14:00:54 <dmellado> #startmeeting kuryr 14:00:55 <openstack> Meeting started Mon Dec 18 14:00:54 2017 UTC and is due to finish in 60 minutes. The chair is dmellado. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:56 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:59 <openstack> The meeting name has been set to 'kuryr' 14:01:07 <dmellado> Hi folks, who's here today for the kuryr meeting? ;) 14:01:26 <leyal> Hi 14:01:26 <dulek> o/ 14:01:38 <dmellado> #chair dulek leyal 14:01:39 <openstack> Current chairs: dmellado dulek leyal 14:02:22 <dmellado> ltomasbo: irenab ? 14:02:32 <ltomasbo> o/ 14:02:53 <leyal> irenab in PTO 14:03:06 <dmellado> oh, true! 14:04:23 <dmellado> seems that some folks are having issues with freenode today, so let's start 14:04:29 <dmellado> #topic kuryr-libnetwork 14:04:47 <dmellado> anything to discuss on libnetwork? hongbin is not here either 14:06:42 <dmellado> going once 14:06:44 <dmellado> going twice... 14:06:50 <dmellado> #topic kuryr-kubernetes 14:06:57 <dmellado> dulek: anything to share? ;) 14:07:41 <dulek> I have an issue with https://review.openstack.org/#/c/527657/ - I was trying to fix occasional OpenShift download failures. 14:07:57 <dmellado> still having those 403 issues with openshift? 14:08:00 <dulek> But now downloading the binary takes ~7 minutes and gate times out. 14:08:16 <dulek> I wonder if we're able to extend the timeout? 14:08:35 <dmellado> hmmm I don't think so by default on infra, but I can bring the topic to the infra meeting 14:08:39 <ryuk_apuimedo> dmellado: here I am 14:08:47 <dmellado> maybe we could try some another mirror 14:08:53 <dmellado> chair ryuk_apuimedo 14:08:59 <dmellado> #chair ryuk_apuimedo 14:09:00 <openstack> Current chairs: dmellado dulek leyal ryuk_apuimedo 14:09:01 <ryuk_apuimedo> where are we at? 14:09:09 <dmellado> we were discussing kuryr-kubernetes 14:09:17 <dmellado> and the issues at the gates with the openshift binary download 14:09:45 <dulek> Do we have an alternative mirror? Currently it's downloaded from GitHub. 14:09:55 <dmellado> hmmm then it's bad! 14:10:14 <dmellado> dulek: did you get to check if it's a github issue? 14:10:19 <dmellado> or just an infra one 14:10:21 <ryuk_apuimedo> yes, I know it should ideally only depend on ost infra 14:10:29 <ryuk_apuimedo> but it was already the case with hyperkube 14:10:33 <dmellado> I mean, how long does it take to download from github from your laptop 14:10:37 <dmellado> +-7 minutes too? 14:10:49 <dulek> dmellado: Hm, good question. 14:10:51 <ryuk_apuimedo> can't be that bad 14:11:32 <ryuk_apuimedo> dmellado: it takes 100s 14:11:40 <ryuk_apuimedo> here on my granny's internet 14:11:44 <dmellado> dulek: let's give that a try and if it takes long too 14:11:55 <dmellado> maybe we can ask the ost infra folks to host an internal mirror 14:12:03 <dmellado> that'd solve the timeouts issues for sure 14:12:04 <ryuk_apuimedo> dmellado: that would be ideal 14:12:12 <ryuk_apuimedo> and speed up the testing! 14:12:15 <dulek> dmellado: ~100 seconds on my laptop as well. 14:12:30 <dmellado> dulek: then I'll ad an 14:12:44 <dmellado> #action dmellado to ask for openshift mirror on ost infra 14:12:46 <dmellado> add 14:13:08 <dulek> Okay, so I just wanted to raise that. Besides I have two patches waiting for merge for quite a long time. 14:13:16 <dulek> https://review.openstack.org/#/c/518024/ - daemon containerization. 14:13:28 <dulek> https://review.openstack.org/#/c/509380/ - CNI daemon docs. 14:13:43 <dmellado> dulek: ack, I'll try taking a look at them 14:13:56 <ryuk_apuimedo> dulek: that'll have to wait for irenab to come back from pto (I think she returns on wed) 14:13:57 <dmellado> ryuk_apuimedo: leave the death node and look at them too if you have the time 14:14:12 <dulek> ryuk_apuimedo: Sure thing. :) 14:14:20 <dulek> So that's it from my side. :) 14:14:25 <dmellado> dulek: thanks! 