22:01:17 #startmeeting containers 22:01:18 Meeting started Tue Jun 16 22:01:17 2015 UTC and is due to finish in 60 minutes. The chair is adrian_otto. Information about MeetBot at http://wiki.debian.org/MeetBot. 22:01:19 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 22:01:22 The meeting name has been set to 'containers' 22:01:35 #link https://wiki.openstack.org/wiki/Meetings/Containers#Agenda_for_2015-06-16_2200_UTC Our Agenda 22:01:41 #topic Roll Call 22:01:44 Adrian Otto 22:01:50 o/ 22:01:51 Perry Rivera 22:01:53 Mike Smith 22:01:54 Rob Pothier 22:01:54 Ton Ngo 22:01:55 OTSUKA, Motohiro 22:01:55 Dane LeBlanc 22:01:56 o/ 22:01:56 Chirag Arora 22:02:06 o/ 22:02:29 hello there 22:02:39 hey all 22:02:45 o/ 22:03:01 hello hongbin juggler_ mismith rpothier Tango yuanying-alt dane_leblanc eghobo chirag_ sdake_ tcammann_ and diga__ 22:03:18 Hi all 22:03:18 adrian_otto do you have a bot to type that in or do you do it manually 22:03:29 o/ 22:03:37 He has the tab button to help! 22:03:42 haha :) 22:03:45 sdake: I type! 22:03:55 but I also use tab completion ;-) 22:04:15 welcome everyone, let's begin here in just a moment. 22:04:18 o/ 22:04:38 hello daneyon and apmelton 22:04:43 hey 22:04:46 yo yo 22:04:52 * sdake needs more minerals 22:04:56 o/ 22:05:11 Our theme for today's meeting is "Adopt the Orphans" ;-) 22:05:27 Martin Falatic checking in 22:05:29 ha 22:05:34 #topic Announcements 22:05:42 o/ 22:06:29 1) adrian_otto will be on vacation starting 2015-06-19 and will be back to work again 2015-07-05. sdake has agreed to chair our team meetings while I am out. 22:06:39 hello thomasem and mfalatic 22:06:52 cheers sdake 22:06:59 hi hi 22:07:02 i need a beer! 22:07:04 any announcements from team members? 22:07:17 * adrian_otto hands an ice cold beet to sdake 22:07:21 haha 22:07:24 a beet 22:07:27 s/beet/beer/ 22:07:39 What a substantial difference 22:07:47 haha pulling beetroots out of the freezer 22:07:50 What a beeting 22:07:52 #topic Review Action Items 22:07:52 apparently beet beer is a thing 22:08:05 apmelton not in my household :) 22:08:05 1) diga__ to contact ewindsich regarding Swarm+Mesos questions 22:08:10 diga__: status? 22:08:47 adrian_otto: yes, I mailed him, I am again sending mail to him today by including you 22:09:01 ya I got an action 22:09:03 ok, should I carry the action item forward to follow up next week? 22:09:03 that needs addressing 22:09:12 didn;t receive response yet from ewindsich 22:09:22 sdake: hang on just one sec, pls 22:09:25 will do 22:09:39 #action diga__ to follow up with ewindsich regarding Swarm+Mesos questions 22:09:49 yep 22:09:51 next action... 22:09:52 2) jay-lau-513 Clarify human name generation algorithm plans in BP 22:10:21 Jay is not here today because this is a ridiculous time of day for him 22:10:34 excuses! :) 22:10:38 anyone have discussions on this topic who can share updates? 22:10:52 I will carry it forward 22:11:02 #topic jay-lau-513 Clarify human name generation algorithm plans in BP 22:11:11 #topic Review Action Items 22:11:14 #action jay-lau-513 Clarify human name generation algorithm plans in BP 22:11:19 that's better. 22:11:32 how about this topic: 22:11:33 3) jay-lau-513 Seek consensus on duplicate names on ML 22:11:43 I think it is decided to discuss this topic in the meeting 22:11:51 the thread concluded that we should auto-generate names 22:12:00 yes 22:12:02 +1 22:12:04 with an option to relax unique constraints, correct? 22:12:17 yes 22:12:25 do we need further discussion on this topic or do we have a clear consensus to proceed? 22:12:36 lgtm 22:12:36 anyone opposed to a #agreed on this point? 22:12:53 jay-lau is following docker algorithm for this 22:13:07 sounds friieghtening :) 22:13:40 :) 22:13:40 #agreed Magnum contributors support unique name requirements by default with an option to disable the constraint. If no name is provided at the time of resource creation, a human readable name will be dynamically generated and automatically assigned. 22:13:49 I can undo that if there are any objecitons 22:14:20 adrian_otto: unique per tenant? 