18:31:43 #startmeeting Networking FWaaS 18:31:44 Meeting started Wed Oct 22 18:31:43 2014 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. 18:31:49 The meeting name has been set to 'networking_fwaas' 18:32:07 seems like everyone is busy preparing for the summit ;-) 18:32:35 #info kilo design summit etherpad: https://etherpad.openstack.org/p/neutron-fwaas 18:33:20 #info neurtron kilo design summit topics: #link https://etherpad.openstack.org/p/kilo-neutron-summit-topics-distilled 18:33:32 lets get started with the standing items on the agenda 18:33:36 #topic Bugs 18:33:45 I dont see anything new popping up 18:33:47 SridarK: ? 18:33:58 SumitNaiksatam: not much to report 18:34:08 i know badveli was going to do some additional testing 18:34:12 but he doesnt seem to be here 18:34:18 lets check back with him offline 18:34:27 SumitNaiksatam: will do 18:35:50 i dont think any of the untriaged bugs are critical 18:36:19 SumitNaiksatam: i will also take a look 18:36:25 SridarK: thanks 18:36:28 #topic Docs 18:36:35 SridarK: any updates? 18:36:42 SumitNaiksatam: not much on this 18:37:25 SridarK: let me know if you need anything from me, happy to jump in… 18:37:35 SumitNaiksatam: will do 18:37:45 ok moving on 18:38:01 #topic Kilo blueprints 18:38:22 SridarK: i noticed that you added a vendor blueprint for the CSR 18:38:26 glebo: hi there! 18:39:18 SumitNaiksatam: yes resubmitted what we had for Juno 18:39:23 SridarK: thats a good thing 18:39:33 SridarK: we need to start refocussing on getting the process going 18:39:40 SumitNaiksatam: lo! 18:39:46 SumitNaiksatam: now our dependencies are in 18:39:54 glebo: hi 18:39:57 SridarK: agreed 18:40:09 glebo: is vishnu planning to repost the service groups blueprint? 18:40:22 natarajk: what about your vendor spec? 18:40:37 i was planning to submit it after the design summit 18:40:40 SumitNaiksatam: absolutely 18:40:45 natarajk: ok sure 18:40:46 I'll ping him on it 18:40:55 suprised he didn't do it already 18:40:58 glebo: yes, i think better to put it in at earliest 18:41:09 glebo: he might have had, and i might have missed it 18:41:10 SumitNaiksatam: ack 18:41:18 glebo: apologies in advance if i missed it 18:41:54 #action: glebo to encourage vishnu to submit service groups BP asap 18:42:20 * glebo surprised vishnu not on the chat here himself 18:42:40 glebo: thanks for adding the AI 18:42:48 #action: glebo to encourage vishnu to submit service groups BP asap 18:43:15 #chairs SridarK natarajk glebo 18:43:32 #chair SridarK natarajk glebo 18:43:33 Current chairs: SridarK SumitNaiksatam glebo natarajk 18:43:43 glebo: i checked, i dont think its in 18:44:08 glebo: but its perfectly understandable if badveli is a bit confused and frustrated here 18:44:13 SumitNaiksatam: ya, me too. Already pinged him on it 18:44:30 SumitNaiksatam: LOL!! 18:44:39 SumitNaiksatam: honestly, if he wasn't, I wouldn't be here 18:44:46 glebo: :-) 18:44:57 i think we all are! 18:45:04 anyway 18:45:11 SumitNaiksatam: moving on 18:45:30 so in terms any additional blueprints to be posted 18:45:42 i think we need the design summit discussion to go through 18:45:45 badveli joining any second now 18:45:54 thanks glebo 18:45:56 sorry 18:45:58 ah badveli just joined 18:46:01 badveli: not worries 18:46:12 *no 18:46:13 badveli: good to see someone is still able to get emersed in coding 18:46:30 badveli: ah… the good ol days… 18:46:41 thanks glebo, sorry today being diwali 18:46:46 badveli: we handed you an AI while you were busy actually working! ;-) 18:47:02 sorry sumit, did not see it 18:47:06 badveli: let's go ahead and have u resubmit the BP for service groups 18:47:14 badveli: ack? 18:47:24 the blue print? 18:47:37 badveli: yeah, the service groups 18:47:39 the blue print is already approved 18:47:49 badveli: and do so asap, yeah, BP, spec and code, all in kilo 18:48:04 SumitNaiksatam: that's the proper process? 18:48:32 glebo: yeah, we have to go by that, until we decide or are told to do something else 18:49:03 sumit, do we submit the bp and code by branch names? 18:49:05 badveli: u need to resubmit spec 18:49:27 thanks sridark, i will try to do that 18:49:32 badveli: resubmit the blueprint (you will have to make minor modifications since the template has changed slightly) 18:49:43 badveli: code can just be rebased 18:49:44 should we get the approvals again? 18:50:00 badveli: yes, we need the approvals again on the spec 18:50:02 for the blue print? 18:50:05 * glebo here it comes, ducking 18:50:08 badveli: yes 18:50:22 glebo: :-) 18:50:42 * glebo feels deja vu for badveli, and for all here, really 18:50:44 badveli: so things are clear on the process? 18:50:53 * glebo lavers, rinses, repeats 18:50:58 glebo: deja vu for the nth time! 18:51:15 SumitNaiksatam: so, is that deja vu vu vu vu vu vu vu 18:51:23 glebo: :-) 18:51:29 SumitNaiksatam: or just more simply deja vu^9 18:51:36 badveli: is the process clear? 