19:01:17 <davidlenwell> #startmeeting refstack
19:01:17 <openstack> Meeting started Mon Jan 26 19:01:17 2015 UTC and is due to finish in 60 minutes.  The chair is davidlenwell. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:01:18 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:01:20 <openstack> The meeting name has been set to 'refstack'
19:02:06 <hogepodge> o/
19:02:09 <davidlenwell> #link https://etherpad.openstack.org/p/refstack-agenda-1.26.2015
19:02:09 <sslypushenko__> o/
19:02:20 <pvaneck> o/
19:02:24 <davidlenwell> The agenda for today.
19:02:44 <catherineD> o/
19:05:17 <davidlenwell> okay.. lets get started..
19:05:42 <davidlenwell> First off.. folks from piston have been ignoring my request to give us control of the domain name.. im sure they're busy and I'
19:05:51 <davidlenwell> m hoping they can get back to me soon..
19:06:09 <davidlenwell> but in the mean time I've gotten frustrated and registered refstack.net
19:06:29 <davidlenwell> this morning I am getting the api server running on it
19:07:18 <davidlenwell> I'll then follow with a commit changing the default api url to reflect the new domain name.
19:07:26 <catherineD> what is our alternative if Piston keeps ignoring us ...
19:07:35 <davidlenwell> we'll just keep using refstack.net
19:07:42 <davidlenwell> they have refstack.org
19:07:46 <davidlenwell> which would be preffered ..
19:07:57 <davidlenwell> but we might as well have all the refstack domains we can ..
19:08:43 <davidlenwell> I'd have picked up .com but its 1800 dollars
19:09:04 * hogepodge wonders who that would be payable to...
19:09:33 * davidlenwell also wonders.. its in a premium domain auction thing
19:09:43 <davidlenwell> so the owner info is private
19:10:41 <davidlenwell> regardless .. we'll just use this one for now.. the domain was 9 dollars.. I'll just eat that and we can keep moving forward.
19:10:50 <davidlenwell> #topic front end website
19:12:02 <davidlenwell> now that we have a domain name.. I will turn my focus on getting the copy we've been working on into html and make it pretty and put it up for review.
19:12:51 <davidlenwell> goal is to get to get to that by the end of the day.
19:12:55 <catherineD> we need at least the front page and navigation to the test results ..
19:13:06 <davidlenwell> wow that was a funny typo
19:13:18 <catherineD> :-)
19:13:27 <davidlenwell> yes .. catherineD it will be a one page site for now.. with information on how to view test results
19:15:55 <davidlenwell> lets move along..
19:16:09 <davidlenwell> #topic pending reviews
19:16:34 <davidlenwell> #link https://review.openstack.org/#/q/refstack,n,z
19:17:29 <hogepodge> #link https://review.openstack.org/#/q/refstack+status:open,n,z
19:17:31 <hogepodge> (open reviews)
19:18:02 <davidlenwell> lets focus on the top 4 in that list
19:18:32 <catherineD> With the instruction in https://review.openstack.org/#/c/149703/  I was able to build a refstack server , run tests with refstak-client , upload and view results
19:19:01 <davidlenwell> hogepodge:  I have already started to review both of your patches from this morning.. expect that to be done shortly after we are free'd up from this meeting..
19:19:26 <sslypushenko__> hogepodge:  If you don't mind I want to implement validation feature:  https://review.openstack.org/#/c/150097/
19:19:59 <davidlenwell> vladiskuz:  I asked you last week to let me know here when you fealt that https://review.openstack.org/#/c/149589/  was ready to test.. what is the status of that?
19:20:27 <hogepodge> sslypushenko__ In general I don't mind, but if I don't get reviews in I won't be able to participate at a higher level in the project (per the meeting last week). But, I value expediency over status, and you'd likely be faster than me.
19:20:38 <hogepodge> s/reviews/code commits/
19:22:27 <hogepodge> sslypushenko__ I see the client work as my main priority right now, since I'm going to be working with vendors on using it.
19:22:50 <davidlenwell> hogepodge: +1  makes sense for you to get as involved with that as possible
19:23:27 <davidlenwell> as you're working with vendors moving forward I've like if we kept documentation on client ussage and trouble shooting up to date as well.
