16:00:27 #startmeeting Mistral 16:00:27 Meeting started Mon Oct 12 16:00:27 2015 UTC and is due to finish in 60 minutes. The chair is rakhmerov. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:28 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:30 The meeting name has been set to 'mistral' 16:01:08 hello 16:01:09 <^Gal^> hi 16:01:15 <_gryf> hi 16:01:21 hi 16:01:23 Hi 16:01:25 ^Gal^: welcome back :) 16:01:32 to you all ) 16:01:33 <^Gal^> Hi :) thanks! 16:01:52 let's start 16:02:07 hi, guys 16:02:09 by the time we get to interesting things I hope all people will join 16:02:11 hi! 16:02:24 #topic Review Action Items 16:02:41 1. rakhmerov, NikolayM: review https://review.openstack.org/#/c/230237/ 16:02:43 done 16:03:00 #topic Current status (progress, issues, roadblocks, further plans) 16:03:15 let's report the status quickly, if you have something 16:03:26 <^Gal^> I'm back in work after a vacation at Vietnam which was great. 16:03:26 <^Gal^> Last week I encountered many problems with my Ubuntu VM, so I've reinstalled the entire partition from scratch (5 times .....) 16:03:26 <^Gal^> Sadly I've lost every of execution update code 16:03:49 oooh, gosh!! 16:03:51 really?? 16:04:02 <^Gal^> yeah :( 16:04:10 F...k 16:04:25 <^Gal^> yeah it sucks, no sectors nothing 16:04:26 ok, let it be the worst problem in your life ) 16:04:49 ok, let's go on 16:04:56 ^Gal^: maybe you can create you own branch in github :-) 16:05:10 <^Gal^> I could even use our git :) 16:05:20 <^Gal^> have me some remote brances 16:05:33 I was looking through our updated docs and filed bug for remarks, also I've tested UI and I've prepared list of things which I don't like, I am going to file bugs 16:05:41 yes, you can actually just send WIP patches 16:05:46 this is a solution as well 16:05:48 <^Gal^> akuznetsova: sounds great 16:06:01 ok, be careful moving forward 16:06:34 my status: 16:06:34 * Fix the error message displayed when authentication failed with auth_enabled=True 16:06:34 * thinking about what Mistral could do in M dev cycle(or later), wrote all my thoughts in the etherpad 16:06:36 my status: getting Launchpad into order (Liberty series, BP statuses etc.), sent a few simple patches, reviewed a bunch of patches and was 50% involved into internal things 16:07:08 I'm about to write the doc for service api in next couple days, but i am not sure whether i can finish it before release day 16:07:09 xylan_kong: I saw what you wrote, looks good. Looking forward to discuss all this in person at the summit 16:07:21 rakhmerov: ok, i hope so 16:07:30 ^Gal^, here is a list http://paste.openstack.org/show/476036/, any objections ? 16:07:59 xylan_kong: it would be great if you completed this doc before release 16:08:13 <^Gal^> akuznetsova: I'm taking a look 16:08:37 btw, I'm planning to make the release approximately on Wed just so we have a couple of days for QA 16:08:40 rakhmerov: i'll try, i hope i won't be interrupted by internal issues :-) 16:08:55 <^Gal^> akuznetsova: looks great!! 16:09:07 :) 16:09:09 <^Gal^> in my eyes all of these are bugs 16:09:18 ^Gal^, ok, will file bugs for every item)) 16:09:31 ^Gal^: seems our UI guys will be busy 16:09:40 <^Gal^> lol :D 16:09:42 :)) 16:09:48 hope so 16:09:58 <^Gal^> hehe :) 16:10:00 ok 16:10:09 akuznetsova: thanks for the testing 16:10:15 hi everyone 16:10:22 any roadblocks, issues in your work you'd like to discuss? 16:10:23 hi, NikolayM 16:10:33 sorry, I'm late 16:10:34 hi NikolayM 16:10:36 ^Gal^: not asking you :) I got your issue 16:10:48 <^Gal^> rakhmerov: lol 16:10:48 NikolayM: what's your status? 16:11:35 my status: Finished working on ssh connectivity problem (now it is possible using proxy VM) 16:12:00 Fixed hardcoded mistralclient version 16:12:08 ok, btw there's an agenda item to discuss your patch 16:12:13 NikolayM: do you have use case to use proxy VM? 16:12:17 ok 16:12:20 for ssh 16:12:37 xylan_kong: yes, let's talk about it tomorrow, I'll explain you 16:12:43 xylan_kong, yes 16:12:51 rakhmerov: ok, all right 16:12:53 ok 16:13:33 in two words: it's one of the ways to solve the problem with accessing guest network from management network 16:13:44 we just have a proxy VM with a floating IP 16:13:57 ooh it's what we talk about in vancouver? 