13:00:05 #startmeeting senlin 13:00:06 Meeting started Tue Nov 15 13:00:05 2016 UTC and is due to finish in 60 minutes. The chair is yanyanhu. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:07 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:00:09 The meeting name has been set to 'senlin' 13:00:22 hello 13:00:52 hi 13:00:59 hi, Qiming 13:01:17 hi 13:01:23 hi, lixinhui 13:01:37 lets wait for a while for other attenders 13:01:43 ok 13:01:43 hi, all 13:01:48 hi, lvdongbing 13:01:57 o/ 13:02:06 hello, elynn :) 13:02:40 ok, lets start 13:02:44 https://wiki.openstack.org/wiki/Meetings/SenlinAgenda#Agenda_.282016-11-15_1300_UTC.29 13:02:59 here is the agenda, please feel free to add items 13:03:03 hi, XueFeng 13:03:08 hi,all 13:03:19 :) 13:03:30 ok, first topic is ocata work items, as usual 13:03:37 #topic ocata work item 13:03:44 https://etherpad.openstack.org/p/senlin-ocata-workitems 13:03:56 performance test, no progress 13:04:23 will consider to move it back to TODO list since may don't have time to work on it in short time 13:04:37 Health management 13:04:37 wondering we can have someone from CMCC work on it 13:04:53 Qiming, that will be great 13:05:03 maybe we can talk with junwei about it? 13:05:26 Junwei told me that he will consider this, but we have to check with him 13:05:32 the basic support has been there, just need to add more context and scenario support 13:05:37 Qiming, ok 13:05:48 yes 13:05:49 how to do this test 13:06:07 it is about rally plugin, XueFeng 13:06:20 got it 13:06:43 XueFeng, you can refer to the wiki of rally to get the basic guide about using rally. Senlin scenario has been added to rally in last cycle 13:07:04 so you can use rally to evaluate the performance of some typical senlin operations like cluster creating/deleting 13:07:29 if you need more information or guide about it, please free feel to ping me :) 13:07:35 OK 13:07:47 ok, lets move on 13:07:54 health management 13:07:58 hi, Qiming, lixinhui 13:08:01 your turn 13:08:10 no update in this thread 13:08:13 I tried mistral in the past week 13:08:23 and found it only 13:08:25 I noticed elynn sent mail to adam to ask for the meeting time of HA team 13:08:35 provide retry 13:08:39 but haven't got response yet? 13:08:41 no response yet. 13:09:03 lixinhui, you mean retry of workflow? 13:09:04 I saw nobody attend their meeting yesterday 13:09:10 ... 13:09:11 elynn, I see... 13:09:16 elynn 13:09:31 did they skip the meeting this week 13:09:34 ? 13:09:46 no people contribute 13:09:56 No, they don't ,but besides chair man, no one attend it 13:10:16 ok... 13:10:20 maybe we can consider to add topics to discuss if needed 13:11:03 adam is the key person 13:11:12 I think 13:11:26 yes, at least we can make some sync with them 13:11:58 you can do it as PTL of senlin :) 13:12:18 then we contribute labor 13:12:19 lixinhui, sure. But I'm not the expert of HA :) 13:12:30 I can just help to bridge you guys 13:12:52 elynn can keep touch adam for the base 13:13:17 yes, so at least, we need get their response I feel :) 13:13:30 no problem lixinhui 13:13:47 :) 13:14:00 will talk with elynn and Qiming about it 13:14:07 to see how to drive this work 13:14:16 ok 13:14:29 the retry of mistral is based on times 13:14:34 of execute some action 13:14:49 so retry is enabled by default? 13:14:50 and time of wait between different try 13:15:05 e.g. when a workflow fails, it will be reexecuted automatically? 13:15:06 no, need configure when define the actiom 13:15:12 I see 13:15:35 how mistrival decides workflow succeeds or not? 13:16:00 they can not judge based on any status 13:16:12 just execute the defined sequence 13:16:31 there are quiet big gap as we expect 13:16:34 um, if so, how to decide whether try is needed? 13:17:02 will it interrupt when some tasks were failed? 