14:03:08 <boris-42> #startmeeting Rally
14:03:09 <openstack> Meeting started Mon Sep  7 14:03:08 2015 UTC and is due to finish in 60 minutes.  The chair is boris-42. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:03:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:03:13 <openstack> The meeting name has been set to 'rally'
14:03:41 <boris-42> oanufriev: ping
14:03:44 <andreykurilin> \o/
14:03:45 <ikhudoshyn> o/
14:03:50 <andreykurilin> ~o~
14:03:54 <boris-42> amaretskiy: ping
14:04:00 <amaretskiy> hi
14:04:41 <e0ne> hi
14:04:57 <yfried> hi
14:05:31 <kun_huang> boris-42: pong
14:05:59 <oanufriev> hi
14:07:30 <andreykurilin> Guys, I have a critical bug-fix for networkcontext+nova-net(v2.1 endpoint) - https://review.openstack.org/#/c/220991/
14:07:41 <andreykurilin> *folks :)
14:08:17 <boris-42> andreykurilin: why this work before?
14:08:42 <andreykurilin> boris-42: nova V2.0 ignores redundant parameters
14:09:14 <kun_huang> andreykurilin: boris-42 seems neutron's python api changes
14:09:22 <boris-42> kun_huang: it's not neutron api
14:09:25 <yfried> andreykurilin: hard to define n-net as critical :)
14:09:26 <andreykurilin> :)
14:09:27 <boris-42> kun_huang: it's nova api
14:09:36 <andreykurilin> yfried: heh
14:09:40 <boris-42> kun_huang: actually our gates are failing
14:09:48 <boris-42> kun_huang: so it's quite critical*
14:10:12 <andreykurilin> yfried: it's critical for our gates:) since we uses it there
14:10:56 <yfried> andreykurilin: reviewed. need jenkins to approve
14:11:13 <andreykurilin> I forgot a link to nova schema - https://github.com/openstack/nova/blob/ae55af90d64fbc0447104c1fb9949501c834085c/nova/api/openstack/compute/schemas/networks.py#L30
14:12:05 <andreykurilin> btw, I have another patch related to V2.1
14:12:08 <andreykurilin> https://review.openstack.org/#/c/220726/
14:12:10 <yfried> andreykurilin: boris-42: so anything else on this subject?
14:12:21 <boris-42> andreykurilin: +2 from me
14:12:32 <boris-42> yfried: I believe we can move on
14:12:36 <boris-42> #topic [yfried] "Compose scenario" vs "new input task format
14:12:55 <yfried> andreykurilin: why can't I see your 2nd patch in Rally dashboard?
14:13:15 <andreykurilin> yfried: hm...I don't know:(
14:13:20 <yfried> boris-42: ^ ?
14:13:57 <yfried> boris-42: I'm reviewing based on these filters. it should be under "Ready for Review"
14:14:22 <boris-42> нyfried topic is another lol=)
14:14:43 <yfried> boris-42: sorry, still hanging on the previous. so - moving on
14:14:48 <yfried> re topic
14:15:24 <yfried> we have 2 specs (one merged one under review) links in the agenda (boris-42 could you please post them?)
