03:01:30 <tfukushima> #startmeeting Kuryr 03:01:31 <openstack> Meeting started Tue Mar 8 03:01:30 2016 UTC and is due to finish in 60 minutes. The chair is tfukushima. Information about MeetBot at http://wiki.debian.org/MeetBot. 03:01:32 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 03:01:34 <openstack> The meeting name has been set to 'kuryr' 03:02:08 <tfukushima> Who's up for Kuryr IRC meeting today? 03:02:16 <baohua> morning&evening, everyone~ 03:02:22 <baohua> hi tfukushima 03:02:22 <banix> o/ 03:02:24 <fawadkhaliq> evening 03:02:31 <fawadkhaliq> o/ 03:03:02 <tfukushima> #info baohua, banix, fawadkhaliq and tfukushima are present 03:03:36 <tfukushima> #link Today's agenda https://wiki.openstack.org/wiki/Meetings/Kuryr#Meeting_March_8.2C_2016 03:04:01 <tfukushima> Ok, let's get started. 03:04:29 <tfukushima> #topic Kubernetes Integration 03:04:49 <tfukushima> #link Spec proposal by irenab https://review.openstack.org/#/c/281132/ 03:06:07 <tfukushima> #action Everyone reviews irenab's spec proposal 03:06:34 <fawadkhaliq> tfukushima: thanks, will do. I did a quick review early on, Will do again on the latest patch 03:06:53 <baohua> +1, will do it 03:07:18 <tfukushima> Yeah, it seems to be updated recently and it'd be again, I guess. 03:07:31 <banix> sure will do too 03:08:30 <tfukushima> It has the descriptions about the API watcher and the CNI driver. 03:08:50 <banix> tfukushima: taku i hear you have a prototupe working 03:09:02 <banix> is it along the lines being described in the spec? 03:09:33 <tfukushima> I think I have to write the devref rather than the higher level spec. 03:10:16 <tfukushima> I'm playing with the apiserver of K8s. 03:10:46 <banix> cool 03:11:27 <tfukushima> The API allows you to get a line of JSON data when you call GET against the certain endpoint with the query string, i.e., /api/v1/pods?watch=true 03:12:02 <tfukushima> You can see what's happening with curl. 03:12:16 <baohua> tfukushima is your prototype a multi-node env? 03:12:44 <tfukushima> At this point one master and two worker nodes. Small cluster. 03:13:04 <banix> sounds multi-node to me :) 03:13:04 <baohua> thanks 03:13:31 <tfukushima> And I'm doing the api watcher-ish stuff, so only the master is my concern so far. 03:14:08 <tfukushima> I'll push back to Kuryr updatream when it gets some shape. 03:14:40 <tfukushima> Oh, I have a question about that. 03:14:51 <banix> great thanks. Mike Spreitzer who dont seem to be around now is also working on a prototype 03:15:45 <tfukushima> Nice. We might need to coordinate our tasks not to be conflicted. 03:16:26 <banix> i think he is doing the OTHER approach discussed for a quick short term solution but should leave it to him to describe what he has 03:16:41 <banix> any of you going to KubeCon? 03:16:45 <banix> in Europe? 03:16:51 <banix> London I believe 03:17:47 <baohua> banix possible not for me, will u go? 03:17:53 <tfukushima> I hope I could. We'd love to see that proposal but yes, it's up to him. 03:18:45 <banix> I am not going. Mike is going. Should let us know how things are in our next meeting hopefully 03:19:24 <baohua> yes, that will be a nice sharing 03:20:02 <tfukushima> Anyways, here's my question: To get a line of JSON data, I'd like to use asyncio, the async I/O library introduced in Python 3.4 not to be blocked for the next event propagation. Does it make sense to drop Python 2.6 support for that? 03:21:16 <banix> 2.7 you mean? 03:21:28 <baohua> to support curl-based api querying? 03:21:30 <tfukushima> Yes. Python 2.x. 03:22:09 <tfukushima> To support better async I/O mechanisms and more features introduced only in Python 3.x. 03:22:30 <tfukushima> curl is blocking actually, I believe. 03:23:29 <banix> i understand what you are saying. is the rest of kuryr ok with 3.4? 03:23:54 <tfukushima> There's Trollius for the equivalent but it's saying it's deprecated and recommending to move Python 3.x. 03:24:00 <tfukushima> http://trollius.readthedocs.org/deprecated.html 03:24:02 <banix> i know we test at gate but beyond that 03:24:45 <baohua> a little worried not full tested yet. 03:25:25 <tfukushima> Maybe we can separate the API watcher from the rest but it'd be a bit bothering. 03:26:45 <tfukushima> Anyways, I'll submit the patch or something and propose this in another way as well but I wanted to know how you would feel about that. 03:27:29 <tfukushima> #link Python3 support in OpenStack https://wiki.openstack.org/wiki/Python3 03:28:15 <tfukushima> Sorry, I took time. Let's move on to the next topic. 03:28:31 <tfukushima> #topic Nested Containers and Magnum Integration 03:28:52 <tfukushima> fawadkhaliq: Any update? 