16:01:17 #startmeeting Solum Team Meeting 16:01:18 Meeting started Tue Mar 11 16:01:17 2014 UTC and is due to finish in 60 minutes. The chair is adrian_otto. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:20 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:24 The meeting name has been set to 'solum_team_meeting' 16:01:28 o7 morning 16:01:36 #link https://wiki.openstack.org/wiki/Meetings/Solum#Agenda_for_2014-03-11_1600_UTC Our Agenda 16:02:16 #topic Roll Call 16:02:19 Adrian Otto 16:02:25 Roshan Agrawal 16:02:25 Pierre Padrixe 16:02:27 Julien Vey 16:02:29 Paul Czarkowski 16:02:30 Ed Cranford 16:02:31 Devdatta Kulkarni 16:02:31 Arati Mahimane 16:02:31 Georgy Okrokvertskhov 16:03:08 Swapnil 16:03:29 hi everyone 16:03:53 Daylight Savings Time (yay) 16:04:35 I'd like to pursue a #agreed to end Daylight Savings time, all time zones will be eliminated, and move everyone to UTC 16:04:42 ;-) 16:04:58 :D .. good luck with that .. which channel you will be discussing it ;) 16:05:09 I agree, so long as we move to metric time 16:05:18 heh, It's an exercise in futility 16:05:31 #topic announcements 16:05:40 our Solum Summit event is growing closer. 16:06:01 if you are planning to attend, be sure you have registered so Dan at Red Hat can get an accurate count 16:06:23 do other team members have announcements to share? 16:06:35 and add topics to discuss here: https://etherpad.openstack.org/p/SolumRaleighCommunityWorkshop 16:07:02 #link https://etherpad.openstack.org/p/SolumRaleighCommunityWorkshop Solum Summit Agenda Topics 16:07:07 thanks devkulkarni 16:07:53 ok, let's look at our action items from last week 16:08:06 #topic Review Action Items 16:08:08 adrian_otto to adjust the wiki to show intent for incubation (adrian_otto, 16:20:35) 16:08:25 this is done. See: 16:08:35 #link https://wiki.openstack.org/wiki/Solum#Summary Solum Wiki 16:08:39 is that that timestamp when you finished it? :D 16:08:52 just in the nick of time 16:09:23 the timestamp above is from the prior meeting chat log, so you can find it in context 16:09:25 "Solum is planning to pursue incubation as an OpenStack project upon achievement of development milestones." — awesome 16:09:32 ah! I see. 16:09:58 anyone is welcome to edit that to improve it if you feel so motivated. 16:10:31 next is the mission statement 16:10:33 adrian_otto to draft mission statement and solicit input form interested contributors (adrian_otto, 16:35:50) 16:10:59 I did not act on this item. I think this would be a valuable exercise for our upcoming Solum Summit event 16:11:18 I plan to come with a draft for us to iterate on then 16:11:35 oh yeah. that is a great idea. discuss it at the summit 16:11:37 who would like to work with me to make that initial draft via etherpad and IRC? 16:11:49 +1 16:12:01 I would like to help 16:12:04 gokrokve: excellent 16:12:50 ok, the three of us can touch base after we adjourn here, and sort out a convenient time. 16:13:04 sounds good 16:13:06 so I will carry this action item forward 16:13:28 #action adrian_otto to draft mission statement and solicit input form interested contributors (gokrokve, devkulkarni) 16:13:34 I'm in ! 16:13:46 julienvey: excellent, I will include you 16:14:08 let's pick an etherpad now. one sec 16:14:37 #link https://etherpad.openstack.org/p/solum-mission Mission Statement Draft work 16:15:13 ok, so those interested are welcome to check that 16:15:28 next topic 16:15:33 #topic Review Blueprints: https://launchpad.net/solum/+milestone/milestone-1 16:15:51 #link https://blueprints.launchpad.net/solum/+spec/api Solum API (aotto) 16:16:06 The API is looking reasonably complete at this point 16:16:11 at least for M1. 