16:01:07 #startmeeting Solum Team Meeting 16:01:07 Meeting ended Tue Dec 17 16:01:04 2013 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) 16:01:08 Minutes: http://eavesdrop.openstack.org/meetings/marconi/2013/marconi.2013-12-17-15.11.html 16:01:09 Minutes (text): http://eavesdrop.openstack.org/meetings/marconi/2013/marconi.2013-12-17-15.11.txt 16:01:11 Log: http://eavesdrop.openstack.org/meetings/marconi/2013/marconi.2013-12-17-15.11.log.html 16:01:12 Meeting started Tue Dec 17 16:01:07 2013 UTC and is due to finish in 60 minutes. The chair is adrian_otto. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:13 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:15 The meeting name has been set to 'solum_team_meeting' 16:01:15 welcome everyone 16:01:20 alcabrera: can you snag those minutes? 16:01:24 kgriffs: will do 16:01:27 #topic Roll Call 16:01:32 Chris Alfonso 16:01:34 Swapnil 16:01:36 tom blankenship 16:01:36 Kurt G. o/ 16:01:36 Adrian Otto 16:01:37 Georgy Okrokvertskhov 16:01:42 Noorul 16:01:43 Paul Montgomery 16:02:03 Devdatta 16:03:29 ok, while we are waiting for a few more participants to join, I'd like you to think about meeting schedule for Dec 24 and Dec 31. 16:03:58 I will be on vacation those days, and think those are good days to cancel 16:04:04 as many of us will be out 16:04:34 rajdeep joined 16:05:28 murali 16:05:33 I would still like to arrange an update that we can send to the ML for those of us who will continue working. You could possibly email all your updates, and I could post them to the ML in a single status update, or we could post status only to the Wiki. How do you all feel about that? 16:06:01 #link https://wiki.openstack.org/wiki/Meetings/Solum Our Agenda 16:06:33 ^^ perhaps we could actually put our status there in lieu of holding meetings on those dates? 16:06:42 +1 16:07:22 any objections to this approach? I know a gap of 3 weeks is pretty long, so I want to be sure you are all happy with this. 16:07:35 +1 to wiki 16:07:53 adrian_otto: +1 on canceling the meetings but posting updates/status 16:07:59 +1 16:08:02 +1 16:08:15 topic #Announcements 16:08:58 #agreed meetings for 2013-12-24 and 2013-12-31 will not be held by IRC, but status will be posted by each stakeholder at https://wiki.openstack.org/wiki/Meetings/Solum 16:09:28 Adrian will be on vacation Dec 23-Jan 8 16:09:41 travelling internationally, reading email. 16:09:58 any announcements other team members would like to make? 16:10:36 ok, if you think of anything we can address them in the Open DIscussion section 16:10:39 #topic Review Blueprints 16:10:49 before I dive in here 16:11:09 I expect the last one on this list will require some meaningful time to discuss 16:11:24 most of our other topics probably have updates that are pretty short 16:11:38 #link https://launchpad.net/solum/+milestone/milestone-1 Blueprints for milestone-1 16:12:01 note that I do plan to split the logging blueprint, which paulmo will mention in a moment 16:12:27 a more narrow scope will be targeted for milestone-1 and a wider scope will be targeted to a future milestone 16:12:31 (I added notes to the logging bp about that) 16:12:37 here we go... 16:12:38 #link https://blueprints.launchpad.net/solum/+spec/api Solum API (aotto) 16:12:45 #link https://wiki.openstack.org/wiki/Solum/API was edited to include plan resources, and eliminate camelCase 16:13:04 note that there is a difference between plan *files* and plan resources 16:13:24 I see that we have documents scattered on etherpad and Wiki 16:13:24 the wiki page above calls out that difference. I will propose a separate blueprint/spec for the plan files 16:13:40 Whiteboard list is huge. 16:13:43 yes, the etherpads will be consolidated into wiki pages 16:14:05 specifically: 16:14:06 #link https://etherpad.openstack.org/p/solum-api-spec-review will yield new blueprints for PATCH, pagination, resource query/parameters, etc 16:14:14 #link https://etherpad.openstack.org/p/solum-demystified will yield a new blueprint for plan files 16:14:49 At the end of this splitting I think we will get a better idea of the big picture from end to end perspective 16:16:37 whoops, network glitch 16:17:12 devkulkarni: did I miss anything after your remark? 