16:00:06 #startmeeting containers 16:00:08 Meeting started Tue Jul 26 16:00:06 2016 UTC and is due to finish in 60 minutes. The chair is hongbin. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:11 The meeting name has been set to 'containers' 16:00:12 #topic Roll Call 16:00:19 Murali Allada 16:00:19 o/ 16:00:21 o/ 16:00:22 o/ 16:00:23 Ricardo Rocha 16:00:26 Jaycen Grant 16:00:28 Madhuri Kumari 16:00:46 o/ 16:01:07 Thanks for joining the meeting muralia strigazi swatson Drago rochaporto jvgrant_ mkrai eghobo 16:01:15 #link https://wiki.openstack.org/wiki/Meetings/Containers#Agenda_for_2016-07-26_1600_UTC Today's agenda 16:01:19 o/ 16:01:21 Anything needs to be added to the agenda? 16:01:55 #topic Announcements 16:02:04 I have no annoucement 16:02:13 Any annoucement from our team member? 16:02:29 #topic Review Action Items 16:02:39 1. Support baremetal container clusters (strigazi) 16:02:45 #link https://blueprints.launchpad.net/magnum/+spec/magnum-baremetal-full-support 16:02:54 strigazi: ^^ 16:03:25 Not much progress since last week, I was trying to setup the gate job properly and replicate the issues locally. 16:03:42 I submitted patches for swarm last week 16:04:01 This week I'll test mkrai's patches 16:04:14 Thanks strigazi 16:04:20 o/ 16:04:24 strigazi: Did the gate work properly? 16:04:48 no, we have only 8GB or ram and 80GB disk 16:05:16 Yes, that is the limitation 16:05:42 another issue 16:05:45 we might add a couple larger nodes at cern to get this going (still needs setup) 16:06:42 the fake ironic nodes were not approachable from ironic conductor 16:07:35 The good thing is 16:08:08 that the issue with the size of insstance_info will be solved in ironic, I'm helping a developer on than 16:08:18 that's all 16:08:23 Thanks strigazi 16:08:35 2. Magnum User Guide for Cloud Operator (tango) 16:08:40 #link https://blueprints.launchpad.net/magnum/+spec/user-guide 16:08:44 tonanhngo: ^^ 16:08:56 The section for mesos was merged 16:09:15 I just revised and update the section for TLS: https://review.openstack.org/#/c/347077/ 16:09:40 The section on bay got some feedback, I will update shortly 16:10:19 I think we got all the major sections, now we can wrap up the minor sections 16:10:41 and do a round of update on various new features. 16:10:55 So generally, we are in decent shape 16:11:03 Yes, we are 16:11:12 That's all I have for now 16:11:25 Thanks tonanhngo . Question? Comment? 16:11:37 After a while we are going to split the guide? 16:11:45 In smaller sections? 16:12:04 strigazi: What do you have in mind? 16:12:31 eg this configuration option is related to performance https://review.openstack.org/#/c/326428/15/doc/source/userguide.rst 16:12:54 we could have a section for futher configuration after the basic installation 16:13:13 add lbaas, add barbican 16:13:16 Like tuning? 16:13:19 yes 16:13:27 tuning and features 16:13:35 That's a good idea. 16:13:41 I mean extra features 16:13:56 Let's put those ideas on the BP so we can track 16:14:02 ok 16:14:05 thanks 16:14:26 OK. Advance topic 16:14:29 3. COE Bay Drivers (jamie_h, muralia) 16:14:44 #link https://blueprints.launchpad.net/magnum/+spec/bay-drivers 16:14:46 muralia: ^^ 16:14:52 I don't any major updates. I need to do a lot more testing. Especially around installing a new driver and using it. 16:15:24 just working on that. Will have more updates soon. 16:15:40 muralia: I saw a few patches were merged 16:15:56 muralia: maybe update the team which parts are done, which parts are TODO 16:16:19 muralia: Could I continue this one https://review.openstack.org/#/c/324470/ ? 16:16:28 It will us clean things up 16:16:32 It will help us clean things up 16:16:52 yes, the patches that got merged move the common and driver files to the driver folder. basically the refactor part is done 16:17:06 whats remaining is the integration with stevedore 16:17:15 strigazi: yes, you can work on that 16:17:20 thanks 16:17:43 about refactoring, I have a question/comment 16:17:48 yes 16:18:09 there is this change up for review https://review.openstack.org/#/c/346209/ 16:18:43 I think it's very large and without planing and discussion with the team 16:19:24 strigazi: I think you can relay your comments to the reviews 16:19:29 moving lbaas was a work done by Drago who wrote this part and hence it was done carefully 16:19:55 I think at the least the patch should be split up a little 16:19:59 That I want to say is, we should start a discussion on who to avoid code duplication with the bay drivers 16:20:16 yes, lets add all these comments in the patch 16:20:18 *on how 16:20:48 ok, I'll add it also as discussion topic for the midcycle 16:20:52 How to avoid code duplication between drivers is a good topic for midcycle 16:21:09 I have added "Supporting the install of bay drivers that use different versions of the common directory" which is very similar 16:21:18 hongbin: yes, let me add that to the etherpad. strigazi, we should be ok to move duplicated files to the common folder 16:21:23 There is some cleaning up we can do also after the bay driver refactoring. 