09:00:24 <gmann> #startmeeting qa 09:00:24 <openstack> Meeting started Thu Dec 1 09:00:24 2016 UTC and is due to finish in 60 minutes. The chair is gmann. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:00:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:00:27 <openstack> The meeting name has been set to 'qa' 09:00:41 <gmann> Hi, who's here today? 09:00:44 <dmellado> o/ 09:00:50 <masayukig> \o 09:01:20 <gmann> masayukig: dmellado hi 09:01:32 <gmann> let's wait for soem time for more people 09:01:53 <dmellado> yeah, let's wait a bit and otherwise maybe we can just wrap-up quickly 09:02:00 <masayukig> +1 09:02:19 <gmann> yea 09:03:54 <gmann> anyway let's start 09:04:09 <gyx_> ./ 09:04:28 <gmann> gyx_: hi 09:04:32 <dmellado> hi gyx_ 09:04:46 <gyx_> hi gmann 09:04:59 <gyx_> hi dmellado 09:05:10 <gmann> #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Agenda_for_December_1st_2016_.280900_UTC.29 09:05:22 <gmann> ^^ Today's agenda 09:05:47 <gmann> first we will go through the Ocata priority 09:05:48 <gmann> #link https://etherpad.openstack.org/p/ocata-qa-priorities 09:06:13 <gmann> I do not have update on my item. still to start :( 09:06:36 <gmann> dmellado: any chance you get to add spec on this - Add and decorator/comment into bug-reproducing tests to know actual bug number easily from test failure 09:06:40 <gmann> or oomichi planning 09:07:07 <dmellado> gmann: I can prepare a spec for that, yeah 09:07:21 <dmellado> I didn't get to work into that yet 09:07:24 <dmellado> need to sync with oomichi 09:07:31 <dmellado> but time zone's being tricky! 09:07:31 <gmann> mdbooth: cool. Thanks 09:07:34 <gmann> ops 09:07:38 <gmann> dmellado: ^^ 09:07:42 <dmellado> no worries gmann 09:07:52 <gmann> mdbooth: just ignore :) 09:08:05 <gmann> dmellado: ok, may be oomichi will check this log 09:08:09 <dmellado> gmann: on the next one Remove deep test class Inheritance for easy debugging 09:08:12 <gmann> next one is - Remove deep test class Inheritance for easy debugging 09:08:15 <dmellado> I'm gonna de-assign myself for that one 09:08:16 <gmann> dmellado: yea 09:08:18 <dmellado> as I'm overcommitted 09:08:29 <gmann> dmellado: ah yea, no issue 09:08:38 <dmellado> I'm syncing with mkoderer on removing the pinning from manila 09:08:39 <gmann> i can start may be next week 09:08:50 <mkoderer> hello all 09:08:52 <dmellado> and creating the tempest plugin for barbican 09:08:55 <dmellado> hey mkoderer 09:08:56 <gmann> dmellado: yea that will be nice if they can use master one 09:09:15 <dmellado> but that'd take *quite* some work, so it'd be one of my main focus on this cycle 09:09:15 <gmann> dmellado: barbican plugin in separate repo right? 09:09:19 <dmellado> gmann: yeah 09:09:24 <gmann> great 09:09:30 <mkoderer> gmann: https://review.openstack.org/#/c/404637/ 09:09:55 <gmann> dmellado: sure thanks 09:10:03 <gmann> mkoderer: cool. thanks 09:10:40 <gmann> other is bug triage we will discuss later 09:10:48 <gmann> #topic Specs Reviews 09:10:57 <gmann> #link https://review.openstack.org/#/q/status:open+project:openstack/qa-specs,n,z 09:11:25 <gmann> policy spec was good just few comment i gave and seems it was updated 09:11:29 <gmann> i will check that 09:11:34 <masayukig> so, what do you think "Show team and repo badges on README"? 09:11:48 <gmann> masayukig: actually i did not read that yet 09:12:16 <dmellado> by repo badges do you mean some ascii art or link? 09:12:39 <gmann> oh it is adding those 09:12:43 <masayukig> dmellado : it's svg 09:13:06 <masayukig> you can see it http://docs-draft.openstack.org/28/402528/1/check/gate-qa-specs-docs-ubuntu-xenial/586394f//doc/build/html/README.html#qa-specs-repository 09:13:34 <masayukig> (it's not on the top page of the doc, though) 09:13:40 <gmann> masayukig: nice, ll check after meeting 09:13:56 <masayukig> gmann : ok 09:14:04 <zhihui> hi 09:14:05 <dmellado> yeah, I'll have a look too 09:14:14 <gmann> other spec we discussed and is up for HA testing 09:14:15 <gmann> zhihui: hi 09:14:18 <gmann> #link https://review.openstack.org/#/c/399618/ 09:14:20 <masayukig> gmann dmellado : thanks :) 09:14:30 <gmann> this one will be good to review 09:15:19 * masayukig added it in my list 09:15:26 <gmann> masayukig: thanks 09:15:41 <gmann> test resource one was almost good to go 09:15:43 <gmann> #link https://review.openstack.org/#/c/173334/ 09:15:56 <gmann> m just waiting for milestone confirmation from oomichi 09:16:43 <gmann> anything else on specs? 