16:00:20 #startmeeting Mistral 16:00:21 Meeting started Mon Jul 13 16:00:20 2015 UTC and is due to finish in 60 minutes. The chair is rakhmerov. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:22 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:25 The meeting name has been set to 'mistral' 16:00:31 hi 16:00:32 hi everyone! 16:00:48 Hi 16:00:54 <^Gal^> Hello 16:01:18 hi guys 16:01:28 hi 16:02:11 ok, I think there's enough people to start 16:02:33 so the agenda is at https://wiki.openstack.org/wiki/Meetings/MistralAgenda 16:02:49 #topic Review action items 16:03:22 1. rakhmerov, others: review and discuss Lingxian's proposal about removing wf_trace module 16:03:36 I haven't done this yet, sorry 16:03:44 :-) 16:03:48 so I'll keep it for myself 16:03:57 #action rakhmerov: review and discuss Lingxian's proposal about removing wf_trace module 16:04:13 2. rakhmerov: start working on deadlock problem in scheduler 16:04:44 on this one, I decided to postpone it 16:05:02 because there's a couple of things that seem to be even more annoying 16:05:08 hi, rakhmerov, is there any detailed info abou this topic 16:05:19 for example: https://bugs.launchpad.net/mistral/+bug/1469926 16:05:19 Launchpad bug 1469926 in Mistral liberty "task with ERROR doesn't post any result" [Critical,In progress] - Assigned to Renat Akhmerov (rakhmerov) 16:05:29 xylan_kong: what do you mean? 16:05:34 which one? deadlock? 16:05:42 yes 16:06:12 i even don't know that, want to know the context 16:06:15 well, before the last Winson's patch our Scheduler (scheduler.py) seemed to have a deadlock 16:06:19 hi 16:06:24 hi m4dcoder! 16:07:13 so something was wrong with transactions in it and we need to investigate it 16:07:45 because the way how it was done in our opinion was correct but it turned to be not 16:08:18 right now what we have in scheduler doesn't provide a guarantee that a delayed call will be eventually made 16:08:29 it doesn't have that deadlock 16:08:54 so next one 16:08:55 3. NikolayM, akuznetsova: finish fixing http://logs.openstack.org/35/193435/4/check/gate-mistral-devstack-dsvm/39237fd/logs/devstacklog.txt.gz 16:09:00 this is done 16:09:07 NikolayM, akuznetsova: well done! 16:09:26 4. rakhmerov, xylan_kong, NikolayM: review https://blueprints.launchpad.net/mistral/+spec/mistral-dashboard-executions-screen BP on Tue 16:09:37 done (at least by myself) 16:09:42 me too 16:09:47 ok 16:09:48 leave my comments 16:09:53 yep 16:09:55 5. rakhmerov, tsufiev: provide all initial links on Merlin and Workbook Builder 16:10:21 I'm not sure this is done but I guess ^Gal^ already started communicating with Timur 16:10:28 ^Gal^: is this correct? 16:10:51 It is correct 16:10:55 ok 16:11:03 <^Gal^> yep 16:11:12 <^Gal^> I'm working on executions atm 16:11:23 <^Gal^> were starting to work on tasks tomorrow 16:11:27 so on Merlin and WB Builder Timur is basically the only one proved source of information :) 16:11:32 ok 16:11:48 6. rakhmerov: make sure to create a BP about integrating WB Builder into Mistral Dashboard project 16:13:04 https://blueprints.launchpad.net/mistral/+spec/mistral-integrate-wb-builder 16:13:13 as you may guess, I just created it :) 16:13:18 so this is done 16:13:35 hehe 16:13:45 :-) 16:13:45 :) 16:13:53 7. [1]melisha: review Lingxian's patches 16:14:08 I don't remember which exact patches though 16:14:20 maybe the service api 16:14:34 so it's up to Moshe, I guess he knows what it means 16:14:35 I hadn't done it yet. Sorry 16:14:38 yeah 16:14:41 no problem 16:14:52 me too :) 16:15:33 sorry xylan_kong, if you're going to Tokyo a beer for you at my expense ) 16:15:43 rakhmerov: np 16:15:46 anyway, I'll do this this week 16:16:00 ok 16:16:16 rakhmerov: I'll remember what you said :-) 16:16:24 abou the beer 16:16:26 sure, you should! 