09:01:26 #startmeeting blazar 09:01:27 Meeting started Tue Jul 10 09:01:26 2018 UTC and is due to finish in 60 minutes. The chair is masahito. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:01:28 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:01:31 The meeting name has been set to 'blazar' 09:01:57 #topic RollCall 09:02:00 o/ 09:02:14 o/ 09:02:24 Today's agenda is 09:02:34 1. Non-client libraries release 09:02:42 2. AOB 09:02:46 3. Review meeting 09:02:51 anything else? 09:02:57 o/ 09:03:02 priteau, hiro-kobayashi, tetsuro: hello 09:04:13 #topic Non-client libraries release 09:05:37 FYI, 20th July is final non-client libraries release 09:06:23 We don't have any, do we? 09:06:31 I'll cut stable/rocky branch if needed and put 2.0.0 tag to blazar-nova, dashboard, tempest 09:07:16 masahito: are those considered libraries? 09:07:29 oops, you're right, priteau. 09:07:30 I thought we would tag them at the same time as blazar 09:07:37 End of August 09:07:59 Non-client libraries are things like oslo 09:08:08 According to the schedule docs it's about oslo and others. https://releases.openstack.org/rocky/schedule.html#final-release-for-non-client-libraries 09:09:02 I was confused in pip libraries. 09:10:53 The next release step for us should be rocky-3 09:11:03 blazar-nova is listed as a library project: https://releases.openstack.org/rocky/index.html#library-projects 09:11:06 rocky-3 and client release 09:11:34 hiro-kobayashi: good catch! 09:11:58 Do we have any pending patches on blazar-nova? 09:12:41 I see the one patch https://review.openstack.org/#/c/555355/ 09:12:53 hiro-kobayashi: thanks. 09:14:04 Looks like the gate log is already removed. Let's wait the gate job result. 09:15:12 any comments? 09:16:31 #topic AOB 09:16:39 less topic today. 09:17:00 I noticed we haven't discussed the Berlin summit 09:17:14 Before starting review meeting, does someone have some thing to discuss or share? 09:17:19 The CFP is July 17 09:17:31 priteau: nice catch. 09:17:50 That's next Tuesday, so if we want to have Blazar presentations that's the last week to work on it 09:18:46 Is anyone planning to submit a talk? 09:18:57 Do you all already have idea for blazar OFFICIAL presentation? 09:20:10 I'm going to submit a talk about the usage of placement from blazar, as a GENERAL presentation. 09:20:28 Is there a track for project updates? 09:20:32 I plan to submit one talk with OPNFV Doctor project. 09:20:39 We can cover new features in Rocky 09:20:45 tetsuro: great 09:21:26 Usually, TC members ask having project update session after the CFP to PTL. 09:22:06 I didn't receive any notice, but I hope there is the update session. 09:22:47 The summit is shorter than others, only 3 days 09:22:57 Maybe some sessions are not happening 09:23:05 It would be good to ask before the CFP closes 09:24:02 right, I'll ask TC member later. 09:24:10 Thanks! 09:26:42 anything else? 09:28:04 If nothing, we can move to review meeting. 09:28:21 I saw a new unit test fails sometimes in local and in gate. http://logs.openstack.org/65/578065/6/check/openstack-tox-py27/555b9af/testr_results.htm 09:28:24 i'm oks 09:28:33 i'm okay 09:28:45 I'll submit shortly for this bug report and fix. 09:29:03 Just to report. 09:29:15 Nothing else from me. 09:29:16 I listed links to series of patches in this meeting agenda page. 09:29:26 https://wiki.openstack.org/wiki/Meetings/Blazar 09:29:48 tetsuro: that link looks invalid 09:29:55 tetsuro: thanks for digging into the bug :-) 09:30:25 hiro: oops, thanks. http://logs.openstack.org/65/578065/6/check/openstack-tox-py27/555b9af/testr_results.html.gz 09:31:42 tetsuro: thanks 09:33:04 priteau: IIRC, you're not available after 10am UTC right? 09:33:22 I can stay longer 09:35:23 got it. So let's start the review from high priority and completed patch. 