22:14:21 ok, sdake you have the floor for action item review 22:14:23 definately a new pattern in openstack ;) 22:14:31 no review 22:14:31 a requet 22:14:34 forwarded from Madhuri 22:14:36 apmelton: yes, per tenant, as proposed in our ML thread 22:14:52 ok, proceed 22:15:00 adrian_otto could you make a non-sharktank introduction to someone responsible in engineering for TLS to Madhuri via email 22:15:05 and somehow get that work prioritized 22:15:18 she is blocked because we think kubernetes is busted 22:15:31 and #google-containers is completely unhelpful forsolving the problem 22:15:39 sdake: you mean a Kubernetes dev? 22:15:49 someone that knows something about the tls side of things 22:15:58 yes, I can take an action for that intro 22:16:01 you may ahve to go thorug hthe sharktank to find the right person :) 22:16:02 thanks 22:16:08 on the same topic of the TLS work, I would really like to see a spec before we start this work formally 22:16:28 sdake: have you heard that kub 1.0 coming within month 22:16:36 have not heard 22:16:38 #action adrian_otto to help Madhuri identify a resource on the Kubernetes team who can assist us with TLS feature implementation 22:16:40 but thats fantastic news :) 22:16:45 sdake: does that look good? 22:16:52 wfm adrian_otto 22:16:57 cool 22:16:58 http://www.kuberneteslaunch.com/ 22:17:36 sorry I'm late 22:17:37 eghobo: let's revisit that one in open discussion 22:17:44 hello apuimedo 22:17:49 hi 22:18:19 okay, so from 22:18 UTC to 22:28 UTC (or less) sdake will lead us in a brainstorming exercise 22:18:25 yay 22:18:28 #topic Collaborative brain storm session on etherpad (#link https://etherpad.openstack.org/p/magnum-functional-test-cases) - time boxed to 10 minutes - sdake to take results and create separate bugs per test case 22:18:30 proceed 22:18:48 #link https://etherpad.openstack.org/p/magnum-functional-test-cases 22:18:49 ok folks log in to etherpad - make sure to put your name in the upper right hand corner so we know who added what 22:18:50 (was the network blueprint discussed?) 22:19:05 I'm going to file sepearte bugs for each of these test cases 22:19:23 write as many test cases as you can think of that we should add to the functional gate 22:19:57 As you can see I bombed the etherpad with a big outline of things 22:20:23 looks like someone was busy here ;) 22:20:23 but that's not use-case specific, it's more of an anti-regression test plan for avoiding mega-problems 22:20:32 ya start fresh at the top 22:20:41 line 12 plz 22:20:45 i'll sort em out later 22:20:53 just write whatever comes to mind 22:21:44 Seems like Adrian's 1-6 are all pretty important and basic. 22:23:07 ok folks, start adding test cases line 12 :) 22:23:18 there is 15 people loggedi nto the etherpad there should be atleast 15 test cases by now :) 22:23:39 apuimedo: we will get to Blueprint discussion after this 22:23:52 good :-) 22:24:03 I'm an admitted lurker…first time at one of these meetings. Pretty cool to see things in action 22:24:09 * adrian_otto indicates 5 minutes remaining for brainstorm (half-way point) 22:24:23 mismith: we are happy to have you 22:24:49 Cool. I had my team at Overstock watch your vancouver video in our last meeting :) 22:25:14 recommend not reading other peoples test cases just put yours in 22:25:32 i'll collate em 22:25:40 is that the right word? 22:25:41 who knows :) 22:25:55 i'm like an all in one fax machine 22:26:05 taking any test cases :) 22:27:33 * adrian_otto indicates one minute remaining for scheduled brainstorming 22:28:31 just a quick note, i'm going to leave hese new wishlist bugs unassigned 22:28:32 ok, this concludes our timebox for the brainstorm 22:28:34 if folks wnat to pick em up feel free 22:28:42 thanks for initiating that one sdake 22:28:44 if you want to add more this week, knock yourself out 22:28:55 i'll file the bugs next monday 22:29:01 next up 22:29:06 #topic Blueprint/Bug Review 22:29:13 Tasks/Bugs 22:29:21 #link http://goo.gl/sIlYla High and Critical Priority Bugs 22:29:30 these are the things we can't afford to miss 22:29:40 I identified 4 