18:31:43 #startmeeting Networking FWaaS 18:31:44 Meeting started Wed Jan 21 18:31:43 2015 UTC and is due to finish in 60 minutes. The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:31:45 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:31:47 hi 18:31:48 The meeting name has been set to 'networking_fwaas' 18:31:53 #info metting agenda https://wiki.openstack.org/wiki/Meetings/FWaaS#Agenda_for_Next_Meeting 18:32:04 #info Kilo-2 is Feb 5th 18:32:34 anything else anyone wants to share/announce upfront? 18:33:00 #topic Bugs 18:33:23 there was a critical bug (UTs breaking) but it was fixed and merged 18:34:02 other than that i dont see any new critical/major bugs 18:34:16 SumitNaiksatam: yes nothing critical present 18:34:18 SridarK: vishwanathj: anything show up on your radar? 18:34:23 SridarK: okay 18:34:53 badveli_: hi 18:34:59 hello sumit and all 18:35:04 badveli_: hi 18:35:08 badveli_: anything show up on your bug radar? 18:35:11 hello sridark 18:35:17 jenkins test fails for our code because we have not uploaded the L3 router plugin code 18:36:17 vishwanathj: on ur review patch ? That is fine now 18:36:24 Ok 18:36:32 vishwanathj: yeah, will come to the vendor patches in a min 18:37:42 badveli_: any progress on #link https://bugs.launchpad.net/neutron/+bug/1386543 after I last posted the message? 18:38:15 i dont see any update on #link https://bugs.launchpad.net/neutron/+bug/1335375 18:38:30 badveli_: just checking in case you happened to spend time on this 18:39:42 ok moving on from bugs 18:39:59 #topic Docs 18:40:02 i did not see any update 18:40:24 SridarK: anything on your radar on this front? 18:40:37 SumitNaiksatam: no updates on this 18:40:43 SridarK: okay 18:40:52 #topic Firewall Insertion 18:41:10 SridarK: good discussion last week 18:41:16 SridarK: any progress? 18:41:27 SumitNaiksatam: have done a first pass on the subtasks 18:41:50 SumitNaiksatam: will need to look at foreign keys from neutron db 18:41:56 (routers) 18:42:27 SumitNaiksatam: i think this should not be a problem - u have probab done this sort of thing for GBP 18:42:54 SridarK: yeah, i can help you with the FK, etc 18:42:55 SumitNaiksatam: most of the work is on the Plugin side, from what i see - the agent should not be too bad 18:43:04 SumitNaiksatam: thanks for that 18:43:13 SridarK: Let me know when you have a handle on how long it'll take to get this in, so I can coordinate refactoring for L3. 18:43:15 SridarK: feel free to post the WIP patch, we can take it from there 18:43:22 SumitNaiksatam: i am trying to see if i can get a WIP patch out next week 18:43:40 SridarK: sooner might be better :-) 18:43:42 SumitNaiksatam: u took the keys out of my keyboard :-) 18:43:55 SumitNaiksatam: will definitely try for that 18:44:01 SridarK: since kilo-2 is feb 5th 18:44:10 pc_m: yes will do so on L3 agent refactor 18:44:19 SumitNaiksatam: yes i agree 18:44:34 pc_m: thanks for bringing that up, yes, we owe you an estimate on that 18:45:00 SumitNaiksatam: there are some details on how we handle state with multiple routers now that we will track the routers 18:45:01 SridarK: so far any blocking issues? 18:45:12 With the WIP, we can probably start to look at what is needed for refactoring. 18:45:13 SridarK: okay 18:45:19 pc_m: agree 18:45:25 SumitNaiksatam: we can see if we need to do this now or it can be outside the scope of this work 18:45:42 SridarK: lets set some time aside tomorrow and white board this 18:45:53 SridarK: does that work for you? 18:45:54 pc_m: yes agreed - on the agent side it seems mostly it is removing code out 18:46:06 SumitNaiksatam: most definitely thanks for the time 18:46:18 SumitNaiksatam: will coordinate with u offline 18:46:29 SridarK: sure, others feel free to join as well 18:46:56 SumitNaiksatam: ok that were the basic things i wanted to bring up 18:47:03 *those 18:47:05 SridarK: so you havent hit any major blockers yet, right? 18:47:11 SumitNaiksatam: not yet 18:47:13 SumitNaiksatam, Please send the webex meeting if you plan to have one 18:47:20 vishwanathj: sure 18:47:37 SridarK: we will need some changes to the CLI for this, right? 18:47:45 SridarK: if so i can add those 18:47:52 SumitNaiksatam: yes will need that ok greate 18:48:04 *great 18:48:37 any other questions for SridarK on this? 18:48:48 SumitNaiksatam: ur help as always is most appreciated 18:49:11 SridarK: thanks for working on this 18:49:23 SumitNaiksatam: no worries at all 18:49:36 #topic FWaaS: L3 Agent restructure 18:49:39 pc_m: hi 18:49:47 we already touched on this a little bit today 18:49:47 howdy 18:50:01 anything you want to add/bring up besides that? 18:50:09 Yeah, once we have a feel for when insertion will be done, we can look at the refactoring. 