16:17:36 Nothing else was typed 16:17:44 ok, next BP then 16:17:46 #link https://blueprints.launchpad.net/solum/+spec/solum-minimal-cli Command Line Interface for Solum (devdatta-kulkarni) 16:17:48 no, there was a network glitch effect on this side as well.. 16:18:03 okay, so there is not much update on this. 16:18:13 I will be getting together with Noorul to understand his PR 16:18:17 perhaps having a few sequence diagrams in addition to demystified doc would help 16:18:19 and take it from there 16:18:37 ok, thanks 16:18:45 #link https://blueprints.launchpad.net/solum/+spec/solum-git-pull Pull integration of Solum from an external Git repo (kraman) 16:19:25 If kraman is not around, I can give an update from the git-integration working group meeting 16:19:48 Okay, here goes.. 16:19:54 funzo: are we expecting kraman? 16:20:12 adrian_otto: I haven't talked to him. I'll see if I can get him 16:20:18 devkulkarni: ok, proceed, and if kraman arrives, he can add to your update. 16:20:23 So the main discussions were whether Zuul can be used for all of Solum's workflow needs.. 16:20:59 kraman and mordered have done some work with regards to understanding Zuul. 16:21:19 adrian_otto: on his way 16:21:26 funzo: tx! 16:21:57 We have agreed to using zuul for git integration (pull and push) 16:21:57 but zuul may not be a right fit for deploying DU workflow 16:22:03 after a POC around zuul we will be able to know for sure though. 16:22:20 that was about it from the meeting 16:22:25 ok, that's promising! 16:22:26 adrian_otto: he's here 16:22:39 hi kraman1 16:22:44 hi 16:22:54 we are doing an update on solum-git-pull 16:23:05 i had a conversation with monty after the git meeting as well 16:23:09 devkulkarni has summarized about the Zuul discussion 16:23:17 we discussed in detail about how zuul will be used 16:23:27 and I will be taking to rest of working group tomorrow 16:23:41 we should have a clear picture about how to proceed after tomorrow atleast for milestone 1 16:23:49 devkulkarni: thanks for summarizing 16:24:01 kraman1: cool 16:24:05 kraman1 and devkulkarni: ok, anything more you want to mention today? 16:24:19 adrian_otto: no, that was it for today. 16:24:27 do we have any code review posted for anything? 16:24:31 git related 16:24:48 or anyone planning to start code soon? 16:24:51 adrian_otto: no code yet. after tomorrow's meeting. will start developing a prototype 16:24:57 ok, tx 16:25:02 next BP... 16:25:08 #link https://blueprints.launchpad.net/solum/+spec/user-authentication User authentication for incoming requests (gokrokvertskhov) 16:25:18 #link https://review.openstack.org/58811 16:25:38 Its done. Just blocked by Nooril's tests patch. 16:25:52 This looks good to merge after adding back test_functional.py I guess 16:25:56 gokrokve: What do you mean by blocked? 16:26:01 ok, please link us to that so we can take a quick peek, please? 16:26:10 You put -1 for some reason. 16:26:18 There is a valid reason 16:26:23 one question on keystone integration 16:26:33 any special roles for solum users? 16:26:52 Nothing special yet for keystone. 16:27:07 Just validation of supplied token. 16:27:08 I think all you have to do is to remove test_functional.py deletion from your patch 16:27:32 ok thanks 16:28:18 rajdeep: we expect to have an RBAC setup for future milestones 16:28:23 noorul: If I remove it I will have to wait for your patch. Which means I cant merge it before your patch. So its blocked. 16:28:33 gokrokve: you set on the direction for overcoming the remaining −1 on the 58811 patch? 16:28:34 That is fine 16:28:53 it sounds to me like there is actually no dispute on it 16:28:56 gokrokve: Why do you have to wait for my patch? 16:29:16 Because test_functional will fail at the gate. 16:29:26 I removed it for a reason. 16:29:53 But I don't see the reason in commit log 16:30:33 If you take a look on my previous path you will se that I rewrote test_functional to work in functional tests. You did the saeme. 16:30:48 Ok. I will add this info to commit log. 16:31:19 ok, so noorul and gokrokve, are you willing to regroup after our meeting (at some mutually convenient time) and hash this out further so we can get to a nice merge point? 16:31:33 Sure. 16:31:46 it sounds like a bit of working together can get us through the last bit of this. 16:31:48 We just need to sync our patches to merge them in proper order. 