22:33:42 <dims_> #startmeeting containers
22:33:43 <openstack> Meeting started Tue Apr  7 22:33:42 2015 UTC and is due to finish in 60 minutes.  The chair is dims_. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:33:44 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
22:33:47 <openstack> The meeting name has been set to 'containers'
22:34:06 <dims_> #link https://wiki.openstack.org/wiki/Meetings/Containers#Agenda_for_2015-04-07_2200_UTC
22:34:17 <dims_> #topic Roll Call
22:34:37 <sdake> o/
22:34:39 <Tango|2> o/
22:34:40 <dims_> so who's here? :)
22:34:46 <rpothier> o/
22:34:47 <juggler> o/
22:34:47 <madhuri_> o/
22:34:50 <yuanying-alt> o/
22:34:57 <dims_> o/
22:35:10 <dims_> giving it a few seconds
22:35:16 <hongbin> o/
22:35:16 <apmelton> o/
22:35:22 <dims_> #topic Announcements
22:35:37 <dims_> Anyone have any announcements? Since i am running this adhoc :)
22:36:07 <sdake> no announcements
22:36:13 <dims_> cool moving on
22:36:15 <dims_> #topic Review Action Items
22:36:41 <dims_> no action items from last week
22:36:56 <dims_> #topic Blueprint/Task Review
22:37:15 <dims_> #link  https://blueprints.launchpad.net/magnum/+spec/python-k8sclient
22:37:22 <madhuri_> About the python-k8s client code
22:37:24 <dims_> madhuri_: want to drive this?
22:37:27 <madhuri_> yep
22:37:31 <dims_> go for it
22:37:53 <madhuri_> Discussion was going about where this code should live?
22:38:14 <madhuri_> I have talked o google guys about some of the issues
22:38:25 <dims_> k
22:38:28 <madhuri_> And they then asked to commit the code in their repo
22:38:42 <madhuri_> So what would be better place for the code?
22:38:57 <madhuri_> A new stackforge project or in google k8s repo?
22:39:11 <madhuri_> sdake suggested new stackforge project.
22:39:13 <dims_> will you or anyone you know have write access to their repo? or do we end up filing bugs and PR?
22:39:16 <hongbin> I think google k8s repo is better
22:39:35 <sdake_> i am not convinced google will actually maintain the repo
22:39:43 <madhuri_> I am not sure about it.
22:39:57 <sdake_> they want it as an example not as an upstream
22:39:58 <hongbin> k
22:40:20 <dims_> madhuri_: any fixes in their code generator itself?
22:40:25 <juggler> i'm guessing it would benefit the community as an upstream, right?
22:40:27 <madhuri_> They have no python client code running
22:40:43 <madhuri_> Yes, I have filed issues for them
22:41:07 <madhuri_> That's why they want it running in their repo.
22:41:28 <dims_> madhuri_: will they publish to pypi as well?
22:41:39 <sdake_> we dont want to pull the kubernetes repo to pull a language binding
22:41:46 <sdake_> that doesn't make any sense
22:41:58 <sdake_> it needs to be a separate repo
22:42:03 <madhuri_> No, I don't think so.
22:42:19 <sdake_> madhuri_ were they talking about a separate repo in the google namesapce or something o ngithub?
22:42:22 <juggler> does the google k8 repo have something gerrit-like?
22:42:36 <sdake_> they have post-commit checks
22:42:40 <madhuri_> No, in k8s repo itself
22:42:49 <dims_> so, they may not maintain it, they may not publish to pypi. which makes the decision easy. we should do it in stackforge
22:42:54 <sdake_> in k8s repo itself is unacceptable
22:42:54 <dims_> my 2 cents
22:43:14 <dims_> +1 sdake_
22:43:15 <madhuri_> in contrib directory
22:43:16 <juggler> based on the previous discussion, +1 stackforge
22:43:39 <sdake_> still unacceptable imo
22:43:44 <dims_> so let's vote then? who is in favor of stackforge?
22:43:46 <sdake_> ;)
22:43:47 <madhuri_> Ok. So will create a new stackforge project for it.
22:43:52 <madhuri_> Thanks,
22:44:04 <madhuri_> The second point is the TLS support.
22:44:24 <madhuri_> They swagger generated code doesn't have it.
22:44:39 <madhuri_> So I think we need to plan our own.
22:44:41 <sdake_> swagger is a bunch of fail
22:45:03 <madhuri_> I agree sdake_
22:45:20 <hongbin> One way is to customize the generator to support TLS
22:45:57 <dims_> is it easy to layer over the generated code?
22:46:24 <madhuri_> No, swagger code is a big mess. So I think apparently not.
