15:00:33 #startmeeting trove 15:00:33 Meeting started Wed Mar 14 15:00:33 2018 UTC and is due to finish in 60 minutes. The chair is zhaochao. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:35 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:37 The meeting name has been set to 'trove' 15:00:46 o/ 15:00:51 #topic roll call 15:00:55 o/ 15:00:57 songjian: hi 15:00:58 0/ 15:01:02 o/ 15:01:27 hi zhaochao 15:02:08 maciejjozefczyk was in the #Trove channel, wait for him to show up 15:03:17 kumarmn said he will also be around, though may not response quickly 15:04:05 ok 15:05:04 ok, maciejjozefczyk may not be around now, let's start 15:05:34 I didn't list many topics today, we will have a lot of time for open discussion, mainly on the rocky goals 15:06:34 as maciejjozefczyk is not here, we'll skip the first topic by now, will come back if he is up 15:06:49 #topic Weekly meetings time schedule 15:07:31 I added a comment on the agendas page last week, the option proposed by songjian and wangyao was taken by Solum 15:07:36 but now is free 15:07:56 yes, I saw the update. 15:08:18 so we could just take it, 1400UTC Wednesdays at #openstack-meeting-alt 15:08:28 zhanggang: yes 15:08:38 perfect 15:08:47 So it start from next week? 15:08:50 do we all agree on this? 15:08:56 zhaochao: Just notice that, so we'll hold our weekly meeting one hour earlier. Sound cool. :) 15:09:09 roger that 15:09:13 agree 15:09:32 zhanggang: if we agreed on that, I'll make a patch to the irc-meetings, to change it offically 15:10:10 what about maciejjozefczyk and kumarmn ? 15:10:37 kumarmn also thought it's ok for him in last week meeting 15:11:04 I also asked him about this in trove channel, but he may not read it yet 15:11:18 [22:04] maciejjozefczyk greetings, for me both ok 15:11:34 ok, sure. Let's do it :) 15:11:38 maciejjozefczyk seems ok with this too 15:11:47 ok 15:12:18 #agreed change the weekly meeting to 1400UTC Wednesdays at #openstack-meeting-alt 15:12:34 then I'll submit the patch for irc-meetings 15:13:04 #topic Set the relative priority of various ongoing efforts 15:13:19 #link https://etherpad.openstack.org/p/trove-priorities-and-specs-tracking 15:13:29 the first draft about rocky goals is here 15:14:12 it should be complete but you all should help to make if better, thanks 15:14:13 a lot of stuff 15:14:22 yes 15:14:37 the first thing is, make priorities 15:14:51 fanzhang has proposed a voting method 15:15:03 https://etherpad.openstack.org/p/BOS-forum-operating-vm-and-baremetal 15:15:46 we should also get a agreement on this, what's your opinions? 15:16:43 hey! 15:16:53 maciejjozefczyk: hi 15:16:58 maciejjozefczyk: yo man. 15:17:00 yes an hour earlier would work for me generally for this cycle 15:17:26 kumarmn: great 15:17:52 for me its ok like I said before 15:18:11 Now it comes to the topic, making priorities. :) 15:18:29 maciejjozefczyk: great 15:18:49 fanzhang: ok, we're going to vote? 15:19:05 yes, we're on goal priorities, https://etherpad.openstack.org/p/trove-priorities-and-specs-tracking 15:19:14 maciejjozefczyk: vote for the task priorities in rocky. LOL 15:19:31 I think it is feasible. There should be no one who will vote badly here. 15:19:32 and fanzhang proposed a voting method, https://etherpad.openstack.org/p/BOS-forum-operating-vm-and-baremetal 15:19:55 In the trove priorities and specs tracking list, we may leave our +1 after H(high), M(middle), L(low) 15:20:27 oks 15:20:27 then it's very easy to get the scores. Then we'll get our task priority list :) 15:20:28 songjian: yes, what do you think, kumarmn maciejjozefczyk wangyao zhanggang 15:20:42 got it 15:20:47 fine for me for leaving +1 and then make calculations 15:20:56 Sounds good to me 15:21:29 zhaochao: should we do it now? 15:22:24 fanzhang: ok 15:23:08 we'll add there lines for voting, H: M: L: 15:23:32 I'm here, sorry I'm late. kumarmn, you invited me to the meeting, do you have anything for me? 15:23:37 yep, not just add +1 after the goals maciejjozefczyk :) 15:23:57 oh, hi smatzek, long time no see. 15:24:01 aa fanzhang ok 15:24:26 Not particularly smatzek, was letting you know that the time change. 15:24:48 If you want to vote on the priorities for Rocky, go ahead. 15:25:14 https://etherpad.openstack.org/p/trove-priorities-and-specs-tracking 15:25:20 seems each of us has a different color :) 15:25:27 smatzek: yes, would be great for helping making goal priorities 15:25:30 in pm, you mentioned that there was a request to document how to debug Trove's gate? Is there a URL of some OpenStack wiki or etherpad you'd like me to put my thoughts in? 15:26:17 Not yet. We should add a link to line 124. 