18:32:57 #startmeeting Networking FWaaS 18:32:58 Meeting started Wed Nov 12 18:32:57 2014 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:32:59 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:33:01 The meeting name has been set to 'networking_fwaas' 18:33:16 probably light attendance today, people still getting back from the summit 18:33:26 SumitNaiksatam: yes seems that way 18:33:28 #topic bugs 18:34:05 glebo: hi 18:34:07 glebo: hi there 18:34:13 'lo 18:34:34 doesnt seem like anything significant has popped up the bugs front 18:34:52 we did a scrub last time, and we have a shorter list today 18:35:03 SumitNaiksatam: yes 18:35:17 #topic Docs 18:35:41 hello all 18:35:41 SridarK: any follow up on #link https://bugs.launchpad.net/openstack-manuals/+bug/1346986 18:35:45 badveli: hi! 18:35:53 hello sumit and all 18:35:54 SumitNaiksatam: not much happened last week i will do this 18:35:58 badveli: hi 18:36:06 hello sridar 18:36:14 SridarK: great, thanks 18:36:22 SridarK: i think rudrajit has unassigned himself 18:36:32 SridarK: perhaps we can add the relevant fwaas part? 18:36:44 SumitNaiksatam: ok i will do take care of that 18:37:18 SridarK: also i think we need to sync up with the DVR team if they covered the services part as well in their documentation (in the context of the FWaaS for DVR) 18:37:55 SumitNaiksatam: i will touch base with Swami & Rajeev on that 18:38:13 * glebo will brb 18:38:43 #action SridarK SumitNaiksatam to follow up with Swami/Rajeev//DVR team on FWaaS documentation in the context of DVR (and N-S firewall functionality) 18:39:20 #topic Kilo blueprints 18:39:24 hi 18:40:32 vishwanathj: hi 18:40:43 hello 18:40:45 vishwanathj: hi 18:40:47 * glebo bac 18:40:56 badveli: you are following up on the service groups/objects spec 18:41:06 yes 18:41:13 badveli: link? 18:41:15 i had some comments and 18:41:47 I have reviewed it privately w/ bauzas, and will add those comments publicly now for others to see 18:41:50 let me post it, we are already in the review 18:41:54 #link https://review.openstack.org/#/c/131596/ 18:42:18 s/ with bauzas/with badveli 18:42:20 only one thing i was not sure was tempest 18:42:35 one of the comment was add a tempest spec 18:43:07 badveli: you had already posted a tempest patch, right? 18:43:21 yes sumit, but they are asking a spec 18:43:27 tempest spec 18:43:36 badveli: ok, perhaps a new requirement 18:43:51 yes, i am trying to check which one needs the tempest spec 18:44:08 and when do we need it 18:44:11 have we id'd who are the MUST have reviewers in order for approval to pass by PTL? 18:44:23 badveli: tempest spec i guess will not get approved until the neutron implementation patch gets approved 18:45:15 glebo: badveli: you can add “neutron-core” to the reviewer’s list 18:45:21 this will add all the neutron cores 18:45:32 ok, i will add it 18:45:38 https://github.com/openstack/qa-specs/tree/master/specs 18:45:41 SridharRamaswamy: hi 18:46:00 i am going through the existing tempest specifications, to know 18:46:31 badveli: do you think you are on top of the comments, or do we need to discuss anything here? 18:46:41 *review comments 18:46:47 i think i am 18:46:52 badveli: okay good 18:47:02 thanks for asking sumit 18:47:04 SumitNaiksatam: yes, we can, but as we learned last time, there are certain cores that matter most, i.e. PTL will not finally approve without those reviewers 18:47:14 badveli: we are not doing any significant changes to the data model correct ? 18:47:22 no 18:47:25 glebo: have you seen any documentation for this? 18:47:25 *from Juno spec 18:47:31 SumitNaiksatam: last time, Mark explained there needed to be a "Security" reviewer, and an API reviewer 18:48:13 SumitNaiksatam: nothing public, only the irc chat transcript between Mark and myself a few months back 18:48:24 glebo: i am not aware who those are, but please update the team with any information you dig up on this front 18:49:08 SumitNaiksatam: Mark said we would identify and assign them specifically once Kilo opened. 18:49:44 the second spec we wanted to put immediately was on the FWaaS insertion 18:49:53 SumitNaiksatam: Now that Kilo is open, we need to close loop on that and identify with Mark & Kyle exactly who must be in approver chain for them to approve it themselves 18:50:17 before we leave the Service Objects & Groups spec, 18:50:36 does anyone understand what needs to be done to satisfy this comment from mestery: 18:50:38 glebo: i believe this should be a documented policy if indeed it is, it should not be function of informal discussions 18:50:53 "I'm particularly concerned about the lack of details of the Open Source implementation. Given the community will be left maintaining this, I'd like more details around this before we consider merging this." 18:51:28 SumitNaiksatam: clearly, and agreed. But it's not. This is what Mark said needed to happen. And he agreed it was not documented. So I'm following that. 18:51:29 glebo: perhaps good idea to ask for explanation in the gerrit review 18:52:02 glebo: to your question on the comment left in the review 18:52:08 SumitNaiksatam: more specifically, which part of "Open Source" implementation might he be referring to? 18:52:29 glebo: i believe a similar comment was made earlier when this was -2’ed but no explanantion was provided 18:52:37 glebo: so i share your confusion 18:53:11 anyone else have any idea to what "lack of details of the Open Source implementation." may be referring to? 18:53:13 yes, there is no open source implementation as such i am thinking to explain a bit of our reference implementaions 18:53:18 glebo: i agree that gerrit is a good place to seek clarification 18:53:45 glebo: i am not sure if somehow the conclusion is being drawn that to implement this fucntionality we need something beyond iptables (in the reference implementation) 18:54:11 glebo: the patch should be covering this on our existing iptables implementation 18:54:16 SumitNaiksatam: & SridarK: ok, will try to get clarification via gerrit. Just didn't want to bug in public if we were missing something obvious 18:54:38 glebo: i agree 18:54:41 Sridar, sumit, if we explain about out refernce implementaions 18:54:42 glebo: i think we just need to clarify about the backend on the ref impl 18:54:42 glebo: which line number on the review? 