09:01:40 #startmeeting blazar 09:01:41 Meeting started Tue Dec 12 09:01:40 2017 UTC and is due to finish in 60 minutes. The chair is masahito. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:01:42 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:01:44 The meeting name has been set to 'blazar' 09:02:23 #topic RollCall 09:02:36 o/ 09:03:12 o/ 09:03:18 today's agenda is 09:03:27 1. PTG schedule 09:03:34 2. Q-2 release 09:03:36 3. AOB 09:03:41 4. review meeting 09:03:49 anything else? 09:04:01 hiro-kobayashi, priteau: o/ 09:05:07 #topic PTG schedule 09:05:43 The foundation asked me how many days Blazar team will have the meeting in the Dublin PTG. 09:06:37 From my experience, I thought two days works for us. 09:06:44 +1 09:07:07 Is that two full days? 09:07:38 I have already replied them with the plan because of short response due date. 09:07:56 My plan was two full days. 09:08:21 Of course, we have less topics. we can finish early. 09:09:11 Sounds good 09:11:32 According to ptg web site, Wednesday to Friday are for single project meeting. So Blazar's meeting could be Wednesday and Thursday. 09:11:36 https://www.openstack.org/ptg/ 09:12:21 I'll share the specific schedule once I 09:12:37 once I've received the details. 09:12:48 thanks 09:12:48 any comments? 09:14:21 ok, let's move on to next. 09:14:27 #topic Q-2 release 09:15:05 Last week was Q-2 release milestone but not yet released q-2 of blazar. 09:15:25 The release patch is still in review board now. https://review.openstack.org/#/c/526616/ 09:16:06 I hope this will be merged soon after zuul job updates. 09:16:16 any comments? 09:16:55 thanks priteau and masahito 09:18:55 masahito: Should blazar-nova be of type library or service? 09:19:58 priteau: I thought it's a library b/c blazar-nova doesn't have its API. 09:20:42 Does zuul use a different release job for a library project? 09:21:01 You must be right 09:21:02 https://releases.openstack.org/reference/deliverable_types.html#library 09:21:55 Libraries are often pushed to PyPI, but I believe my patch configures blazar-nova to just be a tarball release 09:23:16 meaning 'publish-to-pypi' template is good for the job? 09:23:44 I used publish-to-pypi only for python-blazarclient 09:24:10 Should we make blazar-nova installable via PyPI? 09:25:19 I think only python-blazarclient is enough. 09:25:31 Operators will install nova from tarballs anyway 09:25:43 or from packages, etc. 09:26:58 right. blazar-nova don't need to support PyPI. 09:27:57 But I don't see library can't use release-openstack-server job in the above link. 09:28:26 masahito: I didn't say it couldn't. I think our patches are good. 09:28:29 It's not matter to keep current both patch sets, isn't it? 09:28:48 priteau: oh, right :-) 09:29:41 other comments? 09:30:45 #topic AOB 09:31:24 Before moving on to review meeting, we should have AOB topic. 09:31:36 Does someone have something to share/discuss? 09:32:16 Nothing from me 09:32:36 Oh, I have one topic. 09:33:33 I found instance reservation doesn't work with Cells v2 deployment because anti-affinity check does not work 09:33:53 https://bugs.launchpad.net/blazar/+bug/1737676 09:33:54 Launchpad bug 1737676 in Blazar "Enable anti-affinity check for instance reservation" [Undecided,New] 09:35:25 I've been investigating and thinking of a solution. But no good solution yet :-( 09:36:04 I think this bug should be high-prioritized. 09:36:33 The doc says upcalls don't happen when the cloud is deployed as MULTI cell. 09:36:48 right 09:37:18 Does the bug happen when we deploy single cell? 09:37:57 Perhaps it's specific to my test environment. So I've been investigating root cause. 09:38:21 My test env is single cell. 09:38:59 As you said, it's an high priority bug. I'll also check it. 09:39:15 thanks 09:39:47 anything else? 09:40:35 additional info: my env has single cell but uses super conductor 09:41:58 hiro-kobayashi: thanks the info 09:42:43 okay, we can start review meeting from now. 09:43:20 Just one thing before moving on 09:43:27 The other day I was saying that we were running a new version of Blazar on Chameleon and seeing issues with some leases failing to start. 09:43:43 I believe this may have been caused by a badly merged patch in our fork. 09:44:08 I fixed the issue and will be checking if the problem still happens. 09:45:00 priteau: Nice! 09:45:18 Otherwise Blazar is working well. 09:45:40 The issue always happens at start time? 09:46:09 s/happens/happend/ 09:46:48 Always at start time, but only for some leases. It was because a patch that fixes a race condition had been partially reverted during the merge. 09:48:03 got it. 09:48:52 If you find the root cause is in master, please report it. 09:49:39 I believe it's was only in our fork. 09:51:13 Okay 09:51:58 For the review meeting, I'd like to move the openstack-blazar channel because of limited time of this channel. 09:52:35 okay 09:53:10 No problem 09:53:17 okay, 5 minutes break? 09:53:32 okay. 09:54:07 let's start in 5 mins. 09:55:03 thanks for the weekly meerting. I tell the end of the meeting to the meeting bot. 09:55:11 #endmeeting