13:01:23 #startmeeting senlin 13:01:24 Meeting started Tue Dec 1 13:01:23 2015 UTC and is due to finish in 60 minutes. The chair is Qiming. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:01:25 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:01:27 The meeting name has been set to 'senlin' 13:01:37 evening 13:01:43 hello 13:01:46 Hi 13:01:47 hi 13:02:00 please check the agenda and see if you have got things to add 13:02:03 #link https://wiki.openstack.org/wiki/Meetings/SenlinAgenda 13:02:27 #topic mitaka work items 13:02:36 let's start with the etherpad, :) 13:02:40 as always 13:02:53 #link https://etherpad.openstack.org/p/senlin-mitaka-workitems 13:03:09 heat resource types 13:03:28 elynn, yours 13:03:53 hi 13:04:00 Qiming: 13:04:01 Hi 13:04:22 I just update one patch for it and unittests are passed. 13:04:24 I think the patches were blocked somehow by the bugs introduced into senlinclient 13:04:51 I will continue to update other patches ASAP. 13:05:05 we will be doing more code churn to the client code as we are revising the apis 13:05:35 Will it affact senlin resources? 13:05:52 elynn, I believe there will be some impact 13:06:13 for example, when you do a cluster-create, we plan to return a 202 (request accepted) 13:06:42 in that response body, I'm not sure whether we should include a representation of the cluster 13:07:02 Does it return 202 now? 13:07:05 I have checked HTTP spec and api-wg, so far no guidance there 13:07:14 currently not 13:07:34 but in theory, we should return 202, because the cluster creation may take a long time 13:08:01 it is simply wrong to return 200, i.e. to pretend the cluster has been created 13:08:14 what is the topic now, sorry, missed a lot 13:08:30 haiwei_, heat resource type support 13:08:45 ok 13:08:46 Will it return at least a cluster id? 13:08:49 hmmm... i don't think a representation of the cluster is necessary on a 202 13:09:21 in an 202 response, we will add a 'Location' header, where you can get a link to the cluster resource being created 13:10:28 I'm still not sure if we should teach SDK to parse the response header and then do a GET again 13:10:33 ok, I will slow down, after you finish revise APIs, I need do more tests. 13:10:37 it seems the 'Location' header is the http response header 13:11:10 I confirmed it with the api-wg guys 13:11:13 All the resource creation will return 202 only? like profile 13:11:22 yes, most http related packages used today allow you to check the header and behave accordingly 13:11:43 no, elynn, profile-create will return a 201 13:11:52 that creation is for real 13:12:11 https://review.openstack.org/#/c/251771/ 13:12:25 this is the revision, a first step 13:12:25 If SDK don't continue to parse the response header, it might be difficult to handle in heat 13:12:32 yes 13:12:37 that is a big concern 13:13:02 I'd propose something into SDK so that the heat side code doesn't have to look a lot different 13:13:10 If I remember correctly , instance creation only returns 202. 13:13:38 elynn, if that is true, we can check how heat gets instance id today 13:14:10 if necessary, we may have to cap the senlinclient version 13:14:46 https://github.com/openstack/heat/blob/master/heat/engine/resources/openstack/nova/server.py#L740-L763 13:14:47 current version 0.1.8 looks good? 13:15:12 haven't test it yet. 13:15:31 elynn, that code is not telling you the real things happened behind the scene 13:15:39 may need to check novaclient code 13:15:54 Qiming: yes, I think novaclient handles the location. 13:16:12 okay, if that is the case ... 13:16:17 http://developer.openstack.org/api-ref-compute-v2.html#createServer 13:16:35 we may need to check with SDK team to come up with a workaround 13:16:50 personally, I hate all xyzclient packages 13:17:03 :) 13:17:06 moving on 13:17:19 senlinclient, one of the things I hate, :) 13:17:35 haiwei_ just proposed test case part 2 13:17:52 https://review.openstack.org/251798 13:18:02 please help 13:18:13 yes, jenkins failed 13:18:19 will check it later 13:18:32 ya 13:18:50 API response modification 13:19:13 patch #251771 is a global revision 13:19:13 Qiming: https://review.openstack.org/#/c/251771/ - Fix status code returned from API 13:19:28 need eyes on them too 13:19:50 404 vs 400 return code, I think we are done with it 13:20:04 good 13:20:22 health policy 13:20:27 again 13:20:40 i will have a look today 13:20:53 I believe lixinhui is looking into that as well 13:21:06 yes 13:21:13 thx, jruano 13:21:16 we at least need a bp for this 13:21:43 Triggers 13:21:48 What will health policy do exactly? 