13:17:03 ok, so sounds it doesn't work as we expected before... 13:17:16 that is the problem I raise this problem, yanyanhu 13:17:26 yes, elynn 13:17:53 it can switch to some branch if exception is got 13:18:11 but still very easy control nowdays 13:18:17 I see. 13:18:38 qiming, one question is difference between us with workflow based recover 13:18:53 oh, btw, do we need to contribute code to mistral if we want it to support senlin workflow? 13:19:04 Qiming, could you share your mind to compare the two? 13:19:07 e.g. by implementing a plugin/driver for senlin 13:19:20 no idea/experience with mistral ... 13:19:28 yanyanhu, I will investigate more about this 13:19:40 lixinhui, great, thanks 13:20:07 okay, Qiming 13:20:14 that is my part. 13:20:18 it is all about vm ha ... could be an engine/policy-centric solution, could be a workflow centric solution 13:20:39 need some evaluation on the second one to make decision 13:20:49 oh 13:20:57 yes, you are right 13:21:20 will discuss with you when find more 13:21:37 that is my part update 13:21:49 lixinhui, great, thanks for the effort 13:21:58 ok, next one 13:22:05 document, no progress I think 13:22:11 versioned request 13:22:22 me, Ruijie, XueFeng and lvdongbing are working on it 13:23:01 receiver is almost done, only left ones are support for receiver delete and trigger 13:23:13 Ruijie, is policy support done as well? 13:23:21 yes, yanyanhu 13:23:30 policy and policy type have been done 13:23:36 ok, great, will remove it from etherpad 13:23:52 hi, lvdongbing, how about profile support? 13:24:09 it's in good progress 13:24:17 nice 13:24:18 one thing is about the change to object field, the json schema generated now will skip the 'required' property if no property is actually needed 13:24:34 this will help ease some request modeling 13:25:20 Qiming, ok. Currently all objects have at least one field who is not nullable 13:25:44 XueFeng is working on action/event support. I think it is in good progress as well 13:26:08 Yes 13:26:24 A question 13:26:30 action create/delete not support in client 13:26:31 yes 13:27:05 XueFeng, good question :) 13:27:15 And in api layer,delete not support 13:27:43 in engine layer create/delete all support 13:27:44 we do have the plan to let users to customize and create/delete action(s) 13:27:51 or even suspend/resume 13:28:28 but we haven't figured out what is the appropriate way to implement it. 13:28:49 so you can see related service call is left there 13:29:05 but no related API interfaces are exposed to end user 13:29:28 Ok, So in engine layer create/delete need changed to versioned request. 13:29:38 XueFeng, yes, I think so 13:29:42 we can define it now 13:29:55 and use it to implement API interface in future 13:30:00 Ok ,will do this 13:30:09 XueFeng, great, thanks a lot :) 13:30:34 ok, lets move on? 13:30:42 container profile 13:30:55 move 13:30:56 haiwei is not here I think 13:31:09 no progress as well I guess 13:31:11 next one 13:31:18 event/notification 13:31:28 Qiming 's spec has been merged 13:31:50 I believe all of the team have checked it 13:32:37 so lets wait for the patch from Qiming and have further discussion based on it 13:32:55 what spec? 13:32:56 yes working on it 13:33:07 elynn, the spec about notification/event support 13:33:29 https://review.openstack.org/394874 13:33:30 okay~ 13:34:15 looking forward to the patch :) 13:34:28 ok, next one 13:34:34 batch policy 13:34:39 should have been done by Ruijie 13:34:42 will remove it 13:34:54 NFV/baremetal support 13:35:00 no progress I guess 13:35:08 yanyanhu, I want to resume it later.. 