16:21:25 Because that is the issue we face when we start de-duplicating 16:21:26 I found some more code duplicacy in heat templates also. 16:21:37 Drago, didn't see it, thanks 16:21:58 de-duplicating files for the 4 drivers that come by default with magnum should be fine. 16:22:02 may be we can refactor heat templates also. 16:22:58 OK. any other comment? 16:23:10 We should discuss how this is going to happen first 16:23:18 ok 16:23:30 4. Create a magnum installation guide (strigazi) 16:23:30 yes, its been added as a topic to the mid-cycle agenda 16:23:39 #link https://blueprints.launchpad.net/magnum/+spec/magnum-installation-guide 16:23:46 strigazi: ^^ 16:24:27 The guide is complete and merged for the 4 distros documented across all projects. 16:24:45 Yes, I think we can close the BP 16:24:47 Very nice! 16:25:02 strigazi: are you able to close the BP? 16:25:03 we can leave it until the end of the cycle 16:25:16 because I'll have to test it again 16:25:27 thanks strigazi for all your work on the install guides. 16:25:28 the functionality might change 16:25:54 But, we can remove it from the meeting agenda 16:26:04 strigazi: OK. Will do that 16:26:25 I'll work again when the official packages are ready 16:27:02 adrian_otto: It was my pleasure 16:27:03 strigazi: Thanks strigazi 16:27:15 Next topic 16:27:18 #topic Kuryr Integration Update (tango) 16:27:25 tonanhngo: ^^ 16:27:44 I attended the IRC last night but the meeting didn't occur 16:28:02 There is some discussion on the ML for a new meeting time, so I will keep an eye on that 16:28:33 The code refactoring is still not complete, so we can't quite build yet. I am still using the old copy 16:28:56 For us it should be OK, since we will just pull the container image for Kuryr 16:29:27 I do have the openvswitch and L2 agent working in a Fedora 23 container in Swarm 16:29:38 cool 16:29:56 I just register the new BP with some details on the implementation 16:30:24 I will wrap up the prototype and start the patches 16:30:53 One question for the team, I think this was brought up before. 16:31:25 Should we upload custom container image to Docker Hub, for Magnum use? 16:31:58 We can implement the openvswitch container as Dockerfile, but a ready built image would be faster 16:32:23 +1 16:32:24 For me, it is fine. 16:32:29 tonanhngo: This will raise the same issue I had with rexray 16:32:51 Is there a Magnum user account already on Docker Hub? 16:33:03 yes but we don't know the owner 16:33:12 it's openstackmagnum 16:33:29 Should we try to find the owner, or just create a new account? 16:33:39 We can create a new one with different name 16:33:45 We can do both 16:33:59 OK, let me send to the ML to ask first 16:34:17 If no answer, we can create a new account 16:34:39 That's all I have for now 16:34:47 Thanks tonanhngo 16:34:59 I don't know who controls the "openstackmagnum" hub.docker.com account. 16:35:48 me either 16:35:59 but whatever accounts we have, I suggest that it be held only by the PTL, and transferred to the next PTL upon any change in control. 16:36:00 Let's find the owner in ML 16:36:19 adrian_otto: sure 16:36:43 Any other comment? 16:37:01 #topic Other blueprints/Bugs/Reviews/Ideas 16:37:06 and there should be reviewed instructions for how to verify those images, so the fear of tampering can be addressed. 16:37:07 1. Select our project mascot/logo 16:37:20 #link http://lists.openstack.org/pipermail/openstack-dev/2016-July/100011.html 16:37:49 Proposed mascots: 16:38:06 Stallion - http://free-icon-download.com/modules/PDdownloads/singlefile.php?cid=11&lid=44 16:38:11 Snake - http://www.freeiconspng.com/uploads/snake-jungle-22.png 16:38:13 Think we should have a Magnum Moose 16:38:16 Wave - http://www.123rf.com/photo_11649085_set-of-waves.html 16:38:23 Paraponera clavata - https://en.wikipedia.org/wiki/Paraponera_clavata 16:38:28 Bombardier beetle - https://en.wikipedia.org/wiki/Bombardier_beetle 16:38:32 shark - http://www.logoground.com/logo.php?id=10554 16:38:36 kangaroo with a joey in its pouch - http://www.supercoloring.com/pages/red-kangaroo 16:38:42 koala bear with a joey in its pouch - http://web.stanford.edu/~jay/koalas/koala%20mums.jpg 16:38:46 beaver - https://commons.wikimedia.org/wiki/File:Beaver_%28PSF%29.jpg 16:39:00 tcammann: what is Magnum Moose? picture? 16:39:50 majestic moose https://images.indiegogo.com/file_attachments/1328366/files/20150326083908-Mooselaughing.jpg?1427384348 16:39:59 Kangaroo, Stallion, and shark are my favorites so far, though a moose might be pretty cool 16:40:00 My question is, should we have it relate to what magnum does or where it gets its name? Or does it matter at all? 16:40:17 i.e. what significance does the mascot have to Magnum 16:40:17 tcammann: +1 16:40:31 Drago: it can be any animal or natural feature. 