14:06:57 <dmellado> #startmeeting kuryr 14:06:58 <openstack> Meeting started Mon Sep 4 14:06:57 2017 UTC and is due to finish in 60 minutes. The chair is dmellado. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:06:59 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:07:02 <openstack> The meeting name has been set to 'kuryr' 14:07:15 <dmellado> #chair apuimedo 14:07:17 <openstack> Current chairs: apuimedo dmellado 14:07:23 <dmellado> apuimedo: around? 14:07:28 <ltomasbo> o/ 14:07:55 <irenab> hi 14:08:43 <irenab> dmellado: seems you can run the meeting 14:08:49 <dmellado> I can, no worries 14:08:57 <dmellado> #chair irenab 14:08:57 <openstack> Current chairs: apuimedo dmellado irenab 14:09:10 <dmellado> #topic kuryr-kubernetes 14:09:22 <dmellado> did we have any agenda for today on kuryr-kubernetes? 14:09:35 <dmellado> irenab: ? 14:10:00 <irenab> There are patches by ltomasbo 14:10:01 <ivc> i have some concerns about the driver proposed in https://review.openstack.org/#/c/498253 14:10:24 <ltomasbo> I have a couple of patches 14:10:24 <ivc> ^ service type=LoadBalancer 14:10:32 <dmellado> #link https://review.openstack.org/#/c/498253 14:10:35 <irenab> ivc: what is your concern? 14:10:35 <ltomasbo> regarding ports pool population/free 14:10:53 <ltomasbo> and one ongoing one regarding octavia with L2 (not fully working yet) 14:11:00 <irenab> lets first cover lb service type 14:11:03 <ivc> irenab i've left brief comment 14:11:07 <ivc> #link https://review.openstack.org/#/c/498253/4/kuryr_kubernetes/controller/handlers/lbaas.py@63 14:11:22 <ivc> not sure if we need a driver for public ip 14:11:45 <ivc> imo it should be part of lbaas driver itself 14:11:58 <ltomasbo> btw, welcome back to the meetings ivc! nice to have you joining again 14:12:32 <ivc> ltomasbo thnx :) 14:12:34 <irenab> seems the author of the patch is not here 14:12:46 <dmellado> yeah, welcome back irenab! 14:12:47 <apuimedo> ivc: I think it is useful to have a driver 14:12:49 <irenab> welcome back ivc , we missed you 14:13:00 <apuimedo> so it can be used for externalip handling as well 14:13:17 <dmellado> irenab: I was trying to get yossi to attend here, but he's not replying, so let's leave the discussion to the patch itself 14:13:31 <irenab> dmellado: agreed 14:13:41 <ivc> sure 14:14:06 <dmellado> anything more on the topic? ltomasbo, do you have any patch to discuss? 14:14:52 <ltomasbo> dmellado, ivc, irenab: https://review.openstack.org/#/c/498523/ 14:14:53 <ivc> apuimedo my main point about the driver is that if we only want a driver if we expect different implementations to exist 14:15:07 <ltomasbo> would be nice if you take a look at that, apuimedo already did 14:15:16 <irenab> ltomasbo: sure 14:15:22 <ltomasbo> irenab, thanks 14:15:33 <dmellado> #link https://review.openstack.org/#/c/498523/ 14:16:01 <irenab> ltomasbo: is there devstack install doc for the pooled option? 14:16:14 <ltomasbo> yep, there is 14:16:22 <irenab> great, thanks 14:16:34 <dmellado> ok, so besides this, I'm working on a blueprint to add the kubernetes client as a native tempest client and enhance our testing pool, will be posting the link when I've a draft for comments 14:16:42 <ltomasbo> irenab, https://github.com/openstack/kuryr-kubernetes/blob/master/doc/source/installation/ports-pool.rst 14:16:46 <dmellado> the openstack-qa guys will be also reviewing and helping at the PTG 14:16:50 <apuimedo> ivc: that's a good point 14:17:02 <dmellado> so if you've anything you'd like to see there, plese shout out 14:17:26 <dmellado> and I'll be attending a session regarding infra to add zuulV3 multinode testing to our current gates 14:18:16 <dmellado> I still have a TODO on the docs regarding how to write tests, though 14:18:29 <dmellado> is there anyone who would like to help me with that? 14:18:50 <apuimedo> dmellado: I'll help reviewing and can help writing some section if you want 14:19:00 <dmellado> sure, thanks toni 14:19:28 <irenab> dmellado: will try to help by reviewing 14:19:34 <dmellado> thanks irenab 14:19:42 <dmellado> anything more on kuryr-k8s? 14:20:06 <irenab> danil approached me regarding SR-IOV patch 14:20:29 <ltomasbo> it would be nice to have some feedback on this patch too: https://review.openstack.org/#/c/498698/ (related to the other one) 14:20:37 <irenab> as we previously discussed some refactoring to avoid handling different cases inside same handler 14:21:20 <apuimedo> I have nothing today 14:21:40 <apuimedo> too busy trying to deploy on a 30 baremetal nodes for scale testing 14:21:44 <irenab> #link https://review.openstack.org/#/c/498698/ 14:21:47 <dmellado> apuimedo: you tell me... xD 14:21:49 <apuimedo> to spot which areas we should improve during queens 14:22:01 <irenab> apuimedo: any inputs already? 