08:00:06 #startmeeting mistral 08:00:07 Meeting started Fri Apr 20 08:00:06 2018 UTC and is due to finish in 60 minutes. The chair is d0ugal. Information about MeetBot at http://wiki.debian.org/MeetBot. 08:00:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 08:00:11 The meeting name has been set to 'mistral' 08:00:30 It is the Friday office hour! 08:00:32 yay! 08:01:06 d0ugal,rakhmerov,apetrich,bobh, mcdoker18: ping :) 08:01:10 https://etherpad.openstack.org/p/mistral-office-hours 08:01:18 here we are 08:01:24 d0ugal: yep, here 08:02:09 So, the only interesting news I have is that Mistral R-1 has been released! 08:02:23 d0ugal: cool ) 08:03:01 mistral-lib and python-mistralclient releases will happen soon, I made a mistake in the original patches. 08:03:51 With R-1 out, I am going to spend this hour moving all the R-1 bugs to R-2 08:03:58 and if I get time after that I am going to try and do some triage 08:03:59 nice 08:04:23 yep 08:05:30 I made a copy/paste fail - I tried to release mistral-lib with a mistralclient sha and a mistralclient release with a mistral-extra sha :/ 08:05:42 Thankfully the tools block me. 08:06:10 If anyone has anything to discuss, do let me know - there is time for that. 08:06:39 My goal with the office hours is to take care of project/PTL duties so that I am here and active, but if somebody else is here and wants to chat then that is more important. 08:07:24 d0ugal: I don't have much, you've probably seen my memory/performance improvements that have already been merged. That's all I have. Plus I'm investigating other issues now that we're seeing in our env 08:07:39 rakhmerov: other performance issues? 08:07:44 well, yes 08:07:51 I can tell briefly 08:08:01 Sure, if you think they would be interesting 08:08:34 one thing is that we do deletion of resources (wf execs and defs) and for some reason get huge memory footprint during that 08:08:53 I have no idea yet why this happens but it's consistently reproducible 08:09:21 the other thing is that we run a workflow with some subworkflows, so totally we have 4-5 wf executions 08:09:43 rakhmerov: when deleting an object, do we load the full model from the database? 08:09:45 BUT the internal caches that cache WF specifications by executions get full, all 100 entries 08:09:51 which is I'm sure a bug 08:10:11 d0ugal: no, we shouldn't. We just do a delete query 08:10:18 I'll check that anyway 08:10:31 I guess we don't, but I can't think why else memory usage would be high 08:10:43 d0ugal: yeah, me too 08:11:12 I'm very interested to hear what you find. 08:11:24 yep 08:11:37 I need some more time though, hopefully today I'll have more info 08:12:02 because I can't reproduce those issues locally on my mac 08:12:09 only in our distributed env in the lab 08:12:40 that's all I have I guess ) 08:12:44 keep working.. 08:12:47 Thanks for sharing :) 08:13:25 So in R-1 we fixed 13 bugs, but we have 8 in progress and 25 more triaged 08:13:26 but https://review.openstack.org/#/c/562554 makes a really big difference memory wise ) 08:13:35 d0ugal: ok 08:13:49 let me take a look at LP 08:14:02 R-2 already has 15 bugs assigned. We either need to reduce the bug for R-2 or solve lots of things :) 08:14:12 s/solve/fix/ 08:14:40 :) 08:14:48 so looking at R-1.. 08:14:51 https://bugs.launchpad.net/mistral/+bug/1757966 08:14:52 Launchpad bug 1757966 in Mistral "Mistral server consumes more memory over time" [Critical,In progress] - Assigned to Renat Akhmerov (rakhmerov) 08:15:03 my latest patches were addressing it 08:15:17 but I'd prefer not to close it yet and do some more investigation 08:15:22 Sure 08:15:25 so let's move it to R-2 08:15:27 ok for you? 08:15:28 ok 08:16:01 https://bugs.launchpad.net/mistral/+bug/1720797 08:16:02 Launchpad bug 1720797 in Mistral "Migrate Mistral CI jobs to new Zuul v3" [High,In progress] - Assigned to Brad P. Crochet (brad-9) 08:16:13 I thought this was done, no? :) 08:16:15 There has been some recent activity there 08:16:24 ok 08:16:29 I thought... let me check. 08:16:46 https://review.openstack.org/#/c/509184/ 08:16:56 so it seems that mistralclient isn't finished. 08:17:18 ooh, yeah 08:17:18 ok 08:17:35 d0ugal: I'll step out for 5 mins.. Sorry. 08:17:39 rakhmerov: np 08:23:17 i'm here 08:24:05 apetrich: Did you ever hear more from Winson about documentation for the execution events? 08:24:29 d0ugal, not really 08:24:41 apetrich: oops 08:24:46 rakhmerov: ^ that question was for you? 08:24:47 nope 08:24:49 :) 08:24:58 i can ping him 08:25:01 Thanks 08:25:17 Even if he could provide something rough we can then polish that would be useful 08:25:25 yes 08:27:42 sent an email to him.. 08:27:47 Thanks! 08:29:13 yeah, a number of bugs for R-2 is impressive ) 08:29:24 Yeah :/ 08:29:33 https://bugs.launchpad.net/mistral/+bug/1664030 is marked as Fixed 08:29:34 Launchpad bug 1664030 in Mistral "docker image build crashing" [High,Fix released] - Assigned to Brad P. Crochet (brad-9) 08:29:34 At some point I am going to go through that list and remove some 08:29:39 why is it in R-2? 08:30:25 rakhmerov: oh, my mistake. 