13:01:49 #startmeeting senlin 13:01:50 Meeting started Tue Aug 25 13:01:49 2015 UTC and is due to finish in 60 minutes. The chair is Qiming. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:01:52 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:01:54 The meeting name has been set to 'senlin' 13:02:13 Qiming: sorry for end meeting late one minute :) 13:02:16 hello 13:02:22 hi 13:02:23 alex_xu, fine, :0 13:02:29 hi 13:02:45 hi 13:02:55 hi 13:03:00 meeting agenda: https://wiki.openstack.org/wiki/Meetings/SenlinAgenda 13:03:12 feel free to add items 13:03:37 #topic liberty-3 workitems 13:03:46 #link https://etherpad.openstack.org/p/senlin-liberty-workitems 13:04:03 hi 13:04:23 seems we still have quite a lot work left to do 13:04:44 the etherpad is not up-to-date, ... 13:05:14 let's do a quick update on this 13:05:19 yes, the exception handling is almost done 13:05:37 great. pls update the page if so 13:06:02 Triggers implementation is done (server side), not yet committed, waiting for another patch to merge 13:06:03 i am behind, i was out last week. i have time to catch up on my items this week 13:06:19 thanks, jruano 13:06:32 I was out in trip last week as well, will catch up this week 13:06:58 functional test framework is up, that is great! 13:07:15 and some fake drivers are already in place 13:07:27 I think I don't have backlog item now since the cleansing of some importand workflows will be done through functional test 13:07:56 basically, we made all interactions with other openstack services a plugin now 13:08:13 of course there are several big functional test cases need to be added to cover this work 13:08:21 it is a configuration option that can be quickly reloaded (switched), so we can use fake drivers for functional testing 13:08:21 yep 13:09:03 yanyanhu, we need to sync on the functional test design, would be a quick one I believe 13:09:15 yes 13:09:35 client unit test has just started, haiwei and I will jump on to that 13:10:06 unit test coverage is 82%, pretty good a number now 13:10:08 I found the client unit tests are already assigned to some guys 13:10:18 i opened some bugs in the client unit tests 13:10:30 was having some students i was training work on them 13:10:39 thanks, jruano, will confirm and work on them 13:11:02 cool! 13:11:03 yep, some did partial fixes. check with them and reassign as needed 13:11:13 okay 13:11:33 I need to check the container cluster work from Tongji Univ 13:11:38 see where are we now 13:11:57 the mentor of that work is on trip these days, will catch him 13:12:06 please keep me in the loop on that Qiming. i can catch up with that work this week 13:12:21 sure, jruano 13:12:35 those students have been playing with the lb policies and autoscalling policies 13:13:01 they have found quite some bugs, yanyanhu has been helping them fix them 13:13:35 an intern from Germany has helped compiled a test report, covering almost all functionalities 13:14:04 yanyanhu, please share with the team once you have an updated version 13:14:10 oh, sorry I think I forget to send out the updated document 13:14:16 sure 13:14:17 it is an encouraging result 13:14:59 #topic functional testing updates 13:15:11 anything else to share, yanyanhu? 13:15:37 I just merged the test case for cluster creating and deleting today https://review.openstack.org/214090 13:16:08 I plan to add more tests for webhook with scaling policy and lb policy in the coming one or two weeks 13:16:19 great. 13:16:46 then we can use them to verfiy the correctness of some important changes 13:17:18 and also will propose a patch to use Senlin generic driver to manage keystone driver 13:17:27 since functional testing may take some time (though not really interacting with Nova or Heat yet), we need a design that will cover most functions with smallest number of test cases 13:17:37 yes 13:17:57 waiting 2 hours for gate result is not fun 13:18:12 let's move on 13:18:18 #topic exception handling 13:18:30 haiwei? 13:18:35 yes 13:18:55 any outstanding issues in this area? 13:18:59 the exception handling of the drivers are almost done 13:19:29 not that outstanding :) 13:19:48 okay, we need to make sure the exceptions are captured and handled in the right place (abstraction layer) 13:20:07 just like we have designed, we translated all the exception from SDK to internal error and catch them in upper layer 13:20:21 there are still some exception types to be killed from senlin/common/exception.py 13:20:47 yes, will kill them if they are not used 13:20:50 we may need a quick sync on that offline 13:20:57 great 13:21:37 Don't know if xinhui is online 13:21:54 guess not 13:22:02 pinging 13:22:26 send a message on wechat 13:22:26 #topic placement policy 13:22:39 hi, admin__, :) 13:22:42 cool name 13:22:47 hi, Qiming... 13:23:04 anything to share with team about placement policy? 