14:22:57 <apuimedo> irenab: that deployment tools make me want to climb without rope and leap to my death 14:23:14 <dmellado> irenab: apuimedo triple-ouch 14:23:15 <dmellado> xD 14:23:19 <irenab> :-) 14:23:20 <ltomasbo> xD 14:23:29 <apuimedo> irenab: which reminds me, does dragonflow support ml2 dns extension? 14:23:37 <dmellado> let's say it's been quite rough so far... 14:23:43 <irenab> apuimedo: not yet, but planned for Q 14:24:10 <irenab> does ovn? 14:24:33 <dmellado> ovn has a current patch waiting to be merged 14:24:59 <dmellado> more exactly, ovn has, networking-ovn still hasn't implemented it 14:25:45 <irenab> apuimedo: where did you need it? 14:25:55 <apuimedo> irenab: no, but as a first step we made them not crash getting fip when dns extension is enabled 14:25:56 <irenab> part of triple -) 14:25:59 <apuimedo> support will come later 14:26:06 <apuimedo> irenab: no, openshift uses kube-dns 14:26:12 <numans> apuimedo, dmellado, yes that's right. we just need to make use of native dns feature of ovn. it requires some work :) 14:26:13 <apuimedo> but we use neutron dns extension instead 14:27:03 <dmellado> o/ numans, thanks for your work so far on that! 14:27:05 <irenab> in kuryr? 14:27:13 <numans> dmellado, :) 14:27:17 * dmellado will have a look at those errors in the tests 14:30:16 <dmellado> let's hop into the next topic 14:30:42 <dmellado> #topic kuryr-libnetwork 14:31:00 <dmellado> apuimedo: I was just thinking about releases here 14:31:06 <dmellado> did we get to release in the end? 14:31:07 <apuimedo> dmellado: what about? 14:31:12 <apuimedo> no, no 14:31:27 <apuimedo> we have to move to cycle with intermediaries for queens 14:31:33 <apuimedo> too late for pike (obviously) 14:31:54 <dmellado> yeah, we'll have to take care about that for Q 14:32:27 <dmellado> I'll try to have a few words with dough at the ptg too 14:32:38 <dmellado> and I'd propose irenab as release liason xD 14:32:50 <irenab> dmellado: you have a lot on your agenda 14:33:04 <irenab> please not :-) 14:33:14 <dmellado> irenab: yeah, actually I'm trying to figure out how to get to match my schedule there 14:33:25 <dmellado> but let's be optimistic xD 14:33:47 <dmellado> is hongbing around? 14:34:07 <dmellado> anything on fuxi? 14:34:41 <dmellado> doesn't seems like it 14:34:41 <zengchen1> there are several patch ready for review about fuxi-k8s 14:34:51 <dmellado> #topic fuxi-k8s 14:34:56 <dmellado> #chair zengchen1 14:34:57 <openstack> Current chairs: apuimedo dmellado irenab zengchen1 14:35:00 <dmellado> zengchen1: feel free to cover this part 14:35:57 <zengchen1> i have nothing to talk, just needs review on fuxi-k8s 14:36:02 <apuimedo_> meh... my baby daughter unplugged the fiber router 14:36:04 <apuimedo_> :/ 14:36:15 <dmellado> apuimedo_: heh, no worries 14:36:17 <apuimedo_> so I missed all the messages 14:37:07 <dmellado> #link https://review.openstack.org/#/q/status:open+project:openstack/fuxi-kubernetes 14:37:21 <zengchen1> dmellado:thanks 14:37:23 <dmellado> those are the open reviews for fuxi, so pls review if you have the time 14:38:02 <dmellado> #topic open discussion 14:38:20 <apuimedo> i've got nothing 14:38:22 <dmellado> anything on the off-topic that anyone would like to address? 14:38:31 <dmellado> apuimedo: I blame triple-ouch 14:39:47 <apuimedo> I'm seeing double already 14:39:49 <zengchen1> apuimedo:i just said there are several patches ready for review on fuxi-k8s. I hope you talk a look at it. thanks. 14:40:16 <dmellado> all right, if there's nothing more here I'll close the meeting and go get a really needed coffee 14:40:18 <dmellado> ;) 14:40:27 <dmellado> thanks for attending! 14:40:32 <dmellado> #closemeeting 14:40:35 <dmellado> #endmeeting