19:23:29 <hogepodge> re 149589 That's a big patch. Can it be broken up? Not terribly important to me, just asking... you know, for a friend. :-)
19:23:53 <vladiskuz_> <davidlenwell> I worked on the patch. I have some trouble with oslo.config
19:24:55 <davidlenwell> I can spend some time and take a look and attempt to help if you like .. or hogepodge you are welcome to help out with it as well..
19:25:48 <sslypushenko__> hogepodge Refstack-client is very important thing for us. It is quite tricky to use it on real deployment. So your work with vendors will be appreciated
19:25:55 <hogepodge> Yup I'll do a review and test is locally.
19:26:06 <hogepodge> s/is/it/
19:26:18 <davidlenwell> I know you are trying to build up your commits .. so please take any and all oppurtinity as well.. I also value expiediancy over status .. lets just do all we can to get this stuff landed ..
19:26:34 <davidlenwell> the sooner we have that the sooner the api will be to keep deployed and up to date in production
19:26:35 <sslypushenko__> davidlenwell I think we will finish Vlad's tomorrow
19:26:52 <davidlenwell> sslypushenko__:  that is an acceptable timeline
19:27:28 <davidlenwell> So hogepodge your talk of helping vendors is a good segway..
19:27:40 <davidlenwell> #topic refstack-client (defcore) beta testing
19:27:47 <catherineD> one more
19:27:48 <sslypushenko__> I was a little busy for today. Try to run refstack-client on Mirantis-cloud-on-demand
19:27:51 <catherineD> https://review.openstack.org/#/c/149703/
19:28:27 <sslypushenko__> Actual tempest config still is a biggest pain)
19:28:27 <catherineD> please merge this one https://review.openstack.org/#/c/149703/ so sslypushenko__: vladiskuz_: can based on this one ....
19:28:36 <hogepodge> catherineD I'll look over later today and leave a review.
19:28:47 <davidlenwell> yes catherineD I am in the middle of reading it.. have it open.. I think its a no brainer.. but I want to carefully make sure we are not missing any steps
19:28:55 <hogepodge> Or you can just merge. I'll still try it out.
19:29:13 <davidlenwell> I'd like your review hogepodge before merge
19:29:34 <davidlenwell> since catherine wrote it im the merge vote on this one..
19:29:43 <catherineD> it is better that sslypushenko__: vladiskuz_: based on this one with the format change .,, easier later ..
19:31:05 <davidlenwell> okay so lets talk beta vendors..
19:31:20 <davidlenwell> I was in Seattle most of last week.. at the bluebox offices..
19:32:13 <davidlenwell> We're ready to start running tests against our cloud in ci.. we even already have a tempest config worked out.
19:32:15 <hogepodge> Last week I sat down with foundation staff and started working up a list of potential beta testers. This isn't final, some have been contacted, most not, so don't infer commitment from any of the listed names.
19:32:24 <hogepodge> Bluebox is on the list, so glad you got a jump on that.
19:32:46 <davidlenwell> we also have time set asside with q/a at piston to get to run the tests as well..
19:32:52 <catherineD> davidlenwell: hogepodge: great
19:32:58 <davidlenwell> which is why im pushing so hard to have the api up and running on refstack.net today
19:33:08 <hogepodge> Want a mix of public clouds and distros. Based largely on interest, some from this group.
19:33:12 <davidlenwell> I want to have the kinks worked out and vendors posting to it by the end of the week ..
19:33:27 <davidlenwell> we'd love to have a pub cloud to round things out
19:33:42 <sslypushenko__> I can provide result from Mirantis Openstack Express
19:33:53 <davidlenwell> hogepodge:  any and all of those that have a presense in the bay area .. catherineD and I have discussed and agreed we'd be happy to show up and help them get going.
19:34:11 <hogepodge> sslypushenko Mirantis is on the list too.
19:34:28 <catherineD> hogepodge: how many vendors on the list?
19:34:33 <davidlenwell> great.. i think we should classify our own companies as Alpha users.. since we have inside people
19:34:45 <hogepodge> I have a strong interest in working directly with everybody. This is feeding into efforts for defcore/interop.