16:16:27 A beer pm me as well 16:16:38 ok, sounds good 16:16:43 ok, let's continue 16:16:43 #topic Current status (progress, issues, roadblocks, further plans) 16:16:56 as usually, let's quickly tell our statuses 16:17:48 Creating BPs for all CRUD screens 16:17:57 my status: i am still working on service api implementation, have discussed with Winson heavily. not very productive last week, so many internal meetings, prepare for slides, etc. 16:18:10 my status: mostly doing reviews, participating in email and etherpad discussions, started working on https://bugs.launchpad.net/mistral/+bug/1469926 (sent a starting patch today) 16:18:10 Launchpad bug 1469926 in Mistral liberty "task with ERROR doesn't post any result" [Critical,In progress] - Assigned to Renat Akhmerov (rakhmerov) 16:18:37 LimorS is finalizing puppet installation and will start to work on Mistral Qpid NO soon 16:18:58 still working on error-resume 16:19:02 ok 16:19:20 melisha1: about the mistral puppet work, will you contribute back to upstream? 16:19:22 <^Gal^> I'm on execution screen on mistral-dashboard, we're also getting Liat (the fellow member to join) up to speed 16:19:31 xylan_kong: yeah, internal stuff is ok, I have to spend some time on it too 16:19:33 my status: fixed issues in patches related to new RPC, fixed bugs with tests and one related to API 16:19:48 ok 16:19:55 Liat will start the tasks screen 16:20:04 m4dcoder: how is that pause-resume thing going? 16:20:26 I saw that there was non WIP patch about a week ago but then you made it WIP again 16:20:32 still working on the internal mechanism, not at the interface yet. 16:20:48 would you like to discuss anything? 16:20:49 but i can resume from error, with-items, at join. 16:20:55 we can have a meeting if you want 16:20:59 ok 16:21:07 i'm roadblocked for a particular use case but working thru that. no need to meet just yet. 16:21:23 ooh, ok, I see 16:21:37 just keep us posted pls and let us know if you need any help 16:21:52 I started to work on expiration-policies there is a new BP for that https://blueprints.launchpad.net/mistral/+spec/expiration-policies-for-executions 16:21:52 I clearly understand that it's not an easy thing at all 16:22:04 ok. you'll see some WIP patches this week. i promise. 16:22:14 ok, Winson :) 16:22:16 thanks 16:22:34 np 16:22:37 gpaz: I started looking at your spec and left a couple of comments, need to finish though 16:22:38 gpaz is working on expiration policy for executions 16:22:54 gpaz, please see my comments in the etherpad 16:22:56 #action rakhmerov: completely review https://blueprints.launchpad.net/mistral/+spec/expiration-policies-for-executions 16:23:08 ^Gal^: a question to you 16:23:14 <^Gal^> yeah 16:23:23 rakhmerov, xylan_kong, I will, thanks 16:23:49 how are you now with Django? Still on the learning curve or technically skilled to do something? 16:24:23 <^Gal^> it is going well now, took me some time to get things going 16:24:45 because my understanding is that you're mostly an AngularJS guy so I was a little bit concerned about how easy it was for you to switch to Django 16:25:00 ooh, ^Gal^, that's great 16:25:02 <^Gal^> still not sure I'm familiar with all the debug capabilities though 16:25:07 good to hear that 16:25:20 <^Gal^> yeah, lots of new stuff 16:25:27 ping zhenguo, he seems to be very experienced in it 16:25:33 <^Gal^> yeah we talk daily 16:25:34 he can help 16:25:38 cool 16:25:50 <^Gal^> yeah he is 16:25:55 <^Gal^> he's great 16:26:20 ok, next topic 16:26:23 #topic Dashboard progress 16:26:53 I don't have a detailed agenda but would like to find out how it is going 16:27:51 from what I saw (from installed Dashboard) it's now functional and ready to be worked on further 16:28:03 I mean it's completely repaired 16:28:10 thanks to zhenguo, great work 16:28:14 rakhmerov: really a good news! i think we could do a mistral demo in Tokyo, through Horizon. 