15:00:21 <anteaya> #startmeeting third-party
15:00:22 <openstack> Meeting started Mon Jan 12 15:00:21 2015 UTC and is due to finish in 60 minutes.  The chair is anteaya. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:23 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:26 <openstack> The meeting name has been set to 'third_party'
15:00:45 <anteaya> anyone here for the third-party meeting?
15:01:35 <smcginnis> anteaya: Hi
15:01:47 <anteaya> smcginnis: hello there
15:02:31 <anteaya> smcginnis: did you have something you would like to discuss?
15:02:32 <asselin_> hi
15:02:38 <anteaya> asselin_: hi
15:02:46 <anteaya> asselin_: did your patch merge?
15:02:51 <smcginnis> anteaya: Nope, just thought I would listen in in case there was anything interesting.
15:02:56 <asselin_> anteaya, yes it did :)
15:02:59 <anteaya> smcginnis: great thanks
15:03:06 <anteaya> asselin_: congratulations
15:03:18 <anteaya> asselin_: are things in place for the sprint?
15:04:00 <asselin_> anteaya, so far so good. I'd like to do a few more beforehand to iron out any remaining issues
15:04:15 <anteaya> good idea
15:04:27 <anteaya> smcginnis: do you know about asselin_'s work here?
15:05:14 <anteaya> smcginnis: asselin_  has been working to implement this infra spec
15:05:14 <asselin_> hi smcginnis
15:05:20 <anteaya> #link http://specs.openstack.org/openstack-infra/infra-specs/specs/puppet-modules.html
15:05:20 <smcginnis> anteaya: Is that the in-tree CI?
15:05:28 <smcginnis> Hey asselin_!
15:05:34 <anteaya> this is the prep work for the in-tree CI spec
15:05:35 <asselin_> I'm working on both
15:05:47 <anteaya> asselin_: do you want to explain?
15:05:59 <anteaya> perhaps smcginnis can attend the sprint and help out?
15:06:00 <smcginnis> Oh, yes, I had seen some of the puppet work. Very nice!
15:06:17 <asselin_> yes, the puppet splits is straight forward.
15:06:23 <smcginnis> I may be able to a little. Next three weeks are going to be hell for me. :/
15:06:36 <asselin_> move the 'modules' in system-config/modules into their own repos
15:06:38 <anteaya> smcginnis: why is that, if you don't mind me asking?
15:06:39 <smcginnis> I would be willing to do what I can though.
15:06:46 <anteaya> smcginnis: thanks
15:06:53 <smcginnis> anteaya: Just travel and other commitments.
15:07:12 <smcginnis> At least one of the travel weeks is for cinder.
15:07:22 <anteaya> smcginnis: I hear that
15:07:25 <anteaya> yes
15:07:51 <anteaya> asselin_: here is the page for book the virtual sprints channel for a virtual sprint
15:08:00 <anteaya> #link https://wiki.openstack.org/wiki/VirtualSprints
15:08:18 <asselin_> with the recent patch merged, we can now do the other modules sprints much faster with quicker reviews
15:08:21 <anteaya> asselin_: do you want to add the puppet module sprint, so new people can be pointed to it?
15:08:35 <asselin_> anteaya, sure
15:09:04 <anteaya> asselin_: thanks, add something and then ask at infra meeting to see if any details need to be updated
15:09:18 <anteaya> you have the date at least
15:09:33 <anteaya> smcginnis: so it would be helping out doing reviews
15:09:46 <anteaya> smcginnis: have you split out a puppet module?
15:10:07 <asselin_> yes, the hard part is done :)
15:10:11 <smcginnis> anteaya: Definitely can help with reviews. I've worked with puppet modules a small amount.
15:10:15 <anteaya> asselin_: :D
15:10:47 <anteaya> smcginnis: great, perhaps you can help review asselin_'s patches in the next week
15:11:06 <asselin_> smcginnis, tell me more about your openstack background.
15:11:12 <smcginnis> I'll do what I can.
15:11:13 <asselin_> You're in cinder, which driver?
15:11:24 <anteaya> a large patch was recently merged, congratulations asselin_, to reduce the size of the patches to project-config
15:11:37 <smcginnis> asselin_: I work for Dell. My team just implememnted the Dell Storage Center (Compellent) cinder driver.
15:11:51 <smcginnis> asselin_: I set up the CI for the EqualLogic driver.
15:11:59 <smcginnis> asselin_: Adding Storage Center support now.
15:12:15 <smcginnis> Got involved during Icehouse.
15:12:56 <asselin_> nice
15:13:10 <smcginnis> It's been good so far. :)
15:13:21 <asselin_> you have the ci posting results now?
15:14:10 <smcginnis> asselin_: Yep. Well, just on success so far.
