18:03:01 #startmeeting networking_policy 18:03:02 Meeting started Thu Aug 21 18:03:01 2014 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:03:03 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:03:05 The meeting name has been set to 'networking_policy' 18:03:11 rkukura: no, just in time :-) 18:03:21 SumitNaiksatam: Thanks to your reminder! 18:03:40 #info reminder - the feature proposal freeze deadline is today 18:04:03 SumitNaiksatam, eod PST? 18:04:05 anything that we had planned to post and has not yet made it? 18:04:43 hemanthravi: ATE (anywhere on earth) #link http://en.wikipedia.org/wiki/Anywhere_on_Earth ;-) 18:04:53 ATE -> AOE 18:05:20 hemanthravi: so a little more time than PST :-) 18:05:40 perhaps good to plan around PST as the deadline though ;-P 18:05:41 Hello 18:05:53 s3wong: hi, glad you could make it 18:05:54 :) 18:06:16 ok any more thoughts/questions on today’s deadline? 18:06:28 #info agenda: https://wiki.openstack.org/wiki/Meetings/Neutron_Group_Policy 18:06:55 #topic Discussion on the path forward 18:07:43 so we are currently posting the patches for review into main neutron 18:08:05 and we should continue to do that until an explicit change in plans is noted 18:08:39 we are all aware of the discussions around incubator/labs and potentially moving this feature to that 18:08:50 however, that is still in discussion 18:09:10 we can hopefully get closure on that soon 18:09:38 any thoughts/comments or any additional information anyone has on this that they would like to share? 18:10:45 oh before i forget (or get bumped out) - 18:11:08 #chairs rkukura hemanthravi LouisF songole 18:11:50 no questions on this topic? :-) 18:12:09 SumitNaiksatam, waiting for the decision on the path forward 18:12:50 hemanthravi: sure, i definitely feel your frustration and anxiety, and i hope that the decision is something that we collectively take 18:13:05 hence wanted to have the discussion here 18:13:24 we really need to get discussion of this neutron-incubator proposal going, here, and on openstack-dev 18:13:36 rkukura: +1 18:14:09 rkukura: any specific topics that you would want to bring up here? 18:14:20 I think silence is being interpretted as agreement with the details of the current proposal 18:14:33 I had couple of concerns reg how the incubator code will be consumed... 18:14:41 added this as comments in the etherpad 18:14:45 hemanthravi: Thats my main concern as well 18:15:07 Is the etherpad still actively being use for this? 18:15:45 rkukura, not sure but I see some comments posted a few days back 18:15:52 rkukura: hemanthravi: yeah i am not sure why the discussion was moved from the etherpad to the wiki page, even before the etherpad was opened up in the ML for discussion 18:16:09 hemanthravi: Havn’t looked at the etherpad since the wiki pages was created 18:17:11 i believe also that a path forward here is to actually try this out (creating the repo, to posting the patches, to building the packages) and then factor that experience into any process or policies we formalize on 18:17:25 The current proposal doesn’t seem to address making specific features within the repo consumable by OpenStack distributions. 18:18:04 SumitNaiksatam: link to the wiki page? 18:18:08 https://wiki.openstack.org/wiki/Network/Incubator 18:18:14 is this the wiki page? 18:18:23 #link https://wiki.openstack.org/wiki/Network/Incubator 18:18:53 songole: hemanthravi ^^^ 18:19:06 thanks 18:20:01 ok moving on then 18:20:17 #topic patches currently in review 18:20:22 i have not seen any heat patches 18:20:30 #link https://wiki.openstack.org/wiki/Meetings/Neutron_Group_Policy/Patches 18:20:51 LouisF: do heat patches have the same FPF deadline as Neutron? 