08:30:41 Moved it back 08:30:50 I didn't realise thrash|g0ne had opened a new bug. 08:31:29 https://bugs.launchpad.net/mistral/+bug/1765421 08:31:30 Launchpad bug 1765421 in Mistral "docker image build job not working" [Undecided,In progress] - Assigned to Brad P. Crochet (brad-9) 08:31:37 That is the new one, I'll mark that for R-2 08:31:43 ok 08:32:24 Okay, I have moved everything that wasn't complete to R-2 08:32:38 I am going to send an email to the list asking people to check the bugs and blueprints that they care about 08:32:45 and make sure they are all correctly assigned 08:33:39 ok 08:34:47 I can confirm that I should be able to complete those that are assigned to me 08:34:53 Great, thanks 08:34:57 rakhmerov: the blueprints also? :) 08:35:07 and their statues are up to date 08:35:12 on BPs, let me check.. 08:35:54 https://blueprints.launchpad.net/mistral/+spec/mistral-async-workflows probably won't be in my focus 08:36:00 I'd move it to R-3 08:36:09 Sure 08:36:12 I want to get it done in Rocky, but not in R-2 08:36:19 I'll be working on Scheduler 08:36:26 Great 08:36:27 (not even sure if we have a ticket for that) 08:36:35 yes, as we talked at the PTG 08:36:42 and some docs that we also discussed 08:36:55 Good point. We should open a bug/blueprint - whichever you think is most sensible. 08:37:19 ok, I'll do 08:37:41 #action rakhmerov: open a ticket for the Scheduler refactoring work 08:37:49 https://blueprints.launchpad.net/mistral/+spec/mistral-readonly-db-transactions 08:38:10 this one btw is a low hanging fruit that could be given to someone who's getting up to speed 08:38:30 fairly easy but also internally (code wise) useful 08:38:55 rakhmerov: https://review.openstack.org/#/c/529513/ 08:39:02 it is completed alreay? :) 08:39:18 Or is there something else we need to do? 08:39:21 Hah! 08:39:23 Looks like apetrich done it in Queens! 08:39:24 looking... 08:39:35 I even approved it :)) 08:39:40 haha 08:39:43 lol 08:39:50 We are good at this. 08:39:54 :P 08:39:56 ok, but then the status is not up to date :) (who to blame? :)) ) 08:40:09 everything in this meeting :-D 08:40:16 everyone* /me is tired 08:40:27 probably me. I must have forgotten to add the blueprint for the automation to pick it up 08:40:37 ok, np ) 08:40:43 apetrich: blueprints are not automatically closed AFAIK 08:40:58 really? I assumed they were 08:41:01 let's mark it Implemented though? I don't we we can move it back to a closed milestone 08:41:07 and I guess when you did it you didn't have permission on launchpad 08:41:13 I'll update it 08:41:19 oh that's true 08:41:44 I'll just have to leave it as rocky-1 08:41:54 apetrich: so now you can claim credit for it in queens and rocky. 08:41:59 ok 08:42:01 :) 08:42:06 brb, 2 mins 08:42:44 d0ugal: I have reasonable doubts that someone will get https://blueprints.launchpad.net/mistral/+spec/mistral-secure-sensitive-data done ) 08:48:16 back - sorry about that 08:48:24 rakhmerov: Indeed, I agree. 08:48:34 ) 08:49:43 So we actually fixed 22 bugs in R-1, I found a number that had merged but didn't have a milestone 08:50:01 ooh 08:50:27 can we assign a milestone to them now? 08:50:41 Yes, I have done that 08:50:51 https://launchpad.net/mistral/+milestone/rocky-1 08:50:56 The bug list is longer! 08:51:10 ok 08:51:12 awesome 08:51:31 I guess we need to remind someone to set a milestone :) 08:51:40 Indeed! 08:51:40 Hardik probably? 08:51:48 A number of people :) 08:51:50 ok 08:52:00 I'll tell Hardik (my colleague) 08:52:02 but I am not sure that everyone can do that? 08:52:07 bobh had missed a few 08:52:31 I thought for bugs it should be possible, hm.. but may be not 08:52:44 This one for example didn't have a milestone: https://bugs.launchpad.net/mistral/+bug/1761246 08:52:45 Launchpad bug 1761246 in Mistral "cut_list() needs to work the same as cut_dict()" [Low,Fix released] - Assigned to Bob Haddleton (bob-haddleton) 08:53:20 ok 08:53:53 d0ugal: if only a limited circle of people can do assignments maybe we just need to check it on review 08:54:38 Yup, we should look into that. 08:54:53 https://bugs.launchpad.net/mistral/+bug/1765349 08:54:54 Launchpad bug 1765349 in Mistral "doc: missing minor version release notes" [Undecided,New] 08:55:11 We have had a request to go and write historical release notes :/ 08:55:52 We have been getting better are release notes slowly, but I don't think it is realistic to write and backport release notes for previous cycles. 08:56:26 d0ugal: yes, Hardik confirmed that he can't assign a milestone 08:56:50 d0ugal: ooh gosh, no! 08:56:53 rakhmerov: I'll add milestone reviewing to the regular lost of tasks on the etherpad. 08:57:02 I'll eat an entire milestone probably :) 08:57:06 lol 08:57:20 I'll = it'll 08:58:45 rakhmerov: I can't imagine what you take for breakfast :) 08:59:00 :) 08:59:20 pgaxatte: reviews mostly ) 08:59:49 Haha 09:00:37 lol 09:00:47 Okay, we are our of time! That was a good R-1 cleanup 09:00:51 Thanks for the help 09:01:15 Next week I'll start looking at the R-2 goals and making sure they seem realistic. Hopefully I can reduce the work we have planned to someting almost possible. 09:01:18 #endmeeting