16:14:10 I'm not saying it's the best way (there others that we're planning to explore) but we need it for our internal stuff 16:14:12 hmm, sounds like something Zaqar does 16:14:15 LimorStotland: yes 16:14:22 I know it sound familiar :-) 16:14:40 xylan_kong: exactly, Zaqar is another possible way 16:15:32 but we need it fast whereas we don't have a nice integration with Zaqar now and I'm not really aware of Zaqar status at this point 16:15:43 but we will implement Zaqar way as well 16:15:55 ok, I'm actually getting into details )) 16:15:59 let's continue 16:16:04 :-) 16:16:16 it's ok 16:16:23 #topic Official Liberty Release health 16:16:46 so I think we're mostly ok, in short 16:17:33 rakhmerov: seems there are some mistralclient bugs still not set milestone 16:17:34 there are still some bugs in progress but I need to clarify their status 16:17:41 e.g. https://bugs.launchpad.net/python-mistralclient/+bug/1502840 16:17:41 Launchpad bug 1502840 in python-mistralclient "Http Response is shown, In case of 'OS_AUTH_URL' is not set." [Undecided,Fix committed] - Assigned to hardik (hardik-parekh047) 16:18:06 yep, it's because there's not an open milestone at client LP 16:18:18 I'll fix it tomorrow 16:18:24 ok 16:18:42 #action rakhmerov: make sure client bugs are assigned to milestones 16:19:14 btw, this is a Liberty-rc2 page: https://launchpad.net/mistral/+milestone/liberty-rc2 16:19:40 Documentation is mostly fine, akuznetsova fully reviewed it and created a bug with all the remarks about the doc 16:19:51 so doc BP is going to be closed 16:20:40 rakhmerov, correct, here is a link https://bugs.launchpad.net/mistral/+bug/1504582, let's assign it to corresponding milestone 16:20:40 Launchpad bug 1504582 in Mistral "Remarks to updated documentation" [Undecided,New] 16:21:06 so my plan is to have 1 day (tomorrow) make final bugfixes/adjustements, then backport needed patches made after RC1 into stable/liberty and release RC2 which we're planning to test via QA 16:21:20 akuznetsova: yep, thanks 16:21:53 done 16:21:56 thanks 16:22:25 ^Gal^: btw, is https://blueprints.launchpad.net/mistral/+spec/mistral-docs-troubleshooting done? 16:22:49 we recently merged your patch which was related to this but I wasn't sure if it covers this BP completely 16:23:09 <^Gal^> hmm 16:23:18 <^Gal^> what I covered was mistral-dashboard stuff 16:23:58 <^Gal^> by mistral-dashboard side, yep we're done 16:24:18 ok, then I need to check that BP again 16:24:26 <^Gal^> ok 16:24:47 #action rakhmerov: check if https://blueprints.launchpad.net/mistral/+spec/mistral-docs-troubleshooting is completed (it's finished from UI side) 16:25:19 ^Gal^: what about UI BPs? Moving to Mitaka? 16:25:33 <^Gal^> yeah 16:25:39 I guess they won't be completed in 1 day 16:25:41 ok 16:26:29 so, generally I'm planning to have official Liberty release on Thu or Friday 16:26:34 most likely Fri 16:27:06 guys, do you have any concerns or questions? Or anything you want to do before the release? 16:27:10 <^Gal^> wooohoo congrats 16:27:28 it's high time you said if you need something 16:28:18 rakhmerov, what about this bp https://blueprints.launchpad.net/mistral/+spec/mistral-async-actions-doc ? 16:28:40 akuznetsova: it's on my TODO list for this week 16:29:15 rakhmerov, ok, please, set status, assignee and priority 16:29:32 yes, just did 16:29:34 thanks 16:29:50 will do my best to get it done 16:29:58 anything else? 16:30:09 counting to 5 and go the next topic 16:30:10 1 16:30:12 2 16:30:14 3 16:30:16 4 16:30:19 5 16:30:21 sold! 16:30:24 ok 16:30:26 :D 16:30:34 :-) 16:30:40 #topic Discuss https://review.openstack.org/232507 16:31:08 NikolayM, xylan_kong: let's dicuss it 16:31:25 Yes, I want to discuss this patch 16:31:27 yeah, ok 16:31:32 so, xylan_kong, looks like NikolayM addressed most of your concerns 16:31:48 I think we could get rid of tempest at all 16:31:53 except one: these tests can't work w/o OpenStack as long as we use tempest 16:32:09 or we won't be able to run tests w/o OpenStack 16:32:14 NikolayM: get rid of tempest? what's the alternative? 16:33:00 rakhmerov, there were a couple of reasons why these tests were write using tempest 16:33:09 as far as I understand, if we set auth to false then we can run these tests w/o OpenStack, right? 16:33:12 isn't it enough? 