13:22:00 Triggers are history now 13:22:13 ah... yep, i figured that was coming 13:22:28 looking at monasca, and thats what it felt like 13:22:46 elynn, I think that's what we need to discuss in BP :) 13:23:05 health policy is basically a Senlin inherent mechanism to detect node status and take proper actions to fix node failures if told so 13:23:27 yanyanhu, agreed 13:23:50 oh, I C. 13:24:10 Profile 13:24:20 ah, my job 13:24:32 I spent some times on completing related support in sdk side 13:24:50 currently, server rebuild and server resize related function calls has been added 13:24:58 is there a bp for profile? 13:25:02 both merged? 13:25:07 yes 13:25:51 so it can meet the requirement of supporting property update of flavor and some basic properties of os.nova.server prfile I think 13:25:53 \o/ 13:26:21 update of flavor is working too? 13:26:22 and network and image propety update support has been done in senlin 13:26:41 I'm still working on it 13:27:00 but related support in sdk has been there 13:27:05 okay, network and image update are not on the etherpad, so I guess they are all solved 13:27:17 yea, I removed them :) 13:27:31 cool, you may want to paste the patch link into etherpad 13:27:40 Receiver 13:27:46 sure, will add it 13:27:53 jruano, we need a bp on this 13:27:54 im working on the bluprint for receiver 13:28:13 i was out last week. i have a design that im working on now 13:28:38 okay, be it a webhook or a message queue, the parameters are almost the same I think 13:29:05 jruano, happy thanksgiving holidays, to you and your family, :) 13:29:09 yes. i am going to lay out a class hierarchy that abstracts webhook 13:29:13 thanks 13:29:19 yea, happy thanksgiving :) 13:29:36 then we can extend additional implementations from that as needed 13:29:38 jruano, it doesn't have to be a complex class hierarchy I think 13:29:46 yep, i think it is pretty simple 13:29:59 we just need a generalized interface that will be accommodating to future extensions 13:30:02 webhook has probably everything there 13:30:04 yep 13:30:07 great 13:30:17 working on that today 13:30:17 lock breaker 13:30:35 The patches are merged. 13:30:43 all of them? 13:30:54 yes, this afternoon. 13:31:18 Is it better to add functional test for it? 13:31:24 great, that means some bugs on launchpad can be labeled 'invalid' now? 13:31:26 elynn, that will be great 13:31:33 elynn, good idea 13:31:49 I will start to work on them ;) 13:31:55 BTW, there still some failures in current functional test. I will spend some time on it in coming week 13:32:17 to ensure you have a clean basement for new test cases 13:32:20 Will we enable them by default? 13:32:43 elynn, hmm, maybe not now I think 13:33:12 but we do recommand check it manually before merge new patches 13:33:30 recommend, sorry 13:33:36 The command is check experiment? 13:33:40 yep, haiwei_ is still seeing some weird errors from keystoneauth1 13:33:43 check experimental 13:33:56 cool, got it. 13:34:25 Qiming, great, the recent change about authentication caused some trouble :) 13:34:27 anything else about items on etherpad? 13:34:37 Qiming, the error I said this afternoon is my own fault, sorry 13:34:57 haiwei_, so happy you made mistakes 13:35:07 I mean it is not the code's problem, :) 13:35:19 I often made it 13:35:21 moving on 13:35:34 haiwei_, keep on good habits then 13:35:41 #topic bug clearance 13:36:00 #link https://bugs.launchpad.net/senlin/?field.searchtext=&orderby=-importance&field.status%3Alist=NEW&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&assignee_option=any&field.assignee=&field.bug_reporter=&field.bug_commenter=&field.subscriber=&field.structural_subscriber=&field.tag=&field.tags_combinator=ANY&fiel 13:36:00 d.has_cve.used=&field.omit_dupes.used=&field.omit_dupes=on&field.affects_me.used=&field.has_patch.used=&field.has_branches.used=&field.has_branches=on&field.has_no_branches.used=&field.has_no_branches=on&field.has_blueprints.used=&field.has_blueprints=on&field.has_no_blueprints.used=&field.has_no_blueprints=on&search=Search 13:36:01 btw, does anyone notice that when creating/deleting a node TIMEOUT error will happen first and then creation/deletion will still succeed 13:36:27 haiwei_, I didn't 13:36:36 me neither... 