15:14:22 <asselin_> ok :)
15:14:25 <smcginnis> There are still some issues with failure that I want to understand better before I start putting that out there.
15:14:44 <anteaya> smcginnis: did you want to explain what you are seeing?
15:14:47 <asselin_> which ci solution are you using?
15:15:01 <anteaya> asselin_ is pretty good at solving issues
15:15:20 <smcginnis> anteaya: I need to dig in and work with our EQL driver developers. They might have a fix in the works if it actually is that driver.
15:15:34 <anteaya> smcginnis: fair enough
15:15:38 <smcginnis> asselin_: I started with Jay Pipe's stuff. In fact, I think you helped me a little.
15:15:55 <asselin_> ok thanks for the reminder :)
15:16:00 <smcginnis> asselin_: But then I ended up following on Duncan and JohnG's simpler approach.
15:16:03 <smcginnis> https://github.com/stmcginnis/openstack-ci
15:16:23 <smcginnis> It at least helped me understand everything a little better.
15:16:47 <smcginnis> But if we get a common, in tree Jenkins based solution I will switch it over to that when it's ready.
15:16:55 <asselin_> ok nice
15:17:29 <asselin_> yea, I put a spec for that. Still thinking about the best approach. Comments would be appreciated
15:18:00 <asselin_> #link https://review.openstack.org/#/c/139745/
15:18:28 <asselin_> ok I see you as a reviewer
15:18:35 <smcginnis> asselin_: I had given a +1 on the first patch. I'll take another look and see if there's anything to add.
15:18:54 <smcginnis> asselin_: It looked pretty good to me though.
15:20:24 <asselin_> I think everyone agrees with the idea. But what the final solution should look like and how to get there requires more thought.
15:20:44 <smcginnis> asselin_: That's always the tricky part. :)
15:20:57 <anteaya> asselin_: how close do you feel you are to a way forward?
15:21:02 <asselin_> there are some comments here that can help.
15:21:05 <asselin_> #link https://review.openstack.org/#/c/138913/
15:21:26 <asselin_> anteaya, not close enough
15:21:57 <asselin_> anteaya, it would be great to have a discussion on the approach. I haven't heard enough of other ppl's ideas
15:21:59 <anteaya> any points of contention?
15:22:06 <anteaya> sorry I haven't read it through
15:22:22 <anteaya> asselin_: give me the tl'dr
15:22:36 <asselin_> what's tl'dr ?
15:23:31 <smcginnis> Too Long, Didn't Read. "Executive Summary" :)
15:23:55 <anteaya> smcginnis: thanks
15:24:00 <anteaya> asselin_: sorry about that
15:24:23 <asselin_> give me a sec to remember. it's early monday morning pre-coffee for me :)
15:25:52 <asselin_> ok, so currently there are puppet classes in system-config/modules that are being split out into puppet-module.
15:26:02 <smcginnis> Gotta drop off. I'll watch for patches to review. Feel free to ping me on #openstack-cinder if there's anything I can do to help with anything.
15:26:20 <anteaya> smcginnis: thanks, see you around
15:26:47 <asselin_> system-config/openstack-project has additional puppet scripts that configures -infra's ci solution (and more)
15:27:18 <asselin_> there are lots of portions in system-config/openstack-project that are also needed by 3rd party, but difficult to reuse
15:27:38 <anteaya> true
15:27:45 <asselin_> so what the best way to reuse them? and where should the resuable pieces go?
15:28:02 <asselin_> should they go to the new puppet-`module` repo?
15:28:14 <anteaya> good question
15:28:16 <asselin_> or a new e.g. 'openstack-shared' module?
15:28:34 <asselin_> or stay where they are as a reusable class?
15:28:40 <anteaya> what would the differences be?
15:28:49 <anteaya> between the choices?
15:29:05 <asselin_> it's a question of design
15:29:26 <anteaya> what are the ramifications of the different design choices?
15:29:33 <asselin_> so does it make sense to put 'openstack' specific parts in the puppet-'module' repos?
15:29:43 <asselin_> or are those supposed to be more generic
15:30:07 <asselin_> if they're supposed to be generic, then we can eliminate that choice, otherwise, it's probably the best place.
15:30:10 <anteaya> well based on what I am hearing you say, part of this is to make it so openstack too consumes this structure
15:30:23 <asselin_> yes, absolutely
15:30:31 <anteaya> in which case openstack is simply another consumer of the material
15:30:39 <asselin_> yes
15:30:55 <anteaya> so what are you feeling are the openstack specific parts?
