13:00:35 #startmeeting senlin 13:00:36 Meeting started Tue Feb 14 13:00:35 2017 UTC and is due to finish in 60 minutes. The chair is yanyanhu. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:37 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:00:39 The meeting name has been set to 'senlin' 13:00:46 hello everyone 13:01:32 Hi 13:01:37 hi, elynn 13:01:54 hello 13:01:59 hi, Qiming 13:02:32 good evening, I guess somebody may leave for valentine's day :) 13:03:19 lets wait for minutes for other attenders 13:04:46 ok, lets get started 13:05:10 https://wiki.openstack.org/wiki/Meetings/SenlinAgenda#Agenda_.282017-02-14_1300_UTC.29 13:05:19 here is the agenda, please feel free to add items 13:05:27 #topic ocata work items 13:05:41 "Feature Rich" Nova Server 13:06:01 It's ongoing 13:06:03 hi, elynn, I noticed basic support has been done 13:06:06 yes 13:06:10 Three patches merged. 13:06:25 nice 13:06:28 Now it support to create ports with security_groups and floating_ip 13:06:36 ok 13:06:49 so what is the next step? 13:06:52 There're two patches are under reviewing 13:07:08 One is to support jinja2 in user_data 13:07:20 I think that is also important one. 13:07:32 ok 13:07:50 I guess those ones have been enough to support VDU case xinhui described before? 13:08:03 Another one is to handle network update 13:08:20 Not without the jinja2 patch yanyanhu 13:08:28 that is for support vdu node update, right 13:08:46 for deployment, I guess it is enough? 13:08:59 elynn, I see 13:09:06 the network update logic is a mess 13:09:14 And Bran from vmware are testing and migrating their existing templates to use senlin vdu profile 13:09:15 just spent some time reading the code 13:09:27 Qiming, Why do you say so? 13:09:46 I have previously spent a whole week to refactor the code to covery all possible paths and possbilities 13:09:51 It's not easy to handle network update actually... 13:10:02 so I know how tricky it is to get things right 13:10:32 yes, network update is not easy for we need to control the sequence of network resource operations carefully 13:11:11 I said it is a mess because there are some new inline functions introduced, that does nothing at all 13:11:26 it is not easy, that part I agree 13:12:17 You mean I should expand inline functions in _update_network function? 13:12:21 Qiming 13:12:46 you don't need the _get_network call at all, for example 13:13:04 it has been well encapsulated in the sdk exception annotator 13:14:17 it involves a lot of unnecessary DB transactions which can be safely combined 13:15:14 I guess we can define some util functions if those logics are generic enough to reuse, otherwise, we can just define local function in VDU server module 13:15:28 yes 13:15:38 line 1127 to 1154 is now made an inline function, that is making writing test cases for it very difficult 13:15:44 I can refactor that. 13:15:52 the logic is already very complicated ... 13:16:42 Qiming, +1, inline function is very difficult to test... 13:16:48 I was never successful when trying to write an elegant unit test for inline functions 13:16:48 You mean move that logic out? Qiming 13:17:10 it is bad programming style, thus should be fixed 13:17:11 okay, I could do that. 13:18:11 in previous work, I tried very hard to save code lines for validation 13:18:31 that is the reason I introduced 'reason' parameter to _validate_network 13:19:09 I saw that logic Qiming 13:19:11 in this revised version, I need more time to understand how the logic flows and every corner cases covered 13:19:52 After refactoring, 'reason' parameter is used only for a limit part. 13:20:02 so, elynn, please keeping refactoring the code, thanks a lot :) 13:20:11 I can show you the logic behind this patch :) 13:20:37 :) 13:20:39 okay, if possible, keep function complexity below 20 13:20:47 keep function lines below 100 13:21:15 that is a high bar 13:21:17 I can try... 13:21:44 but for maintainability's sake, it is no less important than functional correctness 13:22:55 if you find some logics requently