18:50:13 any developments since we disucssed last? 18:50:46 no. just a holding pattern right now, which is OK, as the router object work is still ongoing. 18:51:33 For VPN, once the router object is ready, we'll do more refactoring. For FW, I don't know, may or may not have to do anything. 18:52:04 pc_m: yes, hopefully we are able to determing that soon 18:52:57 pc_m: thanks, keep us in the loop on anything else that we need to know on the l3-agent refactoring side (and that might potentially affect fwaas) 18:53:04 #topic Service Groups 18:53:10 ok 18:53:25 badveli_: any update? 18:54:44 perhaps not 18:54:53 sumit i am thinking of two patches 18:55:01 badveli_: okay 18:55:26 one as the service objects and the other one as the reference implementation 18:55:46 reference implementation as the firewall 18:55:55 badveli_: okay, sounds reasonable to me 18:56:21 badveli_: just keep in mind that people will probably start reviewing in earnest only after they see both patches 18:56:37 badveli_: for the first one will there be a backend ? 18:56:46 thanks sumit, keeping in mind about the reviews, i will go head and do this kind of split 18:56:59 SridarK: i believe the second patch is the backend for the first patch 18:57:07 no there will be no backend kind of implementation 18:57:23 badveli_:, SumitNaiksatam: ok so one review split as 2 patches 18:57:27 ok got it 18:57:28 in the first patch, the second one is a reference implementation 18:58:37 any questions for badveli_? 18:58:56 thanks for your help, i will go in this direction 18:59:04 badveli_: thanks for the update 18:59:25 oh one thing, it seems like these patches will get split across repos 18:59:32 thanks sumit, all . Let me try 18:59:57 yes one will be in the neutron 19:00:07 so thats a bit of an unchartered territory for us 19:00:53 pc_m: have you attempted any simultaneuos changes to the neutron repo and one of the services’ repos? 19:01:08 pc_m: say for example when you were doing the VPN agent refactor? 19:01:13 no. Not sure if we can. 19:01:17 pc_m: hmmm 19:01:41 pc_m: i was going to ask if one can make a patch dependent across the repos 19:01:45 I created the new stuff in Neutron, committed, and then did change in VPN to use. 19:02:01 SumitNaiksatam: Not sure how to do that. 19:02:06 because i think in badveli_’s case his second patch in the fwaas repo should be dependent on the first patch in the neutron repo 19:02:46 sounds like it will be pioneering in that area :) 19:02:48 pc_m: in this case the first patch will not merge without people seeing the second patch, kind of a chicken and egg situation! :-) 19:03:00 SridharRamaswamy: hi 19:03:14 Sumit we will have the issue if it is one patch or two patches 19:03:19 SumitNaiksatam: No way to decouple? 19:03:27 badveli_: yes thats true 19:03:38 badveli_: i think u have to have 2 patches 19:03:59 the bigger concern is that the second patch is probably not going to pass UTs until the first patch merges 19:04:11 that is if we are not able to put the dependency 19:04:51 badveli_: you are going to be blazing a trail here I guess! ;-) 19:05:22 pc_m: i am not sure how we could completely decouple the second patch 19:05:33 but there could be a chance of atleast getting one patch i guess? 19:05:45 pc_m: since the second patch implements the extensions 19:06:24 Does first patch define the extensions? 19:06:35 pc_m: yes 19:06:57 Is there some way for a temporary stub implementation in Neutron? 19:06:57 in fact even the DB is on the fwaas repo side 19:07:07 * pc_m just thinking out loud 19:07:09 So the extensions will need to go into neutron 19:07:21 so only the extension definition will go into neutron 19:07:38 SridarK: yeah 19:07:45 pc_m: that is a possibility 19:08:04 SumitNaiksatam: so that will need to be a separate patch anyway 19:08:25 pc_m: actually, for the first patch to work, no stub is required 19:08:43 SumitNaiksatam: we will land up with extension and api and no backend or a stub in neutron 19:09:04 if we can figure out how the second patch can make use of the extension definition while the first patch is still being reviewed, we are good 19:09:14 SridarK: correct 19:09:47 SumitNaiksatam: ok this is definitely trailblazing - hope we don't get burned 19:10:03 SridarK: :-) 19:10:16 SridarK: what about your patch, do you anticipate a similar issue? 