16:01:05 #startmeeting containers 16:01:06 Meeting started Tue Apr 18 16:01:05 2017 UTC and is due to finish in 60 minutes. The chair is strigazi. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:07 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:10 The meeting name has been set to 'containers' 16:01:18 #topic Roll Call 16:01:33 Ton Ngo 16:02:05 Perry Rivera o/ 16:03:04 Thanks for joining the meeting tonanhngo juggler 16:03:30 hello strigazi 16:03:31 Adrian couldn't make it today, he will be here next week. 16:03:55 sparse crowd, must be DockerCon 16:03:58 ah, hope things are ok 16:04:28 I hope it is indeed. 16:04:56 #topic Announcements 16:05:17 I don't have any, if you do, shoot 16:05:42 None here currently 16:05:46 none from me 16:05:48 #topic Essential Blueprints Review 16:06:21 #link https://blueprints.launchpad.net/magnum/+spec/flatten-attributes Flatten Attributes [strigazi] 16:08:47 I pivoted a bit the impleemtation towards cluster instead of cluster template. I couldn't figure out the problem I had with update on CTs. My goal is add all attrs to cluster and detach clusters from CTs. This change will allow proper handling of: 16:08:59 #link https://blueprints.launchpad.net/magnum/+spec/cluster-upgrades Cluster Upgrades [strigazi] 16:09:59 Since the attributes we want to change will belong to the cluster and not its CT 16:11:21 I have a prototype working (in devstack) which replaces all the nodes with the new attrubutes. At the moments I didn't do drain first. I'll add it next 16:12:13 I tested upgrading k8s 1.5.2 to 1.5.3 16:12:19 nice work strigazi 16:12:38 how did the tests go 16:13:20 Does k8s provide any support for draining a node? It's probably a common procedure. 16:14:02 Well it worked, but it was more a like a cluster refresh, I need to also talk to k8s to remove the old nodes and maybe add a message to warn users that pod without RCs will be lost 16:14:33 tonanhngo It does, kubectl drain ... 16:15:04 it pop a warning saying the above warning 16:15:17 Nice. But for pod without RC, it's known that it's not guaranteed to live 16:15:33 RC? 16:15:42 a pod without an replication controller will be lost 16:15:47 if a node dies suddenly, that's what would happen 16:15:54 ah, thank you 16:15:57 exactly 16:16:18 tonanhngo GKE does the same 16:16:33 it's a good start I think 16:17:30 Any other comment? 16:17:48 let's skip: 16:17:50 #link https://blueprints.launchpad.net/magnum/+spec/nodegroups Nodegroups [Drago] 16:18:05 since Drago and Jaycen are not here. 16:18:33 #topic Other Ideas 16:19:28 based on the discussion on the last meeting I tied again to containerize the heat agents, I managed to create a 30mb image based in alpine linux 16:19:47 voila: https://hub.docker.com/r/strigazi/heat-container-agent/tags/ 16:20:48 requires some testing though. Check it out. 16:21:24 #topic Open Discussion 16:22:06 thanks strigazi 16:22:21 Did anyone from CERN attend KubeCon? 16:22:29 I don't have anything folks, and I'll be away this week. 16:22:37 Ricardo attended 16:22:54 Anything new that's interesting to us? 16:23:07 (networking, storage, ...) 16:23:26 Added RC to this list: https://wiki.openstack.org/wiki/Magnum/Acronyms#Table_of_Acronyms 16:24:08 juggler: +1, that should be helpful 16:24:09 People were interested in magnum. I have some new intersting features, just a second 16:24:10 Didn't add GKE to this list since it's Google-related... 16:24:34 thx tonanhngo 16:24:43 sure sure strigazi 16:26:54 RBAC for authentication, a few improvements in dynamic storage provisioning but it is already supported for cinder but with annotations 16:27:13 hm...maybe I should add GKE, seeing that K8S is on that list too...I will do so.. 16:27:13 the definition of PVs changes a bit 16:27:41 daemonsets support rolling upgrades 16:27:58 I heard something about V2 for volume to support stateful pod 16:28:36 That is what I was looking for but it is not there yet 16:28:47 how about multitenancy in k8s? 16:29:04 and COREdns instead of skydns and i think kubedns 16:29:55 tonanhngo nothing concrete yet but it is in their roadmap 16:30:17 ok, I think they say 3rd quarter this year 16:30:42 I will chnage a lot how people use k8s 16:31:03 Certainly, and Magnum should plan accordingly 16:31:04 but this year sounds very ambitious 16:32:08 PV -> Persistent Volumes in this context? 16:32:12 yes 16:32:14 thx 16:32:23 There is a large influx of contributors, so k8s should be growing quickly 16:35:17 GKE and PV added to the list... 16:35:30 That is true, but the rate of growth will decrease I guess 16:37:13 We can wrap it up tonanhngo juggler, is there anything else you want to discuss? 16:37:23 That's all from me 16:37:34 good here strigazi 16:37:44 thanks for presiding 16:37:55 See you next week :) You are welcome 16:37:55 Thanks Spyros 16:38:05 Good seeing everyone :) 16:38:05 #endmeeting