16:28:24 yeah 16:28:31 <^Gal^> yeah he fixed the whole thing 16:28:50 <^Gal^> i had some progress with the execution list table 16:28:52 but I didn't actually test all the features myself (e.g. I didn't try to run a wf so far) so please first make sure it works ) 16:29:09 ^Gal^: ok, good 16:29:36 so you're moving on according to https://blueprints.launchpad.net/mistral/+spec/mistral-dashboard-executions-screen ? 16:29:53 <^Gal^> yep 16:29:53 any roadblocks, challenges that the rest of the team could help with? 16:29:57 <^Gal^> it is my workplan 16:30:14 <^Gal^> hmmm none that I can think of atm 16:30:36 <^Gal^> when I come up with something I usually ask Zhenguo 16:30:36 ok, do you have any estimations on when it can be completed? 16:31:02 maybe not on 100% but at a state when it's useable enough 16:31:46 I'm not pushing you at all, just thinking on how to plan/coordinate our further work 16:31:47 <^Gal^> I'll think until the next meeting I'll have the main table pretty solid-ish 16:31:47 :) 16:31:51 <^Gal^> yeah yeah I know :) 16:32:02 ooh, that would be great 16:32:19 <^Gal^> still hard for me to estimate stuff cause I'm not sure of the whole django thingy flow to the ebd 16:32:19 ok 16:32:20 <^Gal^> end 16:32:40 yeah-yeah, but the thing is that we have to provide some estimates 16:32:45 <^Gal^> sure np 16:32:55 even if we make mistakes 16:33:42 so once this is done, do you think it'll be pretty easy to take executions screen and apply most of its tech solutions to other screens? 16:34:05 <^Gal^> yep 16:34:13 by asking this I'm trying to understand what we count for say in a month 16:34:15 ok 16:34:17 <^Gal^> i need to get the notion of how filtering work in horizon 16:34:25 <^Gal^> and it sort most of the stuff there 16:34:36 yes, filtering is actually a problem 16:34:41 We also try to do them in parallel with Liat 16:34:53 let me write a couple of words about it 16:34:55 right now 16:35:04 so that you keep it in mind when working on it 16:35:25 <^Gal^> sure that would be great thanks 16:35:34 right now in Mistral API v2 we don't have a capability to flexibly do server-side filtering 16:35:43 unfortunately 16:36:37 <^Gal^> so guess I'll write me some ajax 16:36:49 ideally we should be able to do something like: /tasks?from_date=123123&to_date=12312423&name=task1 16:37:05 but our endpoints can't do this atm 16:37:43 I wanted to do this but the technology we're now using in for REST endpoints does not allow it easily 16:38:08 <^Gal^> k 16:38:10 so what I'm going to is that you'll probably have to do filtering on a client side somehow 16:38:14 <^Gal^> syep 16:38:24 <^Gal^> I'll come up with something 16:38:26 but I"m not even sure how it's supposed to be done 16:38:30 yeah 16:38:40 <^Gal^> yeah usually its a json file you play with 16:38:43 the obvious downside is performance 16:38:57 server-side filtering would work much faster 16:39:55 but the good news is that it's not forever and we're planning to implement API v3 which will support such capabilities and fix some other problems too 16:40:06 The biggest issue I see here now is