13:35:11 :) 13:35:30 Ruijie, ok, please add it back to etherpad if you further work on it :) 13:35:38 okay :) 13:35:47 I think baremetal is not fully supported by sdk? 13:36:03 elynn, no sure about it... 13:36:29 if so, we may need to add more support for ironic into sdk first 13:36:54 ok, that's all work items in etherpad 13:37:24 any additional ones? 13:37:38 you guys are working on but not on the list? 13:37:52 ok 13:37:55 next topic 13:38:08 #topic new core reviewer announcement 13:38:26 I'm very pleased to announce that Ruijie is now core reviewer of senlin :) 13:38:31 Ruijie, congratulate 13:38:37 thanks yanyanhu. 13:38:43 So happy to work with you all. 13:38:50 me too 13:39:05 looking forward more contribution from you :) 13:39:17 I will ^ ^ 13:39:24 Congratulations,Ruijie 13:39:29 :) 13:39:36 welcome 13:39:46 thanks XueFeng, Qiming 13:39:50 welcome :) 13:40:11 please put +2(-2 if needed:P on patches based on your review :) 13:40:26 thank lvdongbing. 13:40:36 we do need more eyes on patches 13:40:53 yes, yanyanu, will spend more time on reviewing. 13:40:59 thanks a lot :) 13:41:17 ok, open discussion now 13:41:22 #topic open discussion 13:41:47 any topics you guys want to discuss? 13:43:24 oh, BTW, I do want to remind the team that the first miletone of ocata release will pass soon 13:43:26 https://releases.openstack.org/ocata/schedule.html 13:43:37 Nov. 14-18 13:43:59 the second one will be in a month 13:44:33 this is gonna be a short cycle 13:44:34 Today is Nov. 15 13:44:40 so we may need to consider and adjust the plan we made if needed 13:44:42 yes 13:44:59 any high priority bugs to fix? 13:45:26 Qiming, currently no I think. Although there are several important features we want to support after this cycle 13:45:44 i see 13:46:02 I will look through the bug list to reconfirm it 13:46:57 this bug: https://bugs.launchpad.net/senlin/+bug/1594352 13:46:57 Launchpad bug 1594352 in senlin "could not list node sorted by "created_at:desc" " [High,Triaged] - Assigned to Qiming Teng (tengqim) 13:47:19 may need to check if the workaround suggested by milan works 13:47:21 ok 13:48:05 the oslo.db side unit test is not an easy task for me ... haven't got cycles to dig into it 13:48:26 Qiming, I see... I recalled you fix on oslo.db side 13:48:37 about the sort issue we met 13:49:37 ok, any more topics? 13:49:59 if not, I guess we can finish the meeting a little bit earlier? 13:50:44 ok, thanks all you guys for joining 13:50:48 https://bugs.launchpad.net/python-openstacksdk/+bug/1640013 13:50:48 Launchpad bug 1640013 in OpenStack SDK "JSONDecodeError happened in senlin-api when running a senlin command" [Undecided,Confirmed] 13:50:58 this bugs in openstacksdk 13:51:05 s/bugs/bug 13:51:05 ah 13:51:15 this is a bug of sdk actually 13:51:18 need point out to help user 13:51:23 it is solved in 0.9.10 13:51:39 yes 13:51:45 it is an SDK bug 13:51:53 0.9.9 sdk is not working properly 13:52:19 so it influences all senlin cmds? 13:52:22 or part of them 13:52:36 global requirements have been patched to avoid 0.9.9 version: https://review.openstack.org/#/c/397650/1/requirements.txt 13:52:45 Yes,all list comment 13:52:53 I see 13:52:54 s/comment/command 13:52:55 what else do you mean by "point out to help user", XueFeng ? 13:53:16 just without 0.9.9, zaqar message receiver doesn't work... 13:53:30 so we will wait for 0.9.10 I think 13:53:54 Some user may confused if use openstacksdk-0.9.9.Point out in the meeting 13:53:59 it is already there: https://review.openstack.org/#/c/397510/ 13:54:25 great 13:54:39 0.9.10 was created and released, upper constraints raised, 0.9.9 version now forbidden 13:54:56 what else do we want? 13:54:56 if 0.9.9 is now forbidden in requirements, it is ok I think 13:55:12 yes 13:55:19 this thing has been discussed in senlin channel today 13:55:37 ok, I see 13:55:50 great 13:56:15 so this bug report can be closed I guess 13:56:32 since upper constraint has been updated 13:57:07 ok, any more items? 13:57:54 ok, thanks you guys again for joining :) 13:58:00 have a good night 13:58:11 #endmeeting