13:36:46 such a long url 13:37:21 anyway, we still have 1 critical bug, 3 medium (1 in progress), all others are to be revisited 13:37:29 ok, if you meet that error just share to us, maybe something wrong there 13:37:56 in total we have 19 bugs, including 3 already marked 'incomplete' 13:38:00 less than 20 bugs are in open status 13:38:03 seems we need bugs 13:38:05 mine should have a patch checked in today. i was almost done with it heading into last week 13:38:18 jruano, no worry 13:38:31 we will need more bugs 13:38:51 How? 13:39:13 elynn, test and find bugs, certainly 13:39:57 Ha~ Or we can buy some from store ;) 13:40:09 #action Qiming to go through the bug list and triage/confirm bugs 13:40:30 the critical one is not one bug I think, it should be many bugs, but most are fixed now 13:40:34 elynn, :P 13:40:58 haiwei_, the problem is we were not linking back those patches to the bug 13:41:12 moving on 13:41:20 #topic Mitaka release schedule 13:41:21 Seems patch will not automatically link to bug. 13:41:27 yes, it'S better not report this kind of bugs 13:41:29 #link https://wiki.openstack.org/wiki/Mitaka_Release_Schedule 13:41:44 Even I write bug number in commit message. 13:41:51 maybe you are aware of it, mitaka-1 release is happening now 13:41:52 you have to add the Close-bug in commit message 13:41:57 ah 13:42:22 this is week 'R-18' for the mitaka cycle 13:42:39 I don't think we are ready to do a release 13:42:47 agree 13:43:08 so far, in project config, senlin is tag as "release:independent" 13:43:19 that is buying us some time to catch up, but not for ever 13:43:35 I hope we can catch up soon 13:44:26 once the API is stable, we can start the formal release and apply for "release:whatever-official" tags 13:44:30 I think maybe we can try to give the first release after API rework is done 13:44:43 yea 13:44:52 maybe by mitaka-2, we can start practicing a formal release 13:45:15 yes, it will be Jan 19-21, 2016 13:45:29 we won't miss next milestone 13:45:49 moving on 13:46:00 #topic recheck blueprints 13:46:13 https://blueprints.launchpad.net/senlin/+specs?show=all 13:46:48 okay, we need some approvals to the bps 13:47:00 some are outdated 13:47:13 em ... we need more cores 13:47:21 yes. And set some of them to obsolete status 13:47:52 core team please keep your eyes on commits and reviews 13:48:05 yes 13:48:10 yep... the trigger one for monasca i have can be removed now 13:48:13 we need more people 13:48:23 yep, jruano 13:49:02 #action review blueprints and approve/obsolete them as needed 13:49:39 #topic open discussions 13:50:39 can I leverage some functional tests 13:50:46 for the scalability testing? 13:50:52 oh, plz remeber to use reno to generate release note related info when proposing new patch which adds new feature 13:51:31 is there a guideline of how to use reno? 13:51:35 lixinhui, sure. That will be great. 13:51:37 yanyanhu: Is there any example for how to use it? 13:51:40 ah... where can we get the info on that tool? 13:51:41 yep, yanyanhu that is a good habit to develop among the team 13:51:52 haiwei_, there is a sample patch qiming propose to illustrate how to use it 13:52:07 let me find it for you 13:52:20 i remembered it 13:52:39 https://review.openstack.org/#/c/249573/ 13:52:42 yea, this one 13:53:26 http://docs.openstack.org/developer/reno/ 13:54:05 ah ok... looks pretty straightforward 13:54:47 jruano, for first step, maybe just know how to use it to generate release note info with the patch :) 13:54:52 it is spreading the last minute report writing work into daily jobs 13:55:33 cool 13:55:41 by the way, I'm studying how to publish our docs (user and developer) to the official site 13:55:59 Great news 13:56:16 all those docs are very verbose ... believe me 13:56:16 you are a robot Qiming 13:56:26 Qiming, this is very very important :) 13:57:52 anything else? 13:58:06 nope 13:58:08 nope from me 13:58:19 no 13:58:30 not from me 13:59:13 http://snag.gy/5Uox3.jpg 13:59:20 in case you want to take a look 13:59:31 just pasted the generated doc page 13:59:35 very nice 13:59:36 \o/ 13:59:37 hi 13:59:44 Hi 13:59:46 let's get back to #senlin 13:59:46 #endmeeting