19:03:11 <davidlenwell> #startmeeting refstack
19:03:12 <openstack> Meeting started Mon Sep 29 19:03:11 2014 UTC and is due to finish in 60 minutes.  The chair is davidlenwell. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:03:13 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:03:15 <openstack> The meeting name has been set to 'refstack'
19:03:29 <davidlenwell> roll call?
19:04:00 <catheri6> here
19:04:08 <rockyg> o/
19:04:35 <davidlenwell> I'd like to discuss sergeys latest work .. but he doesn't seem to be here
19:05:47 <rockyg> Maybe we can circle back.
19:06:01 <sslypushenko__> Hi all!
19:06:03 <rockyg> Or maybe we don't need to
19:06:23 <davidlenwell> hi sslypushenko__
19:07:16 <davidlenwell> I'd like to discuss what you brought up in channel earlier
19:07:40 <sslypushenko__> Yep
19:08:38 <sslypushenko__> Do we really need run tempest in docker?
19:10:47 <catheri6> So we do not just run tempest in docker ... we can run Ubuntu version of Tempest in either docker or native ... but for other distros support we would like to wrap it in docker so we only need to validate Tempest run with one distro which is Ubuntu
19:11:20 <catheri6> we do not turn refstack-client into integration house where we need to ensure that all distros work
19:11:35 <catheri6> that I think is the idea why we use container
19:11:43 <catheri6> open for discussion ...
19:13:05 <catheri6> of course that makes refstack-client  depend on docker ..
19:13:11 <sslypushenko__> It somewhat strange requirement to run refstack-client in a container
19:13:27 <davidlenwell> agreed
19:13:34 <davidlenwell> it cannot be required
19:13:51 <catheri6> question to every  one:  which distro is OpenStack QA running with?
19:13:53 <davidlenwell> I'd rather automate the installing of dependant libs
19:14:12 <davidlenwell> they work with rehl and centos and susa
19:14:28 <davidlenwell> we are confusing two issues here
19:14:41 <davidlenwell> nobody is suggesting that we force the use of docker
19:14:43 <catheri6> so no ubuntu :-)
19:15:02 <davidlenwell> we can use something like sslypushenko__'s script
19:15:07 <davidlenwell> no reason not to
19:15:22 <davidlenwell> and i don't see a real reason to reduce our binary dependance either
19:15:47 <davidlenwell> lots of apps depend on other apps .. its just how things work .. and honestly we've spent more time arguing about this than it would take to just support rehl
19:16:16 <sslypushenko__> agreed
19:16:30 <davidlenwell> sslypushenko__: if you change your script to bash I will merge it
19:16:36 <davidlenwell> thats my only hessitation right now
19:16:50 <sslypushenko__> why we need bash?
19:17:06 <davidlenwell> its an installer essentially .. belongs in bash
19:17:20 <sslypushenko__> Detect disto in bash is not easy
19:17:21 <davidlenwell> using python for light logic then calling out to shell is just overly complex
19:18:08 <davidlenwell> there is a ton of bash out there already written that does just that
19:18:19 <davidlenwell> if you're not up to it .. fine .. but its actually not that complicated
19:18:41 <sslypushenko__> You can go to devstack repo and look for that code...
19:19:31 <sslypushenko__> I can bring links here, if you want)
19:19:55 <davidlenwell> So I think you are over thinking this
19:20:05 <davidlenwell> all we really need is to detect apt-get vs yum
19:20:32 <davidlenwell> we don't really care what os they are running .. and someone tries it on one our docs say are tested .. I don't actually care ..
19:20:34 <davidlenwell> see .. http://stackoverflow.com/questions/16597342/bash-test-for-yum-or-apt-based-linux
19:21:24 <sslypushenko__> Hmmm... sounds reasonable
19:21:31 <davidlenwell> cool .. lets move on
19:21:40 <davidlenwell> the other major hold back right now is the api
19:21:48 <sslypushenko__> Ok. I will rewrite script tomorrow
19:21:50 <catheri6> If we are going to support multiple distros how about  SUSE  .. it does not use either apt-get or yum?
19:22:08 <davidlenwell> we will cross that bridge later or they can run tcup
19:22:18 <davidlenwell> rehl and ubuntu are the ones I am more worried about right now
19:22:31 <davidlenwell> lets move on ..