14:14:30 <dmellado> ryuk_apuimedo: how about you? 14:15:06 <ryuk_apuimedo> I've been working on openshift-ansible and other deployment questions 14:15:10 <ryuk_apuimedo> also 14:15:17 <dmellado> regarding to that 14:15:21 <dmellado> as soon as shade is released 14:15:22 <ryuk_apuimedo> I've been checking about keystone 14:15:31 <dmellado> you should be able to use the tag on os_stacks 14:15:37 <ryuk_apuimedo> and what we can do to improve the way we deploy 14:15:47 <ryuk_apuimedo> I want to limit the controller access to neutorn 14:15:49 <ryuk_apuimedo> *neutron 14:15:56 <ryuk_apuimedo> just like we do with our access to k8s 14:16:27 <ryuk_apuimedo> dmellado: how?! 14:16:43 <dmellado> I do have a pull request to ansible 14:16:48 <dmellado> waiting for shade to release 14:16:52 <ryuk_apuimedo> ah! 14:16:57 <ryuk_apuimedo> I was gonna do one now XD 14:16:58 <dmellado> it adds heat tag support 14:17:00 <dmellado> xD 14:17:03 <dmellado> no need xD 14:17:06 <ryuk_apuimedo> I have the file open 14:17:10 <ryuk_apuimedo> please, link to it 14:17:22 <dmellado> I have that local, just waiting for shade 14:17:27 <dmellado> but will put it online 14:17:28 <ryuk_apuimedo> and ping mordre, I think he's the author of the module 14:17:40 <dmellado> nope, but I know who is xD 14:17:47 <dmellado> matbu, a red hatter too 14:18:01 <dmellado> but we need first for shade to release 14:18:10 <dmellado> as the module depends on a shade version 14:18:34 <ryuk_apuimedo> true 14:18:58 <ryuk_apuimedo> btw, I plan on doing a Christmas release of kuryr-kubernetes 14:19:07 <ryuk_apuimedo> If we manage to get both dulek's patches merged 14:19:11 <dmellado> heh, what the codename's gonna be 14:19:13 <dmellado> santa? 14:19:15 <dmellado> xD 14:19:35 <dmellado> now seriously, ping me when you do that as I want to make sure rdo rpm goes smooth 14:19:47 <ryuk_apuimedo> Jack Skellington 14:20:02 <dmellado> and I'd link that to a kuryr-tempest-plugin release too 14:20:03 <ryuk_apuimedo> dmellado: will do! 14:20:31 <ryuk_apuimedo> what else do we have? 14:20:34 <dmellado> so from my side, we got gena and jon working on a few tests 14:20:38 <dmellado> pod to pod testing 14:20:49 <ryuk_apuimedo> did we find out why it works on openshift and not on k8s? 14:20:50 <dmellado> and ports_pool 14:20:56 <dmellado> not yet 14:21:02 <dmellado> but I'll try with a newer k8s version 14:21:16 <ryuk_apuimedo> wasn't 1.9 released this week? 14:21:31 <dmellado> perhaps, IIRC the testing was done with 1.6 14:21:32 <ryuk_apuimedo> or did I imagine it in the internet withdrawal syndrome I suffer? 14:21:35 <dmellado> wasn't it, dulek ? 14:21:47 <dulek> dmellado: Gate got updated to 1.8 recently. 14:21:56 <dmellado> hmmm then we need to check 14:22:13 <gcheresh> dulek: the gate that pass? 14:22:15 <dulek> dmellado: https://review.openstack.org/#/c/525502/ 14:22:46 <ryuk_apuimedo> yboaron_: any update? 14:22:49 <dulek> gcheresh: No, gates with OpenShift are passing. I'm talking about K8s gates. 14:23:00 <yboaron_> Please review : https://review.openstack.org/#/c/523900/ - L7 router support and another one - minor documentation update : https://review.openstack.org/#/c/527697/ 14:23:02 <ryuk_apuimedo> leyal: I have to review your specs, sorry I didn't get to it 14:23:10 <dulek> dmellado: You can raise the default in the dummy patch to let gate run 1.9. 14:23:21 <dmellado> dulek: yep, let's try that ;) 14:23:23 <leyal> ryuk_apuimedo , thanks :) 14:24:12 <ryuk_apuimedo> leyal: do you have anything else? 14:24:19 <leyal> I also added work items list to the blueprint - https://blueprints.launchpad.net/kuryr-kubernetes/+spec/k8s-network-policies 14:24:42 <leyal> and upload patch for the first one - https://review.openstack.org/#/c/526916/ 14:24:42 <ryuk_apuimedo> yboaron_: have you tried octavia in HA? 14:24:54 <yboaron_> no 14:24:58 <ryuk_apuimedo> does the l7 routing work everytime or does it still drop packets? 14:25:17 <yboaron_> still dropping packets like crazy .. , 14:25:39 <ryuk_apuimedo> yboaron_: do you have some hint of what could it be? 14:25:45 <yboaron_> I'm trying to check in neutron L7 policy in non-kuryr environment 14:25:50 <ryuk_apuimedo> ok 14:26:05 <ryuk_apuimedo> let us know what's the outcome asap 14:26:15 <ryuk_apuimedo> and we can bring it up with the octavia/neutron teams 14:26:17 <yboaron_> sure 14:26:22 <ryuk_apuimedo> thanks yboaron_ 14:26:37 <yboaron_> and there's the issue of octavia + ova-firewall 14:26:58 <ryuk_apuimedo> did you bring that up with jlibosva? 14:27:19 * ryuk_apuimedo has a meeting that needs attendance in two minutes, so we'll have to cut short 14:27:26 <yboaron_> manage to reproduce it in octavia local.conf , trying to debug it wit the help of good friends (jlisboa nd nmagenezi) 14:27:34 <ryuk_apuimedo> perfect 14:27:38 <ryuk_apuimedo> keep us updated! 14:27:43 <yboaron_> NP 14:27:45 <ryuk_apuimedo> Anything before we close the meeting? 14:28:40 <dmellado> ryuk_apuimedo: yep, merry christmas folks xD 14:29:28 <ryuk_apuimedo> thanks! 14:29:29 <leyal> dmellado, and happy Hanukkah :) 14:29:35 <ryuk_apuimedo> leyal: indeed! 14:29:38 <dmellado> leyal: for sure! ;) 14:29:43 <yboaron_> :-) 14:29:48 <ryuk_apuimedo> #endmeeting