09:00:42 #startmeeting blazar 09:00:43 Meeting started Tue Jun 20 09:00:42 2017 UTC and is due to finish in 60 minutes. The chair is masahito. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:00:44 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:00:46 The meeting name has been set to 'blazar' 09:01:05 #topic Roll Call 09:01:18 o/ 09:01:24 o/ 09:01:25 hi 09:01:42 #chair hiro-kobayashi 09:01:43 Current chairs: hiro-kobayashi masahito 09:01:46 Hello! 09:01:52 today's agenda is 09:01:58 1. instance reservation 09:02:02 2. status update 09:02:05 3. AOB 09:02:08 anything else? 09:02:54 #topic instance reservation 09:03:37 thanks for reviewing and commenting for the spec. 09:03:39 https://review.openstack.org/#/c/459533/ 09:04:36 Today I pushed the new patch set that applied previous comments. 09:04:55 thanks! I've +1ed 09:05:08 Thanks masahito. I will review the latest version. 09:05:36 we can merge it soon. 09:07:06 today's agenda is less. so we can check and merge it at end of this meeting :-) 09:07:27 +1 09:07:33 It's quite a long document, I would like to read it again fully. 09:08:28 What's the change in the latest one is just name of db column. 09:08:36 vcpu -> vcpus 09:10:38 And as I mentioned before, it's ok to merge a spec if we agree the architecture because this team is not big. 09:11:20 we can review small changes in codes. 09:12:39 I thought I had left a comment but I can't find it 09:12:48 What does DR mean in the first paragraph? 09:13:07 disaster recovery 09:13:49 Can you spell it out fully? I would never have guessed that (I was guessing distributed routing but didn't see how it was relevant for the spec) 09:15:01 it's no problem. 09:16:31 masahito: Spec is good to go for me. Thanks a lot 09:17:01 any comments? if nothing move on to the next item. 09:17:40 bertys: np 09:18:33 #topic status update 09:19:00 we skipped the last meeting, so it's good to have status update from each. 09:19:15 someone has updates? 09:19:25 I have 2 topics 09:19:34 1. availability zone 09:19:42 2. horizon plugin 09:19:43 hiro-kobayashi: go ahead :) 09:20:21 about 1. availability zone, I proposed a new way that gets rid of Blazar filter at the last meeting 09:20:47 But now that I think it's not a good solution 09:21:13 So, I've abandoned my proposal https://review.openstack.org/#/c/472000/ 09:22:03 Instead, I've registered a new blueprint: https://blueprints.launchpad.net/blazar/+spec/multi-freepools 09:22:43 I've not yet written the blueprint in detail 09:22:51 what's the difference of the two bp? 09:24:14 for https://blueprints.launchpad.net/blazar/+spec/stop-using-customized-filter , now that there is no solution 09:24:37 the new blueprint is different from that 09:25:07 the new blueprint is for enabling users to reserve resources in specific availability zone 09:25:32 so the new one still use blazar-filter? 09:25:32 e.g. a user reserve 2 hosts in az 1, and 2 hosts in az 2 09:25:37 Yes 09:26:01 Currently this could also be done with host extra capabilities? 09:26:24 priteau: Oh, I didn't noticed it 09:26:59 priteau: I think it's not a common ussage. how do you think? 09:27:09 Yes it's a bit of a hack 09:27:39 It may be a good idea to support AZ natively 09:27:55 It might be useful for cells as well 09:28:17 Then, I will write a spec. Please review it :-) 09:28:32 yes, we need to compare supporting cells or AZ in Nova. 09:28:53 s/in Nova/of Nova/ 09:29:20 ok, will check the spec 09:29:27 thanks 09:29:31 and for #2? 09:29:48 for 2. horizon plugin, I talked with tejaswi 09:30:16 He is too busy to complete the blueprint 09:30:27 So, someone should take over it 09:30:39 If no one, I can take it 09:31:39 I'm ok whoever takes it. 09:32:19 priteau: Is it possible to share patches of Horizon plugin in Chameleon cloud? 