16:16:13 +1 16:16:39 noorul mentioned this morning that we believe everything for the CLI is in there 16:16:40 I have links to chare 16:16:45 yes, we are solving the last problems by doing the functionnal tests 16:17:18 1. Add solum builder client (https://review.openstack.org/#/c/78315/) 16:17:18 This one is required as it fixes a bug. 16:17:19 2. Add plan manager and corresponding tests 16:17:19 (https://review.openstack.org/#/c/72105/) 16:17:19 Takes care of the app commands 16:17:45 3. Connect assembly commands and Rest API 16:17:45 (https://review.openstack.org/#/c/79606/) 16:17:45 Takes care of the assembly commands. 16:18:03 I will focus on making sure these get the attention needed to merge 16:18:16 please review these as well to offer your input 16:18:21 +1 to that adrian_otto 16:18:46 that is a good intro to the related topic 16:18:46 #link https://blueprints.launchpad.net/solum/+spec/solum-minimal-cli Command Line Interface for Solum (devdatta-kulkarni) 16:18:55 you gave the updates already :) 16:18:57 devkulkarni: any additional status beyond the above? 16:19:05 nope 16:19:07 sweet, we can move on 16:19:15 #link https://blueprints.launchpad.net/solum/+spec/solum-git-pull Pull integration of Solum from an external Git repo (kraman) 16:19:23 about this: 16:19:40 we have the code for it in image_handler + build_app. 16:19:52 so we are good for now. the remaining things are: 16:20:26 1) When a plan is registered we need to send back the trigger url to the user so that they can add it manually as a git webhook to their repo. 16:20:50 2) Testing that the git webhook is getting triggered. 16:21:10 3) Invoking the assembly create workflow when trigger is received 16:21:23 that is all 16:21:32 I can do 1. but I don't understand what you mean in 2. ? 16:21:59 There is also some discussions about the trigger 16:22:10 if it should belong to plan or assembly 16:22:12 julinevey: thanks. for 2, I mean just testing that the trigger is actually working (may not be code based test though) 16:22:28 julienvey: yes, I remember that discussion. 16:22:33 don't know where we landed on that. 16:22:41 it is currently in assembly 16:22:50 devkulkarni: on #2, did you mean that we need help from team members, or that this is what will be worked next? 16:23:32 adrian_otto: actually, I might have spoken incorrectly. I don't think we need any code for #2. 16:23:38 we do 16:23:54 because I'm reasonably sure the trigger handler code is just a comment still, right? 16:24:01 for #2, we have an integration test, is that enough ? 16:24:15 for the trigger part, the comment should be replaced by #3 16:24:15 julienvey: yes, I think integration test will be enough 16:24:46 adrian_otto: trigger handler code is kind of there (need to double check). julien, might know the current status on that 16:24:52 ok, seems like #2 depends on #3 16:25:05 #1 is missing 16:25:10 #2 is tested 16:25:12 #3 is definitely missing. 16:25:18 and #3 is missing yes 16:25:19 yes, #1 is missing 16:25:27 ok 16:25:43 we just need to plug it 16:25:49 all the plumbing is here 16:25:51 yep 16:26:15 do we have enough brains applied to that work, or should I help with recruiting more? 16:26:35 I think we have it covered. 16:26:48 of course we will need reviews :) 16:26:51 yes 16:26:56 ok, thakns for the update devkulkarni. Exciting to be so close! 16:27:03 yep!! 16:27:32 #link https://blueprints.launchpad.net/solum/+spec/specify-lang-pack Specify the language pack to be used for app deploy (devdatta-kulkarni) 16:27:42 okay. the updates on this are: 16:27:55 we touched on this on the ML this week too 16:28:05 1) aratim finished the GET lang-pack 16:28:12 2) what adrian_otto said 16:28:31 aratim: !! yay !! 16:28:42 I have some quick notes that touch on both M1 and LP blueprints adter devkulkarni is done 16:28:58 paulczar: please share them :) 16:29:10 I am done 16:29:15 oh wait! 16:29:17 I am also working on POST/PUT/DELETE for LP, to be able to create initial data with the API 16:29:27 yep. was about to add that 16:29:30 thanks julienvey 16:29:33 should have a review tomorrow 16:29:41 +10 16:29:44 If we want to use Docker nova-driver for M1 We need the following reviews merged into their various projects: 16:29:49 julienvey: Thanks!! 