called, you may want to move them into the corresponding driver 13:23:35 that's all from me 13:23:37 okay, I will keep in mind :) 13:23:43 many thanks 13:23:52 great, lets move on 13:24:10 next one, Engine Improvement 13:24:15 xuefeng is not here I think 13:24:24 hi 13:24:28 yanyan 13:24:34 :) 13:24:37 hi :) 13:24:47 I throught you are not here 13:25:02 Come back 13:25:23 About action improvement 13:25:50 Will split the patch 13:26:04 yes 13:26:26 Then you guys help to review 13:26:42 we can split the logic at DB API layer and upper layer 13:26:53 XueFengLiu, thanks a lot 13:27:35 my pleasure 13:27:38 and please keep working on this as well :) 13:27:39 https://review.openstack.org/#/c/424073/ 13:27:44 Will do this tomorrow 13:27:55 thanks 13:28:03 Yes 13:28:06 ok, lets move on 13:28:12 Health Management 13:28:14 Ok 13:28:24 so yanyan's patch won't be revived? 13:28:37 Qiming, you mean the one about action purging? 13:28:42 yes 13:28:45 329338 13:28:46 sorry haven't resumed it... 13:29:01 I guess there won't be difficult to restore it 13:29:07 https://review.openstack.org/#/c/431401/ 13:29:09 I think 431401 is doing the same thing, right? 13:29:09 difficulty 13:29:31 This one 13:29:39 Yes 13:29:56 okay, then 329338 can be abandoned forever? 13:29:57 With recent code 13:30:06 HaHa 13:30:07 that is confusing 13:30:08 Qiming, I guess my patch is more generic solution for action purging 13:30:27 oh, you mean this one 13:30:31 https://review.openstack.org/#/c/431401/ 13:30:39 I guess I can abandon mine 13:30:48 xuefeng didnt continue your patch 13:30:56 and lets keep working on 431401 13:31:00 seems to me he started a new one 13:31:10 The same one 13:31:12 it's ok 13:31:14 okay, let's forget 329338 then 13:31:18 yes 13:31:31 I will abandon 329338 13:31:37 Two patches about action 13:31:40 reassign this to XueFeng now 13:32:04 The first is about cluster check action 13:32:31 And https://review.openstack.org/#/c/431401/ is the same with yanyan's 13:32:45 got it 13:32:56 we can move on then 13:32:57 so please keep working on it 13:32:57 To delete all actions when we delete a cluster/node 13:33:09 OK 13:33:32 nice, lets move on 13:33:35 document 13:33:50 Qiming worked hard on document I think 13:33:55 need eyes on review 13:33:59 and the one for lb policy has been done 13:34:08 I want all policy types documented before final RC 13:34:12 yes, please help to review them, thanks a lot 13:34:15 Yes, it's a hard work 13:34:23 then during the last week, we'll focus only on critical bugs 13:34:51 sounds great 13:35:04 ok, next one 13:35:06 Senlinclient 13:35:23 from release time's point of view, all RCs are just practices for the final release 13:35:31 yep :) 13:35:40 the ultimate goal is to make the final release stable 13:36:21 I'll try get the 3 other policy types documented in the next two days 13:36:21 hi, XueFengLiu, so the functional test for client has not been started? 13:36:35 Qiming, thank you so much, will help to review 13:36:44 Yanyan, sorroy not time to do now 13:36:49 no problem 13:36:50 yanyanhu, I think we need your expertise in setting up the gate job 13:37:18 yes, I think I will try to spend some time on it 13:37:23 to set up the gate job 13:37:37 and then XueFengLiu can keep working on the test case 13:37:51 No problem 13:37:53 will start this work in this week 13:37:56 :) 13:38:00 would be greate 13:38:07 :) 13:38:18 ok next one 13:38:19 wrting test case can be split across the team 13:38:19 profile 13:38:33 Qiming, yes, that can be done in parallel 13:38:42 haiwei is not here? 13:38:44 Yes 13:39:04 ok, lets skip it. Although that is one of the most important TODO for us in next cycle 13:39:10 another two patches need your review 13:39:23 they are about support status at client side 13:39:37 Qiming, yes, saw them, will check them tomorrow 13:39:50 $ senlin policy-type-list 13:39:50 WARNING (shell) "senlin policy-type-list" is deprecated and will be removed by Apr 2017, please use "openstack cluster policy type list" instead. 