18:52:06 maybe we ought to get shirts printed up for sumit, just for FWaaS team to wear, "deja vu^9" 18:52:07 yes, but this will still delay and not able to get why we are doing this again 18:52:26 and maybe we'll bring enough for the LBaaS and VPNaaS teams too 18:52:29 and getting the approvals 18:52:37 again??? 18:52:40 glebo: :-) 18:52:53 since u know they are having the exact same tongue n cheek exchange in their mtgs this week too 18:52:55 badveli: i share your frustration, and we are trying to find a solution 18:53:19 SumitNaiksatam: badveli: +1000 18:53:37 badveli: but until then we have to stay the course 18:53:49 badveli: Let's u and I take it offline. I can coach u thru it tomorrow, k? 18:54:03 fine with me, thanks glebo 18:54:05 badveli: open to your suggestions as well, if you think a different approach might be helpful 18:54:44 Sumit, as you know how hard it is getting the approvals 18:54:51 and going through it again 18:54:53 *badveli pictures blow torches* 18:55:29 badveli: time check. Let's take it offline 18:55:34 badveli: agree, and hence we had team discussions to explore options 18:55:41 ok moving on 18:56:00 #topic Kilo design summit discussion 18:56:01 badveli: it is probably going to be worse this time. only five selected cores can give your spec a +A in Kilo cycle :-) 18:56:20 s3wong: now you are really rubbing it in! 18:56:31 SumitNaiksatam: :-) 18:56:47 s3wong: at least now we know which core's reviews really matter 18:56:53 s3wong: yeah sad but (absolutely) true 18:57:00 thanks s3wong 18:57:03 s3wong: that's a step in the right direction, or bright side 18:57:10 s3wong: btw, welcome to fwaas! 18:57:34 glebo: well, you still need the 2 +2's , but you need an extra step to get the +A in case the two cores that gave you +2's aren't one of the five 18:57:39 * glebo snaps for s3wong 18:58:04 s3wong: ack. Exactly 18:58:13 SumitNaiksatam: just happened to be on #openstack-meeting-3; sorry for the interruption, please move on 18:58:17 s3wong: but at least now we KNOW that 18:58:24 ok Kilo design summit discussion 18:58:28 s3wong: Which, you know, is nice 18:58:55 glebo: :-) 18:59:05 i provided link to the etherpad earlier #link https://etherpad.openstack.org/p/neutron-fwaas 18:59:06 s3wong: actually, very helpful "interruption". Everyone needs to be clear on that process step 18:59:13 please enter your input 18:59:45 s3wong: right, you are welcome to stick around 18:59:47 * glebo scanning epad 19:00:22 SumitNaiksatam: the entries in the kilo epad have been preserved here 19:00:36 brb 19:00:40 BTW, guys, in this morning's neutron-drivers meeting, a new "lighting talks" session is suggested and probably will replace the "retrospective" session 19:00:52 so as we had discussed within the team earlier, the following seem to be community priorities for FWaaS: 19:00:53 i think that does reflect our priorities for kilo 19:01:08 1. Get FWaaS out of experimental status (address router insertion) 19:01:09 2. Support FWaaS for E-W scenario in DVR. 19:01:10 3. Service Groups (since this code is posted and in review) 19:01:11 4. Metering capabilities (Ceilometer dependency) 19:01:12 5. Flavors - we might need to support if this gets implemented 19:01:19 so don't be discouraged if your proposal didn't make the initial cut of design summit, you may still get to do a 5 minutes lighting talk 19:01:50 SumitNaiksatam: +1 19:02:17 i would propose that we keep our focus on the above during the discussions 19:02:25 SumitNaiksatam: +1 19:02:56 on point 1, i myself am not clear as to why FWaaS is qailified as “experimental” 19:03:30 i was consulted on this point in the first release of fwaas, and my feedback was based on the insertion issue 19:03:56 SumitNaiksatam: i think this is probab only open issue 19:04:05 SridarK: yeah 19:04:30 * glebo back 19:04:34 are there any disagreements within the team on the above priorities? 19:05:16 the above obviously does not include vendor specific stuff 19:05:22 SumitNaiksatam: agree, i guess we also have to see how this "services coming out of neutron" story plays out 19:05:27 the above also in the context of Kilo 19:05:33 SridarK: agreed 19:05:35 and how that impacts our list 19:06:08 on the above priorites, 19:06:24 I think we will gain a lot of cred in the community at large if 19:06:31 the key to any “coming out” discussion is who are the cores 19:06:42 we are shown to be talking about pulling out vendor plugins from main 19:07:06 and creating an interface for them to easily be loaded as plugg-able modules 19:07:21 glebo: as i mentioned before pulling out vendor plugins is not a discussion just in the context of FWaaS 19:07:30 as I've mentioned before, it's a huge sticking point for several of the TC, 19:07:37 SumitNaiksatam: understood 19:07:38 but 19:07:52 glebo: this has been on the table for a long time, and i believe it will happen in Kilo 19:07:53 each service element will need to address it for themselves, 19:08:07 so, if we show we are doing our part, it will earn us some cred 19:08:26 SumitNaiksatam: exactly why we need to have it on our list, 19:08:43 so that we are being good citizens to the deeper architectural direction 19:08:50 glebo: yeah, sure we can put it on the list, but we will just be following what the rest of neutron is doing 19:08:53 SumitNaiksatam: make sense? 