16:01:33 #startmeeting containers 16:01:33 Meeting started Tue Feb 23 16:01:33 2016 UTC and is due to finish in 60 minutes. The chair is adrian_otto. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:34 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:37 The meeting name has been set to 'containers' 16:01:41 #link https://wiki.openstack.org/wiki/Meetings/Containers#Agenda_for_2016-02-23_1600_UTC Our Agenda 16:01:50 #topic Roll Call 16:01:53 Adrian Otto 16:01:55 o/ 16:01:57 murali allada 16:01:58 Perry Rivera o/ 16:01:59 Thomas Maddox 16:01:59 Ton Ngo 16:02:03 o/ 16:02:11 o/ 16:02:19 o/ 16:02:31 o/ 16:02:55 o/ 16:03:31 o/ 16:03:42 o/ 16:04:20 o/ 16:04:46 hello coreyb muralia1 juggler thomasem Tango strigazi rpothier dimtruck sew1 dane_leblanc eghobo madhuri Kennan 16:04:52 let's begin 16:04:53 o/ 16:05:00 #topic Announcements 16:05:22 1) Introducing Spyridon Trigazis from CERN who has joined us as a full time contributor. 16:05:39 strigazi: would you like to say a few words? 16:06:05 Hello all, 16:06:20 Glad to know you strigazi: 16:06:28 welcome. 16:06:35 welcome strigazi 16:06:36 I'm Spyros Trigazis and I'm a fellow at Cern in the resource provisioning team 16:06:56 I'll mostly work on Magnum upstream 16:07:11 Some bio notes: 16:07:14 hello strigazi! we're glad you're here. 16:07:34 strigazi: we are really happy to have you as part of the team, and we look forward to working with you. 16:07:42 Bsc and Msc in CS on Information Sysystem from www.csd.auth.gr 16:07:57 Information Systems 16:08:15 please let us know if there is anything we can do to help you 16:08:17 I have previously contributed in www.ganeti.org 16:08:46 I stared to feel comfortable with Devstack 16:09:19 And I'm currently work on adding docker storage drivers in baymodel 16:09:26 o/ 16:09:39 I'll send tomorrow a WIP patch 16:09:54 Thanks strigazi. We look forward to that! 16:09:57 Welcome strigazi ! 16:09:59 o/ 16:10:38 2) We had a very productive midcycle meetup at HPE in Sunnyvale last week. I will be putting together a summary of the outcomes there to share with the team. 16:11:06 the largest of which has to do with our decision to handle COE specific code in drivers. 16:11:32 so we will be planning some refactoring work to make that possible. 16:11:38 That concludes our prepared announcements. Are there announcements from team members today? 16:12:12 ok, let's continue 16:12:18 #topic Review Action Items 16:12:28 1) hongbin create a BP for creating an installation guide 16:12:58 #link https://blueprints.launchpad.net/magnum/+spec/magnum-installation-guide 16:12:58 hongbin, do we have a link to this we can share? 16:13:01 thanks!! 16:13:08 Status: COMPLETE 16:13:30 #topic Blueprint Review 16:13:40 Essential Blueprint Review 16:14:32 hi adrian_otto 16:14:40 there is another review item 16:14:41 #link https://blueprints.launchpad.net/magnum/+spec/magnum-troubleshooting-guide (Tango) 16:15:04 We have 2 patches under review 16:15:08 #link https://blueprints.launchpad.net/magnum/+spec/user-guide (Tango) 16:15:19 I noticed further progress on those BPs 16:15:38 https://review.openstack.org/#/c/275034/ 16:16:12 we still have 11 sections that needs to start from scratch with no author assigned for the troubleshooting guide 16:16:26 Kennan: I'd be happy to address that one in just a moment. 16:16:35 For the user guide, I am writing the section on image management 16:16:50 will be coordinating with the new COE driver refactoring 16:17:24 I am using the discussion from the midcycle to write about image 16:18:07 For the user guide, there are 6 sections that need to start from scratch with no author assigned 16:18:22 8 sections are done, have existing doc, author assigned, or in progress 16:18:54 So overall, we are about half way on both the user guide and troubleshooting guide 16:19:18 That's all I have 16:19:42 ok, what's the best way to get contributors to adopt each of the remaining sections? 16:19:56 perhaps pull them into an etherpad during today's meeting? 16:20:02 The BP's have a list of section as TODO 16:20:18 Anyone can pick up any section 16:20:27 just put your name on the TODO 16:21:03 Tango: so just a PR with your name there or in the blueprint? 