15:00:12 #startmeeting heat 15:00:13 Meeting started Wed Jun 21 15:00:12 2017 UTC and is due to finish in 60 minutes. The chair is ricolin. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:17 The meeting name has been set to 'heat' 15:00:21 #topic roll call 15:00:42 o/ 15:00:47 o/ 15:00:59 o/ 15:01:04 \o/ 15:01:08 o/ 15:01:21 hi 15:01:29 o/ 15:01:56 #topic adding items to agenda 15:01:58 #link https://wiki.openstack.org/wiki/Meetings/HeatAgenda#Agenda_.282017-06-21_1500_UTC.29 15:03:16 #topic weekly report 15:03:44 so python gate since been broken twice 15:03:49 this week 15:04:49 current is cause by tempest 15:05:04 wonder if we should just turn if off for a period of time and wait for it to become stable 15:05:09 yeah the py35 job 15:05:21 Thomas actually got a review for it now 15:05:23 #link https://review.openstack.org/#/c/476138/ 15:05:49 So, shall we disable it :) 15:05:57 but we still don't understand why the py2 job is not broken by the devstack-gate change 15:06:11 no a single clue 15:06:33 ricolin: what do you mean by disable? I think the patch therve has is making it non-voting 15:06:50 ramishra, sorry, non-voting:) 15:06:58 not "disable":) 15:07:13 though it seems like an issue with our way of running the tests and we should fix that 15:07:48 but yeah, we'll make it non-voting to buy some time 15:07:58 actually tempest's py35 job also very unstable if you give it a check 15:09:01 +1ed:) 15:09:04 Hmm.. It's been there for 2/3 months though 15:09:27 you mean for tempest repo? 15:09:58 I've not seen any issue specific to py3 other than neutron OOM issue and the current issue 15:10:06 no I mean the the py35 job 15:10:17 anyway, let's make it non-voting to start 15:10:48 sounds like all agree:) 15:11:07 move on 15:11:10 #topic Add q-trunk to heat job https://review.openstack.org/#/c/473700 15:12:16 it look good to me(not consider chances to fail in future) 15:13:34 please give it a review, and I will +1 on it if it fine to all 15:13:49 Not sure, I don't think enabling the plugin enables the neutron services like q-svc,q-dhcp 15:14:02 so the only-for-master thing is going to break as soon as we branch for stable/pike 15:14:05 I had put a comment in the previous patchset for that 15:14:17 zaneb, yep 15:15:24 yeah, the check should be if [[ ! "stable/newton stable/ocata" =~ $ZUUL_BRANCH ]]; then 15:15:25 ramishra, Lajos has update the patch after your comment 15:16:38 ricolin: L60 is still removed in the latest patch https://review.openstack.org/#/c/473700/3/jenkins/jobs/heat.yaml 15:18:15 zaneb, do we still have to put mitaka on it? 15:18:30 ricolin: opinions differ ;) 15:18:42 okay 15:18:52 as long as the branch exists, I think we should 15:19:17 other folks have said that we "shouldn't" be submitting patches to that branch so it doesn't matter 15:19:47 shouldn't != won't for my money ;) 15:20:51 zaneb, got it! 15:21:03 After a branch is EOLed normally the references to the branch cleaned up in the jobs, it has not happened for mitaka yet I think. 15:22:01 ramishra, will give it a clean by then:) 15:22:34 Lajos around? 15:23:06 okay, move on 15:23:15 #topic Support rolling upgrade https://review.openstack.org/#/c/407989/ (specs) and https://review.openstack.org/#/c/475853/ (doc) 15:23:32 o/ 15:23:33 hi 15:23:37 hi:) 15:23:51 it's yours 15:24:17 after having discussion with ricolin at Boston, we want to continue the works from ramishra in asserting rolling-upgrade tag for Heat 15:24:21 #link https://review.openstack.org/#/c/407989/ 15:24:42 so we're here for requesting your approvals :) 15:24:48 btw, we are from Fujitsu 15:25:23 and we also pushed a patch for upgrading guideline 15:25:26 #link https://review.openstack.org/#/c/475853/ 15:25:39 hieulq_, sure we will be happy to get help 15:26:09 IIRC, all the works seem complete 15:26:20 It was a surprise for me (upated patchsets). I had the impression that we're holding on to it for one more cycle. 15:26:22 hieulq_, maybe you should talk to ramishra more about that plan 15:26:35 yes 15:26:40 Has the boston discussions captured somewhere? 15:27:13 I'll have a look 15:27:13 ramishra, more like a quick discussion at break time:) 15:27:19 after Heat project updates session, we asked ricolin about rolling-upgrade works 15:27:21 yep 15:28:11 ramishra: can you elaborate the reason why we hold on to it for one more cycle? 15:28:24 lack of contributors or something else? 15:29:21 I'm not sure if any db changes landed this cycle after properties_data stuff 15:29:59 I had the impression that we wanted to make some more changes this cycle too. 15:30:48 The approach I had outlined earlier does not take care of DB changes. Though tbh, I've not looked at the new patchsets yet. 15:32:00 thanks, using vhost seems a little bit different with other projects, but it could avoid DB changes that break upgrading 15:32:20 ramishra: isn't the question whether there are changes we want to make in _future_ cycles? 15:32:31 hieulq_, I won't worry about it, since we only got one mile stone left for feature freeze, so will be better if we can got more discussion and plan on this, and make the implement out around PTG time point, and we shall give it a quick land if everything sound like a good way to go 15:32:48 ramishra: i.e. if we think all of the worst DB changes are done in Pike, then we can assert the tag for Pike? 15:33:11 zaneb: +1 15:33:30 ricolin: it seems only documentation works left 15:33:52 we are testing rolling-upgrade Heat in Kolla and VM environment 15:34:11 if we could find any bugs, will report and try to fix them 15:34:12 zaneb: if the changes happend in Pike then we can assert rolling upgrade for Pike->Queens 15:34:55 so the assert would be in Queens I suppose. May be I've forgotten everything abbout it;) 15:35:34 ramishra: right, yeah. I'm not sure which release the tag attaches to (the one you can upgrade from or the one you can upgrade to) 15:36:00 IIRC, it's the one you can upgrade to 15:36:34 hieulq_, you should just start for test and find bug in anytime, and you talking about current shape 15:37:17 I mean if you talking about the current shape 15:37:56 ricolin: yeah, we just set up a Heat cluster in Ocata and try to upgrade it to master 15:38:12 with few stacks 15:40:14 anw, if you core have times, please review the doc patch and spec patch 15:40:53 hieulq_, sure 15:40:54 we will have 2 guys dedicated for this work 15:41:04 hieulq_, nice:) 15:41:48 hieulq_, also Multi-version Interoperability might contain some hard works 15:42:41 hieulq_, anyway, let's keep update on this weekly:) 15:43:01 ricolin: sure, thanks :) 15:43:17 hieulq_, thank you!:) 15:43:22 #topic Open discussion 15:43:35 Anything to discuss? 15:43:53 Nope 15:44:13 LanceHaig, you still here!? 15:44:21 yup 15:44:27 :-) 15:44:42 LanceHaig, have a nice dinner than:) 15:45:04 Thanks 15:45:19 If no other thing to discuss, shall we close meeting:) 15:46:45 That sounds like all agree to me! 15:47:00 Okay, thanks all for join:) 15:47:00 :) 15:47:07 please do more review:) 15:47:20 #endmeeting