14:15:49 <boris-42> #link  https://github.com/openstack/rally/blob/master/doc/specs/in-progress/new_rally_input_task_format.rst
14:15:58 <boris-42> #link  https://review.openstack.org/#/c/215762/
14:16:32 <boris-42> yfried: ^
14:17:26 <yfried> mkrcmari and myself would like to make sure that #2 is not included in #1 and that #2 is generally accpetable because this is something that will get a lot of our manual testing team to use rally
14:17:55 <yfried> btw mkrcmari is aix
14:18:04 <boris-42> yfried: so I don't think that spec about new task fromat is related to this new spec
14:18:16 <boris-42> aix: ping*
14:18:16 <boris-42> =)
14:18:34 * aix is here
14:18:35 <boris-42> yfried: there are 2 different things
14:18:50 <yfried> boris-42: let me draft the difference as I understand it and tell me if you agree
14:18:52 <boris-42> yfried: so format just allows you to execute series of ready scenarios*
14:19:03 <boris-42> yfried: and this thing should be able to combine parts
14:19:12 <boris-42> yfried: and these parts can't be scenarios
14:19:26 <boris-42> yfried: I believe the good thing is to combine current atomic actions*
14:19:31 <yfried> so format spec will allow you to say run "A, B, A(different input), C, D, B(same input),..." and also (maybe) run A and at the same time run B
14:19:50 <yfried> boris-42: ^ 1sec
14:20:31 <boris-42> yfried: so what I think is that this should be done as a part of spec that refactors rally scenario utils
14:20:47 <yfried> the 2nd spec is about concatenating multiple scenarios to create some high level scenario (build network, build volume, boot VM into previous resources, collect Ceilometer input about network, volume and VM)
14:22:06 <yfried> boris-42: how?
14:23:05 <boris-42> yfried: nope they are not
14:23:31 <yfried> boris-42: please explain, as I think we got out of sync
14:23:51 <boris-42> yfried: new input task format is about running N benchmarks one by one or in parallel in the same context
14:24:11 <yfried> boris-42: which is the same as I've said, isn't it?
14:24:13 <boris-42> yfried: mostly for load + monitoring + HA
14:24:18 <boris-42> yfried: nope it is not
14:24:30 <boris-42> yfried: there is no idea to use output of firs benchmark as a chain
14:25:00 <yfried> boris-42: the last part is about "compose"
14:25:02 <boris-42> yfried: + you are running fully first benchmark, then fully second benchmark, then fully third benchmark
14:25:07 <boris-42> yfried: nope it is not
14:25:20 <boris-42> yfried: there is big difference
14:25:56 <boris-42> yfried: between running 3 different Scenarios (fully A, fully B, fully C) and running load (A, B, C n times)
14:26:32 <yfried> boris-42: and you say the 1st option is what new format is about?
14:26:37 <yfried> or second?
14:27:03 <boris-42> yfried: first is about input task
14:27:08 <yfried> boris-42: yep
14:27:09 <boris-42> yfried: second is comopsing benchmark
14:27:29 <boris-42> yfried: so in one case you are running 3 scensarios in second you are running 1 scenario
14:27:40 <boris-42> yfried: and that is huge difference
14:27:43 <yfried> boris-42: so about input task - it will allow us to run A(args1), B, A(args2)
14:27:49 <yfried> boris-42: ^?
14:28:00 <boris-42> yfried: like it is described...
14:28:04 <yfried> boris-42: ok
14:28:09 <yfried> boris-42: and about "compose"
14:28:11 <yfried> :
14:28:45 <boris-42> yfried: in case of input task each scenario has own sla/context/runner
14:28:53 <yfried> boris-42: ^ agreed
14:29:08 <boris-42> yfried: in case of compose you have one runner and one context for all frankenstain *
14:29:19 <yfried> boris-42: ^ agreed as well
14:29:19 <boris-42> all/whole*
14:29:35 <yfried> boris-42: and you are not ok with "compose", are you?
14:29:52 <boris-42> yfried: I am ok with compose
14:30:01 <boris-42> yfried: I am not ok with making it from scenarios*
14:30:06 <yfried> boris-42: because I like it and I've seen demand for it around my colleagues
14:30:39 <boris-42> yfried: I like as well functionallity but it should be implement in another way
14:31:00 <yfried> boris-42: using atomic-actions would require (IMO) to make atomic a plugin
14:31:02 <boris-42> yfried: as I said it should be done as a part of scenario utils refactoring
14:31:17 <boris-42> yfried: and that is okay
14:31:38 <yfried> boris-42: could you please suggest the general idea of how?
