21:00:38 <strigazi> #startmeeting containers 21:00:39 <openstack> Meeting started Tue Oct 2 21:00:38 2018 UTC and is due to finish in 60 minutes. The chair is strigazi. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:40 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:43 <openstack> The meeting name has been set to 'containers' 21:00:50 <strigazi> #topic Roll Call 21:00:53 <strigazi> o/ 21:00:54 <brtknr> o/ 21:00:58 <imdigitaljim> o/ 21:02:01 <strigazi> Thanks for joining the meeting brtknr imdigitaljim 21:02:07 <strigazi> #topic Announcements 21:02:28 <brtknr> glad to be finally joining! 21:02:29 <openstackgerrit> Theodoros Tsioutsias proposed openstack/magnum-specs master: [WIP] Instoduce magnum nodegroups https://review.openstack.org/607363 21:02:40 <ttsiouts> o/ 21:03:03 <strigazi> We cherry-picked a couple of patches in rocky with flwang https://review.openstack.org/#/q/status:merged+project:openstack/magnum+branch:stable/rocky 21:03:39 <strigazi> and btw the ttsiouts pushed a WIP of the nodegroups spec just now, as you can see. 21:03:50 <strigazi> and btw ttsiouts pushed a WIP of the nodegroups spec just now, as you can see. 21:03:58 <ttsiouts> It's really a WIP 21:04:09 <strigazi> no worries, thanks 21:04:18 <ttsiouts> I need to update the patches also.. 21:04:31 <strigazi> regarding the cherry-picks, 21:05:19 <cbrumm> o/ 21:05:33 <strigazi> we want the self-hosted flannel patch in too and me and me and flwang will push for the health check patch too, 21:05:51 <strigazi> hey cbrumm 21:06:13 <strigazi> #topic Stories/Tasks 21:07:26 <strigazi> Apart from the cherry-picks and the release that will follow them, I only have two patches for swarm-mode 21:08:04 <colin-> sorry i'm late 21:08:13 <strigazi> I pushed one to allow the configuration of the overlay networks cidr, probable brtknr is interested 21:08:22 <strigazi> colin-: o/ 21:08:54 <strigazi> s/probable/probably/ 21:09:39 <strigazi> and one more to set the socket where dockerd listens to. Set it to /var/run/docker.sock 21:10:19 <strigazi> this last one, came up after our L1TF reboots 21:10:23 <strigazi> docker didn't start 21:10:45 <strigazi> it couldn't create the socket on boot 21:11:18 <strigazi> That is all from me for last week 21:11:43 <imdigitaljim> we've got our openstack cluster auto scaler poc working well and we auto deploy services (dash, prom, etc) with a remote helm in the cluster creation, we've also been heavily weighing in on a new data flow pattern for bootstrapping. 21:12:21 <imdigitaljim> the new pattern could easily be staged for supporting the old version and new version together 21:12:39 <strigazi> I didn't get the last part, what new flow 21:12:59 <imdigitaljim> to take some burden off the cloud-init user_data (which we are capping out) 21:13:29 <imdigitaljim> strigazi its similar to some of the data flow we discussed in cern 21:14:03 <strigazi> for prom, dash, coredns, calico we aren not using cloud-init 21:14:10 <imdigitaljim> yeah we are not either 21:14:12 <strigazi> ok, got it 21:14:26 <imdigitaljim> sorry that was like a word vomit of things 21:14:29 <imdigitaljim> 3 things not 2 21:14:33 <strigazi> deploy services with helm pointing to a remote repo 21:15:33 <imdigitaljim> 1) autoscaler 2) deploying services with helm instead of software deployments 3) considering redesign to reduce cloud-init payload and decouple the dynamic magnum template data 21:16:18 <strigazi> I get the insentive for helm, but cloud-init is not an issue since queens 21:16:24 <imdigitaljim> how so? 21:17:09 <strigazi> software deployments are not bound by the user_data limit in nova 21:17:34 <strigazi> we can have as many SD as we wont 21:17:36 <strigazi> we can have as many SD as we want 21:17:45 <strigazi> no? 21:17:54 <strigazi> you reached a limit? 21:18:15 <imdigitaljim> i suppose the heat-container has issues with certain SDs 21:18:23 <imdigitaljim> if you wanted to do everything via SD 21:18:31 <imdigitaljim> we'll need to revisit the heat container 21:19:25 <strigazi> if there is a limit we can look into it. deploying anything with the k8s API is not an issue. 21:20:06 <imdigitaljim> its not a limit 21:20:13 <imdigitaljim> its the extent of what the heat container has access to use 21:20:21 <imdigitaljim> on the host OS 21:20:45 <imdigitaljim> since they heat-agent is atomic mounted into the host OS 21:21:08 <imdigitaljim> there are permissions issues, binaries unreachable, etc 21:21:18 <imdigitaljim> due to mounting 21:21:26 <strigazi> this model has not limits: https://review.openstack.org/#/c/561858/1/magnum/drivers/common/templates/kubernetes/fragments/configure-kubernetes-minion.sh@16 21:21:35 <strigazi> s/not/no 21:22:02 <strigazi> it is like ansible 21:22:38 <imdigitaljim> :) not entirely 21:22:38 <strigazi> Do you have a use case in mind? 21:23:26 <imdigitaljim> https://github.com/openstack/magnum/blob/master/magnum/drivers/common/image/heat-container-agent/config.json.template#L25 21:23:32 <imdigitaljim> anything that doesnt fall here 21:23:38 <imdigitaljim> in the hostOS 21:24:22 <strigazi> in the patch I sent the heat-agent is talking to the host over ssh. 21:24:38 <strigazi> so any binary is reachable 21:24:42 <imdigitaljim> well check it out 21:25:05 <strigazi> in the line that I shared it creates inodes in the host's fs 21:27:10 <strigazi> imdigitaljim: you're typing? 21:27:16 <imdigitaljim> no 21:27:20 <imdigitaljim> im checking out the PR 21:27:36 <imdigitaljim> we can move on! 21:27:44 <imdigitaljim> thanks for the info on the PR 21:28:12 <strigazi> ok, if you have a concern with the heat-agent please share 21:28:30 <strigazi> brtknr: you have something? 21:32:05 <strigazi> brtknr: Are you still there? 21:32:31 <strigazi> I don't have anything else to add, so 21:32:41 <colin-> nothing from me but i'm still here :) 21:32:55 <strigazi> imdigitaljim and @all have a look in ttsiouts WIP on nodegroups 21:33:17 <strigazi> colin-: :) 21:33:50 <imdigitaljim> yeah thats looking imo 21:33:51 <ttsiouts> imdigitaljim: there are still lots of changes especially in the patch but you can get the general idea from the spec 21:34:13 <imdigitaljim> we should be able to support testing it out as well 21:34:26 <ttsiouts> awesome! 21:34:34 <ttsiouts> thanks a lot!!! 21:34:44 <imdigitaljim> np! 21:34:47 <imdigitaljim> good work! 21:35:47 <ttsiouts> :D 21:36:07 <strigazi> let's wrap then? 21:37:13 <strigazi> said once 21:37:24 <strigazi> said twice 21:38:09 <strigazi> see you in the channel or next week :) 21:38:23 <ttsiouts> bye! 21:38:37 <strigazi> #endmeeting