic Ocata Release Schedule - Update 18:05:33 #link https://releases.openstack.org/ocata/schedule.html 18:05:39 so there's a deadline tomorrow for ocata 18:05:46 I'm hoping to get a couple of things merged before that 18:06:09 thx johnma for merging a couple of things yesterday and today 18:06:23 what deadline? 18:06:26 nothing for nov 17 18:06:29 amrith: so is the spec deadline this week or next. from the schedule it looks like next week 18:07:16 amrith: I am going to be working on some more reviews today and will try to get them merged today 18:07:24 johnma, mvandijk the deadline is o-1 18:07:31 it is a release that we are required to cut 18:07:41 I'll send you an email about that; from the release team 18:08:31 so we have to tag a release tomororw 18:08:52 #link https://review.openstack.org/397768 18:08:54 #link https://review.openstack.org/397771 18:09:01 ./ 18:09:15 mvandijk, ok? 18:09:20 yes 18:09:41 so the primary thing that I'd wanted to merge was the API/RPC versioning thing 18:09:45 I see that peter +1'ed it 18:09:48 thanks peterstac 18:09:54 johnma, I'd still appreciate a review 18:10:08 if something is broken we can still (must still) try and fix it if possible 18:10:37 johnma, thanks; anything(s) we can merge today would be a plus 18:10:53 WARNING TO ALL: We are running the gate with nova networking as a temporary workaround for some neutron problems. 18:11:06 yes amrith. will do 18:11:09 we'll revert that once we can figure out what's going on with neutron. peterstac is working on that. 18:11:16 stay tuned for updates about that. 18:11:25 that's about all I had re: ocata release schedule. 18:11:34 @amrith I have had success with neutron 18:11:45 in the gate? 18:11:45 let me know if you need help 18:11:46 And ... as mvandijk and johnma said the deadline for specs is next week. 18:11:53 mvandijk, speak with peterstac 18:12:00 he's up in canada I believe :) 18:12:10 peterstac, meet mvandijk 18:12:13 mvandijk, meet peterstac 18:12:18 you folks should chat :) 18:12:45 Anything else on the subject of ocata release schedule ... 18:12:47 mvandijk, running it locally seems to work fine - the issues are with gate setup and what looks like ipv4/v6 issues 18:13:20 ok we can talk later 18:13:32 sure 18:14:36 #topic Specs that are up for review (if any) 18:14:42 moving back up the agenda 18:14:52 #link https://review.openstack.org/#/q/project:openstack/trove-specs+status:open 18:14:59 these are the specs currently in discussion ... 18:15:13 Ignore https://review.openstack.org/391637 18:15:25 that one may not move fast enough for ocata 18:15:36 mvandijk, is https://review.openstack.org/#/c/389760/ still WIP? 18:15:44 yes 18:15:52 would you please WF-1 it. 18:16:14 if y'all have time please review these and provide feedback. 18:16:26 vgnbkr, I see a couple that you've not reviewed recently :) 18:16:45 if there are other specs forthcoming, your time is rapidly running out for Ocata 18:17:17 That's all I had on specs, anyone have anything else to add? 18:18:02 #topic Code Reviews 18:18:06 There has been a proposal on https://review.openstack.org/#/c/388100/ to make the metadata file always present on the guest. I think it's a bad idea. It would certainly simplify everything. We should however all agree. I think this can be discussed offline though,,, 18:18:58 pmalik_, now's as good a time as any other 18:19:39 ideal time actually ... so are you referring to trevormc's suggestion? you can go ahead and discuss it here; he's here ... 18:19:45 or later 18:19:49 y'all decide what you want to do 18:20:07 I think it was vgnbkr 18:20:21 well you can discuss it face to face with him :) 18:20:25 You think it's a bad idea because it would simplify everything? 18:20:26 and let us know what you think 18:20:45 missed not :D 18:20:55 "not a bad idea" 18:22:04 Ah - Too many fingers on keyboard error (for you old Commodore Pet users). 18:22:12 We just need to document really well that the guest-agent expects this file and what properties it requires in it. We can update our elements to have that. 18:22:50 18:23:06 The existing images should work just fine. This would only affect new guests that use those properties. 18:23:23 over 18:23:34 well, it would also affect anyone who updated the guest on existing instances ... 18:23:41 so you think it is a good idea. "not a bad idea" 18:23:59 so if that's the case, update the spec with that proposal and let's re-review it ... 18:23:59 right 18:24:11 upgrade implications are an issue 18:24:21 we'd have to make sure the documentation clearly stated that upgrading the guest would require injecting that new file as well 18:24:35 but that is only certain update paths, right? Our trove upgrade should work just fine. 18:24:36 also the current code assumes that the conf directory will be empty (last I checked). 18:24:46 not sure how things like cloud init will like it if they are not 18:24:48 but worth testing 18:24:53 sure, but not all providers use that route :) 18:25:43 s/sure/right - 'trove upgrade' will work/ 18:25:45 any other code reviews to discuss? 18:25:55 It's not necessary to put it in conf, and it would not be injected. I'm sure oslo_config would be happy to read it from /opt/stack/trove/... 18:26:00 I'd like to see the Region stuff go in 18:26:23 I need to rebase my module stuff to it, but have been holding off in case there were other changes 18:26:34 (plus I think the client stuff has already merged ...) 18:27:03 (nope - client is still sitting there too) 18:27:17 #link https://review.openstack.org/#/c/382493/ 18:27:25 yes peterstac, i am reviewing the multi region one. hopefully will get those in today 18:27:30 #link https://review.openstack.org/#/c/382495/ 18:27:37 johnma, great, thx! 18:28:33 np peterstac. 18:28:45 and I'd like to see the scenario tests run in less than 2h 46m 18:29:03 how long does it take now 18:29:08 why on earth does pxc have a beef with https://review.openstack.org/393478 18:29:13 it has failed consistently 18:29:14 if you pin them to the OSIC cloud they will run in well under 2h 18:29:18 johnma, there is an issue with the clouds 18:29:25 depends on which cloud we get for our CI jobs 18:29:27 like dougshelley66 says 18:29:34 if you are lucky and get osic, it works well 18:29:38 i'm assuming we don't have that level of control 18:29:42 but if you get rax it is a bit slower 18:29:45 oh ok 18:29:46 yes, we don't have that control 18:29:54 but https://review.openstack.org/393478 always ends up on rax 18:29:55 and fails 18:31:02 so peterstac pmalik_ thoughts about splitting scenario tests in half 18:31:25 we could do it pretty easily 18:31:41 I guess the infra guys wouldn't like it if we made cron jobs that just kept resubmitting our jobs 'til they ran on OSIC and passed :-) 18:31:46 have _supported_[12] 18:32:47 vgnbkr, go for it; peterstac the big time sink up front is image build. I'm working to not use the cloudimage but instead use a locally cached cloudimage 18:33:05 that would be a change to a base element that would pull an infra cached copy 18:33:21 that should save 15 or so min on each of those two jobs. 18:33:31 but what if the elements change in the trove repo (since they're there now)? 18:34:05 well, the ubuntu base element is ooo 18:34:05 peterstac i think he is just talking about the base image? 18:34:15 dougshelley66 ++ 18:34:49 ok 18:35:41 #topic Open Discussion 18:35:48 that's called a smooth transition ... 18:36:00 anyone got anything? 18:37:24 ... going once ... 18:38:08 ... going once more ... 18:38:40 Thanks everybody! 18:38:41 #endmeeting