19:04:23 <davidlenwell> #startmeeting refstack
19:04:24 <openstack> Meeting started Mon Sep 22 19:04:23 2014 UTC and is due to finish in 60 minutes.  The chair is davidlenwell. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:04:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:04:27 <openstack> The meeting name has been set to 'refstack'
19:04:35 <davidlenwell> roll call ?
19:04:49 <catherine_d> here
19:05:08 <sslypushenko__> +
19:06:36 <davidlenwell> Okay.. looks like its just us chickens ;)
19:06:39 <pvaneck> o/
19:06:59 <davidlenwell> So lets dive right in..
19:07:05 <davidlenwell> First..
19:07:26 <catherine_d> <davidlenwell> do you have any way to contact Rob?
19:07:42 <davidlenwell> this is a hard meeting for him to make
19:08:41 <davidlenwell> So the first thing I want to talk about is the fact that we have a new core reviewer!
19:08:45 <davidlenwell> Congrats to catherine_d
19:08:57 <zehicle_at_home> o/
19:08:59 <catherine_d> Thank you!!!
19:09:05 <davidlenwell> there he is !
19:09:07 <zehicle_at_home> +1
19:09:11 <rockyg> o/
19:09:21 <rockyg> Congratz Catherine_d
19:09:22 <davidlenwell> the gang is all here!
19:09:37 <catherine_d> Thanks!
19:09:52 <davidlenwell> So lasat week catherine_d, rockyg and my self.. sat down at oenstack sv
19:10:15 <davidlenwell> the primary item on our list was dealing with uuids for tests
19:11:32 <davidlenwell> we agreed that we would use a portion of todays meeting to toss out ideas for getting around this issue.
19:11:50 <davidlenwell> and that we'd then spend the next few weeks exploring those options
19:12:42 <zehicle_at_home> should we start with Julia's list
19:12:42 <juliashapovalova> o/
19:13:15 <catherine_d> whatever options we choose we need to sync the version of tempest that we used to create the list and using the same version to test ...
19:13:56 <juliashapovalova> it should not be a big problem
19:14:15 <juliashapovalova> generally Tempest is enhanced with new tests
19:14:15 <catherine_d> 2 options are being considered: 1) Using a specific SHA 2) Git Labels at a specific time/state
19:14:37 <juliashapovalova> old obes are reorganazed rarely
19:14:39 <davidlenwell> I don't think git labels is even worth discussing
19:14:59 <davidlenwell> juliashapovalova:  that isn't exactly accurate
19:15:19 <davidlenwell> maybe lately they've just been reorginizeing .. but they have been moving things around
19:15:21 <zehicle_at_home> I'd like to consider the the ID being Date based.
19:15:29 <zehicle_at_home> so, they'd be 20140922AA
19:15:35 <juliashapovalova> yeah, i know//
19:15:56 <zehicle_at_home> we'd start from 20140000000 for the existing body of tests
19:16:16 <davidlenwell> lets start an etherpad and list these ideas
19:16:26 <catherine_d> +1
19:16:28 <davidlenwell> we can use that doc to define the ideas over the next week.. then we'll vote
19:16:30 <zehicle_at_home> +1
19:16:32 <rockyg> I'm trying to find the link to the etherpad....
19:16:43 <davidlenwell> rockyg:  im not sure we made one yet for this
19:16:45 <davidlenwell> did we?
19:17:12 <rockyg> #link https://etherpad.openstack.org/p/Refstack-TestID-Approaches
19:17:40 <rockyg> needs lots of fleshing out...
19:17:50 <davidlenwell> okay .. so lets use this pad over the next week.. we'll discuss again in detail next monday.
19:20:17 <zehicle_at_home> do we also need to discuss the way we track them?  or was the suggestion to have an xref file ok?
19:20:56 <davidlenwell> each method will have different problems in that regard .. bullets should be added under each
19:21:06 <davidlenwell> then we can argue about it and come up with the right solution
19:21:57 <zehicle_at_home> can we talk about priorities for the selection?
19:22:08 <davidlenwell> can you elaborate on that?
19:22:14 <zehicle_at_home> like human readable and able to be used on tempest metadata
19:22:33 <davidlenwell> I think the etherpad should open with a description of the problem and what we need o that its very clear
19:23:11 <zehicle_at_home> ok
19:23:14 <rockyg> yeah.  list of requirements
19:23:28 <davidlenwell> you can't propose solutions to a problem that isn't clearly spelled out
19:23:35 <davidlenwell> so we'll start with that and move out
19:26:18 <juliashapovalova> so do we stick on the xref file for uuid<->full_test_name pairs?
19:28:42 <zehicle_at_home> juliashapovalova, I think that's the best plan
19:28:57 <zehicle_at_home> I have a pull for that already
19:30:02 <davidlenwell> we aren't sticking to anything until we've discussed the options
19:30:12 <davidlenwell> sorry guys my dsl has been flakey today
19:30:43 <catherine_d> so we should be active in describing the options in the ehterpad for next week's discussion ...
