14:06:09 <andreykurilin> #startmeeting Rally
14:06:11 <openstack> Meeting started Mon Dec 12 14:06:09 2016 UTC and is due to finish in 60 minutes.  The chair is andreykurilin. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:06:12 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:06:15 <openstack> The meeting name has been set to 'rally'
14:06:59 <andreykurilin> ok
14:07:09 <andreykurilin> let's start from regular topic
14:07:15 <andreykurilin> #topic status of gates
14:07:41 <andreykurilin> last several weeks we had problems with keystone, heat, nova-net and etc
14:08:05 <andreykurilin> Currently, all failures should be resolved
14:08:35 <amaretskiy> hi
14:08:49 <andreykurilin> hi
14:09:18 <andreykurilin> so, if you see any new problems with gates, please let me know about them
14:09:47 <andreykurilin> #topic keystone v3
14:10:08 <andreykurilin> another one great news, we merged latest patch for keystone v3 support
14:10:25 <andreykurilin> all scenarios are ported to support both keystone v2 and v3
14:10:41 <stevemar> andreykurilin: great to hear
14:11:02 <andreykurilin> :)
14:12:38 <andreykurilin> there is only one case is not supported by rally yet - domain-scope. I heard that some guys have problems with it(maybe it is wrong use case). In case of domain-scope they have empty service catalog and we need to switch to project-scope for obtaining it
14:13:09 <andreykurilin> stevemar: is it valid use case? ^
14:13:45 <stevemar> andreykurilin: i believe thats valid
14:14:00 <andreykurilin> ok, so we need to implement this "switch"
14:14:30 <andreykurilin> But in general, we support domain-scope if service catalog is not empty :)
14:15:00 <andreykurilin> ok
14:15:06 <andreykurilin> let's move to the next topic
14:15:23 <andreykurilin> #topic Nova-Network is dead. Long live Neutron!
14:16:03 <andreykurilin> all our jobs don't have nova-net at all
14:16:12 <andreykurilin> it was disabled by nova team
14:16:34 <andreykurilin> and it will be removed soon
14:16:46 <hai_shi> including nova-net to networking service?
14:17:29 <andreykurilin> In future yes.
14:17:45 <hai_shi> got it.
14:17:52 <andreykurilin> But now we need that our NetworkService( aka NetworkWrapper) suppor tit
14:17:53 <andreykurilin> *it
14:18:16 <andreykurilin> since we have a lot of users with old openstack releases
14:18:29 <andreykurilin> I do not know when we will able to remove support of it :(
14:18:45 <hai_shi> ok. my job's process is slowly :(
14:19:05 <andreykurilin> I heard about guys who uses latest rally to test openstack grizzly
14:19:15 <andreykurilin> hai_shi: do you need any help?
14:20:45 <hai_shi> maybe I clould commit the patch firstly(just little work). you could review my jobs.
14:21:04 <hai_shi> and give me some advise:)
14:21:19 <andreykurilin> ok, feel free to post any "work-in-progress" patches and we will try to help you
14:21:47 <hai_shi> ok. thanks a million.
14:21:55 <andreykurilin> #topic get rid of validation_results field in database
14:22:04 <andreykurilin> ok. it is a topic to discuss
14:22:06 <andreykurilin> :)
14:22:57 <andreykurilin> currently we have special status "validation_failed" with additional data
14:23:30 <andreykurilin> for this additional data we have separate db field
14:24:04 <andreykurilin> https://github.com/openstack/rally/blob/master/rally/common/db/sqlalchemy/models.py#L184
14:24:31 <andreykurilin> it is good information to store
14:24:37 <andreykurilin> but it is not "unique"
14:25:22 <andreykurilin> we use sla_results to store some other information - information about aborting, about runtime errors, about failures in contexts
14:25:55 <andreykurilin> Maybe we should extend validation_results to something bigger?
14:26:09 <andreykurilin> amaretskiy ^
14:27:04 <andreykurilin> also, I prefer to move it to subtask level
14:27:04 <amaretskiy> andreykurilin no ideas right now
14:27:37 <andreykurilin> astudenov ^
14:27:46 <chenhb__> Our subtask is available?
14:28:05 <astudenov> I think we need to leave something for validation result in task level
14:29:04 <astudenov> and add something for error messages in subtask level
14:29:44 <andreykurilin> chenhb__: astudenov is working on its support. he can share his status:)
14:30:03 <andreykurilin> astudenov: I dislike validation result at task level
14:30:04 <rvasilets> hola
14:30:10 <andreykurilin> hola
14:30:31 <chenhb__> Hi
14:30:45 <astudenov> subtask exists in db at least :) I have a path adding some basic usage of subtask https://review.openstack.org/#/c/395554/
14:31:00 <andreykurilin> astudenov: I want to have ability to launch all valid subtasks and workloads, even if one workload is invalid
14:31:19 <andreykurilin> so it makes thanks to store validation results per subtask or even per workload
14:31:43 <andreykurilin> also, it would be nice to display this validation results(in case of failure) in our html-reports
14:32:35 <astudenov> lets try
14:32:45 <andreykurilin> nice
14:32:58 <andreykurilin> ok, let's move to the next topic
14:33:09 <andreykurilin> #topic get rid of enums in db
14:33:13 <chenhb__> Thanks,@astudenov
14:33:21 <andreykurilin> why we need them at all?)
14:33:40 <andreykurilin> I don't see any use cases for them
14:34:49 <rvasilets> this is just validation at db level
14:35:33 <andreykurilin> but we do not change them regularly
14:35:37 <andreykurilin> I mean statuses
14:35:53 <andreykurilin> so we should miss the change when we add new status
14:36:28 <andreykurilin> Also, we do not have places where we transmit status strings without usage rally.consts module
14:36:44 <andreykurilin> so, imo, we do not need that validation at db level
14:37:09 <andreykurilin> it requires additional migrations each time when we want extend statuses
14:37:59 <rvasilets> reasonable
14:38:04 <rvasilets> at first sign
14:38:16 <andreykurilin> :)
14:38:39 <andreykurilin> ok, let think more about it and make final decision at the next meeting
14:38:56 <andreykurilin> #topic Open discussion
14:39:30 <andreykurilin> we finished fixing gates and I switched to reviewing patches... We have a long queue now
14:39:57 <andreykurilin> Any topics to discuss?
14:40:38 <rvasilets> Nope
14:41:58 <chenhb__> Nothing from me
14:42:07 <andreykurilin> ok
14:42:14 <andreykurilin> bye
14:42:19 <andreykurilin> #endmeeting