orphans 22:29:54 let's see about adopting any that are still orphaned 22:29:58 Orphans 22:30:05 #link https://bugs.launchpad.net/magnum/+bug/1412380 Magnum Documentation link report 404 error 22:30:07 Launchpad bug 1412380 in Magnum "Magnum Documentation link report 404 error " [High,In progress] 22:30:15 in progress, but with no owner 22:30:19 documentation 404 makes me sad :( 22:30:21 I can fix that one 22:30:36 next is 22:30:38 #link https://bugs.launchpad.net/magnum/+bug/1444368 Bay status is still CREATE_IN_PROGRESS after stack create failed 22:30:40 Launchpad bug 1444368 in Magnum "Bay status is still CREATE_IN_PROGRESS after stack create failed" [High,Confirmed] 22:30:43 That 404 comes up for me :) 22:31:11 any takers for this one? 22:31:25 seems like convergence may actually be the answer 22:31:35 this might be fixed now? 22:31:45 possible. 22:32:01 need some proper logs to reproduce it etc 22:32:27 I can set it back to incomplete to ask for steps to reproduce and logs 22:32:35 +1 22:32:41 need heat logs 22:33:10 done 22:33:34 tcammann_: please add additional remarks on that bug to guide the submitter to get us exactly what is needed 22:33:45 sure 22:33:49 next is 22:33:50 #link https://bugs.launchpad.net/magnum/+bug/1451118 magnum throws db exception on failed bay deletion 22:33:51 Launchpad bug 1451118 in Magnum "magnum throws db exception on failed bay deletion" [High,Triaged] 22:34:34 this one wants an owner 22:35:11 the last orphan I have on the agenda was bogus, so I won't cover that. 22:35:33 basically we thgouht that scale down events in ResourceGroups caused all nodes to be replaced 22:35:41 I will take this 22:35:45 that was an issue in Heat long ago that has since been fixed 22:35:51 thanks diga__ 22:35:56 yjsy odm 22:35:56 please assign yourself 22:35:56 oh fantastic :) 22:36:01 that isn't what i thought 22:36:04 if you need help let me know 22:36:06 adrian_otto: welcome! 22:36:12 i thought kubernetes will implode if you delete a minion underneath it :) 22:36:13 I proved today that it works properly 22:36:30 kubernetes needs to be set up with a proper rc configuration 22:36:37 and if you do that, it will heal properly 22:36:50 but there is not a heat bug in there 22:37:01 Blueprints 22:37:16 heal properly as in the container persistent data will be moved? 22:37:50 sdake: no, persistent data is not a feature of k8s so you can only autoscale 12 factor apps 22:38:05 #link https://blueprints.launchpad.net/magnum Magnum Blueprints 22:38:24 each of the essential bp's should get an update from the assigned owner or delegate 22:38:34 #link https://blueprints.launchpad.net/magnum/+spec/async-container-operations 22:38:34 what is a 12 factor app? :) 22:38:43 http://12factor.net/ 22:38:44 sdake: all state is stored in a db 22:38:55 got it wasn't familiar with that term 22:39:14 apmelton: any update on async-container-operations? 22:39:18 no update, I might be able to start it before k1, but it definitely won't be finished before k1 22:39:23 er l1 22:39:29 ok, I will re-target 22:39:45 done 22:40:00 next is 22:40:02 #link https://blueprints.launchpad.net/magnum/+spec/objects-from-bay 22:40:11 sdake: update on this? 22:40:12 i bounced that to l2 22:40:17 ok, thanks 22:40:21 i got some code 22:40:24 far from complete 22:40:29 next 22:40:31 investigating atm 22:40:31 its a big job 22:40:31 #link https://blueprints.launchpad.net/magnum/+spec/secure-kubernetes 22:40:33 xl 22:40:50 sdake: we should break it down into smaller bits 22:40:52 blocked on kumari intro 22:41:08 Also think we should put a formal spec together for this bp 22:41:09 Madhuri is not here to address the secure-kubernetes blueprint 22:41:11 adrian_otto once I figure out how to decompose it i'll file smaller blueprints for it if you like 22:41:17 is anyone here as her delegate? 