13:23:37 I am adding the across AZ functions into base placement policy 13:24:05 Have read the ticket of heat and think implements in senlin should be more easier 13:24:15 I will send out review for your comments 13:24:27 great 13:24:33 just an update for you 13:24:44 another thing is about vspherepolicy plugin 13:25:02 yes? 13:25:11 we may need to envolve it with new vio feature on affintiy/anti-affinty 13:25:43 do you want to check it in as well? 13:26:04 any help needed from Senlin side? 13:26:05 not just load balancing aware placemet provided and also host level palcement 13:26:10 yes 13:26:40 do you think it is proper for us to check in the vsphere related policy? 13:26:42 em, sounds interesting 13:26:53 sure 13:27:14 though host level placement is mostly transparent on public clouds 13:27:28 there are definitely use cases for private clouds 13:27:42 just personal view on this 13:27:49 agree 13:27:49 would be interested in seeing the details 13:28:07 okay, will send out review 13:28:35 thanks. it will be at least helpful to trigger ideas when people develop their own policy plugins 13:28:52 one thing to update with you is a call I just had with TOSCA people 13:28:54 yes 13:29:22 we may need to study the standard and do some work uniforming the terms 13:29:29 haha 13:29:59 what are the first priority? 13:30:14 I mean the cooperation 13:30:15 when looking from a standard's perspective, our sample policies today are still a little bit adhoc 13:30:38 we will go in parallel 13:30:55 the standard is under-development, so is our code 13:31:24 could you give more introduction on TOSCA 13:31:38 its mission and imapct 13:31:43 we will sync with the team regularly, to make sure we are walking closer to each other 13:31:47 by this opportuntiy 13:31:55 tosca is gaining a lot of visibility with vendors and cloud customers as well 13:32:15 redhat i believe has adoped it for their platform 13:32:29 admin__, it will take too much time to go through all the details, I can share with you some docs and links offline 13:32:47 jruano, I searched some cooperation between TOSCA AND cloudstack 13:33:17 in OpenStack, at least Heat has a sub-project called heat-translator translating TOSCA files to HOT 13:33:44 Qiming, seems so complex and can not summerize with several words 13:33:59 yes, for now openstack is the main platform 13:34:38 i have some overview decks i can share as well. ping me to remind me to send them out 13:34:44 admin__, if you go through the TOSCA standard quickly, you will find it is not that complicated 13:35:07 Okay, thanks, Qiming, thanks, jruano 13:35:19 it is really targeting real-world problems, not just some docs for you to archive, :) 13:35:26 good 13:35:34 I will read it 13:35:51 #topic senlin-dashboard project 13:36:09 please check the meeting agenda wiki, I have attached two links 13:36:17 #link https://launchpad.net/senlin-dashboard 13:36:25 #link http://git.openstack.org/cgit/stackforge/senlin-dashboard/tree/ 13:36:55 Zhenguo from Huawei has kindly offered to build a Horizon plugin in a sub-project senlin-dashboard 13:37:08 we had a discussion on this when I was in Xi'an last week 13:37:16 this is great 13:37:33 he is also "recruiting" some developers from 99cloud to work on this 13:37:42 :) 13:38:02 it will help people get familiarized with Senlin quickly 13:38:14 yep, definitely 13:38:45 I'm not a web-UI person, but I believe Zhenguo can help lead that effort well 13:38:58 is there any milestone for senlin-dashboard 13:39:03 it can also help us to understand how horizon works with backend service 13:39:06 we may need to help them understand Senlin's concepts and APIs 13:39:19 haiwei, not yet 13:39:37 we are ready for question anytime 13:39:49 according to Zhenguo, developing horizon plugin is pretty a straightforward work 13:40:10 django web app 13:40:11 we just need the right persons to do this 13:40:12 it's senlin's sub-project, does it matter if senlin wants to become a openstack project 13:40:48 haiwei, it doesn't matter, afaict 13:41:00 that's good 13:41:48 as liberty-3 milestone is drawing near, we will need to work hard to do a release this time 13:42:25 sounds ambitious? 13:42:45 it does, but i think it needs to happen this time 13:42:48 we should try our best I think 13:43:06 senlin1.0? 13:43:19 0.1 IMO :) 13:43:20 yep, maybe senlin 0.1, ;) 13:43:36 0.2? 13:43:47 hoho, like that 13:43:49 0.2 13:43:57 lol 13:43:57 why 2 13:44:22 the project wasn't started yesterday, :) 13:44:27 2 is special in China 13:44:39 :) 13:44:40 admin__, LoL 13:44:47 it means dumb 13:44:52 yes, very special 13:45:25 ah... 0.1 sounds like really early beta code 13:45:33 I think jruano is right 13:45:40 something other than that i think better describes the current state of senlin 13:45:41 agreed 13:45:44 0.2 is good 13:45:59 0.2 is not 2 13:46:13 stay foolish, stay hungry 13:46:38 no problem 13:46:48 #topic open discussion 13:47:16 oh, the voting result will be published tomorrow 13:47:26 I don't have more to share 13:47:30 bless your guys 13:47:31 voting for topic 13:47:41 let's pray tonight 13:47:44 thanks ;p 13:47:48 ok 13:48:48 anything else? 13:48:55 nope from me 13:48:56 or time to pray? 13:49:09 nothing for me 13:49:26 okay, let's end this 13:49:31 #endmeeting