16:33:22 alternative is to create our own library for requests (like tempest) 16:33:28 akuznetsova: I know, yes 16:33:29 NikolayM: no, if we continue to be in OpenStack big tent, we should use Tempest for our functional testing 16:33:34 no, it is just kind of workaround 16:33:52 xylan_kong, that's not quite right 16:33:56 especially, one day, if we want mistral test case to be included in DefCore :-) 16:34:05 yep, then we don't need to get rid of tempest 16:34:25 I agree with xylan_kong here, we should try to be aligned with OpenStack practices 16:34:36 unless it conflicts with common sense :) 16:34:36 yes 16:35:02 rakhmerov, as far as I know, heat and sahara do not use tempest for real functional tests 16:35:02 xylan_kong, what about your question about running tests without OpenStack? 16:35:11 for mistral functional tests, if it must work with OpenStack, we should define skip_check method 16:35:32 if OpenStack is not installed, then just skip the tests 16:35:57 xylan_kong: how does a test know that? 16:36:05 if OS is installed or not 16:36:09 just a config option? 16:36:13 via configuration 16:36:15 yes 16:36:16 right 16:36:24 ok 16:36:24 like tempest does now 16:36:27 tempest does not work without Openstack 16:36:52 but tempest can work without a perticular service 16:37:00 akuznetsova: you mean there's no config option to disable OpenStack in tempest? 16:37:20 hm.. sorry guys, I'm not that familiar with tempest 16:37:23 can you clarify this? 16:37:24 tempest is created for OpenStack tests :) 16:37:42 rakhmerov, correct, tempest can't work with OS, it authorize in keystone 16:37:46 akuznetsova says tempest can't be used w/o OpenStack, xylan_kong says somehow we can do it 16:37:48 *without 16:37:52 where is the truth? :) 16:38:02 akuznetsova:let me make it clear 16:38:12 yes, please do 16:38:14 I and i think this flug can be useful not only for testing but for all the stand alone feather 16:38:21 tempest has config options meaning that one or another service is available 16:38:39 if we want to run mistral functional tests, we must install tempest first, right 16:38:46 LimorStotland: I didn't quite understand you 16:39:53 xylan_kong: ok, and? 16:40:25 then we got tempest configuration, and we can config items to decide if the services mistral needs is there, if no, we just skip the tests which need OpenStack installed 16:40:51 am I right? akuznetsova NikolayM ? 16:41:14 if i am following this conversation (and i am not so sure i am any more :-) ) xylan_kong want to add fulg for the config if we have openstack or not 16:41:23 is this correct? 16:41:26 take https://review.openstack.org/232507 for example, you need nova and neutron 16:41:28 ok 16:41:49 if anyone of them is not installed, then we skip the tests 16:41:56 neutron is not necessary here, I think 16:42:16 Can you imagine OpenStack without compute service? 16:42:48 or, that is correct case? 16:43:16 xylan_kong: ok, when we use this decorator we just need to specify a dependency on particular services, right? 16:43:31 rakhmerov: yes 16:43:35 LimorStotland: yes, sort of ) 16:43:36 NikolayM: hmm, let me think 16:44:30 NikolayM: so, you just want to decide whether OpenStack is installed or not, right? don't care about the specific service? 16:44:53 xylan_kong, generally yes 16:45:05 as far as I understand such flags work only for tests which located in tempest repo 16:45:12 then we can just set a dependency from Nova, right? 16:45:31 meaning that Nova always exists if OpenStack is installed 16:45:32 if you want to run tests just without OpenStack (even without auth), take a look at run_functional_tests.sh script 16:45:36 can we do this? 16:46:34 guys, our main problem is not nova, neutron, etc, keystone - tempest always use authorization 16:47:10 I'm confused. Someone of you told that auth can be disabled 16:47:26 if so we can disable it and run w/o OpenStack, no? 16:47:26 rakhmerov, yes, using some hacks 16:47:32 what hacks? 16:47:44 we can disable auth in api tests 16:48:02 can we make it configurable? 16:48:05 but how we can disable auth in tests for openstack actions ? 16:48:17 but keystone should always be installed because all services are using authorization.so i don't understand the problems kuznetsova 16:48:45 LimorStotland: that's exactly what we're trying to agree on 16:48:59 akuznetsova sorry google translate :-) 16:49:03 xylan_kong, what tests you want to skip ? 