17:59:00 #startmeeting trove 17:59:00 Meeting started Wed Jun 10 17:59:00 2015 UTC and is due to finish in 60 minutes. The chair is SlickNik. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:59:01 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:59:03 The meeting name has been set to 'trove' 17:59:16 hello yall 17:59:22 hey cp16net! 17:59:32 hello SlickNik, cp16net 17:59:34 Giving folks a few minutes to trickle in. 17:59:42 o/ 17:59:46 hey everyone 17:59:47 hi hi 18:00:02 Meeting agenda is at: 18:00:06 #link https://wiki.openstack.org/wiki/Meetings/TroveMeeting 18:00:10 o/ 18:00:37 o/ 18:00:40 o/ 18:01:04 Hello 18:01:22 Okay, let's get started. 18:01:25 o/ 18:01:32 #topic Trove pulse update 18:01:41 #link https://etherpad.openstack.org/p/trove-pulse-update 18:01:57 o/ 18:04:22 i was close to being up there... 18:04:36 challenge accepted 18:04:40 cp16net: gotta try harder next time :P 18:04:46 ?/ 18:04:50 Total velocity still around ~200 reviews this week 18:05:16 which is about what our average velocity has been like. 18:05:31 So nothing new here. 18:05:51 I did have an action item from last week to follow up on the auto-abandon job and what happened to it. 18:05:56 we added 4 more reviewers this week upto 25 from 21 last time 18:06:58 sushilkm: That's because folks from other projects were reviewing the "Protected instance support" spec. 18:07:12 :) 18:07:25 ah that makes sense 18:07:31 I wouldn't necessarily count on them to be Trove reviewers. 18:07:32 i assume the core team size should have shrunk to 8? 18:08:05 might need to take an action item on that SlickNik 18:08:38 dougshelley66 / cp16net: did we lose someone else from core? 18:09:04 while i hate to have seen Vipul leave, he did resign from Trove Core, no? 18:09:15 Ah yes - of course. 18:09:35 what is the link to the trove core group in launchpad? 18:09:44 maybe i'm overlooking it 18:09:52 #action SlickNik to submit patch to revirewstats to remove vipul from core. 18:09:55 :P 18:10:10 cp16net: We're using gerrit groups now, so it's not associated with LaunchPad any longer. 18:10:19 OH.... 18:11:12 I've removed vipul from the gerrit group — haven't done it from reviewstats yet, so the numbers aren't reflected. 18:11:36 Will do that. 18:11:58 Back to the auto-abandon job. 18:12:52 The OpenStack infra team explicitly stopped running that. 18:13:11 More context of the discussion at: 18:13:13 #link http://lists.openstack.org/pipermail/openstack-dev/2015-March/058104.html 18:14:29 Basically they found that the negative impact of the job outweighed the positive impact, and that folks had other tools to prioritize looking at reviews — so they killed the job. 18:16:32 That explains why we haven't seen it running for a while. 18:16:40 Any questions regarding any of this? 18:16:59 . 18:17:01 Sounds though like we still wanted a solution to the view issue 18:17:20 so that means that over time our backlog of reviews will just grow? 18:17:23 we might want to look into some of the tools that were mentioned 18:17:31 i.e. if there are reviews that just get virtually abandoned? 18:17:51 dougshelley66: yes. that's the case — you will have crufty reviews in the system. 18:18:29 you also run across the issue of crufty reviews that don't make any sense based on where trove has moved since they became idle 18:18:43 ashleighfarnham: ++ I can talk to the infra folks to figure out how to get a view to the reviews that can filter out some of this cruft. 18:19:02 might be able to run a crazy query to find some of them 18:19:33 i just saw something about gertty and seeing how it works 18:19:46 cp16net: We did have some crazy queries that I had set up before — but I think there is now support for native gerrit dashboards that you can set up, so we should probably look into that. 18:20:11 oh yeah they were moving to the templates or something right? 18:20:23 I think more improved views are a good thing but if we are using those weekly reviewstats to monitor our progress... 18:20:33 views won't clean that up 18:21:48 yeah but if we can tell that a review is not been touched and not going to be touched again as core we can abandon it 18:22:17 dougshelley66: Some stats will probably not be useful to monitor (eg. wait times for open reviews) but it would still make sense to track review and merge velocity. 18:22:17 if that was a mistake it can be reverted. 18:22:52 SlickNik, cp16net sounds good 18:23:42 so if slick can find a view of that we can quickly go over which reviews are still valid and prune the list 18:24:43 i *assume* that other teams might do something similar 18:25:12 cp16net: I think the idea (if you read that mail thread closely) is that we shouldn't be abandoning other folks' changes just because they're out of date since it's a pretty strong thing to do, and can have negative consequences. 18:25:52 I'll try to get us a good dashboard that will filter old changes out, so that it doesn't affect our workflow as reviewers. 18:26:04 ahhh yeah true 18:26:12 what if we contact them? it would be a lot of work... but still 18:26:29 vkmc — it's okay to get them to abandon their own changes. 