16:21:07 ugh, never mind 16:21:13 ok, so that's at https://blueprints.launchpad.net/magnum/+spec/user-guide and https://blueprints.launchpad.net/magnum/+spec/magnum-troubleshooting-guide 16:21:15 you just said "BP's have a list..." 16:21:26 it would be great if we could each find one to help with 16:21:40 At the bottom of the BP, there is a list of work items, some of them are in TODO state 16:22:02 Tango: yup, sorry about that. understood 16:22:42 So if you know about a topic, please feel free to pick it up 16:23:55 Blueprints, Bugs, Specs, and other work items to be discussed as a team 16:24:04 back the Kennan 16:24:10 … (getting the link) 16:24:26 #link https://review.openstack.org/#/c/275034/ 16:24:59 so this one one I placed a -2 vote on a while back because it disabled selinux 16:25:27 hi adrian_otto: I have added comments for you concern. In short 16:25:57 ok, I'm happy to revisit this because there are new patchsets. 16:26:03 volume driver is optional fearure, 1) we add todo to track that in upstream (docker) side in issue 16:26:18 2) if user not enable such volume driver 16:26:23 selinux is still enabled 16:26:31 so not impact any components 16:26:41 docker is still working on that 16:26:47 can't we implemnet the storage driver in a way that allows selinux to remain enabled? 16:26:55 and we are working with them 16:27:17 I generally don't like the idea of having insecure features in magnum. I think we shoudl wait until it works securely before implementing it 16:27:30 right now, volume driver plugin can not work because of docker issue 16:27:45 it is not means magnum like to be insecure 16:27:54 but as experimental feature 16:27:58 but disabling selinux makes magnum less secure 16:28:01 it is OK for such volume dirver 16:28:07 you are telling me that docker has a feature that requires selinux to be disabled in order to use it? 16:28:09 no coreyob: 16:28:24 let me give you link 16:28:37 I doubt it. I think this is rather a matter that requires a new selinux policy to be created 16:28:47 https://github.com/docker/docker/issues/18005 16:28:56 it is known issue 16:29:42 specifically with the rexray driver right? 16:29:51 coreyob, experimental feature not means perfect. as we are working with docker, but it not means becasue docker issue, we can not experimental volume feature in swarm 16:29:54 so can we implement the feature in magnum with something other than rexray? 16:29:57 no coreyob 16:30:20 right, I'm advocating that we don't have experimental features in magnum that reduce the overall security of the system 16:30:42 i like adrian_otto's idea of having a custom selinux policy that allows rexray to function without compromising the rest of the bay (if that is possible) 16:30:48 coreyob: I did not think right now it is all production ready for magnum 16:31:05 we're trying really hard to make magnum production ready 16:31:43 yes coreyob: as we know, even docker itself can be say secure, like user namespace just enabled in 1.10.0 16:31:52 so all is progress 16:32:02 ok, I'm willing to review the individual docker bug, but if there is a fundamental security problem in a docker feature, it needs to be solved there. 16:32:16 yes adrian_otto 16:32:24 we are pushing docker 16:32:25 I'm not comfortable poking a huge security hole in Magnum to add a buggy feature. 16:32:44 I hear your argument that the feature is optional. 16:33:04 but I'm not fully persuaded 16:33:08 adrian_otto: I understood that. you know, if user not like to use volume feaure management in COE 16:33:19 my preference is to wait until this is settled upstream 16:33:19 he not need to use volume_driver 16:33:22 then integrate it. 16:33:42 so volume_driver is optional feature. 