09:00:59 #startmeeting blazar 09:01:00 Meeting started Tue Jun 6 09:00:59 2017 UTC and is due to finish in 60 minutes. The chair is masahito. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:01:01 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:01:03 The meeting name has been set to 'blazar' 09:01:10 #topic RollCall 09:01:20 o/ 09:01:34 o/ 09:01:36 o/ 09:01:50 hi hiro-kobayashi, hrito, tejaswi 09:02:02 today's agenda is 09:02:10 1. gate failure 09:02:15 2. instance reservation 09:02:22 3. next weekly meeting 09:02:29 4. AOB 09:02:32 anything else? 09:02:35 o/ 09:02:45 hello, priteau 09:02:46 masahito: Big Tent? 09:03:01 #chair hiro-kobayashi priteau 09:03:01 Current chairs: hiro-kobayashi masahito priteau 09:03:08 priteau: right. 09:03:14 4. BigTent 09:03:42 let's move on to first item. 09:03:48 #topic gate failure 09:04:30 The gate failure in tempest tests appears again. 09:05:23 The log of blazar-a shows oslo.messaging fails to send a message to blazar-manager. 09:05:24 http://logs.openstack.org/64/467064/6/gate/gate-blazar-devstack-dsvm/97b6df3/logs/screen-blazar-a.txt.gz#_Jun_05_19_13_57_996255 09:06:04 Could it be a regression from a requirements lib? 09:06:28 This issue seemed to be fixed last week, but show up again. 09:06:33 I don't think we have changed anything related to messaging recently 09:06:37 priteau: maybe 09:07:04 I guess that's caused by the update of oslo.messaging https://review.openstack.org/#/c/470975/ 09:08:12 But I'm not sure the root cause 09:08:27 if the change is a root cause, we can check it by reverting the requirements.txt. 09:08:42 There was a blacklist added for 5.25.0: https://review.openstack.org/#/c/469539/ 09:08:46 s/reverting/revicing/ 09:09:22 But gate fails even with this blacklist: https://review.openstack.org/#/c/470084/8 09:10:35 another package could have upper constraints with 5.25.0. 09:10:48 if so, it hits the bug. 09:12:07 we should add !=5.25.0 to our requirements.txt to avoid the error. 09:12:13 does it make sense? 09:12:34 And 5.26, maybe 09:13:04 right. 09:13:06 masahito: that's what is proposed by the bot: https://review.openstack.org/#/c/470084/8/requirements.txt 09:13:31 maybe 5.26 is buggy too but hasn't been noticed? 09:14:11 we can try a WIP patch forcing oslo.messaging to an older version 09:14:14 I think it hasn't 09:16:17 according to the commit message, we can avoid the bug with config in devstack. 09:16:18 https://review.openstack.org/#/c/469539/ 09:17:16 anyway, check the result of the patch hiro-kobayashi is pushing now. Then if it works, we can do quick fix. 09:17:36 s/do quick fix/apply the fix/ 09:17:41 I pushed the patch: https://review.openstack.org/#/c/471177/5 09:18:29 I'm sorry but it also contains another changes. But I think they doesn't effect the result 09:18:42 What are the non-requirements changes for? 09:19:56 They are the remains of my investigation 09:19:57 i think it's different change he is working on now. 09:20:07 OK, I'll ignore :-) 09:20:22 thanks 09:20:23 Let's see what happens with the gate then 09:21:05 let's jump on next :-) 09:21:13 #topic instance reservation 09:22:12 I saw some of patches related to instance reservation has CR+2 and WR+1 09:22:15 https://review.openstack.org/#/c/467064/ 09:23:08 but spec of the feature is still in review status. so the above patch will wait the spec is merged. 09:23:17 spec: https://review.openstack.org/#/c/459533/ 09:23:45 Should I remove my W+1? 09:24:07 or add CR+2 and/or W+1 to spec :-) 09:24:17 Need to review still :-) 09:24:44 zuul handles the dependency. so you don't need to remove W+1. 09:24:53 I don't see a way to remove it anyway 09:25:42 I just happened to see this patch in my inbox and it was a quick one to review 09:26:12 got it. 09:26:35 priteau: please review the spec. 09:26:41 OK 09:27:06 any comments about instance reservation? 09:27:32 if nothing, move on to next. 09:28:37 #topic next weekly meeting 09:29:24 hiro-kobayashi and me will attend OPNFV summit next week. So we can't chair the meeting. 