16:00:33 <strigazi> #startmeeting containers
16:00:34 <openstack> Meeting started Tue Jan 23 16:00:33 2018 UTC and is due to finish in 60 minutes.  The chair is strigazi. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:35 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:37 <openstack> The meeting name has been set to 'containers'
16:00:40 <strigazi> #topic Roll Call
16:00:48 <strigazi> Spyros Trigazis
16:00:51 <oikiki> o/
16:00:57 <clenimar> o/
16:01:26 <ricolin> 0/
16:02:07 <slunkad> hi
16:02:45 <strigazi> Thanks for joining the meeting oikiki clenimar ricolin slunkad
16:02:49 <strigazi> Agenda:
16:02:53 <strigazi> #link https://wiki.openstack.org/wiki/Meetings/Containers#Agenda_for_2018-01-23_1600_UTC
16:03:05 <strigazi> #topic Announcements
16:03:09 <strigazi> None
16:03:13 <strigazi> #topic Review Action Items
16:03:28 <strigazi> strigazi to draft the PTG schedule [DONE] https://etherpad.openstack.org/p/magnum-ptg-rocky
16:03:45 <strigazi> #link https://etherpad.openstack.org/p/magnum-ptg-rocky
16:04:09 <strigazi> strigazi to look for a new meeting time APAC friendly [incomplete]
16:04:31 <ykarel> o/
16:04:35 <strigazi> Sorry for this ricolin, I'll try to move it this week
16:04:47 <ricolin> strigazi, yeah, I though we're moving this week:)
16:04:58 <strigazi> hi ykarel
16:05:08 <strigazi> #topic Blueprints/Bugs/Ideas
16:05:22 <strigazi> Cluster Federation https://review.openstack.org/#/q/status:open+project:openstack/magnum+branch:master+topic:bp/federation-api
16:05:25 <strigazi> [ready for review]
16:05:29 <ricolin> strigazi, NP, remember to change wiki and openstack-infra/irc-meetings
16:06:16 <strigazi> clenimar do you want to say a few thing for the work on federation? What is the status etc
16:06:27 <strigazi> ricolin: yes, thanks
16:06:36 <clenimar> strigazi, sure
16:07:39 <clenimar> we've got two patches ready for review (db layer and api controllers). I'm currently working on the fedora-atomic driver implementation, which is also published as a wip
16:08:42 <vabada> hi, sorry for the delay
16:08:45 <clenimar> and I'd love some feedback, especially on how to use softwaredeployments to manage the federation (create, join members etc)
16:09:12 <strigazi> clenimar: how kubefd is going to be fetched on the node? Are there any restrictions on version compatibility?
16:09:33 <strigazi> clenimar: all the work will happen on master, right?
16:09:40 <strigazi> * master node
16:09:51 <clenimar> strigazi, yes, all the the work is done in the master node of the hostcluster
16:10:16 <clenimar> we're currently downloading the binary and we can set the version
16:10:55 <strigazi> Did you use the RBAC patch I have written? Is ti working for you?
16:11:40 <strigazi> Does kubefed the same version  as kubernetes?
16:11:51 <strigazi> Does kubefed follow the same version  as kubernetes?
16:12:11 <clenimar> haven't tried the RBAC patch yet, but it looks good.
16:12:20 <clenimar> yes, kubefed follows the version
16:12:44 <clenimar> we can set it to a specific release
16:13:02 <clenimar> to match the cluster
16:13:25 <clenimar> we're currently creating clusters on which version? 1.7.4?
16:14:37 <strigazi> The version comes from the conainers that fetched using the kube_tag label. We shoud get he version from the api server and fetch the proper version
16:15:19 <clenimar> yes, we can do that
16:15:37 <strigazi> We can have a patch that finds the version and also returns it to heat
16:15:50 <clenimar> cool
16:15:50 <strigazi> ricolin: makes sense? ^^
16:16:14 <ricolin> yes
16:16:27 <strigazi> I think we can have dependencies in software deployments
16:17:37 <strigazi> clenimar: do you need anything else apart from feedback? Are there any show stopper atm?
16:19:13 <ricolin> strigazi, also about os-collect-config, working on a patch to change the config, which should be able to target some current waste of connection issue
16:19:47 <strigazi> clenimar: are you there?
16:19:48 <clenimar> I've got some doubts on the driver implementation itself... should we create a separate template definition for the federation? or can we just trigger the actions from the exisiting cluster template?
16:20:22 <strigazi> ricolin: I'll get to that in a bit
16:21:00 <strigazi> clenimar: sinse federation has its own api, we can act on the heat stack directly
16:21:47 <strigazi> clenimar: the cluster temlate won't change
16:21:56 <clenimar> updating the stack whenever we want to perform an operation?
