14:02:39 <dulek> #startmeeting Kuryr
14:02:40 <openstack> Meeting started Mon Jun  3 14:02:39 2019 UTC and is due to finish in 60 minutes.  The chair is dulek. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:02:41 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:02:43 <openstack> The meeting name has been set to 'kuryr'
14:02:50 <ltomasbo> o/
14:02:55 <dulek> Hi folks!
14:03:06 <dulek> Maybe I can ping dmellado as well…
14:03:06 <maysams> Hi!
14:03:18 <dmellado> hiya fellows
14:03:24 <dmellado> meeting and meeting all over again
14:03:41 <dmellado> so, how's life and patches? ;)
14:04:30 <dulek> Pretty okay, getting merged. ;)
14:05:04 <dulek> We recently merged some SR-IOV stuff from Samsung folks.
14:05:31 <dulek> This got as a bit of an issue, as we now have 2 new dependencies.
14:05:51 <dulek> But this seems to be solved by RDO folks now, I only need to add it to the RPM spec.
14:06:10 <dulek> Besides that I started rewriting kuryr-cni plugin to golang.
14:06:38 <ltomasbo> nice!
14:06:49 <dulek> This is to be able to run kuryr-cni when host doesn't have Python. It's needed for OpenShift as well as OpenStack Magnum.
14:07:44 <dulek> So in the next few days I should have a patch ready. At the moment I have a version that partially works, but I need to clean it up and add support for routes and DNS settings.
14:07:55 <dulek> Interestingly the worst part is reading the o.vo.
14:08:35 <dulek> That's all the updates from my side, I think.
14:08:36 <ltomasbo> dulek, ohh, I thought it was only the trigger what was needed in golang
14:08:44 <dulek> ltomasbo: trigger?
14:08:45 <ltomasbo> dulek, not the complete daemon
14:09:09 <ltomasbo> the part that is called from kubernetes
14:09:32 <dulek> ltomasbo: Yes, but still - kuryr-daemon returns the o.vo and kuryr-cni is reading it and returning e.g. IP's to kubernetes.
14:09:51 <ltomasbo> dulek, if running containerized, why can't we add python to the container?
14:09:51 <dulek> I didn't wanted to change format of communication to keep compatibility.
14:10:14 <dulek> ltomasbo: kuryr-cni is called by kubelet, so it's running on the host.
14:10:55 <dulek> ltomasbo: We use python in kuryr-cni to find the correct pod to do `docker/runc exec`.
14:11:40 <dulek> ltomasbo: Even if I'd rewrote that JSON parsing to bash, we're still not guaranteed we have bash in the container (Magnum's case IIRC).
14:12:06 <ltomasbo> dulek, got it
14:12:43 <ltomasbo> dulek, it just been a while since I looked at cni side code
14:13:18 <dulek> ltomasbo: Yeah, it's a bit complicated. It should get simpler now with just a binary.
14:13:50 <maysams> Me too. I actually just took a quick look on the code to refresh my mind
14:13:59 <dulek> Any other updates/requests? dmellado, ltomasbo, maysams?
14:14:41 <dmellado> how does the work of migrating cni to go goes?
14:14:46 <dmellado> I've been a little bit out of that
14:14:57 <ltomasbo> lol, that is what we just discussed...
14:15:29 <ltomasbo> not much to report on my side, the namespace/NP subnet pre-population patch was recently merged
14:15:38 <ltomasbo> as well as some maysams bug fixes for NPs
14:15:40 <dulek> dmellado: I have a working prototype, will propose the patch in a few days, once it's cleaned up.
14:16:26 <dmellado> ltomasbo: just now I did get to the backlog xD
14:16:31 <dmellado> thanks dulek!
14:16:50 <dmellado> maysams: great work on the NP devref, left a few comments, but mostly fine with it!
14:17:10 <ltomasbo> dulek, this is what needs to be moved to go, right? https://github.com/openstack/kuryr-kubernetes/blob/master/kuryr_kubernetes/cni/main.py
14:17:52 <dulek> ltomasbo: Yes!
14:18:05 <maysams> dmellado: Thank you for taking a look on it
14:18:18 <dulek> ltomasbo: The good thing is that CNI provides a framework to build plugins in golang.
14:18:27 <dulek> So a lot of the code is already written.
14:18:30 <ltomasbo> dulek, nice!
14:18:44 <ltomasbo> and I think that is even CNCF supported...
14:18:44 <dulek> And we get request validation, automatic CNI version handling, etc.
14:18:50 <dulek> ltomasbo: Yup!
14:18:56 <maysams> cool
14:18:58 <ltomasbo> that's cool
14:20:18 <dulek> Anything else or do we close the meeting?
14:20:24 <maysams> From my side, the NP issues found got merged
14:21:04 <maysams> as dmellado pointed, I've proposed a devref doc for NP
14:22:12 <maysams> and I continue maintaining the named port patch for NP
14:22:18 <dulek> Yeah, I think I already reviewed it.
14:22:20 <maysams> that's all
14:22:23 <maysams> yup
14:22:29 <maysams> Thanks!
14:23:23 <dulek> Okay, so I guess that's it for today.
14:23:28 <dulek> Thanks all!
14:23:40 <dulek> #endmeeting