09:30:09 And Gerald is on vacation 09:30:11 Does it make sense to skip next meeting? or we'll have the meeting? 09:30:20 priteau: right. 09:30:26 Maybe skip due to few people attending? 09:30:57 In other words, who can join next week? 09:31:40 I can. but I dont have any topic 09:31:44 I can too 09:33:39 So maybe skip if it's just the two of us? 09:33:57 yes. I think it's ok. 09:34:31 priteau hrito, if either of you have topics, we'll have. 09:35:23 I don't have any topic for now 09:35:23 I'm thinking to send 'skip meeting mail' this weekend. If you have topics, please update the agenda wiki page. 09:35:47 ok 09:37:00 #topic BigTent 09:37:31 priteau: thanks the review of the proposal. 09:37:40 https://etherpad.openstack.org/p/blazar-project-definition 09:38:20 I'll try to push the patch in this week. 09:39:39 What is the tag "team:single-vendor "? 09:40:25 https://governance.openstack.org/tc/reference/tags/team_single-vendor.html 09:40:31 And I asked TC of what we need to BigTent. then thierry replied the mail. 09:40:51 The conditions of the team tag are listed on this page 09:41:32 In principle we are not single vendor, as there are three organizations in the active core reviewers, and more in the active non-core reviewers 09:41:40 the definition's of the tag is written here. https://github.com/openstack/governance/blob/master/reference/tags/team_single-vendor.rst 09:42:12 However, for the past 6 months lots of patches came from NTT, so maybe the script will detect us as single vendor 09:42:18 masahito: Have you been able to run the script? 09:42:42 Got it 09:42:57 priteau: I run the script and it doesn't say "blazar shouldn't have the tag". 09:43:29 What does it say? For me the script throws an exception 09:43:43 the script just check whether each project shouldn't have tags. 09:44:17 if the project doesn't have the tag, the script says nothing. 09:44:48 sorry, wrong sentence. 09:45:04 if Blazar doesn't have single-vender tag, the script says nothing. 09:45:17 if Blazar have single-vender tag, the script says nothing. 09:46:08 it means the script only shows "wrong tags" in projects.yaml. 09:47:01 So do you think we should keep the tag for proposal? 09:47:48 IMO, it's not good to hide the fact. 09:48:49 I agree 09:49:12 but we can also write the truth that we have some contributers from many organizations in the commit message. 09:49:28 +1 09:49:44 and the history of the project shows the tag, too. 09:50:04 Hopefully we can remove the tag in the future 09:50:39 definitely 09:51:16 right. 09:52:03 any comments for the proposal? 09:52:56 +1 for the proposal. I have no additional comment 09:53:09 thanks masahito! 09:53:10 hiro-kobayashi: thanks. 09:53:17 #topic AOB 09:53:30 last 7 mins. 09:53:42 The gate still fails... 09:53:56 any update or info? 09:54:14 Looks the same log from blazar-a 09:54:44 BTW, I have an idea for getting rid of blazar-nova filter 09:54:55 for the host reservation 09:55:43 The solution is using AggregateMultiTenancyIsolation just like the new instance reservation feature 09:56:40 Reservation owners specify the reserved resource my the availability_zone parameter when creates servers 09:56:59 Not to use sheduler-hint 09:57:12 Interesting approach 09:57:25 sounds nice. 09:57:44 I'll right details in Gerrit in this week maybe. So please review it! 09:57:52 could you write down the idea in the bp? 09:58:00 mashito: yes 09:58:06 of course, spec is ok :-) 09:58:23 one fyi from my side 09:59:07 OPNFV Promise team will show a quick demo that shows blazar's reservation feature in next week. 09:59:36 The scenario is L34-44 in this etherpad. https://etherpad.opnfv.org/p/doctor_poc_at_summit_2017 10:00:21 it only shows a host reservation barriers non-reserved instance can't be scheduled to the reserved hosts. 10:00:45 the scenario is quit similar to the tempest scenario we have. 10:01:10 that's my fyi. 10:01:43 oops, running out of time. 10:02:03 thanks all! 10:02:09 thanks 10:02:23 see you in next two week. 10:02:30 thanks! see you! 10:02:34 bye 10:02:41 #endmeeting