16:40:43 it can relate but it doesn't have to 16:41:01 The actual story of the magnum name can't be connected in this case. Our name came from a Magnum of Champagne. 16:41:13 I like the waves one as it relates to whales (docker) 16:41:27 +1 wave 16:41:29 I like tonanhngo idea of kangaroo carrying joey in its pouch 16:41:32 waves, kangaroo, moose (in this order) 16:41:35 they have an artist who will make an icon that has a similar style to all the others in Openstack 16:41:53 Shark, Waves, Kangaroo, Moose 16:42:00 if what they have on the site is the style they will be cartoony like drawings 16:42:05 something related to the sea since we have bays 16:42:08 As it can relate to kangaroo being magnum carrying containers(joey) 16:42:36 more "badass" waves http://images.metmuseum.org/CRDImages/as/original/DP130155.jpg 16:43:03 How are we going to choose with so many choices? :) 16:43:17 we only need to identify a short list 16:43:21 Waves makes me think Hawaii (Magnum P.I.) 16:43:23 by tomorrow 16:43:49 Yes, let's get a short list first 16:44:22 so maybe if three of us can agree on an item being on the list, then it makes it in? 16:44:46 and we can use whatever formal voting process to narrow it down form there 16:45:27 Looks like Waves, Kangaroo, Shark, and Moose are top so far 16:45:50 jvgrant_: yes 16:45:55 so if you feel strongly about adding to that list, make your case. 16:46:45 Yes, let's assume here is the list : Waves, Kangaroo, Shark, and Moose 16:46:58 If you want to add to this list, let me know by the end of today 16:47:14 Sounds good. 16:47:21 The, I will send the list to OpenStack foundation 16:47:54 OK. Anything else for this topic? 16:48:13 2. Midcycle topics 16:48:20 #link https://etherpad.openstack.org/p/magnum-newton-midcycle-topics 16:48:53 For everyone, please feel free to add your topics, and +1 on the existing topics if you want 16:49:13 We can collaborate on the etherpad offline 16:49:23 #topic Open Discussion 16:50:06 my small query: for magnum there are lot of bugs that are open and those seems invalid. Some bugs are there which do not have description at all. Can someone from core reviewers mark them invalid/won't fix. This will narrow down the bug list of magnum and this will result in closing the 16:50:28 +1 16:50:29 +1 16:50:32 yatin: good suggestion 16:50:38 +1 16:50:49 I think the core team can do that 16:51:02 Maybe a good session for the midcycle 16:51:07 +1 16:51:07 The pod/container bugs should be easy 16:51:56 i would like to simplify the user commands to create a baymodel/bay by making most of the params optional (external-network, keypair, image, dns-nameserver). in most cases these are not needed, coe (bay driver) specific or better defined by the operator (in the service config). blueprint first or should i go for a patch? 16:52:41 I would suggest BP first, since this involves the user experience 16:52:42 rochaporto: I think both bugs and BP will work 16:52:53 ok 16:53:06 I think it is a good idea 16:53:24 We need to hard-code the default value to read the default value from config file 16:54:06 but the direction is OK for me. 16:54:16 ok i'll work on this 16:54:47 If nobody else has a topic, I want to raise this one: #link https://review.openstack.org/#/c/331614/ 16:55:40 In the patch, there is a discussion about whether Magnum needs a feature to disallow duplicate bay/baymodel name 16:56:05 Some folks think this feature is unnecessary, others think the opposite 16:56:24 all openstack services allow duplicate names today, so i prefer to not merging this patch 16:56:36 muralia: ack 16:57:03 muralia: +1 16:57:19 My concern is that it can cause confusion and degrade the user experience with Magnum. 16:57:41 I have a question, I' having trouble creating bay behind proxy 16:57:58 http://paste.openstack.org/show/542148/ 16:58:09 tonanhngo: can you detail the scenario(s) you are concerned about? 16:58:16 It says create in progress and times out 16:59:01 akhilakishore: Please ping us in the #openstack-containers channel offline. I believe someone will help you to troubleshoot that 16:59:11 akhilakishore: that's often the callback to heat (for the wait condition) failing. check the heat nodes are reachable from your vms 16:59:12 akhilakishore, It would be better if you show us baymodel parameters 16:59:14 Oh ok 16:59:53 1 minute left 17:00:08 https://bugs.launchpad.net/magnum/+bug/1605416 i want to discuss this bug 17:00:08 Launchpad bug 1605416 in Magnum "neutron loadbalancer uses incorrect security group" [Critical,Triaged] - Assigned to rajiv (rajiv-kumar) 17:00:24 Time is up. Let' move the discussion to the project channel 17:00:28 time up 17:00:29 Thanks everyone 17:00:31 adrian_otto: I mentioned in the comment. Basically different clouds can be configured to behave differently with respect to duplicate name, and which baymodel is referenced. 17:00:33 #endmeeting