03:29:44 <fawadkhaliq> tfukushima: yes, the spec is in good shape 03:29:52 <fawadkhaliq> have several +1s and a +2 03:30:01 <fawadkhaliq> I am hoping we merge it soon 03:30:23 <fawadkhaliq> And in parallel, we are planning to kick off the discussion on Kuryr Agent 03:30:28 <fawadkhaliq> there is a separate spec for that 03:30:38 <fawadkhaliq> that's the current update and plan moving forward. 03:30:45 <tfukushima> Nice. I seriously should look at the spec! It's my shame. 03:31:05 <fawadkhaliq> tfukushima: thanks 03:31:48 <tfukushima> fawadkhaliq: BTW, have you tried Magnum? 03:32:09 <fawadkhaliq> tfukushima: yes, I have in general usage sense 03:32:20 <fawadkhaliq> are you referring to some specific? 03:32:28 <fawadkhaliq> s/some/something/ 03:32:50 <tfukushima> fawadkhaliq: My friend tried it but he told me he got stuck in the middle of creating a bay. 03:33:18 <fawadkhaliq> tfukushima: okay, maybe he can drop an email 03:33:54 <tfukushima> Yeah, he'd but I was just wondering how stable Magnum is now. 03:34:29 <fawadkhaliq> tfukushima: its reasonable, the base infra has to be perfect to make it work 03:34:55 <fawadkhaliq> tfukushima: sine Heat is invovled and sometimes Heat stacks are just stuck forever 03:35:12 <fawadkhaliq> and thats normally tied to the base infra 03:35:46 <tfukushima> fawadkhaliq: I see. That makes sense. Thanks. 03:35:55 <fawadkhaliq> tfukushima: npp 03:36:05 <fawadkhaliq> so thats all on the nested container topic for today 03:36:14 <tfukushima> #action Everyone review fawadkhaliq's spec again. 03:36:28 <tfukushima> Good. 03:36:30 <fawadkhaliq> thanks, hoping to get it merged soon. 03:37:09 <tfukushima> #topic Existing Networks 03:37:36 <tfukushima> banix: You might want to share something about that. 03:37:46 <banix> yes sure 03:37:59 <banix> so the Neutron tags got merged for neutron networks 03:38:24 <tfukushima> Yay! 03:38:30 <banix> in that light, we can get rid of the use of neutron network names 03:38:30 <tfukushima> Finally. 03:38:34 <banix> yes 03:38:52 <banix> so i have a patchset to do that. Will update it based on the comments 03:38:58 <tfukushima> I heard you tried it and it worked? 03:39:08 <banix> this will just change the kuryr to use tags 03:39:19 <banix> yes it works 03:39:39 <banix> Then I am going to update your existing network spec Taku 03:39:54 <banix> I had made changes based on our stop-gap planning 03:40:01 <tfukushima> banix: Yes, please. I left it forever. :-( 03:40:07 <banix> but now, I can uodate it to use tags 03:40:24 <tfukushima> That's sweet. 03:40:34 <banix> tfukushima: no worries. Gal and Toni mentioned you are busy 03:41:06 <banix> So that will get rid of the using name ugliness and add the support for existing networks in a clean way 03:41:41 <banix> That’s all for me unless there are questions 03:42:19 <tfukushima> Cool. 03:43:01 <tfukushima> Ok, we shall go to the next topic. 03:43:14 <tfukushima> #topic Packaging 03:43:50 <tfukushima> Actually I have no idea about the situation of the packaging. 03:44:24 <banix> We shall leave this for next week then :) We need Toni to get the status 03:44:57 <tfukushima> Yeah, he should know better and I saw hi submitted a few patches. 03:45:16 <tfukushima> #topic Fullstack and Rally integration 03:45:48 <tfukushima> baohua: Would you like to share something? 03:45:53 <baohua> yes 03:46:01 <baohua> two patches under review, welcome anyone's comments, thanks! 03:46:03 <baohua> https://review.openstack.org/#/c/265105/ 03:46:08 <baohua> https://review.openstack.org/#/c/286531/ 03:46:17 <baohua> that's it :) 03:47:13 <baohua> tfukushima, pls go ahead :) 03:47:23 <tfukushima> baohua: Thanks. I have to review them. 03:47:30 <tfukushima> #action Everyone reviews haohua's Rally tests. 03:48:16 <tfukushima> baohua: The patch looks well documented and I like it. 03:48:25 <baohua> thanks :) 03:49:02 <tfukushima> Ok, let's wrap up the meeting. 03:49:09 <tfukushima> #topic Open Discussion 03:49:29 <tfukushima> Does anyone have anything else to discuss? 03:50:04 <tfukushima> banix; Oh, I submitted a patch. https://review.openstack.org/#/c/289147/ 03:50:32 <tfukushima> I'm sadly a OSX user and I wanted to run unit tests. :-p 03:50:44 <banix> cool 03:51:20 <banix> will have a look 03:51:53 <tfukushima> Good. I think we can finish today's meeting. 03:51:58 <fawadkhaliq> nothing else from my side 03:52:07 <fawadkhaliq> yay! we save 9 min :-) 03:52:09 <fawadkhaliq> :P 03:52:14 <banix> thanks taku 03:52:18 <baohua> great! thanks everyone~ 03:52:21 <banix> and everybody :) 03:52:37 <tfukushima> Thanks for attending, guys. 03:52:42 <baohua> :) 03:52:44 <fawadkhaliq> thankjs 03:52:46 <fawadkhaliq> bye 03:52:50 <baohua> bye 03:52:56 <tfukushima> #endmeeting