18:21:19 SumitNaiksatam: dont know 18:21:25 LouisF: okay 18:21:56 my understanding was that songole and hemanthravi had also signed up for this earlier 18:22:02 LouisF, SumitNaiksatam I think susaant posted the heat patches will check with him 18:22:17 hemanthravi: thx 18:22:17 hemanthravi: my bad, yes Susaant posted a patch 18:22:27 hemanthravi: can you please inform Susaant to update the wiki page 18:22:30 and generally they can lag the feature patch 18:22:41 SumitNaiksatam, will do 18:23:03 LouisF: this was done more than a week back and right after the client patches in place 18:23:09 it kind of slipped my mind 18:23:31 hello 18:23:34 sorry I am late 18:23:38 s3wong: welcome back 18:23:43 s3wong: hello 18:23:50 struggled to get onto the network at Big Switch :-) 18:24:17 heat patches: https://review.openstack.org/#/c/111417 & 111419 18:24:26 songole: great thanks 18:24:37 songole: thx 18:24:39 songole: Can you add it to https://wiki.openstack.org/wiki/Meetings/Neutron_Group_Policy/Patches 18:24:45 heat patches #link https://review.openstack.org/#/c/111417 https://review.openstack.org/#/c/111419 18:24:48 songole: (Thanks) 18:25:04 mandeep: yes, I will. 18:25:19 LouisF: please review and follow up with Susaant on the missing pieces :-) 18:25:30 LouisF: i know you wanted to get involved with this from the beginning 18:26:05 #topic Resource Model/API/DB/Plugin Update 18:26:26 some last week i rebased the whole series in Neutron, add the db migration, etc 18:26:38 SumitNaiksatam: i will review 18:26:39 at that point i think all patches had passed jenkins 18:26:48 LouisF: thanks 18:27:06 point being that you should consider the series to be fairly current 18:27:14 SumitNaiksatam: +1 18:27:23 we dont have the name changes in the series yet 18:27:59 but we first want to make sure that the downstream patches like s3wong’s are not affected immediately 18:28:22 did not want to disrupt s3wong’s work prior to aug 21st 18:28:58 and of course the old story is that the -2 is still present on the first patch (but i dont think that excites anyone anymore!) 18:29:13 #topic Mapping Model/Driver Update 18:29:21 rkukura: anything new happening here? 18:29:40 i did rebase those patches as well 18:29:50 Nothing new here - I’ve been working on trying to get an ML2 BP implemented for today’s FPF. 18:29:59 rkukura: ok thanks 18:30:07 #topic Security Groups mapping update 18:30:09 s3wong: ? 18:30:12 Yes 18:30:15 SumitNaiksatam: Thanks for doing the rebasing! I think there may be one or two comments I’ll want to incorporate. 18:30:28 So I have been trying to get db migration going for my patch 18:30:29 rkukura: sure, i just did a mechanical rebase 18:30:48 unfortunately my patch was prior to Ronak's fix on --mysql-server option, thus it fails 18:30:50 s3wong: ok 18:31:05 is ronak here? 18:31:22 and thus I was trying to rebase to SumitNaiksatam 's latest Contract patch, but couldn't get that to work neither 18:31:49 (git review -d 103789 keeps on trying to get revision 1 instead of revision 7) 18:31:51 s3wong: i will work with you on that after this 18:32:05 s3wong: i need to understand the steps that you were following 18:32:16 s3wong: lets keep that issue aside 18:32:22 s3wong: are you set otherwise? 18:32:31 and the last resort I attempted was to just get latest neutron on devstack, checkout SumitNaiksatam 's contract revision 7, and manually patch my stuff in to get db migration to work 18:32:42 but for that I got "Database not up to date" 18:32:47 s3wong: ok 18:32:53 so exhausted all options at this moment :-) 18:33:05 s3wong: but are you able to run UTs in your setup? 18:33:15 SumitNaiksatam: yes, UTs are fine 18:33:19 SumitNaiksatam: 18:33:22 SumitNaiksatam, when you get some time can you send out an e-mail for the list of steps for the rebase.. our driver patch ended up posting on all the ancestors 18:33:24 s3wong: ok good 18:33:41 a week ago 18:33:42 SumitNaiksatam: but can't get patch off of WIP unless I get DB migration 18:33:43 hemanthravi: that cannot be avoided (i dont think you did anything wrong) 18:34:00 s3wong: ok we will work on that after this meeting, sound okay? 18:34:02 SumitNaiksatam, :( 18:34:12 hemanthravi: but we will discuss that as well 18:34:33 SumitNaiksatam: also, my current patch does NOT have rule/classifier/actions update support yet 18:35:12 s3wong: you mean only “update”, right? 