03:04:22 <tfukushima> #startmeeting kuryr 03:04:23 <openstack> Meeting started Tue Jun 14 03:04:22 2016 UTC and is due to finish in 60 minutes. The chair is tfukushima. Information about MeetBot at http://wiki.debian.org/MeetBot. 03:04:24 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 03:04:26 <openstack> The meeting name has been set to 'kuryr' 03:05:05 <tfukushima> Hi. Who's up for the Kuryr weekly meeting? 03:05:31 <tango> Hi, this is Ton Ngo from the Magnum team 03:05:43 <vikasc_> o/ 03:05:58 <vikasc_> hi Ton 03:06:04 <tfukushima> tango: Hi, welcome! 03:06:44 <tfukushima> #info vikasc_, tango and tfukushima are present 03:07:21 <tfukushima> So we don't have the agenda for today. Let's start with the IPAM issue. 03:07:36 <tfukushima> #topic IPAM improvements 03:08:18 <tfukushima> vikasc_: Do you want to share something on this? 03:08:51 <vikasc_> tfukushima: I have pushed a devref as we discussed in last to last meeting 03:09:04 <tfukushima> Great. 03:09:28 <vikasc_> plus there are PRs in docker for more than 2 weeks, i am not getting any response, good or bad 03:09:29 <vikasc_> :) 03:09:59 <tfukushima> #info Vikas' devref for overlapping CIDRs https://review.openstack.org/#/c/326894/ 03:10:19 <tfukushima> vikasc_: Could you give us a link, please? 03:11:06 <vikasc_> https://github.com/docker/docker/pull/23163 03:11:32 <vikasc_> its there, in devref also in "references" and in "dependency" section 03:12:17 <tfukushima> Ok, I missed it. Thanks. 03:13:33 <tfukushima> Hmm, in my opinion, Docker team is super busy and not so responsive sometimes. 03:13:46 <tfukushima> Actually that's my actual experience. 03:14:00 <tfukushima> So it might require some additional patience. 03:14:08 <vikasc_> ok 03:14:44 <tfukushima> In the meantime, we'd better to review your devref and patches for Kuryr upstream . 03:14:47 <vikasc_> this is my first time o not sure if i am missing something to get their attention. I mean pinging them on irc or ml 03:15:28 <vikasc_> yes, once devref is approved i can start pushing kuryr side changes 03:16:02 <tfukushima> vikasc_: It might be helpful but in my experience, we would not be able to expect the response and the feature or the patch is suddenly added to Docker. 03:16:23 <vikasc_> ok 03:16:40 <vikasc_> i will follow up 03:16:47 <tfukushima> Awesome. 03:17:08 <vikasc_> done on Ipam update from my side 03:17:30 <tfukushima> Let's see this as the long term improvement and wait for a while patiently relaxing. 03:17:38 <tfukushima> vikasc_: Thanks, Vikas. 03:18:37 <tfukushima> So I'd like to move forward to another topic but I'm wondering if tango had anything he wanted to raise. 03:19:08 <tfukushima> #action Everyone (especially tfukushima) reviews Vikas's patches. 03:19:16 <tango> Hi Taku, Vikas, for myself, I started to work on integrating Kuryr to our Swarm bay 03:19:40 <tango> I do have a few quick questions that maybe you can help with 03:19:58 <tfukushima> Nice. Let us know if you have anything required from our side. 03:20:10 <tango> I found the docker image Kuryr from docker hub, is that a supported way to deploy Kuryr? 03:21:18 <tfukushima> I believe you're mentioning the following Docker image. 03:21:34 <tfukushima> #link kuryr/libnetwork on DockerHub https://hub.docker.com/r/kuryr/libnetwork/ 03:21:50 <tango> Yes, this would fit nicely into Magnum's framework for Swarm bay 03:22:40 <tfukushima> It's built from the current upstream master and it can be used by Magnum for sure. 03:23:04 <tango> That's great, I just want to confirm. 