16:30:18 devstack fixes for docker driver install - https://review.openstack.org/#/c/78322/ ( we have an easy localrc workaround for this, so not a huge problem ) 16:30:52 paulczar: let's revisit this one in Open Discussion today. I have news to share about this. 16:30:54 olso-common libraries fail out when glance images have no name - https://review.openstack.org/#/c/78642/ ( this is BREAKING ) 16:30:57 oh ok 16:31:12 tx 16:31:31 devkulkarni: any more on specify-lang-pack BP? 16:31:42 adrian_otto: that will be all. 16:31:46 #link https://blueprints.launchpad.net/solum/+spec/logging Logging Architecture (paulmo) 16:32:02 I think this is done, right? 16:32:15 paulmo is out today 16:32:22 structured logging is in scope for M1 16:32:30 but I don't recall seeing code for that 16:32:46 let's ask paulmo when he returns 16:32:50 #link https://blueprints.launchpad.net/solum/+spec/deploy-workflow Workflow outlining deployment of a DU (asalkeld/devdatta-kulkarni) 16:32:58 Paul's patch did get merged. Georgy is using that right? 16:33:08 ok. about the deploy-workflow. 16:33:19 Patch was merged 16:33:29 Angus has started on adding the Heat client 16:33:55 devkulkarni: did we get the unit/func test code added for that? 16:34:07 I have not looked at that review thread since Thu 16:34:10 you mean for the Heat client? 16:34:18 will check. 16:34:33 but I did see that asalkeld posted a series of newpatchset actions 16:34:54 if that's not yet done, we might seek volunteers to assist him with that 16:35:01 yeah. I need to also go back and check that.. 16:35:19 okay. let me take a look soon and get back to you. 16:35:20 ok, one moment 16:36:01 ok, we can revisit that later 16:36:10 I think we are in pretty good shape with it 16:36:11 sounds good 16:36:18 yeah, I thought so too. 16:36:47 #topic Open Discussion 16:36:49 that will be all on that bp 16:36:54 so paulczar 16:36:59 yes 16:37:12 I met with members of the Nova team this past week 16:37:32 and got a sense of clarity with respect to where our containers support should come from 16:37:49 early on we evaluated options for using Docker, or similar technologies 16:37:58 our favorite one was to use the Nova Docker Driver. 16:38:21 was? 16:38:38 that driver is somewhat incomplete, and has some open bugs, although Sam Alba is continuing to work on it 16:38:58 there is also work on an equivalent solution that plugs in container support underneath the libvirt driver 16:39:21 so Nova would interact with libvirt just like it does to get a KVM instance 16:39:31 but would get a Docker container back 16:39:52 this also provides an integration point for other emerging container technology 16:40:07 so we should expect that driver to mature. 16:40:12 That sounds similar to how lxc is implemented 16:40:18 exactly 16:40:50 it may be difficult to do for docker due to the complexity of the union filesystem 16:40:59 to the implication to Solum is that we can couple to that driver without necessarily coupling to any singe container tech. 16:41:15 s/singe/single/ 16:41:29 how would this affect Heat? As in, in the HOT that Solum generates would we have to worry about the compute? 16:41:32 we will still need to use a container image format 16:41:50 and Docker's is suitable, so I doubt we have a good reason to deviate from that 16:41:50 *type of compute 16:42:06 devananda: great question 16:42:14 s/devananda/devkulkarni/ 16:42:25 sorry devananda (again) 16:42:44 so.. 16:42:54 we ask Heat for a compute instance 16:43:05 and Nova is configured to use the libvirt driver 16:43:19 and produces a Docker container 16:43:34 or potentially a KVM instance 16:43:47 depending on hints passed through 16:43:59 so that is very convenient for us 16:43:59 devkulkarni, adrian_otto: the image type being set to 'container' would tell the scheduler to use a nova-compute host in an AZ that supports that container type 16:44:21 that looks nice 16:44:23 so yeah, we should get that for free 16:44:33 paulczar: yes, from the appropriate compute aggregate 16:44:43 so the hint is actually in the form of glance metadata 16:45:03 so, what to do for m1 ? stick to docker driver for speed of instantiation, swap to kvm for compat ? 