18:35:54 well, creating rules and attaching to contracts --- in that sequence --- works 18:36:05 s3wong: ok 18:36:11 SumitNaiksatam: but I don't have mapping of policy-rule => SG rules 18:36:34 s3wong: i am a little confused 18:36:47 s3wong: so what are you mapping to what? 18:36:55 SumitNaiksatam: thus can't update / delete individual ones, since the mapping driver doesn't maintain the mapping 18:37:19 s3wong: so you mean you are not persisting the mapping association? 18:37:49 SumitNaiksatam: when you create classifier/actions -> rules and associate with contract, then create contract with those, that sequence works as contract has reference for me to create those entries for SG 18:38:14 SumitNaiksatam: but since I don't have the mapping of individual rules -> SG rules, update doesn't work 18:38:22 banix: welcome, thought you were going to be MIA today ;-) 18:38:44 SumitNaiksatam: I am :) 6000 miles or so sawy 18:38:52 away 18:39:12 s3wong: okay maybe this is a longer discussion for the lenght of this meeting 18:39:23 SumitNaiksatam: sure 18:39:31 s3wong: lets sync up on this after we get to resolving your rebase issues 18:39:39 banix: europe? 18:39:41 SumitNaiksatam: sure 18:40:01 #topic API Intercept 18:40:05 kevinbenton: there? 18:40:15 kevinbenton: were you able to sync up with rkukura? 18:40:16 SumitNaiksatam: hi 18:40:31 SumitNaiksatam: 6000 miles from New York. Tehran, Iran to be more specific. Visiting my parents. 18:40:31 SumitNaiksatam: yeah, we just decided to lift the WIP to get feedback on it 18:40:50 banix: oh wow! 18:41:15 kevinbenton: ok nice, thanks for staying on message on this, and ploughing through 18:41:22 kevinbenton: +1 18:41:38 SumitNaiksatam: no prob 18:41:47 rkukura: so you are okay with the current state of the patch? 18:41:57 #link https://review.openstack.org/#/c/109901 18:42:06 Its been a long time since I looked at it. Need to re-review. 18:42:17 rkukura: ok 18:42:23 But it is in by the FPF! 18:42:43 rkukura: yes, thanks to kevinbenton, yay! 18:42:53 #topic CLI/Client Update 18:42:58 songole: hi 18:43:07 any new updates? 18:43:19 you got some feedback last week 18:43:23 how is that coming along? 18:43:32 Naming change? 18:44:09 yeah, but that hasn’t made it to the API 18:44:19 I would wait for plugin changes 18:44:21 but things like the “gp-“ prefix 18:44:58 prefix change I haven't incorporated. I would all the changes at once 18:45:09 songole: okay makes sense 18:45:31 Magesh is adding unit tests 18:45:33 #topic Tempest tests 18:45:43 songole: nice, thanks for the update 18:45:54 i dont have any updates on the tempest tests 18:46:11 will add more API tests at the earliest 18:46:24 #topic Vendor/Open Source Controller Drivers 18:46:42 so at this point i know one convergence has posted their driver 18:47:08 will do a rebase and remove the wip today 18:47:09 hemanthravi songole: nice work you guys in at least meeting the FPF 18:47:17 hemanthravi: great 18:47:28 anyone else want to discuss their drivers? 18:47:49 i am guessing that the incubator discussion sucked the momentum out of this effort a little bit 18:48:11 We are going to wait and see how the incubator discussion pans out 18:48:17 banix: ok sure 18:48:36 great to see the driver by hemanthravi though! 18:48:42 i think the code will be useful regardless of where it lands 18:49:10 #topic Horizon update 18:49:31 not sure if ronak is here 18:49:54 but i think he updated the patch a week ago 18:50:08 and i think abishek was reviewing it from the horizon side 18:50:20 #link https://review.openstack.org/#/c/93590/ 18:50:32 LouisF: did you get a chance to take a peek at this? 18:51:24 perhaps not 18:51:28 #topic open discussion 18:51:37 we might actually finish a little earlier today 18:51:50 anything anyone else wants to discuss? 18:52:09 SumitNaiksatam: yes 18:52:12 otherwise i can go back to s3wong and work with him on the rebase issue 18:52:14 LouisF: thanks 18:52:47 I’m curious, if we go with the incubator option, whether we would want to base our development on the juno version of the neutron repo for a while, rather than tracking master? Would that enable vendors to complete their drivers and make them available to early adopters using the juno release? 