16:33:50 or to come up with another workaround that allows selinux to remain enabled 16:34:22 right now, because docker side issue, seems volume plugin can not work if enable selinux 16:34:41 Kennan: the issue is on Atomic only or not? 16:34:52 Kennan: If not, you could use the CoreOS template 16:34:54 I tried on Atomic, it exist 16:34:58 not sure other OS 16:35:19 Kennan: maybe you can try CoreOS 16:35:30 Maybe we can do some more investigation 16:35:47 Kennan: you can merge the feature back to Atomic later 16:36:10 sure hongbin: but that would means a new work on that templates. Anyway, I will work with docker upstream and at same time check coreos 16:36:23 seems coreos have just raw implemented 16:36:29 TLS seems not ready 16:36:42 so still it is not secure :0 16:37:03 It will be soon. The TLS patch is up for review 16:37:50 Kennan: Alternatively, you could try the feature in other COE first 16:37:55 ok hongbin: two ways for me 1) still track upstream docker 2) check coreos 16:38:47 Kennan: sure 16:38:48 I will push related docker commmunity to solve that general issue 16:39:22 Kennan, as soon as we can add the feature in Magnum without disabling selinux, I will be happy to remove my -2 vote. 16:39:50 adrian_otto: I will track that. aslo I learnt Atomic 23 is working in progress 16:39:59 I need to check if that new images ok for that 16:40:18 seems that work in still in review 16:40:58 yeah i keep running into gate issues so I haven't been able to get it though 16:42:53 hongbin: I saw your email yesterday about the recent Heat commit you called out 16:43:10 was there any confirmation that Hat was the source of the gate trouble, or was that ruled out? 16:43:31 adrian_otto: Yes, they seems to confirm that, and proposed a revert 16:43:56 adrian_otto: Unfortunely, the reverted patch is not merged yet, since the Heat gate is also broken 16:44:46 hongbin: ceilometer working on that, so heat would soon ready for ceilometer issue 16:44:58 what's wrong with the Heat gate? 16:45:11 is it related to the same issue, or something different? 16:45:15 different 16:45:58 #link https://review.openstack.org/#/c/283297/ 16:47:26 we may run into #link https://bugs.launchpad.net/ceilometer/+bug/1548634 after heat is working too 16:47:26 Launchpad bug 1548634 in Aodh ""openstack role add" command failure in devstack installation" [Undecided,In progress] - Assigned to Liusheng (liusheng) 16:48:01 coreyob: ceilometer had work-around for that. 16:48:50 yep 16:49:43 as part of the open discussion - my sincere request to the team to draw conclusion for https://review.openstack.org/#/c/275003/ , https://review.openstack.org/#/c/267134/ 16:49:50 ok, any other work items for team discussion before we proceed to open discussion? 16:49:54 adrian_otto: ^^ 16:49:57 thanks suro-patz 16:50:33 there is the bp I'm working on 16:50:47 https://blueprints.launchpad.net/magnum/+spec/support-for-different-docker-storage-driver 16:51:30 It says that we must have: 1) A user seeking this, and 2) A developer willing to add it 16:51:44 adrain_otto : regarding the quota work ; would be great if i can get some help from the team just for next month or so 16:52:31 https://review.openstack.org/#/c/259201/10 16:52:57 i was planning to submit the controller logic changes so need help specifically there 16:53:13 adrian_otto : ^^ 16:53:14 ok, suro-patz we can vote on those spec patches to allow that to move forward 16:53:29 thanks adrian_otto 16:54:29 adrian_otto: additionally I will register a BP/spec for the asynchronous mode of operation for bay updates 16:55:25 vilobhmm11: thanks for the callout on that. To be clear you have some downstream work you will be focused on for a month, so you are seeking team assistance to continue your upstream work in the mean time. 16:55:39 #topic Open Dicsuccion 16:55:41 adrian_otto : thats correct 16:55:49 and thats only for a month or so 16:55:58 #topic Open Discussion 16:56:09 vilobhmm11: understood 16:57:10 adrian_otto: thanks! 16:57:17 if anyone is interested in working on quotas, please let vilobhmm11 or me know. Otherwise, I can ask around and see what's possible. 16:57:53 adrian_otto: is it ok to proceed with the bp that I mentioned? 16:58:53 strigazi: you are welcome to pick up any approved blueprint. We have some really critical work that came from the midcycle that I'd like you to consider as well. 16:59:37 adrian_otto: ok 16:59:45 I'll get my write-up done to convey the decisions, and how those impact our next steps, and where we will need to focus attention to succeed with them. 17:00:04 Our next team meeting is 2016-03-01 at 1600 UTC. 17:00:19 thanks everyone for attending today. I look forward to seeing you next week. 17:00:20 thanks for presiding, and thanks all 17:00:22 #endmeeting