19:22:43 <davidlenwell> #topic api
19:23:20 <davidlenwell> I'm working now to break up the api that I have now working into logical commits.. those should start rolling in today
19:23:49 <davidlenwell> I'll want everyone to review them .. keep an eye out I will ping you all with a description of how I have it broken up and why I did some of the changes
19:24:14 <sslypushenko__> Good news!
19:24:27 <catheri6> seems like we will be using a specific tag of Tempest .. should we add a field in the result model to save this info
19:25:25 <davidlenwell> catheri6: two steps ahead of you
19:25:27 <davidlenwell> ;)
19:25:33 <davidlenwell> that is one of the changes that I made
19:25:51 <catheri6> davidlenwell: ha ha
19:26:11 <davidlenwell> feild is called tempest_version its a varchar feild . so we can use it for either a sha or to name the tagged release
19:27:21 <davidlenwell> So I'll be pretty busy the next couple of days making sure the api lands and then working on getting it hosted
19:27:43 <davidlenwell> So there is another topic that I want to discuss..
19:27:46 <davidlenwell> its kinda important
19:27:53 <davidlenwell> and I with rob was here for it
19:28:14 <davidlenwell> #topic refstack beyond defcore
19:29:02 <davidlenwell> anyone keeping track of the political landscape in openstack will know by now that defcore is going a bit sideways and the tc is working on making some serious changes..
19:29:22 <rockyg> yup.
19:29:26 <catheri6> will that affect refstack?
19:29:34 <davidlenwell> which raises the question what life does refstack have if defcore is no more .. or doesn't need us as a validation tool for enforcing trademark
19:29:47 <davidlenwell> the simple one word answer to that is interop
19:30:09 <davidlenwell> the community and the tc still very much care about having interop data
19:30:13 <rockyg> My take on the discussion seems to put refstack in a more critical position because it tests interop
19:30:38 <davidlenwell> rockyg: thats exactly what I am saying
19:30:44 <rockyg> ++
19:30:47 <catheri6> ++
19:30:52 <rockyg> you got there first ;-)
19:30:56 <davidlenwell> :)
19:31:29 <davidlenwell> So what I am suggesting is that we adjust the wording in our mission statement in the wiki to be more interop and play down defcore
19:31:39 <rockyg> good idea
19:32:06 <davidlenwell> okay .. cool .. well I've covered everything on my mind this week.. does anyone have anything else?
19:32:10 <davidlenwell> #topic open discussion
19:32:14 <catheri6> rockyg: will you take care of that for us?
19:32:29 <davidlenwell> rockyg: actually lets start an etherpad
19:32:31 <catheri6> that is the wiki
19:32:38 <davidlenwell> to work the text before putting it in the wiki
19:32:47 <catheri6> good idea
19:32:48 <davidlenwell> we can start with the current wiki text
19:32:53 <catheri6> who in tc do we work with?
19:32:56 <davidlenwell> and I'll go through and slash things out
19:32:56 <rockyg> ++
19:33:21 <davidlenwell> we don't need to work with the tc yet
19:33:48 <davidlenwell> but several of them will be open to working with us when we're ready .. I've already had a lot of offline conversations aligning us the right way
19:35:02 <catheri6> defcore used to be our parent :-)  ...
19:35:16 <davidlenwell> I've never really seen it that way
19:35:39 <davidlenwell> more like one of our customers
19:35:42 <rockyg> #link https://etherpad.openstack.org/p/refstack-wiki-draft
19:35:55 <davidlenwell> awesome . thanks rockyg
19:36:39 <catheri6> next ... do we need to discuss this https://etherpad.openstack.org/p/Refstack-TestID-Approaches .. or we are set to use tagged release for now ..
19:36:57 <davidlenwell> tagged release
19:37:03 <davidlenwell> thought I covered this last wek
19:37:04 <davidlenwell> week
19:37:05 <rockyg> catheri6:  I think both Josh and Randy Bias were agitating for what is now Refstack long before Defcore got created
19:37:21 <davidlenwell> yep
19:37:34 <davidlenwell> it was josh and monty actually
19:37:45 <davidlenwell> but randy is a huge ally in all things interop
19:38:12 <davidlenwell> the main goal is going to be to creat an interop map
19:38:15 <catheri6> but radny has not joined us...
19:38:31 <rockyg> he's too busy running his company
19:38:31 <davidlenwell> he's got other priorities
19:38:35 <davidlenwell> yep
19:38:54 <davidlenwell> but I Assure you he is paying attention
19:39:08 <rockyg> Once we have the framework up, I bet some of his employees will be adding interop type tests to tempest
19:39:10 <catheri6> and agree with what we do?