18:55:29 mestery review Nov 7 09:13, general comment (i.e. not inline to the text) 18:55:31 cwe can explain more about our reference implementations 18:55:40 we 18:55:52 badveli: i think so too 18:56:15 badveli: is it clear to you what is needed to address this comment? 18:56:21 okay, so to me it shows that lack of understanding of the spec 18:56:29 yes glebo, 18:56:43 either the review was done in a hurry or may be we didnt do a good job of explaining 18:57:09 badveli: perhaps clearly mentioning that this relies on the existing iptables library might help 18:57:21 badveli: great. Then lets get the change made asap and we'll validate w/ mestery that it suffices, and if not, we'll get clarification on exactly what he wants to see 18:57:30 Yes sumit, i am not sure why do we need to explain about this 18:57:47 glebo, they also requested i file a tempest spec 18:57:54 i am working on that 18:57:59 badveli: glebo: the other thing we can also do is bring this up for discussion in next week’s neutron IRC meeting 18:58:02 badveli: yup, understood 18:58:36 it is an on demand agenda, so given the amount of time this has been in review, i think it makes sense to get some focussed attention on it 18:58:46 SumitNaiksatam: let's get a touch of text to clarify in place, ask mestery to review, and if it's not addressed, then we can bring it up in neutron. Good idea 18:59:08 ok 18:59:20 glebo: yeah, i meant it in that order; if things are resolved before the next meeting, we dont have to bring it up :-) 18:59:26 let's get the ball back into reviewers' courts asap 18:59:41 fine glebo 18:59:43 SumitNaiksatam: ack 19:00:12 SridarK: regarding the firewall insertion spec, is there anything new we need to discuss? 19:01:08 SumitNaiksatam: i am still on the wall about router_id as the proposal from Mark, did not get a chance to discuss more with him, we can take this up in gerrit 19:02:49 SumitNaiksatam: in thinking more about this and also in discussion with folks that handle networks - i am getting feedback that L2 is also a strong requirement 19:03:02 SridarK: so we decided that we will put the spec in as is (per demand by mark) and take it from there? 19:03:17 SumitNaiksatam: yes i think that is best 19:03:56 ok great 19:04:45 #topic Open discussion 19:05:04 we all met and discussed plenty in person over the last few days 19:05:11 back on fw insertion 19:05:15 anything we missed or needs follow up today? 19:05:31 did we send email to mark yet asking for clarification on L3 req? 19:05:37 that was to be step 1, 19:05:39 right? 19:05:54 * glebo behind on OS emails at present 19:06:23 glebo: no i have not sent an email yet - i am in some discussions with deployment and customer facing folks 19:07:25 Working on writing up use-cases I've learned from users, per our discussion on Monday. Will send those to the group this week. 19:07:37 if my notes are correct, SumitNaiksatam and SridarK say we already have published text on why L3 is not sufficient, right? 19:07:40 BrianTorresGil: great, thanks for following up 19:07:44 so we just need to point to those? 19:07:44 BrianTorresGil: hi & thx 19:08:00 glebo: i will send pointer to existing docs 19:08:16 BrianTorresGil: same here. 19:08:36 BrianTorresGil: let's sync offline on which cases u r doing, so we don't dup efforts 19:08:46 SumitNaiksatam: perfect 19:08:59 glebo: no harm in churning again - we want to make sure we have wide consensus 19:09:11 glebo: sure, sounds good. Though I'm interested in what we come up with separately, too. 19:09:22 BrianTorresGil: yes, i agree 19:09:38 so glebo: we also had you on the hook for the use cases ;-) 19:09:51 and arvind, but i dont think arvind is here 19:10:00 so will follow up separately 19:10:05 BrianTorresGil: yes that would be good - lets not influence each other - to take a look at this fairly independently and arrive at consensus 19:10:43 perhaps i did not articulate this perfectly - but i think u see what i mean too 19:11:02 #action BrianTorresGil glebo arvind to exchange use cases within the fwaas team 19:11:20 #action BrianTorresGil glebo arvind to exchange use cases within the fwaas team by end of the week 19:11:31 we will consolidate these based on follow up discussion and proceed from there 19:12:39 SumitNaiksatam: great. We can map out roughly what they cases are, and then fill in the detailed description following 19:13:16 glebo: perfect! 19:13:29 okay, anything more for today? 19:13:37 SumitNaiksatam: enikanorov_ reached out in Paris for someone to handle the bugs that he does a prelim triage on - i have volunteered for that 19:13:49 SridarK: nice! 19:13:50 and i will reach out to others to help 19:13:56 anyone blocked on anything, or any other pressing concerns? 19:14:20 vishwanathj: welcome to the ur first fwaas irc 19:14:30 thanks 19:14:40 looking forward to guidance from you guys 19:15:18 indeed, vishwanathj a warm welcome! 19:15:31 Thanks SumitNaiksatam 19:15:47 okay so if nothing more, on that happy note, lets get back 15 mins of our time! 19:15:54 thanks all for joining 19:15:55 bye 19:15:59 bye all 19:16:02 awesome...bye all 19:16:10 #endmeeting