the execution list 16:40:21 my best hope is that it'll happen in the beginning of Sept 16:40:34 melisha1: yes, you're 100% right 16:40:35 After a few weeks the list will be huge and takes a long time to load 16:40:40 yes 16:41:02 Tasks had a filter by execution id so it is okay 16:41:10 what I'm thinking is that we'll probably have to come up with some server-side workaround for now 16:41:13 <^Gal^> I guess I would write some patch 16:41:17 dunno exactly what it will be 16:41:24 yeah 16:41:36 <^Gal^> ill post some json and do it on the clientside 16:41:40 It must probably need to be server side 16:41:48 so, just letting you know in advance cause I see it as one of the biggest issues 16:41:54 from UI perspective 16:42:04 <^Gal^> yeah I guess so 16:42:32 <^Gal^> I'll be smarter when I'll get to that 16:42:37 <^Gal^> sooner than later 16:42:49 ^Gal^: yes, that's fine but as melisha1 said imagine hundreds of thousands executions (maybe millions?) that you'd have to fetch over the net every time and process at the client side 16:42:49 :) 16:42:52 <^Gal^> will update asap 16:42:55 doesn't sound funny 16:43:07 <^Gal^> yeah I know 16:43:12 ^Gal^: ok 16:43:12 <^Gal^> hmmm 16:43:43 so once you will get to that point we'll need to discuss it again and think how to fix it in the server 16:44:01 otherwise it won't be usable in production 16:44:17 ok 16:44:26 <^Gal^> k sure 16:44:48 alright, let's move forward 16:44:49 True. We will also try to think of ideas to present 16:45:05 yes, please do 16:45:19 #topic Liberty 2 progress 16:46:32 so on this topic I don't have much to say except "please update the statuses of all bugs and BPs you happen to be working on" 16:47:13 so Liberty-2 milestone is scheduled for July 29 (~2 weeks left) 16:47:32 so I think most of the BPs will be implemented on time 16:47:57 please also go over a list of bugs and see if you could take some of them 16:48:40 Okay 16:48:51 btw, just to let you know: I'll be on vacation from July 20 till Aug 2 16:48:57 NikolayM will be replacing me 16:49:06 rakhmerov: ok 16:49:15 so ping him any time, even at night 16:49:15 have fun 16:49:16 :) 16:49:25 m4dcoder: yeah, thanks 16:49:26 :-) 16:49:28 <^Gal^> lol 16:49:50 maybe i'll stay at home when you are on vacation 16:50:14 so I'll do my best to complete things assigned to me this week and I'll get back for just one day to release Liberty-2 on July 29 16:50:52 NikolayM: please be in touch with the team on a daily basis and keep LP up to date 16:51:12 yes, sure 16:51:19 xylan_kong: what do you mean at home? :) why? 16:51:51 rakhmerov: paternity leave 16:52:03 :)) 16:52:05 congrat! 16:52:13 !! 16:52:13 rakhmerov: Error: "!" is not a valid command. 16:52:15 m4dcoder: thanks 16:52:27 :) 16:52:28 cool 16:52:48 #topic Open discussion 16:53:15 we have ~7 mins so please go ahead if you have other questions 16:53:22 <^Gal^> congrats!! 16:53:30 if not we can shut down the meeting 16:53:40 ^Gal^ :) 16:54:00 xylan_kong: I have a question actually 16:54:14 do you have a boy or girl? 16:54:15 :) 16:54:26 rakhmerov: i don't know 16:54:29 :-) 16:54:33 ooh really! 16:55:09 is it a chineese tradition not to find out a gender in advance? 16:55:10 :) 16:55:12 Cool. A surprise 16:55:20 no, just depends on you 16:55:29 ok :) I see 16:55:44 so no more questions? 16:55:44 melisha1: yes 16:56:27 no from me 16:56:35 ok guys, have a good week! Come to our channel #openstack-mistral whenever you want to discuss something 16:56:46 thanks a lot for coming 16:56:54 ok, nice meeting, see you 16:56:58 <^Gal^> bye all 16:57:01 bye! 16:57:04 bye 16:57:06 #endmeeting