13:39:51 +--------------------------------+---------+----------------------------+ 13:39:51 | name | version | support_status | 13:39:51 +--------------------------------+---------+----------------------------+ 13:39:52 | senlin.policy.affinity | 1.0 | SUPPORTED since 2016.10 | 13:39:54 | senlin.policy.batch | 1.0 | EXPERIMENTAL since 2017.02 | 13:39:56 | senlin.policy.deletion | 1.0 | SUPPORTED since 2016.04 | 13:39:58 | senlin.policy.health | 1.0 | EXPERIMENTAL since 2017.02 | 13:40:00 | senlin.policy.loadbalance | 1.0 | SUPPORTED since 2016.04 | 13:40:02 | senlin.policy.region_placement | 1.0 | EXPERIMENTAL since 2016.04 | 13:40:04 | | | SUPPORTED since 2016.10 | 13:40:07 looks nice 13:40:08 | senlin.policy.zone_placement | 1.0 | EXPERIMENTAL since 2016.04 | 13:40:10 | | | SUPPORTED since 2016.10 | 13:40:12 +--------------------------------+---------+----------------------------+ 13:40:14 this is the output 13:40:26 sorry, I broke the line :) 13:40:38 this looks great 13:40:50 you broke nothing at my side 13:40:57 :P 13:41:11 it is all caused by connection delay 13:41:18 yes 13:41:54 XueFengLiu, elynn, please help to review them as well if you guys have time :) 13:42:03 ok, those are all items in the list 13:42:14 OK 13:42:16 #topic Ocata RC3 for senlin and final release for senlin-dashboard 13:42:24 will 13:42:33 so this week will be the time for senlin rc3 and final release for senlin-dashboard 13:42:57 plan to cut the release by this Tuesday 13:43:08 come on 13:43:17 today is tuesday 13:43:19 today is Tuesday 13:43:20 so if there is any critical fix, please push it 13:43:24 sorry... 13:43:27 misclicking 13:43:31 Thursday 13:43:59 ate too much chocolate 13:44:19 :) 13:44:44 bad for your brain, boy 13:44:44 ok, so open discussion now 13:44:50 Qiming, agree... 13:44:59 #topic open discussion 13:45:25 I'm really curious about the mistral workflow progress 13:45:25 any more want to discuss? 13:45:34 oh, btw, our final mascot has been ready 13:45:42 o? 13:45:46 but ... I think the girl may be dating ... 13:45:51 Qiming, I think xinhui will talk with mistral guys in PTG 13:45:58 hope she can bring back more news 13:46:05 https://www.dropbox.com/sh/b8kpj5k29eaf65t/AADoIo9VUz14A81c--UTRtLNa?dl=0 13:46:05 WoW 13:46:13 good for her :P 13:46:40 elynn, will forward you guys the mail about Mascot 13:46:43 it's cool 13:46:50 the link I pasted above 13:46:55 What's that? 13:47:01 that is the mascot for Senlin 13:47:04 An unknown webside 13:47:06 the final version 13:47:16 sigh,I can't open that link... 13:47:21 at home 13:47:23 need proxy I think 13:47:27 yes 13:47:30 seems so 13:47:59 can you access https://snag.gy/ ? 13:48:02 That is the final version? 13:48:08 elynn, think so 13:48:32 Qiming, yes 13:48:34 that is accessible 13:48:45 Looks sweet, looks like tree family. 13:48:53 okay, I'm pasting to that site 13:48:56 haha 13:49:14 the link will be .. 13:49:17 .. 13:49:22 (uploading ) 13:49:24 .. 13:49:27 .. 13:49:30 https://snag.gy/ZUNIkV.jpg 13:49:54 saw it, haha 13:50:04 haha 13:50:29 so many .... trees 13:50:31 :D 13:50:35 thanks a lot for the effort of Heidi and her team 13:51:08 Where will we put that logo? 13:51:16 anywhere 13:51:23 when you need to tell a story about it 13:51:38 you can print it on your T-shirt 13:51:39 :) 13:51:57 That's a good idea :D 13:52:01 For example senlin wiki 13:52:14 right, right 13:52:18 I should have done that 13:52:56 noted, as a work item in etherpad 13:53:15 thanks :) 13:53:34 ok, so that's all topics for today 13:53:38 any more for discussion? 13:53:45 no from me 13:54:00 nope 13:54:15 ok, thanks all you guys for joining 13:54:19 In wiki we'd better add two things:1.tell peoplesome link need vpn 2.add some Frequently Asked Questions 13:54:45 XueFengLiu, +1, FAQ is important 13:54:59 :) 13:55:15 added to etherpad 13:55:24 thanks, Qiming 13:55:38 No need much ,just several FAQ 13:56:13 ok, so happy valentine's day to you all and have a good night:) 13:56:36 u 2 13:56:37 good night, all:) 13:56:39 bye 13:56:44 bye 13:56:45 good night 13:56:48 #endmeeting