19:09:05 SumitNaiksatam: sure, and 19:09:16 it indicates are are "cheefully complying" and 19:09:30 if we can add some value there, the community will like that too, and 19:10:07 we need to ensure that the hooks, plugs, interfaces, structures, etc. that we need in order to succeed in such an extraction are present 19:10:12 glebo: yes we will definitely comply, i dont think we will have a choice, and we definitely prefer that option 19:10:43 SumitNaiksatam: so, it becomes a pretty big work item for us in kilo, 19:10:48 I'll add it to the epad 19:11:13 glebo: “interfaces” for plugging into neutron is not a FWaaS-specific topic 19:11:29 glebo: that becomes a general services’ discussion 19:11:43 SumitNaiksatam: agreed, and 19:11:48 glebo: feel free to add to the etherpad, since it reinforces the discussion 19:11:51 FWaaS has it's side of it 19:12:15 it's our part of the service insertion model, our domain-specific elements 19:12:26 glebo: but i wanted to make sure that we dont get lost in that disucssion and hence stall on the features for FWaaS that are sort of independent of that aspect 19:13:20 by “that” i mean the “service insertion” related conversation 19:13:30 SumitNaiksatam: not to be sarcastic, but to then go ahead and be sarcastic… those features aren't really "getting in" right now anyway 19:13:31 since this has happened in the past two cycles 19:14:07 SumitNaiksatam: job 1: unstick the system, whatever it takes. Job 2: make FWaaS feature rich 19:14:23 SumitNaiksatam: doing 1 w/o 2 is wheel spinning, no? 19:14:56 also, in the past we have tried justify the value/need of the service insertion framework using FWaaS as an exmaple/case study 19:15:40 however the net result of that was that we could neither get the service framework in, and we couldn’t solve the problem for fwaas either (and thus continue to be branded experimental) 19:16:12 glebo: so most of what you are saying is deja vu to me ;-) (and not trying to be sarcy there) 19:16:40 SumitNaiksatam: totally empathetic 19:16:41 okay so the meta point was, please put your comments on the etherpad :-) 19:16:55 SumitNaiksatam: right. Will do 19:17:01 ;-) 19:17:04 glebo: ;-) but could be :-( just as well! 19:17:22 (some of my points are already there, prepended with [glebo]) 19:17:28 glebo: nice 19:17:48 so what the above list means is that we miss out very critical features like zones and address groups 19:18:32 and logging and hit counts 19:19:12 the fact that none of features were allowed to go through in the last two cycles has created a huge backlog of items 19:19:40 and that will take a long time to clear up in the current process/system (if at all) 19:19:58 (hit counts is a very cool feature. We have them in our product, and customers LOVE them) 19:20:12 glebo: absolutely agree 19:20:29 glebo: its critical for firewalls 19:20:43 so any way the point i wanted to make is that we havent lost track of these features 19:21:16 but we will unlikely have time for them with the current constraints 19:21:38 since we are veering more into open discussion territory 19:21:43 #topic Open Discussion 19:21:53 SumitNaiksatam: re: backlog, that's sort of ok, for the short term, because we aren't trying to build an open source FW here, just trying to define policy blobs. In reality, all the products can define their own policy blobs in their config / mgmt systems, so there is a work around in the short run, for operators, automatable through vendor API 19:22:41 glebo: people prefer a single pane of glass 19:23:28 SumitNaiksatam: any update from the service teams strategy session last week. As you know I left early. Saw the write up. Commented. But not sure if any response came back after that. ? 19:23:40 glebo: i responded to your emails 19:23:54 glebo: i did not notice your emails until yesterday 19:23:55 SumitNaiksatam: ok, must have missed it 19:24:02 SumitNaiksatam: will check 19:24:18 glebo: for some reason the thread got buried and popped only because of s3wong’s email yesterday 19:24:24 SumitNaiksatam: re: 1 glass: for sure. Wouldn't be here otherwise. 19:24:30 glebo: :-) 19:24:53 ok if nothing else to discuss we can get 5 mins of our time back 19:25:20 SumitNaiksatam: perhaps we can all start thinking about the priority list 19:25:32 SridarK: sure 19:25:53 SumitNaiksatam: so in terms of spec etc we can get some early discussion at the summit 19:26:09 i would also propose that we popularize this etherpad and make sure that we get a good critical mass of folks for the F2F discussions in the summit 19:27:21 ok thanks folks for attending today 19:27:26 bye! 19:27:27 bye all 19:27:41 #endmeeting