03:23:38 <tango> The Kuryr container seems to work OK, I see Docker being able to connect to Kuryr 03:23:56 <tango> I still have some problem connecting Kuryr to Neutron 03:24:00 <tfukushima> If you wanted to run your own libnetwork plugin modified from the upstream, I think you need to launch it by yourself with the script in the repo or you might want to build your own Docker image. 03:25:08 <tfukushima> tango: Ah, environment variables are used to configure the Neutron endpoints and credentials, so you might want to validate them. 03:25:59 <tango> Yes, I thought I have the variables set up correctly, but maybe I am still missing something 03:26:13 <tfukushima> If you had any problem, let us know on #openstack-kuryr channel. And apuimedo is managing kuryr/libnetwork image AFAIK. 03:26:27 <tango> Is there a Kuryr log that I can check 03:26:29 <tango> ? 03:27:43 <tfukushima> It should emit the log to stdout and the file depending on the config but let me confirm it. 03:28:48 <tango> I see a volume being mounted for log, but maybe I don't have it set up correctly yet 03:29:18 <tango> Anyway, I will continue debugging and will ask for help if I get stuck 03:29:55 <vikasc_> tango: you are welcome 03:30:29 <tfukushima> tango: Ok. I'll make sure the logging will be better. 03:31:07 <tfukushima> #topic kuryr/libnetwork image on DockerHub 03:31:26 <tfukushima> #action tfukushima improves the logging capability 03:32:07 <tfukushima> #action tfukushima asks apuimedo if kuryr/libnetwork is the official image built from the upstream 03:32:22 <tango> Thanks Taku, that would be very helpful. 03:32:58 <tfukushima> tango: Do you have anything else you might want to share at this moment? 03:33:14 <tango> That's all I have for now. 03:33:25 <tfukushima> Good. Thanks. 03:33:55 <tfukushima> #topic K8s integration 03:34:33 <tfukushima> So I submitted a series of patches to improve the devref for K8s watcher. 03:35:17 <tfukushima> #link watcher devref improvement patches https://review.openstack.org/#/q/project:openstack/kuryr+branch:master+topic:watcher-devref-fix 03:35:48 <tfukushima> Please take a look at it when you had some time guys. 03:36:50 <vikasc_> tfukushima: i already provided my review comments. btw are these patches related to this bug https://bugs.launchpad.net/kuryr/+bug/1583972 03:36:50 <openstack> Launchpad bug 1583972 in kuryr "The style of Raven devref is broken" [Undecided,New] 03:37:36 <tfukushima> vikasc_: Actually this patch is related to it. https://review.openstack.org/#/c/319125/ 03:38:36 <vikasc_> somehow its not getting updated on launchpad 03:38:38 <tfukushima> Hmm, somehow the link is missed even though I included Closes-Bug: #1583972 in the commit message. 03:38:39 <openstack> bug 1583972 in kuryr "The style of Raven devref is broken" [Undecided,New] https://launchpad.net/bugs/1583972 03:38:58 <vikasc_> i think Closes-bug will work 03:39:01 <vikasc_> small b 03:39:18 <tfukushima> Oh, my bad :-( 03:39:20 <vikasc_> not sure 03:39:44 <vikasc_> and please assign it to yrself on launchpad also. 03:40:22 <tfukushima> Done. 03:40:23 <tfukushima> https://wiki.openstack.org/wiki/GitCommitMessages#Including_external_references 03:40:40 <tfukushima> "Closes-Bug: " looks correct. 03:41:06 <tfukushima> Meh, sometimes it goes wrong. 03:41:10 <vikasc_> hmm .. not sure then.. 03:41:17 <limao> yes，sometimes it just not work 03:41:19 <tfukushima> I'll resolve it manually. 03:41:28 <vikasc_> thanks 03:41:48 <tfukushima> limao: Hi limao. 03:41:54 <limao> I get similiar problem before... not sure why 03:42:11 <vikasc_> limao: I remember :) 03:42:34 <limao> Hello tfukushima vikasc 03:42:50 <tfukushima> limao: Do you want to share anything? I was thinking to finish our meeting otherwise. 03:43:25 <limao> any one know how to run full test in local devstack？ 