22:41:24 i can speak for her 22:41:32 sdake: yes, please, or bug tickets that link to the BP. 22:41:32 we spoke last night on this topic 22:41:46 she expressed the need to speak with someone in the know with tls 22:42:01 i intro'ed her ot the guy at csco who wrote the guide for tls setup 22:42:07 she needs an intro to a tls cat in google 22:42:16 the three of them should be able to sort it out from there 22:42:16 ok, so I have the action for that 22:42:22 what sdake said :) 22:42:31 atm we believe kubernetes is busted 22:42:36 ok, any more discussion on this BP? 22:42:49 next 22:42:51 #link https://blueprints.launchpad.net/magnum/+spec/external-lb 22:42:57 Madhuri again 22:43:07 Tango 22:43:09 sorry 22:43:13 Hi :) 22:43:13 Tango: update? 22:43:18 I have several patches for the Heat temlates to configure Kubernetes services in the bay to talk to Neutron 22:43:26 I am adding a few more patches for the conductor to compute the parameters for the templates 22:43:46 these are upstream patches, correct? 22:43:56 can you please reference them in the BP whiteboard? 22:44:10 I think they are linked already 22:44:16 ok, good 22:44:20 I have been testing the feature with the wordpress example: 22:44:23 if there are any new ones, please add them there 22:44:29 https://github.com/GoogleCloudPlatform/kubernetes/tree/master/examples/mysql-wordpress-pd 22:44:37 So for the functional test, I will see if we can run a variation of this. 22:44:44 excellent 22:44:46 The OpenStack plugin seems to be still in development. 22:44:50 upstream patches being #link https://github.com/stackforge/heat-coe-templates 22:44:53 this is targeted for kilo 22:44:57 There was a patch included in the latest release 0.19, and I see at least another WIP patch. 22:45:05 I would like to pull it into liberty-2 22:45:10 is that realistic? 22:45:18 when's liberty-2 end? 22:45:27 2 months form now 22:45:31 july31 22:45:31 *from 22:45:44 For the external LB? 22:46:12 yes, Tango 22:46:20 #link https://wiki.openstack.org/wiki/Liberty_Release_Schedule 22:46:20 We can get everything from our side 22:46:35 we need it from kube side before they pull triggeron 1.0 22:46:44 but there will be some dependency on Kubernetes code 22:46:59 FYI, I'm interested in taking on https://bugs.launchpad.net/magnum/+bug/1444368 22:47:00 Launchpad bug 1444368 in Magnum "Bay status is still CREATE_IN_PROGRESS after stack create failed" [High,Incomplete] 22:47:23 mfalatic: please touch base with diga__ on that and split the work as needed 22:47:32 Currently, when I try Kubernetes version 0.18 or 0.19 on our Fedora Atomic VM, the services crash with: 22:47:40 panic: runtime error: invalid memory address or nil pointer dereference 22:47:44 unles he was volunteering for something else 22:47:52 Not sure if the image needs updating, so I am building a new Fedora Atomic image with tutelage from sdake 22:47:57 oh 22:48:09 Tango: let's follow up on the ML about this 22:48:16 ok 22:48:20 best way to deal with situations like this is to make them highly visible 22:48:26 can I assign you an action? 22:48:32 sure 22:49:01 #action Tango to start and ML discussion to explain trouble with Magnum/K8s integration for LB (https://blueprints.launchpad.net/magnum/+spec/external-lb) 22:49:02 thanks 22:49:03 next 22:49:23 #link https://blueprints.launchpad.net/magnum/+spec/magnum-swarm-scheduling (diga__) 22:49:34 yes 22:49:39 I think this is actually solved, and needs to be withdrawn, correct? 22:49:47 or is there more to it? 22:50:25 yes, but never got time to work on this, please set target for l2 22:50:32 ok, retargeting 22:50:37 i htink it is solved by swarm 22:50:40 At the summit didn't we discuss to only use native scheduling? 22:50:46 we agreed at summit we wouldn't do custom scheduling 22:50:50 done 22:50:55 sdake: +1 22:51:00 at least not at this stage 22:51:23 next 22:51:28 #link https://blueprints.launchpad.net/magnum/+spec/secure-docker (apmelton) 22:51:41 Ok, so sounds like diga__ is working https://bugs.launchpad.net/magnum/+bug/1444368 and just isn't names in launchpad for some reason. 