09:17:22 <gmann> ok, let's move to next 09:17:26 <gmann> #topic Tempest 09:17:32 <shuyingya> Hi all, I want to consult a question. I use tempest/test.py get_client_manager method and build a auth_provider client. How can I use method addCleanup to delete the resource? 09:17:58 <gmann> shuyingya: hi 09:18:08 <gmann> shuyingya: you want to use on your test right? 09:18:22 <shuyingya> I found when I execute tempest run everytime, the test will create a user and project in my env 09:18:23 <shuyingya> yes 09:18:38 <shuyingya> gmann 09:18:50 <shuyingya> thanks for your reply ^^ 09:19:25 <shuyingya> self.client_manager_admin = BaseTestCase.get_client_manager('admin') 09:19:25 <shuyingya> auth_provider = self.client_manager_admin.auth_provider 09:19:37 <gmann> shuyingya: https://github.com/openstack/tempest/blob/32a9e308e65b99a80575354532dbca98dea5ea76/tempest/api/compute/admin/test_aggregates.py#L68 09:19:42 <gmann> shuyingya: this is one of the example 09:20:14 <shuyingya> ok, thanks. So I need delete the user_id? 09:20:20 <gmann> shuyingya: those are if any specific resource you want to delete automatically when test complete 09:20:21 <shuyingya> and project_id? 09:20:28 <masayukig> shuyingya , gmann : I think this kind of questions should be in #openstack-qa not here? 09:20:44 <gmann> shuyingya: ok, so let's discuss after meting 09:20:49 <gmann> masayukig: yea :) 09:20:53 <shuyingya> thanks 09:21:02 <shuyingya> gmann 09:21:03 <gmann> so on tempest 09:21:05 <gmann> #link https://review.openstack.org/#/q/project:openstack/tempest+status:open 09:21:11 <gmann> ^^ open review 09:21:24 <gmann> currently gate if blocked for layer4 job failure 09:21:45 <masayukig> yup 09:21:53 <gmann> thanks masayukig for digging into this 09:22:16 <masayukig> gmann : np :) 09:22:28 <gmann> thats mainly becasue of this -https://review.openstack.org/#/c/343023/ 09:22:39 <gmann> so we are reveting that as of now - https://review.openstack.org/#/c/405183/ 09:22:44 <dmellado> +1 09:22:46 <gmann> and ll check real cause 09:22:49 <masayukig> #link https://review.openstack.org/#/c/405183/ 09:23:01 <gmann> so devstack core check this soon :) 09:23:18 <gmann> oomichi: mtreinish ianw ^^ 09:23:38 <gmann> FYI- currently gate is block so recheck would not work 09:24:23 <gmann> anything else on Tempest side or we jump to bug ? 09:25:40 <gmann> this week, it was my turn on Bug triage 09:25:40 <gmann> #link https://etherpad.openstack.org/p/ocata-qa-bug-triage 09:25:40 <gmann> i did not get much time to cleanup more i tought 09:25:40 <gmann> #link https://etherpad.openstack.org/p/tempest-weekly-bug-report 09:25:40 <gmann> few of them i triaged or cleaned up 09:26:06 <masayukig> great :) 09:26:19 <dmellado> gmann: is there any bug that would need urgent attention? 09:26:38 <gmann> we have some ssh one 09:26:42 <gmann> but not sure as of now 09:26:51 * andreaf o/ 09:27:03 <gmann> one i brought earlier also and found bug also for that 09:27:07 <gmann> #link https://bugs.launchpad.net/tempest/+bug/1569562 09:27:07 <openstack> Launchpad bug 1569562 in tempest "Tempest plugin support for available services mechanism" [Medium,Confirmed] - Assigned to Ghanshyam Mann (ghanshyammann) 09:27:07 <gmann> andreaf: hi 09:27:11 <dmellado> hi andreaf 09:27:19 <andreaf> hi - sorry I'm late 09:27:37 <chandankumar> hello all! 09:27:38 <gmann> i will push a patch to make service available config option as list and get feedback 09:27:47 <gmann> that came during baremetal test removal also 09:27:50 <gmann> chandankumar: hi 09:27:52 <dmellado> oh, gmann 09:27:59 <masayukig> gmann : we should have filed the layer-4 job failure. 