19:10:37 SridarK: since you need to update the fwaas extension definition as well 19:10:45 sumit, sridar, please let me know if we need a discussion, we can take it offline also 19:10:56 SumitNaiksatam: yes exactly - 19:11:21 :-) 19:11:44 SumitNaiksatam: extensions will need to go into neutron but nothing much more - will need to use routers resource from neutron (but that is not an issue) 19:12:00 SumitNaiksatam: not as bad as service objects case 19:12:27 SridarK: perhaps 19:12:30 SumitNaiksatam: also on the topic of extensions - will have some impact from the pecan framework changes 19:12:45 SridarK: yeah whenever it lands 19:12:58 SumitNaiksatam: ok 19:14:02 SridarK: it will be good if you can identify all the changes you need to make in the extension, DB, and plugin, that way we can work through this surgically 19:14:09 SridarK: we need to do this at the earliest 19:14:14 SumitNaiksatam: ok will do so 19:14:19 Sumit, Sridar thanks for your updates. I will sync up offline and think about this 19:14:37 SridarK: the agent is on the fwaas repo side, so dont foresee as many issues there 19:14:42 SumitNaiksatam: yes this needs more thought for sure 19:14:58 SumitNaiksatam: yes the agent is quite straightforward not much in terms of changes 19:15:26 this is the kind of blocker i was asking about, and was hoping that we had not run into! :-( 19:15:52 anyway, perhaps also a good idea to send to the -dev mailer and check if someone else has thought about this already 19:16:09 SumitNaiksatam: yes need to get used to thinking in terms of multiple repos - glad we had this conversation today - had not really thought about that 19:16:27 +1 19:16:37 #topic Vendor Plugins/Drivers 19:16:45 SumitNaiksatam: so although the extensions are defined in neutron - hopefully the migration will stay in fwaas repo 19:16:46 vishwanathj: your turn, now 19:17:01 SridarK: the DB is on the neutron-fwaas side 19:17:15 SridarK: so yes, migration is not in neutron 19:17:21 I have uploaded the code and can be viewed at https://review.openstack.org/#/c/147600/ 19:17:28 SumitNaiksatam: yes 19:17:39 vishwanathj: thanks, sorry havent had a chance to look 19:17:55 vishwanathj: same here - need some time to take a look 19:18:02 ok 19:18:16 badveli_: perhaps you can also help with reviewing vishwanathj’s patch 19:19:11 SumitNaiksatam: we will also upstream a vendor patch once our vendor repo is set up for the dependency (L3) 19:19:16 fine sumit 19:19:28 vishwanathj: do you have any blockers at this point (apart from getting reviewer attention)? 19:19:38 SridarK: yes, i was coming to your patch 19:19:47 oops sorry :-( 19:19:55 SridarK: np, :-) 19:20:21 SumitNaiksatam, nothing that i can think of, the meeting with you, SridarK, pc_m last week was beneficial, a big thanks to you folks 19:20:29 vishwanathj: the way i understand your patch is based on the current framework and implementation, right? 19:20:37 yes 19:20:56 vishwanathj: no worries, we appreciate your sincere effort, and for getting conversation started 19:21:25 vishwanathj: as i mentioned we need at least one more core reviewer to help review 19:21:41 ok 19:21:53 so lets try to identify who can help with that 19:22:20 SridarK: sorry, go ahead 19:22:36 SumitNaiksatam, any suggestions on who I can approach 19:22:53 SridarK: so you wont be submitting the patch to neutron-fwaas? 19:23:11 SumitNaiksatam: no worries - just wanted to add that we will have 2 patches(plugin, agent) pushed up once we have our vendor repo setup for L3 19:23:23 SumitNaiksatam: it will be to neutron-fwaas 19:23:40 vishwanathj: i believe all the neutron-drivers are cores on fwaas, and then there are a couple of more 19:23:42 SumitNaiksatam: we will need to pull out our Router implementation from neutron 19:23:54 vishwanathj: i will add them to your review, you can proceed from there, sound okay? 19:23:59 ok 19:24:17 SridarK: ah, so the external repo dependency for the l3 plugin being pulled out? 19:24:23 SumitNaiksatam: exactly 19:24:30 SridarK: got it 19:24:43 SumitNaiksatam: i think we have figured out the mechanics for this and should happen this week 19:24:46 SridarK: but you can still post the patch in neutron-fwaas, right? 19:24:55 WIP, i mean 19:25:02 SumitNaiksatam: yes we can - did not want to risk a -2 19:25:14 okay 19:26:08 SumitNaiksatam: ok - can do that - our code is done - wanted to put a note on the vendor repo dependency (point to a wiki or something) and get our stuff out so folks realize that there is a plan in place for vendor repo 19:26:36 SumitNaiksatam: we may be good by this week or worst case will get it out by early next week with a WIP 19:26:49 SridarK: yes sure, i think WIP should insulate against the possibility of an overzealous -2 ;-) 19:26:58 SumitNaiksatam: ok :-) 19:27:06 ok we are hitting the hour mark 19:27:10 #topic Open Discussion 19:27:22 and i was thinking we would have had a shorter meeting today! 19:27:30 anything else folks? 19:27:33 SumitNaiksatam: :-) 19:27:49 vishwanathj: i have added the reviewers 19:27:56 SumitNaiksatam, Thanks 19:28:16 alrighty, lets call it a wrap for this meeting 19:28:21 Bye all 19:28:22 thanks all for joining! 19:28:24 bye 19:28:27 bye 19:28:35 #endmeeting