16:31:49 thanks guys. 16:32:00 #link https://blueprints.launchpad.net/solum/+spec/specify-lang-pack Specify the language pack to be used for app deploy (devdatta-kulkarni) 16:32:20 I have a WIP for this. 16:32:28 #link https://review.openstack.org/62548 16:32:33 ^^ the WIP 16:32:47 Clayton had some comments about the attributes of a LP 16:32:50 Will be syncing up with him today/tomorrow 16:33:05 The WIP will change based on comments to Murali's WIP 16:33:29 Also, I need to integrate with the nova object model from Clayton's recently merged patch 16:34:14 ok, any more on this one? 16:34:16 Angus mentined that datamodel is not versioned. 16:34:25 Thats pretty much it 16:34:27 I did not understand that remark 16:34:34 devkulkarni: lets work together on that. i need to integrate with nova objects too. 16:34:41 gokrokve: yep. saw that comment on muralia's wip 16:34:46 muralia: +1 16:35:07 the datamodel was moved out of api/v1 16:35:20 in my WIP. we need to move that back under v1 16:35:21 will work with muralia on nova objects and datamodel versioning changes 16:35:38 for versioning. 16:35:41 Ok. 16:35:50 muralia: yep 16:35:55 oh, don't we want it to be versioned 16:36:07 ok 16:36:09 yes we do. 16:36:13 ready for the next one? 16:36:17 Why do we need additional handlers and dispatchers for services? 16:36:17 we do want it to be versioned. 16:36:34 It looks overcomplicated. 16:36:39 although, we also want to separate the wsme objects from Solum 16:36:44 internal objects 16:36:57 gokrokve: we will discuss the api worker blueprint in just a moment 16:37:07 ok 16:37:12 so let's revisit that in just a moment 16:37:15 #link https://blueprints.launchpad.net/solum/+spec/logging Logging Architecture (paulmo) 16:37:35 adrian_otto: yes, ready for the next one 16:37:40 I am proposing that we split the logging requirements into an M1 blueprint and a potential future >=M2 blueprint 16:37:45 devkulkarni: tx 16:38:11 As long as everyone agrees to the following logging rules (that we continue to follow as Clayton already has examples of): https://wiki.openstack.org/wiki/Solum/Logging 16:38:13 paulmo: do you want to split it, or would you like me to? 16:38:30 I see you made a comment in the whiteboard 16:38:44 I can (if we need). The reason for the split is to not hold up M1 and to give me some time to track down the potential changes to Oslo log that I keep hearing rumors of. I want to sync and not duplicate work. 16:39:05 seems reasonable to me 16:39:26 The final part is that we will need unique user identification in the logs (tenant/project ID, etc). Do we want to shoot for M1 for that since RBAC and database isn't fully there for a while? 16:39:28 ok, let me know if you'd like my assistance at all 16:39:57 next one 16:39:57 #link https://blueprints.launchpad.net/solum/+spec/api-worker-architecture API service/worker architecture for async operation (murali-allada) 16:40:15 did everyone get a chance to take a look at my WIP. https://review.openstack.org/#/c/62466/ 16:40:23 if not, please do and give me your feedback. 16:40:43 The basic idea is to get us to a point where api requests are reading from the DB 16:40:58 I've already got some nice comments 16:41:10 one interesting comment from angus 16:41:14 Any specific files that we should look at for Clayton's examples? 16:41:32 Does WIP really gets reviewed? 16:41:48 WIPs should get reviewed. 16:41:56 to comment on the general direction of the work 16:42:15 devkulkarni: the handlers is where we would integrate with claytons work 16:42:43 one comment I got from angus was that there is a lot of indirection in this code. with no value. 16:43:11 i agree that we dont see it right now, but theres a reason to break it up into a controller/dispatcher/handler model. 16:43:14 https://review.openstack.org/#/c/62466/ 16:43:18 That has -1 from Jenkins 16:43:50 muralia: thanks. I was actually asking for specific files to look at for logging examples. 16:43:51 right now all method calls are just pass throughs, but we have place holders for adding queues 16:44:03 authentication and db integration 16:44:36 devkulkarni: let me look at some files and give you some pointers. 