22:46:27 <dims_> hongbin: then we have to maintain the generator patch
22:46:39 <hongbin> dim_: true
22:46:42 <madhuri_> I think it is good to customize the generator
22:47:33 <sdake_> here is a proposal
22:47:34 <dims_> madhuri_: either that or monkey patch their generated code if layering over their code is not simple
22:47:38 <sdake_> we got swagger code that works
22:47:47 <sdake_> we hack it until it doesn't look like garbage anymore
22:47:51 <sdake_> and has test cases
22:47:59 <sdake_> and if new apis come out we re-swagger and add the new objects
22:48:40 <madhuri_> So we should customize it?
22:48:42 <madhuri_> Right?
22:48:52 <sdake_> that would be my proposal
22:49:00 <sdake_> it should be easy to tell if a new object comes out of the api
22:49:09 <madhuri_> Ok, sure. I will work on it.
22:49:35 <dims_> sdake_: we can try that, and revisit if it becomes headache
22:49:58 <dims_> my fear is they may fix bugs in existing objects and we'll have to revisit the garbage
22:50:06 <madhuri_> I will look into it and will let you know how it takes to be.
22:50:24 <sdake_> I think the soution is to let madhuri_ propoose the path forward
22:50:39 <sdake_> because she has got us this far, which is further then anyone else of the 3 or 4 cats that have tried this project :)
22:50:46 <juggler> +1
22:50:48 <dims_> +1 sdake_
22:51:05 <madhuri_> I will try
22:51:11 <dims_> go madhuri_!
22:51:22 <madhuri_> Ok and then the third point.
22:51:43 <madhuri_> v1beta1 and v1beta2 is deprecated
22:51:52 <madhuri_> So we should move to v1beta3
22:52:06 <madhuri_> I will regenerate the code for it and test it.
22:52:07 <sdake_> agree
22:52:10 <dims_> +1
22:52:23 <juggler> thx madhuri!
22:52:31 <dims_> ++
22:52:35 <dims_> #topic Open Discussion
22:52:46 <dims_> floor is open
22:52:47 <madhuri_> Ok then I have got my points clarified so will work on this,
22:52:48 <sdake_> tango any luck
22:52:50 <madhuri_> Thanks
22:53:07 <juggler> there's a potential bug in devstack, for those building it. please note: https://bugs.launchpad.net/magnum/+bug/1440982
22:53:08 <openstack> Launchpad bug 1440982 in Magnum "devstack fail with noattribute error" [Undecided,New]
22:53:10 <sdake_> on the load balancer thingy
22:53:18 <Tango|2> sdake_: Been reading a lot of docs and heat templates.  I got a decent idea now
22:53:31 <sdake_> nice
22:53:32 <juggler> should we bother mentioning it in our quickstart or hope that infra fixes it?
22:53:33 <Tango|2> I will ping you to clarify a few things
22:53:42 <sdake_> I can give a quick update on kobject
22:54:03 <dims_> +1 sdake_
22:54:07 <sdake_> juggler i suspect we hope people will be usinga working devstack
22:54:20 <sdake_> i've studied the  curent object code about 20 hours
22:54:30 <sdake_> and think it will be pretty easy to merge the pod/rc/service objects into one object
22:54:37 <sdake_> a kobject (kubernetes object)
22:54:48 <sdake_> kubernetes has other types of objects as wel)
22:55:00 <sdake_> but they all take a file manifest
22:55:02 <sdake_> (or dont)
22:55:19 <sdake_> their rest api doesn't actually take arguments
22:55:43 <dims_> sdake_: using oslo.versionedobjects?
22:56:01 <sdake_> i am not going to port to oslo.versioned objects yet
22:56:06 <dims_> k
22:56:07 <sdake_> I am not sure if I have time in the cycle
22:56:12 <sdake_> but the current coe is versionedobjects
22:56:19 <sdake_> just not the oslo version
22:56:22 * dims_ nods
22:56:39 <sdake_> if I have time I will port us as well
22:56:52 <sdake_> I woudl be odne but fell and in alot of pain
22:56:59 <sdake_> kids shit on the floor
22:56:59 <juggler> ouch
22:57:01 <sdake_> bad for business
22:57:16 <dims_> hope you feel better soon sdake_!
22:57:22 <sdake_> tx
22:57:26 <juggler> feel better S
22:57:45 <dims_> +1 for kobject as laid out sdake_
22:57:58 <juggler> +1 here too
22:58:11 <sdake_> we are out of time but goog doesn't wnt us to wrap their rest api
22:58:17 <sdake_> we can discuss tht in channel if you ike ;)
22:58:31 <dims_> +1 sdake_
22:58:41 <dims_> rest of open discussion please move over :)
22:58:54 <sdake_> dont forgt #endmeeting
22:58:56 <dims_> thanks everyone
22:58:58 <dims_> #endmeeting