19:34:56 <hogepodge> This is not a set list, but:
19:35:02 <davidlenwell> so marantis ibm and bluebox will probably get more attention than anyone else.. piston will be our first real beta user
19:35:28 <hogepodge> public cloud providers: HP, Rackspace, Bluebox, Dreamhost
19:35:41 <davidlenwell> bluebox isn't a pub cloud .. just fyi
19:35:47 <davidlenwell> hosted private cloud
19:36:00 <davidlenwell> dedicated hardware.
19:36:01 <hogepodge> private distros: Metacloud/Cisco, SuSE, Mirantis, IBM, VMWare
19:36:14 <hogepodge> public/private cloud. Thanks for the correction.
19:36:42 <hogepodge> Bluebox, Mirantis, IBM appear on the list because of our affiliations with them.
19:37:37 <hogepodge> I had a meeting the Metacloud two weeks ago, they're on board. I'll be contacting the rest. Please send me contact information for the person in your company so I can start working with them and aggregating results. Will make my life a lot easier for collecting experiences and pain points.
19:37:56 <davidlenwell> okay .. well I'll keep everyone apprised of the status with the folks we've already reached out to.. in the meantime ..
19:38:35 <davidlenwell> for bluebox that is going to be me for the time being.. our team is really small (only 7 work on our openstack product) So I am the only spare resource at the moment.
19:39:39 <davidlenwell> Steve Maxwell <steve.maxwell@pistoncloud.com> is who we're setup to work with at picton
19:39:41 <davidlenwell> piston
19:40:07 <davidlenwell> he's head of Q/A .. they plan on putting a refstack-client run in the nightly build process possibly..
19:40:47 <davidlenwell> tenitively I'll be going into there office on friday to help them get a working tempest config
19:41:19 <davidlenwell> if we don't hear back about our domain in the time period I'll sit on someones desk in their office until I get an answer.
19:42:05 <davidlenwell> lets move on..
19:42:08 <davidlenwell> time is short..
19:42:11 <davidlenwell> #topic new spec: Add refstack-client capabilities mode.  https://review.openstack.org/150114
19:42:18 <davidlenwell> I wanted to give this spec a little of its own time..
19:42:36 <davidlenwell> I consider it an important change and I'd like everyones attention on the review..
19:42:47 <davidlenwell> my comments will be added today.. I've already got several in draft.
19:43:03 <davidlenwell> hogepodge:  do you have anything to add to that?
19:43:15 <hogepodge> My motivation is this: Right now the total number of api tests far exceeds the capabilities tests.
19:43:27 <hogepodge> We have two use cases: run all api tests to see what people are using.
19:43:37 <sslypushenko__> I think that is a realy good idea
19:43:43 <hogepodge> run capabilities test to see if you conform to some interoperability standard.
19:44:12 <davidlenwell> hogepodge: it was our intention early on to get as many results from as many tests as we can to provide data to defcore about any given cloud installation beyond what is currently in the core capabilites file
19:44:33 <hogepodge> Right, both things are important.
19:44:36 <davidlenwell> to get a real world look at what is actually being used to help weigh future additions to the list
19:44:39 <davidlenwell> sure
19:45:00 <catherineD> hogepodge: on the results display page, one can filter to only look at core ...
19:45:01 <hogepodge> In the short term we have recommended capabilities and there's interest in proving that clouds can pass them.
19:45:05 <davidlenwell> but I do like the idea that the client should have the abillity to get a focused short list of tests to run
19:45:27 <davidlenwell> catherineD: the ability to shrink the list of tests will improve run times of the test
19:45:40 <davidlenwell> so in certain conditions this is important ..
19:45:54 <hogepodge> catherineD the problem is there are over a thousand tests that are run. This can take a while and has the potential to pollute the cloud it's run against. Plus makes people worry about negatives that may not matter.
19:46:02 <catherineD> davidlenwell: agree from test time point of view and not from results point of view
19:46:15 <davidlenwell> catherineD: agreed
19:46:39 <davidlenwell> It hink we can satisify this need by simply allowing the client to operate from a targeted list of tests..
19:46:46 <hogepodge> I have a goal of demonstrating defcore interoperability testing, and this is going to make that testing palatable for vendors.
19:47:05 <sslypushenko__> We also can speedup test run if we remove -t key from ./run_tempest.sh script
19:47:15 <hogepodge> Also helps us in debugging tempest tests and reaching the goal of generalized API reporting, which I think is very important for identifying capability updates.