09:35:44 And continuing to next week. 09:36:33 Let's start from cleaning time 09:36:46 https://review.openstack.org/#/c/575021/ 09:37:07 I think the discussion point is default value for the cleaning_time option 09:37:45 I set 5 min but priteau and Stig set it's a big change 09:37:56 s/set/said/ 09:38:43 I think it's reasonable and I'm okay to set zero as a default value. 09:39:01 Now that I'm thinking about it, no default change is easy to upgrade and less confusing. 09:39:34 Agree. 09:39:40 Stig also makes a good point that it can make reservations unaligned with time boundaries 09:40:35 Right. It's important for operators. 09:40:46 For example, if a user reserves a node for a full day (00:00 to 00:00 the next day), with cleaning time the next user would only be able to reserve from 00:05 09:41:44 This example looks fine for me. 09:42:58 The Blazar's reservation is similar to hotel booking mechanism. The cleaning time should be placed out of each reservation. 09:43:18 Yeah, I don't see any problems on that example either. 09:44:29 It's just that if users want to keep reservations aligned, they would need to know the cleaning time 09:46:05 Anyway, I think hiro-kobayashi was agreeing to set default to zero? 09:46:06 ah, it's a case that same one user want to try create two reservation continuously, isn't it? 09:46:22 priteau: Do you mean the cleaning time value should be exposed to users? 09:46:32 priteau: yes, i agree with zero. 09:46:33 Same user or two or more users working together 09:47:12 Alice and Bob plan together so that Alice books one node on Monday, Bob the same node on Tuesday 09:47:28 I now understand the situation you and Stig mentioned. 09:47:39 But Bob is surprised when his reservation starting at 00:00 on Tuesday is rejected 09:47:54 When we have resource availability API, that will be easier to handle 09:48:04 Blazar will be able to say, reserve from 00:05 09:48:13 make sense 09:48:19 But at the moment it is trial and error 09:48:25 I see. 09:48:34 From my experience, this type of "user effected config" should be announced by cloud operator. 09:49:41 That's right, though if we make default 5 minutes, some operators won't think about announcing it 09:49:43 List of AZ names, what reservation plugin is supported in the cloud, and etc 09:49:54 right. 09:50:23 So setting the default to 0 makes sense to me. 09:50:23 hiro-kobayashi: I think the cleaning time will effectively be exposed to users by the resource availability API 09:51:15 cool 09:51:44 So we are all in agreement to update the patch to set default to 0? 09:52:02 yes, of course. 09:52:07 defaulted to zero makes sense. 09:52:08 yes. I'll update and push the patch. 09:52:53 Thanks! 09:53:05 if nothing more, let's move on to Placement API support 09:53:19 spec: https://review.openstack.org/#/c/576343/ 09:53:34 patch #1: https://review.openstack.org/#/c/578073/ 09:53:42 patch #2: https://review.openstack.org/#/c/578065/ 09:53:48 patch #3: https://review.openstack.org/#/c/578637/ 09:53:55 patch #4: https://review.openstack.org/#/c/578017/ 09:54:48 I am sorry, I haven't had the chance to look at the spec or any of these patches yet 09:54:54 I don't plan to add anything to the spec anymore. So, if you think it's okay, let's merge it. 09:55:01 For Rocky cycle, I'm happy the spec is merged at least. If the patch series are merged, it's great. 09:55:04 priteau: np :-) 09:55:59 I will have some minor comments on the spec 09:57:11 The spec has concrete description, so I hope it's easy to merge. 09:57:29 I haven't seen the latest PS of the spec, sorry, but the PS before the latest was all okay to me. 09:59:27 thanks. Anyway, tetsuro's placement related patches will not be affected by spec changes. So, let's review them. 09:59:48 Running out weekly meeting time. 10:00:10 Let's move to #openstack-blazar 10:00:27 ok 10:00:32 thanks all 10:00:38 thanks! 10:00:39 #endmeeting