09:28:01 <dmellado> on that I've got a patch ongoing for adding that to the ironic plugin 09:28:15 <dmellado> #link https://review.openstack.org/#/c/379256/ 09:28:26 * masayukig waves andreaf chandankumar 09:28:32 <gmann> basically idea is to support plugin service available also so that plugin installation is not needed to check service availability 09:28:56 <dmellado> o/ chandankumar 09:29:04 <gmann> masayukig: ah, yea i will do after meeting 09:29:20 <masayukig> gmann : great, thanks :) 09:29:22 <dmellado> +1 09:29:28 <gmann> dmellado: but that you are adding in plugin 09:29:48 <dmellado> gmann: yeah, I just want to bring the point that not all the plugins support that 09:29:58 <dmellado> that's something that maybe we'll have to take into consideration as well 09:30:09 <dmellado> but +1 on your approach 09:30:30 <gmann> dmellado: yea 09:30:56 <gmann> dmellado: even when separate repo for plugin then it make very bad 09:31:32 <gmann> dmellado: anyways you can check thsi too - https://review.openstack.org/#/c/404436/ 09:31:39 <andreaf> gmann, dmellado: so if I understand the problem right, there's need to a "service available" flag that can be used by plugins to discover which services are available? 09:31:53 <gmann> andreaf: yes 09:31:59 <andreaf> gmann, dmellado: I kind of assumed that if a plugin is installed the service is there 09:32:16 <gmann> andreaf: issue is when they need other service availability which are under plugin config 09:32:37 <gmann> andreaf: yes, but if plugin in separate repo then it has to be installed 09:32:49 <dmellado> andreaf: in my usecase basically is in order to skip tests if the plugin is in-tree 09:33:01 <dmellado> but the user is using tempest with system-site-packages enabled 09:33:07 <dmellado> and don't want to trigger those tests by some reason 09:33:13 <gmann> andreaf: say ironic is there but other plugin nees ironic service check need to install ironic plugin if it is in separate repo 09:33:42 <tosky> yep, this may be solved with splitting all the plugins from the repos (as we hope it will happen for Pike) 09:33:51 <tosky> but right now it's more complicate 09:33:54 <dmellado> o/ tosky yeah, that will make things easier 09:33:55 <tosky> complicated* 09:33:57 <gmann> tosky: thanks 09:34:28 <dmellado> along with that, and sorry for the offtopic 09:34:30 <andreaf> ok I see - well it's up to the plugins to implement a "service available" flag on the plugin side - we can add a note in the docs for that 09:34:42 <dmellado> is there any plan to hand a cross-project session for this at the PTG? 09:35:02 <dmellado> andreaf: yeah, and I was thinking maybe at the cookicutter template level too 09:35:07 <gmann> andreaf: but they have to add their side as tempest would not have currently 09:35:32 <gmann> andreaf: and other plugin need their service availability for their scenario tests 09:36:08 <chandankumar> dmellado: gmann what about finding the service available based on service client python module used? 09:36:13 <gmann> basically service availability should not be tight to plugin if those are in separate repo 09:36:31 <andreaf> gmann, dmellado: sure there is a list of service clients avaialble that plugins can use 09:36:39 <dmellado> chandankumar: hmmm what if the plugin doesn't extend any service client? 09:36:47 <dmellado> you wouldn't be able to discover anything in that case 09:36:54 <chandankumar> i mean a service depend on how many other openstack services 09:37:00 <gmann> chandankumar: but for using service client we need service availability right 09:37:10 <chandankumar> gmann: yes 09:37:12 <dmellado> yeah, that'd be a first step too 09:38:01 <gmann> andreaf: chandankumar dmellado: yea but how we will know whether ironic is available through service clients list 09:38:08 <andreaf> we could make a helper in the lib with a two steps check: if service available flag exists in config for the service use that, else rely on the service client being available 09:38:51 <andreaf> that would work for plugins that do not have a "service available" flag, and it would give the option to plugins to implement that config option over time 09:38:52 <gmann> but in case Ironic is there, and i do not install their plugin then nor service flag is there nor servcie clients 09:38:54 <dmellado> yeah, that would work, and also adding that to both the docs and cookiecutter pugin template won't hurt 09:38:58 <gmann> but still ironic there 09:39:36 <gmann> andreaf: but problem is when plugin implement service availability flag only 09:39:39 <andreaf> gmann: for the ironic case, the ironic plugin would have to implement a "service available 09:39:48 <gmann> because thats goes to separate repo than service repo 09:39:57 <dmellado> andreaf: that feature should be merged soon 09:40:08 <dmellado> so that wouldn't be an issue 09:40:11 <gmann> andreaf: yes but to get to know that we need to install ironic plugin 09:40:20 <gmann> which should not be needed actually 09:40:48 <gmann> my point here is why we should install plugin if service is independetly available 09:41:18 <dmellado> gmann: I see your point, but as of now it'd be an issue as there a ton of in-tree plugins still around 09:41:50 <andreaf> gmann: yeah that would be a work-around to help plugins that are in tree 09:41:58 <gmann> dmellado: i agree but as we recommended that plugin should be in separate repo we should solve this 09:42:18 <dmellado> gmann: then should we wait for pike to land that feature? 09:42:28 <dmellado> because the tc proposal hasn't even been approved yet... 09:42:44 <gmann> andreaf: for that if we make tempest cinfig flag as list and allow tester/devstack to add plugin service into list 09:43:09 <gmann> so that any plugin can see available service from tempest config only 09:43:30 <gmann> dmellado: yea that is still on the way 09:44:29 <gmann> anyways we are running out of time seems, we can discuss this on qa ? 09:44:32 <dmellado> gmann: in any case I'll have a look at your patch and yeah 09:44:45 <dmellado> worst case we can speak about that later or propose a generic blueprint 09:44:49 <dmellado> should we go on? 09:44:51 <andreaf> gmann: I don't think we need to do that - if plugin A is not installed, then plugin B won't be even able to see A's options 09:44:54 <gmann> dmellado: yea may be that is better. i will tr to push early next week 09:45:12 <andreaf> ok let's continue this discussion later 09:45:22 <gmann> andreaf: yea 09:45:30 <gmann> thanks anyways for all input 09:45:35 <gmann> so we were on bug things 09:45:42 <gmann> next week seems andreaf ? 09:45:51 <dmellado> yeah 09:46:14 <gmann> then chandankumar : thanks for that, it will be really helpful 09:46:15 <andreaf> yes I guess my bug triage duty starts today :) 09:46:23 <gmann> andreaf: :) 09:46:29 <dmellado> \o/ 09:46:33 <dmellado> xD 09:46:39 <gmann> next 09:46:42 <gmann> #topic DevStack + Grenade 09:46:46 <chandankumar> gmann: you are welcome :-) 09:47:17 <gmann> so we need to merge this to unblock gate - https://review.openstack.org/#/c/405183/ 09:47:39 <gmann> other we can skip seems if nothing from anyone 09:47:55 <masayukig> +1 09:48:01 <gmann> #topic openstack-health 09:48:06 <gmann> masayukig: ^^ 09:48:41 <gmann> #link https://review.openstack.org/#/q/project:openstack/openstack-health+status:open 09:48:46 <masayukig> I made one bug fix patch for graph. 09:48:56 <masayukig> #link https://review.openstack.org/#/c/402333/ 09:49:29 <masayukig> But there's few people in openstack-health right now.. :( 09:49:47 <gmann> masayukig: humm 09:50:13 <gmann> masayukig: let's see if we cna get more in Atlanta :) 09:50:39 <dmellado> ;) 09:50:40 <masayukig> heh, yeah 09:50:41 <masayukig> And I'll file bug/bp from https://etherpad.openstack.org/p/ocata-qa-openstack-health 09:50:51 <andreaf> masayukig: you get my +1, but my js skills are not so good :P 09:51:17 <masayukig> andreaf : heh, me neither 09:51:31 <gmann> masayukig: andreaf you both are expert :) 09:51:34 <dmellado> masayukig: much better than me, at least xD 09:51:47 <masayukig> I hope so :) 09:51:53 <gmann> anyways anything more on dashboard 09:52:02 <masayukig> that's it from my side 09:52:07 <gmann> ok 09:52:15 <gmann> let's move next 09:52:21 <gmann> #topic Critical Reviews 09:52:32 <gmann> any critical review currently ? 09:52:53 <andreaf> #link https://review.openstack.org/#/c/404981/ 09:53:14 <andreaf> this is a devstack fix needed to get the personality tests back to life when ssh checks are on 09:53:33 <andreaf> which is needed if we are to make the ssh job ever voting :) 09:53:40 <gmann> andreaf: nice, 09:54:03 <masayukig> andreaf : great, thanks 09:54:09 <andreaf> it's not the whole story though, there's a nove patch as well, https://review.openstack.org/#/c/401366/, and perhaps more 09:54:44 <gmann> yea was searching that 09:55:09 <masayukig> I think my skip patch will not be needed after andreaf 's one will be landed 09:55:09 <gmann> ll check thsoe tomorrow 09:55:24 <dmellado> added to my backlog 09:55:40 <gmann> masayukig: may be but its upto you :) 10:07:05 <tosky> #endmeeting