03:44:12 <limao> currently，i just push patchset to 03:44:22 <vikasc_> limao: you just need to point kuryr to actual running neutron 03:44:24 <limao> let it trigger 03:44:27 <vikasc_> and tox -e fullstack 03:44:36 <limao> ok 03:44:41 <limao> 3q 03:44:43 <tfukushima> https://github.com/openstack/kuryr/blob/master/tox.ini#L20-L22 03:44:59 <tfukushima> Yes, tox -e fullstack should work. 03:45:00 <limao> I do not have other thing，nice to 03:45:10 <limao> see you 03:45:15 <tfukushima> baohua might know better. 03:45:25 <vikasc_> limao: nice to see you too 03:45:26 <tfukushima> limao: Ok, thank you. 03:45:38 <limao> OK，i will reach him,3q 03:45:46 <vikasc_> tfukushima: i recently ran full stack tests locally 03:46:21 <vikasc_> i added a fullstack test recently , https://review.openstack.org/#/c/328236/ 03:46:41 <tfukushima> Very nice. 03:47:09 <vikasc_> but tox -e py34 is failing somehow 03:47:20 <tfukushima> #link Vikas's IPAM fullstack test patch https://review.openstack.org/#/c/328236/ 03:47:20 <vikasc_> i raised a bug also on launchpad 03:47:51 <tfukushima> vikasc_: Oh yeah, I saw that. Do you have any detailed log? It seemed python3.4 was failed somehow. 03:48:08 <tfukushima> vikasc_: Do you have Python 3.4 installed in your environment? 03:48:36 <vikasc_> tfukushima: honestly, i dint check :) 03:49:03 <tfukushima> It's worthwhile checking, I bet. :-) 03:49:05 <vikasc_> i thought tox will install whaterver requirements are 03:49:15 <vikasc_> tfukushima: many thanks for the tip 03:49:45 <vikasc_> tfukushima: one more thing 03:49:56 <tfukushima> Yeah, it would but it could just create the separated environment with venv. 03:50:37 <vikasc_> tfukushima: we are struggling with fullstack tests failure on gate, and not able to reproduce locally 03:50:42 <vikasc_> especially banix 03:51:13 <vikasc_> tfukushima: just in case, you could suggest something so thought of bringing it to your notice 03:51:29 <tfukushima> OK. Do you have any idea what is the flaky part? 03:51:45 <vikasc_> https://review.openstack.org/#/c/265105/45 03:52:06 <vikasc_> somehow kuryr server is crashing 03:52:25 <vikasc_> http://logs.openstack.org/05/265105/42/check/gate-kuryr-dsvm-fullstack-nv/1b4aea6/logs/screen-kuryr.txt.gz 03:52:30 <vikasc_> have a look please 03:54:43 <tfukushima> Hmmm, that's weird. It looks like the error or Flask or Werkzeug. 03:55:57 <tfukushima> It's happening after the legitimate Kuryr error. There might be the problem on error handling. I'll take a deep look later. 03:56:35 <vikasc_> tfukushima: thanks, one clue i got just now. not sure if that is the reason, i will try 03:56:56 <tfukushima> vikasc_: "Unexpected error happened: No subnet is found using pool 6ca8600c-1d35-43f9-ab95-435a2ac5596c and pool_cidr 10.0.0.0/24" Is it the expected error in the test? 03:57:07 <vikasc_> no 03:57:47 <vikasc_> in release_address ipam api we should be just logging that subnet already deleted 03:57:54 <vikasc_> rather raising exception 03:58:07 <vikasc_> i will try with this change 03:58:23 <tfukushima> Ok, I need some time to see it closely. 03:58:46 <tfukushima> #action tfukushima see the cause of the failure in https://review.openstack.org/#/c/265105/45 03:58:53 <vikasc_> if my experimentaion will not work, i will let you know 03:58:58 <tfukushima> #link Unexpectedly failing fullstack test https://review.openstack.org/#/c/265105/45 03:59:13 <tfukushima> vikasc_: Ok, thanks. 03:59:25 <tfukushima> I think we're running out our time. 03:59:40 <tfukushima> Does anyone have other words? 03:59:47 <vikasc_> nope. 03:59:59 <tango> Nope 04:00:13 <tfukushima> Good. Then I'm closing the meeting. Thanks for attending guys, vikasc, tango and limao. 04:00:23 <tango> Thanks everyone 04:00:23 <tfukushima> #endmeeting