22:51:42 Launchpad bug 1444368 in Magnum "Bay status is still CREATE_IN_PROGRESS after stack create failed" [High,Incomplete] 22:51:42 this is in new status 22:51:57 not started 22:52:10 adrian_otto: yes 22:52:21 apmelton there is a guide kicking around somewhere how to do tls 22:52:31 apmelton recommend reading before beginning 22:52:31 want to pull in volunteers to help? 22:52:51 owning a blueprint only means that you have to respond to status inquiries and coordinate the effort 22:52:58 you can get as many contributors to help you as you want 22:53:13 *can convince :) 22:53:15 actually, this piece is going to be pretty dependent on how k8s handles tls 22:53:34 https://bugs.launchpad.net/magnum/+bug/1444368 & https://bugs.launchpad.net/magnum/+bug/1451118 22:53:36 Launchpad bug 1444368 in Magnum "Bay status is still CREATE_IN_PROGRESS after stack create failed" [High,Incomplete] 22:53:37 Launchpad bug 1451118 in Magnum "magnum throws db exception on failed bay deletion" [High,Triaged] - Assigned to Digambar (digambarpatil15) 22:53:45 apmelton: +1 22:53:48 so who ever is working on that can probably modify the swarm template pretty similarly 22:53:58 mfalatic: both are same I guess 22:54:15 ok, let's link them together for now 22:54:25 okay 22:54:29 apmelton: can you do that please? 22:54:48 associate secure-docker and secure-kubernetes 22:55:00 adrian_otto: sure, let me figure out how 22:55:09 that concludes review of the Essential BPs 22:55:16 we have 5 min remaining 22:55:18 sdake: do you have the recommended reading link handy? thanks 22:55:30 normally we do Open Disucssion here. 22:55:34 juggler define handy 22:55:43 I didn't see the discussion about the network one 22:55:49 as in an available URL :) 22:55:51 but I have one orphan in High priorty: 22:56:03 https://blueprints.launchpad.net/magnum/+spec/native-docker-network 22:56:06 juggler i think tis on the wiki 22:56:08 I would love to hear about any plans for RDO packaging for Magnum 22:56:09 if you cant find it there ping me 22:56:10 apmelton: yes, that one 22:56:22 I need a volunteer to drive that one 22:56:31 mr smith i packaged magnum for rdo 22:56:33 hey mismith welcome 22:56:35 The quickstart documentation is all devstack centric - not very helpful for us operators 22:56:40 #link https://blueprints.launchpad.net/magnum/+spec/native-docker-network 22:56:41 adrian_otto: which one? 22:56:44 sdake: Oh yeah? 22:56:45 ah 22:57:07 I gave a presentation yesterday at the OpenStack Israel Conference 22:57:09 sdake: Where can one find the RPMs or srpms? 22:57:21 http://presentation.midokura.me/openstack_israel_2015/#/ 22:57:22 i am in the process of adding #link https://github.com/contiv/netplugin to #link https://github.com/stackforge/heat-coe-templates 22:57:26 I will start Open Discussion now. If you can help with native-docker-network, please update the BP, or see me. 22:57:33 #link Open Discussion 22:57:35 Any plans on the midcycle meetup adrian_otto? 22:57:38 http://presentation.midokura.me/openstack_israel_2015/#/23 22:57:39 mismith https://copr.fedoraproject.org/coprs/sdake/openstack-magnum/ 22:57:43 starts the magnum part 22:57:48 mismith the systemd stuff is untested 22:57:50 tcammann_: I will begin planning for that this week. 22:57:54 feel free to modify and send me patches :) 22:57:56 mismith: just curios is it just for RHEL or Fedora/CentOS too? 22:57:57 I think it could be a good fit for Magnum networking... meeting a lot of the needs mentioned in #link http://lists.openstack.org/pipermail/openstack-dev/2015-June/066805.html 22:58:10 sdake: Great I'll check it out. We use CentOS 22:58:25 adrian_otto: I would be willing to put a lot of time into the networking part 22:58:26 #adrian_otto to circulate a poll for dates to hold Magnum MidCycle Meetup for Liberty 22:58:40 apuimedo: excellent 22:59:03 about the three options in the blueprint 22:59:09 t-1 minutes 22:59:23 sorry for the super full agenda this week 22:59:26 I feel like we should go for 2 and/or 3, 22:59:39 sdake: Do you know how the RDO folks pick up packages like yours and add it to the official RDO ones? 22:59:55 our next meeting is Tuesday 2015-06-23 at 1600 UTC, sdake will be our chair 23:00:01 mismisth there is a rdo mailing list for such questions 23:00:05 ok thanks all 23:00:10 thanks everyone for attending today! 23:00:14 #endmeeting