19:47:28 <davidlenwell> I am going to say yes we want this feature.. but we also want to encourage vendors to run more than just the defcore tests
19:47:36 <hogepodge> yes to both.
19:47:39 <hogepodge> absolutely.
19:47:45 <davidlenwell> yes .. okay .. we are all in agreement ..
19:47:59 <davidlenwell> lets review the spec and get it landed and get the code in as fast as we can
19:48:15 <hogepodge> One is for now, the other is for the future and health of interop.
19:48:23 <davidlenwell> lets move on..
19:48:32 <catherineD> sslypushenko__: The reason we have -t on run_tempest is for test repeatability
19:48:40 <davidlenwell> #topic existing spec: meta and uuid data
19:48:52 <davidlenwell> catherineD:  lets go back to that point in open discussion
19:49:06 <catherineD> sslypushenko__: this is the only way that the test is repeatable from run to run ...
19:49:54 <hogepodge> The uuid/metadata spec is absolutely important to get through.
19:49:59 <davidlenwell> its been a few weeks since I've posted an update to this spec.. last week I didn't have enough time to meet my commitment to get it posted.. I am going to attempt to make up for that and get it in by tomorrow
19:50:18 <hogepodge> Tempest tests fail on setting up networks. We can't pin and fix tests.
19:51:06 <hogepodge> I have a good working relationship with Matt, and I'd be happy to take this over and work with Sergey for getting his work in front of the qa team.
19:51:28 <davidlenwell> hogepodge:  give me until tomorrow before we do that.
19:51:28 <sslypushenko__> catherineD  Will tests failed if we run them without -t flag?
19:51:54 <davidlenwell> hogepodge:  I'm fine with off loading things when I don't have time.. but I've set asside today and tomorrow to catch up on refstack
19:52:16 <davidlenwell> hogepodge: I'd also like your input .. I'll get an etherpad going later today and share it with you
19:53:16 <sslypushenko__> davidlenwell Any ideas how to fix issue with autogenerated tests?
19:53:35 <hogepodge> If you don't have the cycles it's ok to offload that work. I'm very interested in it, and it's been three weeks since an update, almost two since the last comment.
19:54:01 <davidlenwell> hogepodge:  I'm on the same page as you .. but I'd like to get in this last cycle before passing it on ..
19:54:33 <hogepodge> I don't want to keep getting into this cycle of "rush for an update then wait three weeks". It really needs continuous attention.
19:54:53 <hogepodge> I'm not trying to be a jerk. We have a very limited timeframe before Vancouver.
19:55:09 <davidlenwell> I'm with you 100% .. lets work together to get it finished
19:55:15 <davidlenwell> I just ask that you work with me and not around me
19:55:17 <davidlenwell> thats all
19:55:46 <hogepodge> I'll work with you, but that means getting work done.
19:56:03 <hogepodge> We can't afford to spin our wheels any more.
19:56:08 <davidlenwell> we're pretty short on time .. catherineD will you talk with sslypushenko__ in #refstack
19:56:22 <davidlenwell> hogepodge: yes.. lets do it .. today and tomorrow are the days that I have the time ..
19:56:37 <davidlenwell> I was thinking tomorrow morning is a good time to dedicate to that
19:56:55 <davidlenwell> today I am a little distracted by getting the website up and api server deployed
19:57:06 <davidlenwell> (almost done with getting the api online now)
19:57:31 <catherineD> hogepodge: davidlenwell: while you work on the UUID spec, please consider the case of autogenerated test cases as mentioned by sslypushenko__:
19:57:51 <jaosorior> greghaynes: no +2, sorry dude :/
19:58:10 <davidlenwell> we will .. thank you catherineD
19:58:14 <jaosorior> wrong chat, sorry guys
19:58:33 <davidlenwell> okay .. lets end this and get out of the next teams way.. I'll update #refstack with status on the api server
19:59:38 <catherineD> BTW I have created this doc https://docs.google.com/document/d/12GUhUphBSzPuQqp7WreZ-DXaINnUWSrYAYb4CyPmpAw/edit
19:59:47 <catherineD> to track tempest issues
19:59:52 <davidlenwell> #endmeeting