18:26:53 cool 18:26:57 or even take over the changeset if it still applies 18:27:04 that would be a decent solution so we can have a more representative stats 18:27:47 +1 for take over of changeset which are overdue and waiting for long 18:28:23 presumably you should communicate with the owner to take over a changeset 18:28:44 sushilkm: Of course, but please talk to the folks you're taking over from so that you don't end up duplicating work / stepping on each other's toes. 18:28:54 SlickNik++ 18:29:02 true 18:29:09 #agree 18:29:13 Okay — Let's move on. 18:29:32 SlickNik: I dont agree. If somebody does not touch a patch set for a month or so we should abandon it. You cannot just take it. If they want to proceed working on it they can always restore it. 18:29:48 LGTM a requirement for take-over :) 18:31:00 pmalik: Have you looked at http://lists.openstack.org/pipermail/openstack-dev/2015-March/058104.html 18:31:35 That lays out some very good reasons why one shouldn't simply just abandon someone else's review just because it's been sitting for a while. 18:32:03 Well, it does not mean they are right... 18:33:05 Well, the only reasons I've seen for abandoning it so far is that it affects some of the stats we measure — and I have to admit the reasons in that thread seem far more compelling to me than that. 18:34:13 yeah there are some good points in it 18:35:22 We can have more of this discussion offline pmalik - let's move on with the meeting agenda. 18:35:55 #topic Open Horizon Reviews 18:36:37 So dougshelley66, shayneburgess and a few others noted a couple of days ago that we have quite a few Trove related changes in horizon seem to be making slow progress with reviews. 18:37:29 #link https://etherpad.openstack.org/p/trove-reviews-in-horizon 18:37:51 yeah i think some of these should have been in since the beginning like the database/user calls 18:38:15 One of the reasons for this — talking to horizon folks — is that many of them don't have context into whether some of these changes are correct from a trove perspective. 18:38:35 I can pass this list to couple of horizon cores and see if they can help. I can also work with them to view these patches if it helps 18:38:56 So I'd like to start getting some more trove folks to start looking at these patchsets, and +1 them if they look good from a trove perspective. 18:38:57 johnma, thanks! 18:39:03 so we should review them for validity at least it sounds like 18:39:07 johnma: that would be super helpful — thanks! 18:39:12 cp16net: yes ++ 18:39:14 SlickNik: Sounds reasonable 18:39:28 sure, that sounds like a plan. I can help with it 18:39:48 yes please review them from the trove standpoint 18:40:23 working on a model in horizon where trove reviews could go through with less review overhead from horizon 18:40:44 basically trove reviewers worth one +2 and horizon core the other +2 18:40:57 hopefully decreasing the bottleneck 18:40:58 i think i can help as well but i dont have much exp with the horizon pieces 18:41:12 david-lyle would that be to have the trove dashboard external to the horizon repo as a plugin? 18:41:19 or some such 18:41:34 david-lyle: The new model sounds interesting - you have an eta on that? 18:41:37 dougshelley66: that may be a long term solution, but testing becomes more difficult 18:41:39 do we know whether these patches involve ny angular changes as well or is it mostly django 18:41:56 david-lyle: ++ 18:42:15 right now working on moving trove and sahara to a contrib directory as a first step 18:42:30 the review model can work that way now 18:42:38 I can follow up with you to figure out how to ease the bottleneck with reviews 18:43:02 johnma: it's mostly django — I haven't seen any angular based reviews yet. 18:43:19 so please just validate the correctness from the trove interaction standpoint and the horizon reviewer will tackle the horizon part 18:44:07 david-lyle: To me that sounds like a good, workable plan for the near term. 18:44:13 thats good. I think I can handle the Django part, havent been up-to-date with the angular changes happening. THanks SlickNIk 18:44:14 we just don't have enough trove expertise to properly validate that side 18:44:30 makes sense to me 18:44:31 great, I'd really like things to get in a lot faster 18:44:52 Awesome, thanks david-lyle! 18:45:19 david-lyle thanks for the info and help 18:45:29 #topic Open Discussion 18:45:39 Just a little suggestion for open discussion. I think the project description on Launchpad (https://launchpad.net/trove) should be updated to reflect the fact Trove now supports other datastores too. 18:46:05 pmalik: ++ 18:46:07 maybe it should have the current mission statment? 18:46:11 who can update that 18:46:13 I can update that with the mission statement 18:46:21 slicknik ++ 18:47:16 #action SlickNik to figure out how to update LaunchPad Trove messaging to be current and update it with Trove's mission statement 18:47:56 Other items for open discussion? 18:48:43 i'd like to catch up on clustering that was chatted about at the summit but that doesnt have to be here. 18:49:02 cp16net: Let's catch up on that after the meeting. 18:49:08 we can catch up in #trove after the meeting 18:49:10 ok 18:49:23 Looks like we're all done for the meeting. 18:49:24 Thanks all! 18:49:27 #endmeeting