15:31:13 <asselin_> that's the goal, have openstack & 3rd party consume the same stuff
15:31:22 <anteaya> great
15:31:27 <anteaya> makes a lot of sense
15:32:04 <asselin_> the openstack specific parts will also be the 3rd party specific stuff
15:32:20 <anteaya> can you give me a concrete example?
15:32:24 <asselin_> which is basically the configuration
15:32:29 <anteaya> so I know we are thinking of the same thing
15:32:39 <asselin_> for example, the log server split I did
15:32:39 <anteaya> ah configuration
15:32:44 <anteaya> right
15:33:14 <asselin_> ok...that's not the best example
15:33:50 <asselin_> a lot of the openstack configuration also include ip table & core ssh key setup
15:34:01 <anteaya> right
15:34:13 <asselin_> I think those parts should be separated
15:34:20 <anteaya> okay
15:34:30 <asselin_> because those are deployment specific.
15:34:33 <anteaya> third party will need those parts too, will they not?
15:34:44 <anteaya> ah so deployment specific material
15:34:54 <anteaya> now I am understanding better
15:35:04 <asselin_> right, b/c 3rd party ci would like to install everything on a single machine
15:35:12 <asselin_> b/c the scale is much smaller than -infra
15:35:27 <anteaya> well ideally deployment specific should be configuration files, should they not?
15:35:44 <anteaya> asselin_: yes, one would hope a smaller scale :D
15:36:28 <anteaya> have you seen examples of deployment specific code that is not currently in a configuration file?
15:36:47 <asselin_> yes, perhaps via configuration. not sure how that would work right now
15:38:13 <anteaya> perhaps you could add an agenda item to tomorrow's infra meeting?
15:38:37 <asselin_> I can
15:38:43 <anteaya> identify files that have configuation that is not yet in a separate configuration file
15:39:15 <anteaya> ask for some direction and explain this needs to be address for work on the in-tree CI spec to proceed
15:39:22 <anteaya> does that sound fair?
15:40:07 <asselin_> ok I will investigate that today
15:40:12 <anteaya> awesome
15:40:28 <anteaya> since based upon what I am hearing you say, this appears to be a blocker
15:40:40 <asselin_> well that + where to put the new files
15:40:44 <asselin_> e.g. log server
15:40:45 <anteaya> and in order to get passed it you need to hear from infra
15:40:50 <anteaya> right okay
15:41:04 <anteaya> well you could just randomly pick places until you get consensous
15:41:21 <asselin_> lol :)
15:41:27 <anteaya> or you could ask and either we get consenseus in infra or jim makes a decision
15:41:41 <anteaya> either approach, depending on how much energy you have
15:41:43 <asselin_> I was trying to get that in the spec
15:41:48 <anteaya> me? I'd just ask
15:42:00 <anteaya> fair enough
15:42:21 <anteaya> so far I haven't seen an infra core review it yet
15:42:46 <anteaya> might be nice perhaps to start by requesting a place in core's review queues
15:42:56 <asselin_> no, that's what I like first
15:43:01 <anteaya> and specifically point out where you need direction/feedback
15:43:15 <anteaya> yup, a good place to begin
15:43:21 <asselin_> ok, so ask in -infra or the meeting?
15:43:33 <anteaya> well you can see how busy things are today
15:43:49 <anteaya> if you can get some interaction by asking, by all means, go for it
15:43:57 <asselin_> ok
15:44:07 <anteaya> if folks are busy and you don't get a response, to the meeting
15:44:30 <anteaya> many times meetings are items we just have been too busy to respond to in channel
15:44:39 <asselin_> sounds good.
15:44:43 <anteaya> great
15:44:49 <anteaya> wonderful work here
15:45:05 <anteaya> you are showing good followup, that is the most important thing
15:45:10 <anteaya> good ideas are good
15:45:27 <anteaya> relentless followup is better
15:45:58 <anteaya> asselin_: what else shall we discuss?
15:46:10 <anteaya> you are driving a lot of work right now
15:46:16 <asselin_> well really joined to help answer other's questions
15:46:30 <anteaya> wonderful, I am glad you did
15:46:32 <anteaya> thank you
15:46:40 <anteaya> we seem light on questions today
15:47:03 <asselin_> yes
15:47:34 <anteaya> everyday is different
15:47:51 <anteaya> are you coffeed yet?
15:48:02 <asselin_> no, after the meeting
15:48:12 <anteaya> ah cool
15:48:31 <anteaya> well we have 12 mintues remaining
15:48:37 <anteaya> what would you like to do?
15:48:49 <anteaya> I can end the meeting now and you can get coffee if you like?
15:49:00 <anteaya> I will be heading back to bed myself
15:49:02 <asselin_> no objections :)
15:49:11 <anteaya> very good
15:49:18 <anteaya> thanks for a great meeting today
15:49:23 <anteaya> #endmeeting