09:32:23 I might be able to provide some support, but can't lead this effort 09:32:25 hiro-kobayashi: do you know whether tejaswi has some pieces of GUI he developed or not? 09:32:42 hiro-kobayashi: I had shared them with tejaswi, I will send you the link 09:32:51 priteau: Thanks! 09:33:00 masahito: I'll check it to tejaswi 09:33:18 Then I'll take it over 09:34:28 hiro-kobayashi: I just sent it to you 09:34:31 hiro-kobayashi: thanks volunteers! if he had, it's good to add him as 'co-author' statement in its commit message. 09:34:49 What a quick :-) thanks! 09:35:09 masahito: ok 09:35:56 priteau: I've received your mail. thanks! 09:36:07 That's all from my side 09:37:04 I'll do next. 09:37:36 I've been working on instance reservation feature. 09:38:26 the patches of its series are available in https://review.openstack.org/#q,topic:bp/new-instance-reservation,n,z 09:39:25 that's from my side. 09:39:52 Thank you for a lot of works! I'll review them 09:40:09 priteau: just a question for your comment. 09:41:13 yes? 09:41:31 Do you think it's good to raise an error and fails to start blazar-manager when a deployer specifies old instance reservation plugin in his/her config? 09:42:19 Yes, I think it is. I prefer over using a dummy plugin that the deployer might not notice right away 09:42:34 Of course there should be an upgrade release note as well 09:43:21 I'm in neutral. so either raising an error or ignoring the plugin looks good to me. 09:44:15 Current implementation ignores invalid plugin name and doesn't raise an error. 09:44:45 ok, I try to raise an error. 09:44:49 Thanks! 09:45:04 any updates? 09:46:20 I tested instance termination when leases end recently, and couldn't get it to work. 09:46:28 I think it's a bug and I have submitted a patch 09:46:43 https://review.openstack.org/#/c/474709/ 09:46:52 Thanks for the review, I will address the comment 09:47:09 thanks the patch 09:47:15 thanks 09:47:33 priteau: if possible, could you report the bug in the launchpad? 09:47:36 I am also continuing to review our Blazar patches in Chameleon, some will be submitted to Gerrit 09:47:38 masahito: OK 09:47:59 it's easy to check it later. 09:48:11 The behavior may have been changed because of trust related patch 09:49:24 hiro-kobayashi: maybe. But now we have Tempest checking for it :-) 09:49:31 The progress of the upgrade is well? 09:49:57 priteau: nice! 09:50:26 masahito: Blazar hasn't changed that much so it should be OK. 09:50:37 priteau: nice! 09:51:58 cross my fingers for you 09:52:29 me too :-) 09:52:42 I am also looking at a fix for the issue with host changing when started leases are updated 09:52:51 this one: https://bugs.launchpad.net/blazar/+bug/1692805 09:52:53 Launchpad bug 1692805 in Blazar "Avoid a reallocation of hosts if a lease has already started" [Medium,Confirmed] 09:53:02 It should be fairly easy 09:54:26 priteau: Yes, it should be easy 09:55:13 the idea written in the bottom looks good to me. 09:55:34 any updates? 09:55:46 if nothing, I changes the topic. 09:56:14 #topic AOB 09:57:04 I and hiro-kobayashi attended OPNFV summit last week. and we had a demo of Blazar. 09:57:45 In term of OPNFV, we had a demo of Promise project. Then 60-70 people came to our booth. 09:57:54 Nice! 09:59:24 some of people had an interest for blazar. So I hope we have new contributers or new deployers. 09:59:40 masahito, hiro-kobayashi: thanks 09:59:53 I hope so! 10:00:54 That would be great 10:01:05 From HPC or related area, they're struggle with noisy neighbors. so Blazar can relax it. 10:01:41 oops, less time. 10:01:50 anything else? 10:02:42 thanks, all! bye 10:02:51 Thanks everyone. Bye! 10:02:52 Thanks! Good bye! 10:02:53 #endmeeting