16:45:11 paulczar: I see. And we don't need to worry about placement of containers because Heat + Autoscale will take care of that for Solum. 16:45:47 orrrr, switch to LXC and get the instantiation speed, but lose the unionfs ( which only affects the time taken to copy the image to the nova host ) 16:46:01 deployment speed is important to reduce the CI/CD cycle time, so we want to use the Docker driver until we have a working libvirt driver option to substitute there. 16:46:11 adrian_otto: do you have some links (reviews, bps..) to share about the libvirt container support ? Maybe we could help them 16:46:29 right. we need to get at least two bugs pushed through to use docker for m1 16:46:32 -1 to switching to LXC (since we already have Docker based builds) 16:46:33 julienvey: not yet, but I would be happy to track them down 16:47:25 I talked to the Docker team recently, their current focus is on getting the CI working in nova, that's why there is not much work on the driver 16:47:29 #action adrian_otto to locate Nova blueprints/reviews on libvirt driver support for Docker 16:47:54 Eric Windisch is leading that 16:48:01 julienvey: I got the same story talking with Eric 16:48:04 is the Docker team also working on the libvirt option? 16:48:22 devkulkarni: I will check with them to confirm 16:48:24 they're using my dockenstack project as a base which is pretty cool. 16:48:39 paulczar: link please :) 16:48:47 I do know that RAX is contributing to that effort 16:48:58 adrian_otto: awesome. thanks! 16:49:15 we have two blocking bugs in oslo and nova that will stop us from using docker driver 16:49:36 paulczar: each has open reviews in Nova, right? 16:49:38 I have patched oslo, and fixed the patch for nova 16:49:39 yes 16:49:47 https://review.openstack.org/#/c/78642/ 16:49:51 https://review.openstack.org/#/c/65393/ 16:50:18 and this is a nice to have - https://review.openstack.org/#/c/78322/ 16:50:57 devkulkarni: https://github.com/ewindisch/dockenstack 16:51:30 heh, sudden flurry of +1's :) 16:51:48 paulczar: cool.. thanks! 16:52:28 paulczar: feel free to add me to any such reviews as above... (https://review.openstack.org/#/c/78322/) 16:52:58 hardcoding a specific username in https://review.openstack.org/78322 is likely to raises concerns 16:53:24 specific username ? 16:53:25 if other solutions are possible, we should explore those 16:53:39 the patch adds "samalba/registry" as a default 16:53:42 paulczar: ewindisch 16:53:53 adrian_otto: that's just the name of the trusted build 16:54:27 ewindisch: thx 16:54:31 paulczar / adrian_otto: that's off the master branch. What is used right now is the cut release. 16:54:44 if there is a bug in the release we should cut a new one 16:55:32 ewindisch: I gave up asking for that! ( before you started ) ;) 16:55:33 ok, seeing something named after a person jumped out at me 16:56:18 https://review.openstack.org/#/c/78642/ wants a unit test … does a real dev want to help out with that ? 16:56:46 paulczar: sure 16:57:25 paulczar: thanks for raising those reviews to our attention 16:58:04 np 16:58:16 paulczar: Doug Hellman asked for a unit test on https://review.openstack.org/78642 16:58:37 adrian_otto: julienvey just offered to help with that 16:58:49 lagged. thanks! 16:59:27 anyone else that would like to be recorded in the attendance, please chime in 16:59:30 with those merged in I think we'll be good to M1 on docker 16:59:33 as long as we're linking reviews... I've got a pile for Docker stuff: https://review.openstack.org/#/q/status:open+project:openstack/nova+owner:ewindisch,n,z 16:59:53 ewindisch: thank you! 17:00:10 thanks everyone 17:00:15 #endmeeting