18:53:14 rkukura: that makes sense to me 18:53:20 rkukura, +1 18:53:27 rkukura: however would that mean that we would have to wait for juno to release? 18:53:31 rkukura: I agree, that is very helpful 18:53:32 rkukura: i hope not 18:53:52 SumitNaiksatam: I mean once Juno is tagged, and open for Kilo development 18:53:54 can we base it off J3, for instance 18:53:54 well that seems the logical way of doing this 18:54:11 SumitNaiksatam: I assume that by "tracking Juno" you imply "tracking Juno Milestones" 18:54:13 rkukura: yes, but that seems a long way away 18:54:38 mandeep: i did not mention about tracking Juno 18:54:46 mandeep: i think rkukura mentioned about tracking master 18:54:49 If this is what makes most sense, we need to influence the neutron-incubator proposal to support this sort of thing. 18:55:18 rkukura: so lets start we had to start on the incubator/labs today 18:55:24 SumitNaiksatam: Ooops ;-) 18:55:31 I think vendors will have different approaches, some try to follow all cycle milestones some stick to the final release 18:55:34 rkukura: we would still need to base it off the master, right? 18:55:36 We could also do ongoing development on labs master, tracking neutron master, and back-port to stable-juno 18:55:44 rkukura: we dont have J release tagged yet 18:56:30 I’m making the assumption we will have something merged to the labs repo before juno ships, but it will need more work 18:57:07 rkukura: SumitNaiksatam: is the labs package going to be released the same time as Juno release? 18:57:28 So does consuming that ongoing work mean using neutron-master post-juno, or can it be consumed as an add-on to juno neutron? 18:57:37 s3wong: i think that is still worked on 18:57:54 rkukura: good point 18:58:00 If this is something we care about, we need to get discussion going on the openstack-dev ML 18:58:05 rkukura: that would depend on what strategy we adopt 18:58:25 rkukura: so we will need to start with the dependency on the master 18:58:53 rkukura: but i think per your suggestion, we eventually get to a point where base the development on the labs off the milestones 18:58:54 rkukura: Yes, I agree 18:59:05 rkukura: i think you are explicitly pinning that to the stable branch 18:59:14 rkukura: (on starting a larger discussion quickly) 18:59:28 rkukura: i am wondering if this should be more granular, and we can base that of the milestone releases as well (like K1) 18:59:59 I’m just trying to get everyone thinking about the options for how we’d go forward with the labs repo, how we’d deliver stuff, and what might need changing in the current proposal. 19:00:09 rkukura: so in that sense the labs releases can lag the neutron master releases by one milestone 19:00:45 rkukura: and i am not sure that anyone should agree to any one proposal as is 19:00:59 If the goal is getting beta versions of new features into the hands of early adopters that are using packaged juno releases, lagging might make sense 19:01:06 rkukura: i believe this is going to be a learning and evolving experience (and we would need to factor that feedback in) 19:01:38 true to our tradition, we have gone over now :-) 19:01:45 lets wrap it here 19:01:57 thanks everyone 19:01:58 I’d hate to see the proposal adopted with its focus on PyPI as the release mechanism, supporting multiple neutron versions simultaneously (without branching), etc. 19:02:05 rkukura: +1 19:02:14 thanks SumitNaiksatam! 19:02:39 agree but that discussion seems to be happning outside the ML I am afraid …. 19:02:52 rkukura: my suggestion would be that we should treat this as an evolving process and formulate the policies as we go along 19:03:06 banix: I’m not sure the discussion is really happening at all ;) 19:03:18 but we need to get started, else we are discussing in a vaccuum 19:03:20 banix: Yes, let us follow up on that issues on the ML thread 19:03:26 SumitNaiksatam: +1 19:03:37 rkukura: i wont be surprised if that is the case. 19:03:58 ok thanks all 19:04:07 sorry for logging in in the middle of the call. talk to you all later. 19:04:08 #endmeeting