19:31:30 <davidlenwell> yes .. everyone should jump in
19:31:52 <catherine_d> should we move to the next discussion item --> refstack-client for RH and CentOS
19:32:23 <zehicle_at_home> I may not be able to make meeting next week
19:33:35 <davidlenwell> yes
19:34:07 <davidlenwell> So catherine_d and I were discussing this last week and we both agreed that if we get back to making tcup working and a thing that we don't need to worry about adding support for all the os's in the client
19:34:24 <zehicle_at_home> +1
19:34:35 <davidlenwell> not that we don't want redhat support .. but we also need to pick our battles .. and the docker thing gives us two birds with one stone.
19:35:04 <davidlenwell> zehicle_at_home:  Do you think this would be good work for sslypushenko__?
19:35:36 <zehicle_at_home> I can ask him tomorrow, possibly
19:35:56 <sslypushenko__> I'm here)
19:36:00 <davidlenwell> he's here now
19:36:13 <davidlenwell> sslypushenko__: what do you think?
19:36:41 <sslypushenko__> using docker - it is a good idea
19:37:02 <davidlenwell> Do you think you are up to doing the work?
19:38:12 <sslypushenko__> Maybe I will need a details
19:38:25 <sslypushenko__> But I can make it
19:38:35 <davidlenwell> zehicle_at_home: is the brain trust on this effort so I'll let him guide you on it
19:39:28 <zehicle_at_home> we can try and get d_qoi on it too
19:39:47 <zehicle_at_home> we can scope out the work and see about it.  I agree on the benefits
19:39:59 <rockyg> That's what I was thinking.  I gotta work on getting salary turned back on for him this week.
19:40:01 <zehicle_at_home> do we want it to work like before, or will it be more interactive?
19:40:08 <zehicle_at_home> I'm thinking interactive env
19:40:24 <zehicle_at_home> rockyg, he will be delighted to work on it.  he has some time
19:40:49 <rockyg> Yeah, but better if he gets paid;-)
19:40:57 <davidlenwell> I think we want it to work both ways
19:41:07 <davidlenwell> silent running and interactive
19:41:29 <zehicle_at_home> let's do interactive first
19:41:29 <rockyg> ++
19:42:21 <davidlenwell> yep
19:42:26 <davidlenwell> okay thats agreed
19:42:35 <davidlenwell> anyone have anything else to discuss?
19:43:02 <zehicle_at_home> board meeting update....
19:43:43 <zehicle_at_home> basically not much to report.  Board backed away from DefCore recommendations.  We're going to be spinning on multiple trademarks instead of a single Core mark
19:44:42 <zehicle_at_home> so, it's not certain if OpenStack will be a single thing for commercial or a lot of projects.  Right now, it's looking like the later
19:45:11 <davidlenwell> cool
19:45:17 <davidlenwell> I actually think that is better
19:45:18 <zehicle_at_home> that does NOT change Refstack - we still need the data to make decisions
19:45:23 <davidlenwell> yep
19:45:29 <catherine_d> does Board still support DefCore, Refstack ...as the tools for trademark
19:45:35 <davidlenwell> I was just typing that i do not think it effects our mission
19:45:35 <catherine_d> got it
19:45:48 <zehicle_at_home> catherine_d, yes.  very much
19:46:24 <zehicle_at_home> davidlenwell, I think that's easier for the TC but commecially setting us back
19:46:30 <rockyg> From conversations at OpenStack SV, there is much interest in Refstack for validating interoperability on existing installs.
19:46:47 <davidlenwell> zehicle_at_home: I'd recomend reading montys recent blog post on the subject
19:46:48 <rockyg> And new installs
19:46:51 <catherine_d> In IBM we have 3 different products to start collecting data ...
19:46:59 <davidlenwell> rockyg: indeed
19:47:11 <zehicle_at_home> davidlenwell, will do
19:47:46 <davidlenwell> anything else?
19:47:55 <catherine_d> do we target to have some data for Paris ?
19:48:04 <rockyg> Also, if we can get operators to test against full test sets, we can get aggregate data for future releases
19:48:24 <rockyg> If they test against full tempest and submit, we can see what is being run.
19:49:05 <catherine_d> rockyg: +1 ... so will Huawei test with refstack-client?
19:49:14 <rockyg> That's an important option fro Refstack.  Test group selection by config file, uuids, whatever.
19:49:49 <rockyg> catherine_d  I need to talk to our guys and one is currently travelling to China, but with you volunteering tech support, I think so.
19:50:15 <zehicle_at_home> davidlenwell, link to Monty?
19:50:25 <davidlenwell> http://inaugust.com/post/108
19:50:45 <davidlenwell> This is also a good read .. http://blog.leafe.com/simplifying-openstack/
19:50:56 <davidlenwell> both refer to this .. https://dague.net/2014/08/26/openstack-as-layers/
19:51:46 <davidlenwell> read all three of those and you'll have a pretty good idea of how the tc is thinking about things
19:54:40 <davidlenwell> okay .. if nobody has any other topics to discuss I am going to kill it .. as always feel free to jump in #refstack if you need further discussion on something.
19:54:44 <davidlenwell> #endmeeting