15:32:26 Do we need to add py3 as a goal? 15:32:56 Its not needed, for now mox is crucial as community goal 15:33:29 but anyway python3 should be working and It should be tested on gates 15:34:17 I start working on the mox goal, and came up with the first patch for Trove, https://review.openstack.org/#/c/552021/ 15:34:26 I've added some info to the etherpad on nova file injection. 15:34:44 I think a second patch will remove all mox from the trove repo 15:34:51 smatzek: thanks a lot for that 15:35:13 zhaochao: I'll check it 15:35:22 zhaochao: thx! 15:35:26 As some of you may know, I had an abrupt and unexpected job change in November and haven't been working on Trove, OpenStack, or any cloud tech since. So my memory of exactly how to debug Trove gate has faded and will continue to fade. 15:35:40 for python3 support, trove repo seems almost ok, the last one exception will be removed by my last patch 15:36:11 on python3, your last patch is for the guest agent I assume? I think that was the last part that wasn't compliant. 15:36:13 smatzek: about nova file injection, so you mean that you've done some fix work in Trove like months ago? 15:36:37 zigo from debian was asking for supporing python 3 in trove-dashboard, there're a lot of problems there, I'm working on that for the last two days 15:36:45 fanzhang: I didn't do any fixing but investigated the impact and possible mitigation in Trove after the Sept Denver PTG when file injection deprecation became known. 15:37:04 smatzek: right, the last one is for guestagent unittests 15:39:31 smatzek: ok, my simple solution for that is like what you said, using user-data, cloud-init, maybe config-drive, but recently, got no time to think through. I would like to commit some patches for this issue. 15:41:01 +1 for user-data 15:41:19 ok, I think I myself have done the voting 15:41:21 I think the priority there would be to verify what I found before is still true, that the code remains in Nova and works if you pass the earlier microversion, which Trove was/(is?) doing. Then next, find out if and when they are going to remove support for the older microversion. That will tell you how long you have to do the user-data code. You may have a couple cycles. 15:41:50 smatzek: about gate issues, I wrote down some notes after that day you shared your experience with me. And thanks so much. If you don't have time or are not convenient to write it down, I'd like to do that. 15:42:00 if there are anything missing, please append them to the end, and let everyboy know about it 15:42:22 smatzek: ok, thanks for noticing. :) 15:42:30 zhaochao: for me the list is enough big :) 15:42:43 so if it is time consuming an expensive fix for userdata you could probably defer it to another release if there are higher priority items for this release and Trove can still work with file injection. 15:43:01 maciejjozefczyk: yes, it's really a long list ;) 15:44:21 smatzek: ok, I see :) Thanks again . 15:44:35 smatzek: good for me to check and first verify 15:44:48 And for the Octavia way of API-Guestagent communication, I also did a little investigation, I think it could be a possible solution for trove 15:45:14 zhaochao: I think its also possible, and as we know, its already working (somehow) :) 15:45:53 could be simpler than deploying another service (zaqar), just only to have trove working 15:46:20 a good news is one of my colleague who is focusing on the Neutron will help on this 15:46:39 that's awesome! 15:46:41 maciejjozefczyk: no, it may not be simpler 15:47:08 so we all finish voting ? 15:47:24 fanzhang: for me yes 15:47:50 zhaochao: I'll figure out :) 15:47:55 then I can help do the math tomorrow LOL. 15:48:26 Ok, do we all finish voting? 15:48:42 fanzhang: thanks, please do help 15:48:59 zhaochao: simple math for fun ;-) 15:49:23 zhaochao: yea 15:49:28 I have finished. 15:49:41 finished 15:50:41 #action All of us finish goal priorities voting, fanzhang will do the math tomorrow. 15:50:57 ok, that's all topics from the agenda. 15:51:18 Thanks. 15:51:20 do we have any other topics to discuss? 15:52:25 no topics here. 15:52:42 Done voting 15:52:47 For aboveI have setup testing environment for python3.4.3, and test troveclient by tox -e py34, was passed. Seems we don't need do hard work for it.But trove-dashboard failed for py34 test. So we may fix it. Like zhaochao said. 15:53:35 wangyao: good news. yes, I'm working on trove-dashboard py3 supporting 15:54:41 Done voting,it just seems to be dropped. 15:54:56 zhaochao:great 15:56:11 then we could end the meeting a little earlier, I think 15:56:46 So have a good night :) 15:56:51 and just a reminder, the next meeting should be an hour earlier, 1400UTC Wednesdays at #openstack-meeting-alt 15:56:59 bye guys :) 15:57:01 thanks for everyone 15:57:07 bye 15:57:07 bye 15:57:19 bb :) 15:57:21 good night :) 15:57:36 #endmeeting