16:44:42 we do need decoupling and separation between controllers and things that will actually query db. 16:45:10 The "service" layer would be the place where we do all the logic, if any. 16:45:36 So at a minimum I do see the following arch: controllers -> services -> db 16:45:37 devkulkarni: what sort of coupling are you imagining? 16:45:37 noorul: yes, jenkins −1'd is. I'm still working on the code, but wanted to put it out there as devkulkarni was working on some work that needs this. 16:45:47 devkulkarni: compositional style calls? 16:46:04 Command pattern style services? 16:46:20 muralia: did you follow up with asalkeld about his remarks? 16:46:34 claytonc: just separation of concerns actually. controllers should not do any more logic than basic error checking 16:46:44 adrian_otto: not yet. i plan to do that today 16:46:51 I encourage you to chat with him in #solum and fully explore the concern. 16:47:00 yup. 16:47:06 and reach some consensus. Thanks. 16:47:08 muralia: If you can split this into multiple patches then it will be easy for reviewers 16:47:10 devkulkarni: for simple CRUD it's probably reasonable we should be doing a bare minimum of work in controller, and let the object model handle the details 16:47:16 for more complex flows completely agree 16:47:19 noorul: agreed. 16:47:24 for an abstraction around "process" 16:47:35 claytonc: either compositional or command/control.. don't know which pattern is the best for us yet. 16:48:01 claytonc: was about to say that.. based on your merged work, we need to revisit this 16:48:23 devkulkarni: +1 16:48:49 ok 16:48:52 #topic Open Discussion 16:49:01 I have one thing for this 16:49:12 I have 3, hopefully quick items whenever I can fit in 16:49:23 Gate job for pypy fails on singledispatch. noorul, you mentioned we contacted an upstream somehow about this? 16:49:53 adrian_otto: I meant, I submitted an issue in the tracker 16:50:26 * noorul is searching 16:50:37 general question - what constitutes a working group, and how does one become a part of one? 16:50:40 https://bitbucket.org/ambv/singledispatch/issue/1/failed-to-install-on-pypy-202 16:52:30 I got dropped off the network again, sorry 16:52:41 adrian_otto: https://bitbucket.org/ambv/singledispatch/issue/1/failed-to-install-on-pypy-202 16:52:50 noorul: thanks!! 16:52:51 briancline: just attend the meetings 16:52:55 they're discussed on the public ML 16:53:11 it's a subgroup of folks working to hammer out specific details for blueprints 16:53:13 briancline: and you can joing the blueprints for update notifications 16:54:06 The working group meeting times are on the Wiki, right adrian_otto? If not, we should add them. 16:54:48 devkulkarni: yes, on the Breakout Meetings wiki page 16:55:18 paulmo: did you get in everything you wanted to mention? 16:55:28 I had a few items 16:55:50 First, I just wanted to link the new security requirements page that I've been working on (and continue to do so): https://wiki.openstack.org/wiki/Solum/SecurityRequirements 16:55:53 #link https://wiki.openstack.org/wiki/Solum/BreakoutMeetings meeting schedule 16:56:01 ^^ devkulkarni 16:56:09 I'll likely be reviewing using that as a source in the future. 16:56:29 paulmo: that's a really impressive wiki page 16:56:33 2nd, is it possible to get an agreed on logging requirements? https://wiki.openstack.org/wiki/Solum/Logging 16:57:02 Thanks, on the first page there, if you want to review and link a specific topic, use the links on the first column like this: https://wiki.openstack.org/wiki/Solum/SecurityRequirements#git_pull_regex 16:58:13 Just for clarity, relinking the 2nd wiki, the logging requirements: https://wiki.openstack.org/wiki/Solum/Logging 16:58:27 Have we discussed this enough to get an agreed 16:58:29 ? 16:58:38 I like the Solum/Logging requirements 16:58:42 anyone disagree? 16:59:17 I think folks will need time to go through these before voting to agree 16:59:18 (last week, nobody disagreed so trying my luck a little farther this week) :) 16:59:18 Looks great. 16:59:36 quick question..anything we are leveraging from oslo for logging 16:59:45 Yes, all Oslo log 16:59:47 ok, paulmo we can revisit that at our next IRC meeting 16:59:52 Thanks! 17:00:07 ok,thanks 17:00:10 next irc meeting is 2014-01-14 17:00:22 please put updates on wiki page between now and then, thanks! 17:00:24 #endmeeting