16:49:05 we need to be able to run tests w/o OpenStack 16:49:12 we're just trying to understand how 16:49:19 rakhmerov, which tests ? 16:49:33 sorry i am behind :-( 16:49:35 functional tests 16:49:43 api / cli / tests for actions ? 16:50:19 akuznetsova: i mean, if we have test case needs openstack installed, and we want to run the tests without openstack, there must be a mechanism to skip 16:50:24 ok, guys, we're running out of time so let me please drive the following conversation 16:50:43 the goal we're trying to achieve: 16:50:49 xylan_kong, rakhmerov, we have run_functional_tests script 16:50:59 wait a sec 16:51:05 which can run API functional tests without auth at all 16:51:09 1. We have functional tests 16:51:23 2. Part of the func tests are not related logically with OpenStack 16:51:34 3. Part of them are related with OpenStack 16:52:03 yes, API tests generally are not related to OpenStack 16:52:20 xylan_kong, problem is that now we should skip all tests ) because they use tempest, tempest uses authorization in keystone before tests running 16:52:23 4. We want to be able to run functional test with OpenStack (full set of tests) and without OpenStack so that in this case all test dependent on OpenStack are just skipped 16:52:46 xylan_kong, akuznetsova, NikolayM: do you all agree with my points? 16:52:50 is this our goal? 16:53:45 rakhmerov, yes, but I don't see now how we can run tests which use tempest without auth and without hacks 16:53:52 for me 'run_functional_tests' is fine 16:53:59 rakhmerov: it's aslo my understanding, akuznetsova NikolayM we can discuss this in another time 16:54:15 NikolayM: how does it work? that script 16:54:32 how does it bypass the problem that akuznetsova describes 16:54:40 as akuznetsova said, with some kind of hacks 16:54:52 rakhmerov, using that https://github.com/openstack/mistral/blob/master/mistral/tests/functional/base.py#L220 16:55:33 overring and mocks 16:55:36 NikolayM: if we run this script what happens tests that depend on OpenStack? 16:55:59 rakhmerov, script runs only api tests 16:56:07 It doesn't run tests depending on Openstack 16:56:13 only API tests 16:56:20 how does it distinguishes them? 16:56:20 yes, thanks akuznetsova 16:56:28 by tests path 16:56:46 only tests/api directory 16:57:00 oooh... 16:57:04 I see now 16:57:13 xylan_kong: do you think it's acceptable for us? 16:57:23 NikolayM: how our gate distinguishes them? 16:57:25 or you disagree with this approach? 16:57:50 xylan_kong: he just told about it, it just looks at their path 16:58:11 gate does not run them 16:58:28 our dsvm gate run all tests 16:58:57 https://github.com/openstack/mistral/blob/master/run_functional_tests.sh#L17 16:58:58 yep, but it runs all tempest tests (without this script) 16:59:03 rakhmerov, xylan_kong we do not use run_functional_tests in our gates 16:59:10 I know, yes 16:59:35 I am just asking Lingxian if this is ok to run this script if we need to run func tests w/o OpenStack 16:59:37 <_gryf> maybe the question is: are there any tests which are not run on the gate, but ran by the sctipt? 16:59:55 _gryf, no 16:59:56 _gryf: now 16:59:58 no 17:00:01 akuznetsova, NikolayM you mean OpenStack related tests won't be run in our gate? 17:00:11 brrrr.... 17:00:13 :))) 17:00:15 no 17:00:26 they run at our gate 17:00:35 ok, guys, sorry, we ran out of time 17:00:36 ok, i got 17:00:41 <_gryf> so, what's the point of discusing the script on the first place? 17:00:42 no, I mean on gate that script is not used 17:00:46 xylan_kong: let's continue tomorrow in IRC 17:00:56 :-) 17:01:03 <_gryf> it;s role is to mock the lack of openstac, no? 17:01:12 rakhmerov, NikolayM, akuznetsova thanks all of you 17:01:18 for patience 17:01:32 _gryf: we just need an ability to run tests w/o OpenStack 17:01:42 those that don't depend on OpenStack at all 17:01:47 <_gryf> rakhmerov, on the gate? 17:01:53 that's what that script does 17:01:58 _gryf, locally 17:02:00 no, not on the gate 17:02:03 locally 17:02:17 <_gryf> hm. ok. 17:02:28 cause mistral can work w/o OpenStack 17:02:30 ok, let's end the meeting. We're occupying the channel 17:02:47 thanks for joinging! see you tomorrow in IRC 17:02:50 bye 17:02:51 <^Gal^> bye have a good week 17:02:52 thanks guys, appreciate your explanation 17:02:54 bye 17:02:55 <_gryf> bye 17:02:56 bye 17:02:57 sure 17:03:03 #endmeeting