14:31:41 <boris-42> yfried: validation and types will be plugin as well
14:32:58 <boris-42> yfried: so just make them plugins that are easy to use in context and scenarios and compose scenarios
14:33:23 <yfried> boris-42: ok, so compose is way down the road and would require me to work on the "scenario utils refactoring" spec. I'd appreciate your help on that
14:33:25 <boris-42> yfried: in case of atomic we can use output
14:33:31 <yfried> boris-42: ok
14:33:39 <boris-42> yfried: yep I will help
14:33:56 <boris-42> yfried: but I believe if we desing everything well it won't take forever to cover both cases
14:34:05 <boris-42> yfried: or even 3
14:34:19 <yfried> boris-42: so please post a summary of this to as review on that spec and let's combine both specs
14:34:27 <boris-42> yfried: making it simple to use atomics in context, remove wrappers and do compose scenario
14:34:52 <boris-42> yfried: so I put actually pretty the same information in that spec
14:35:03 <yfried> boris-42: also, can/do we have a way to see approved specs in readthedocs?
14:35:14 <boris-42> yfried: so it can be done
14:35:24 <boris-42> yfried: like we are doing for feature request
14:35:31 <yfried> boris-42: exactly
14:35:38 <boris-42> yfried: but it is not done yet
14:36:37 <boris-42> #topic  [boris-42] Rally Info & Plugin reference
14:36:58 <boris-42> Okay guys I did a bunch of work related to removing rally info and creating rally plugin
14:37:06 <boris-42> that contains same functiaonllity
14:37:18 <boris-42> and in future will manage rally plugins
14:37:59 <boris-42> #link https://review.openstack.org/#/c/217005/
14:38:01 <boris-42> ^ here is the chain
14:38:03 <yfried> boris-42: it makes it clear that ctx and runners have bad doc
14:38:11 <boris-42> yfried: yeeeppp
14:38:24 <boris-42> yfried: as well that we don't have good docs for base classes
14:38:31 <yfried> boris-42: I'll open bugs about it
14:38:37 <boris-42> yfried: like context.Context, scenario.Scenario, scenario.Runner
14:38:39 <andreykurilin> boris-42: what about deprecation rally info or removement?
14:38:53 <boris-42> andreykurilin: so I did the next, i removed all commands
14:39:11 <boris-42> andreykurilin: except rally info find, rally info list where I wrote it's deprecated use rally plugin find
14:42:19 <boris-42> any questions?
14:42:49 <ikhudoshyn> nope
14:42:52 <boris-42> #topic [boris-42] What is important for Rally 0.1.0 release
14:43:10 <boris-42> so I believe there are 3 things
14:43:18 <boris-42> that we should do in this release more
14:43:25 <boris-42> 1) rally info stuff
14:43:31 <boris-42> 2) rally task abort
14:43:36 <boris-42> 3) rally reports
14:43:47 <boris-42> ^ yfried ikhudoshyn andreykurilin please review these patches
14:43:57 <ikhudoshyn> sure
14:44:56 <ikhudoshyn> btw, which is for 'task abort'
14:45:20 <ikhudoshyn> I've seen one which looked abandoned for me
14:45:43 <ikhudoshyn> ?
14:46:04 <andreykurilin> about 'task abort': I'll try to check it locally and post new change
14:46:15 <boris-42> ikhudoshyn:https://review.openstack.org/#/c/161636/
14:46:46 <ikhudoshyn> what is 'aborted' in gerrit btw?
14:48:19 <boris-42> ikhudoshyn: ?)
14:48:50 <ikhudoshyn> boris-42: the patch that you pointed is marked as 'aborted' ))
14:49:00 <boris-42> ikhudoshyn: ;)
14:49:06 <ikhudoshyn> looks like something not really good for me
14:50:11 <ikhudoshyn> haha, got it, I feel especially dumb today))
14:50:12 <boris-42> ikhudoshyn: it's name of patch lol
14:50:19 <boris-42> ikhudoshyn: =)
14:50:25 <boris-42> okay so let's end this meeting
14:50:27 <boris-42> see you next week
14:50:33 <boris-42> #endmeeting