16:22:11 <strigazi> clenimar: yes
16:23:13 <clenimar> and how would it work for join/unjoin operations?
16:24:14 <strigazi> clenimar: the api will need to do actions against the heat stacks that it  needs.
16:24:40 <strigazi> clenimar: Let's continue on this in the reviews or in #openstack-containers right after
16:24:51 <clenimar> ok, thanks
16:25:10 <strigazi> ricolin: what change are you doing for os-collect-config?
16:25:28 <strigazi> about the deploy url?
16:26:09 <ricolin> a config in os-collect-config will change the polling list
16:26:43 <ricolin> `collectors` config in config file
16:27:02 <strigazi> why do we need this? What is the problem?
16:27:02 <ricolin> so we can avoid poll from other sources
16:27:16 <strigazi> to avoid hitting the metadata server?
16:27:30 <ricolin> strigazi, that will be one of the issue
16:27:50 <ricolin> also protentially https://bugs.launchpad.net/magnum/+bug/1744362
16:27:50 <openstack> Launchpad bug 1744362 in Magnum "heat-container-agent fails to communicate with Keystone." [Medium,Confirmed]
16:28:46 <strigazi> ok, is there a patch up?
16:28:57 <strigazi> ricolin: ^^
16:29:17 <ricolin> will send it out today or tomorrow(one hour left for today)
16:29:35 <strigazi> ok
16:30:51 <strigazi> I'm not sure why the internal_url is given to the nodes, Ill
16:30:57 <strigazi> I'm not sure why the internal_url is given to the nodes, I'll check it
16:31:10 <strigazi> ok, next
16:31:21 <strigazi> Cluster healing spec: https://review.openstack.org/#/c/529897/
16:32:19 <strigazi> The spec is in good state, so please have a look, I'll add my vote as well
16:32:41 <strigazi> ricolin ykarel have you seen it ^^
16:32:57 <ykarel> strigazi, not yet
16:33:05 <strigazi> ricolin: Are you aware of it?
16:33:29 <ricolin> strigazi, actually no, but will give a look now!lol
16:33:52 <strigazi> ricolin: speard the word to your team
16:34:12 <strigazi> slunkad: your team  might also be interested
16:34:32 <ricolin> strigazi, surly will
16:34:52 <slunkad> yes
16:35:56 <strigazi> cool, next: kubernetes patches: RBAC, config fixes, CI fix https://review.openstack.org/#/q/status:open+project:openstack/magnum+branch:master+topic:bug/1726482
16:36:56 <strigazi> ykarel: is the redirection working now for centos and ubuntu as well? I didn't have to look at it again
16:37:08 <ykarel> strigazi, yes
16:37:16 <ykarel> we can take it
16:37:20 <strigazi> cool
16:37:59 <strigazi> ykarel: did you check the RBAC+(minor fixes) patch?
16:38:10 <ykarel> strigazi, yes, i left a comment there
16:38:23 <ykarel> can you check
16:38:27 <strigazi> ykarel: ok, thanks
16:38:47 <strigazi> and from me is:
16:39:16 <strigazi> Cluster Upgrades: spec: https://review.openstack.org/#/c/433728
16:41:27 <strigazi> Since the release is 3 weeks away, I'll rush this and next week, I think we should take a first iteration, ykarel what do you think? Since it doesn't break existing functionality and adds new API paths, it is safe. thoughts?
16:42:34 <ykarel> strigazi, will look tomorrow morning, after that i am on PTO, but it's ok to take the first iteration
16:42:56 <strigazi> ok
16:43:37 <strigazi> Let's move to:
16:43:40 <strigazi> #topic Open Discussion
16:43:52 <strigazi> oikiki: slunkad do you have anything?
16:44:16 <oikiki> strigazi: My two patches are ready for your review
16:44:36 <oikiki> i finished up the cache barbican certs fro periodic tasks yesterday
16:45:05 <oikiki> https://review.openstack.org/534364
16:45:12 <strigazi> oikiki: ok I'll have a look
16:45:22 <oikiki> and send cluster metrics config param is ready: https://review.openstack.org/529098
16:45:50 <strigazi> great! thanks
16:45:53 <oikiki> :)
16:47:14 <strigazi> others, do you have anything else?
16:48:26 <strigazi> ok, we can end the meeting then
16:48:33 <ricolin> not from me:)
16:48:36 <strigazi> Thanks everyone
16:48:40 <ricolin> strigazi, thx
16:48:46 <strigazi> #endmeeting