19:39:45 <rockyg> Yeah.  Refstack lets him prove the work he does is really OpenStack
19:39:55 <davidlenwell> he doesn't strike me as the type to be silent when in dissagreement
19:40:30 <davidlenwell> but we're not going to worry about him or anything else right now .. we need to stay on track
19:40:54 <catheri6> yes get refstack and restack-client to work ...
19:41:03 <davidlenwell> +1000000000000000
19:41:13 <catheri6> hopefully before Paris so we can show to as many people as possible at the summit
19:41:20 <davidlenwell> so right now I think we are on track for that goal
19:41:35 <davidlenwell> api to land .. then I'll turn my focus on the front end of the data views
19:41:57 <catheri6> I am happy with the progress in refstack-clent for the last week ... I think we are in th right direction ...
19:42:05 <catheri6> davidlenwell: +++1
19:42:14 <davidlenwell> good
19:43:34 <catheri6> only 2 outstanding reviews in refstack-clent .. Sergey's .. will merge after his change ... how about this one https://review.openstack.org/#/c/120926/
19:45:13 <davidlenwell> I'll review it now
19:45:18 <davidlenwell> well after this meeting
19:46:02 <davidlenwell> also .. just fyi .. catherine learned last week how effective emailing me is to get my attention on a review
19:46:14 <catheri6> :-)
19:46:52 <davidlenwell> if you cant' catch me here and you want my thoughts on something or to review some code . try irc .. then email me
19:47:09 <davidlenwell> my email address for those who don't know is dlenwell@gmail.com
19:47:16 <catheri6> we need to work on the reviews in refstack .. but I guess they are more on the defcore side
19:47:19 <rockyg> good to know.  for me, email,email, text
19:47:47 <davidlenwell> catheri6: indeed .. I cannot merge defcore things without reviews from someone on defcore
19:50:43 <davidlenwell> we have an agreement not to merge without a review from a defcore commitie member
19:50:47 <catheri6> if we do not have any thing else ... we can discuss a technical topic ...
19:51:09 <catheri6> I think we need a defcore committe member ....
19:51:28 <davidlenwell> we have to be prepared for a world without a defcore commitie
19:51:54 <davidlenwell> what is your technical topic?
19:52:28 <catheri6> I found that the git version in Ubuntu 12.04 does not support git with branch ...
19:52:55 <davidlenwell> oh that is a hicup
19:52:59 <davidlenwell> like not at all ?
19:53:37 <catheri6> not at all because of the git version but saucy is ok ... but we can work aroind the issue by not using git ...
19:54:03 <davidlenwell> use http
19:54:08 <catheri6> just unpack the zip like we did before for stable Havana ...
19:54:10 <davidlenwell> I think thats fine
19:54:20 <catheri6> that would work for all version
19:54:27 <catheri6> versions ...
19:54:40 <davidlenwell> only real reason to use git itself is if we want to push code back .. otherwise grabbing the tagged release via http is totally fine
19:54:55 <catheri6> davidlenwell: agree
19:55:28 <davidlenwell> just make sure we validate the md5 of the zip file we grabbed before unzipping it and running the code inside
19:55:46 <davidlenwell> you know .. just to keep things secure
19:55:54 <catheri6> yep
19:56:05 <davidlenwell> okay .. so who is going to make that change?
19:56:13 <rockyg> sounds simpler and possibly more reliable.
19:56:15 <davidlenwell> It hink we should create a story for it
19:56:32 <davidlenwell> and assign it to someone
19:56:39 <davidlenwell> can do that if nobody else wants to
19:56:50 <davidlenwell> I mean create the story not make the change to the code
19:57:07 <catheri6> davidlenwell: makle sense ...
19:57:33 <catheri6> for now sslypushenko__: will code is set_env as is ..
19:57:58 <catheri6> davidlenwell: can you create story?
19:58:24 <catheri6> or I can do that too..
19:58:35 <davidlenwell> catheri6: why don't you give it a try
19:58:45 <catheri6> ok
19:58:46 <davidlenwell> pm me the link to it when you do and I can help you adjust it
19:59:08 <catheri6> davidlenwell: sure
19:59:15 * davidlenwell is trying to include catheri6 in part of the project managment so he doesn't get overwhelmed
19:59:34 <davidlenwell> okay .. that is all our time for today .. we can still chat in #refstack if anyone needs anything
19:59:38 <davidlenwell> #endmeeting