17:00:28 <cathy_> #startmeeting network_common_flow_classifier 17:00:29 <openstack> Meeting started Tue Jun 14 17:00:28 2016 UTC and is due to finish in 60 minutes. The chair is cathy_. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:30 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:32 <openstack> The meeting name has been set to 'network_common_flow_classifier' 17:00:41 <davidsha> Hi 17:00:41 <cathy_> Hi everyone 17:00:48 <s3wong> hello 17:00:50 <LouisF> hi cathy_ 17:00:55 <ihrachys> o/ 17:00:57 <jlibosva> o/ 17:01:17 <asselin_> is this meeting time double-booked? 17:01:31 <cathy_> I have posted the agenda in the wiki page https://wiki.openstack.org/wiki/Neutron/CommonFlowClassifier#Contributors 17:01:43 <mohankumar_> Hi cathy_ 17:01:44 <cathy_> asselin_: not as far as I know 17:01:44 <mmedvede> this was usually third-party meeting spot 17:01:48 <ja3> asselin: looking that way 17:02:04 <jlibosva> asselin_: is third-party weekly mtg? 17:02:22 <jlibosva> or bi-weekly? 17:02:26 <mmedvede> jlibosva: it is bi-weekly 17:02:44 <asselin_> mabye it's that biweekly bug? 17:02:45 <s3wong> this is approved at openstack-infra, which the software would ensure no double booking 17:02:48 <asselin_> for ical 17:03:32 <asselin_> link? 17:03:40 <mohankumar_> Neutron Common Classifier meeting booked for odd weeks 17:03:54 <mmedvede> ok, for third-party folks, lets convene in #openstack-third-party-ci channel 17:04:17 <cathy_> mmedvede: sorry about this. Thanks. We will check the time slot again 17:04:18 <mmedvede> and figure it out after, as we do not have a big agenda 17:04:20 <ihrachys> looking at http://eavesdrop.openstack.org/#Neutron_Service_Chaining_meeting 17:04:28 <ihrachys> I see that it's Thursdays for SFC 17:04:47 <cathy_> ihrachys: this is not for service chaining specific 17:04:59 <ihrachys> oh sorry 17:05:01 <cathy_> ihrachys: it is for the Neutron Common Flow Classifier 17:05:02 <jlibosva> ihrachys: it;s Neutron Common Classifier meeting 17:05:16 <cathy_> ihrachys: np 17:05:29 <cathy_> OK, let's start 17:05:32 <ihrachys> as per .ical file, it's next week 17:06:26 <mohankumar_> ihrachys , yes , classifier meeting on odd weeks .. there is some confusion 17:06:40 <cathy_> ihrachys: yes, some confusion here. how is it considered odd week? 17:06:57 <cathy_> This is the 3rd week of June, isn't it? 17:06:58 <asselin_> i remember and even/odd week bug for ical where jan 2016 was 2 odd weeks in a row 17:07:04 <ihrachys> cathy_: how do you count? from the start of the year? 17:07:12 <ihrachys> cathy_: it's annual, not monthly 17:07:31 <mohankumar_> This is Week 24 in this year 17:07:38 <cathy_> So count from the start of the year? 17:07:44 <ihrachys> aha 17:08:00 <ihrachys> or just rely on .ical to calculate it for you 17:08:14 <cathy_> ihrachys: OK, my bad. Sorry folks. I will correct this starting from next meeting 17:08:54 <cathy_> ihrachys: what do you mean by .ical? 17:09:13 <ihrachys> cathy_: http://eavesdrop.openstack.org/#Neutron_Common_Classifier_meeting 17:09:14 <davidsha> cathy_: file for outlook that saves an appoinment 17:09:20 <ihrachys> see the "ICS file for this specific meeting" 17:09:28 <ihrachys> download it and add to your calendar app 17:09:50 <ihrachys> it's .ics actually, sorry 17:09:52 <cathy_> ihrachys: got it. Thanks! 17:10:06 <mohankumar_> cathy_ , you can download @ http://eavesdrop.openstack.org/#Neutron_Common_Classifier_meeting 17:10:17 <cathy_> mohankumar_: yes, thanks 17:10:30 <cathy_> now let's start with the first topic 17:10:35 <ihrachys> +1 17:10:39 <cathy_> #topic Bug Status: developed as a RFE over neutron-core? 17:11:23 <cathy_> there are two bugs associated with this feature 17:11:30 <cathy_> #link https://bugs.launchpad.net/neutron/+bug/1583299 17:11:30 <openstack> Launchpad bug 1476527 in neutron "duplicate for #1583299 [RFE] Add common classifier resource" [Wishlist,In progress] 17:11:45 <cathy_> #link https://bugs.launchpad.net/neutron/+bug/1476527 17:11:45 <openstack> Launchpad bug 1476527 in neutron "[RFE] Add common classifier resource" [Wishlist,In progress] 17:11:51 <cathy_> one is duped to the other 17:12:27 <cathy_> In our last meeting, the consensus is to developed this feature as a RFE over neutron-core. 17:13:27 <cathy_> Does anyone know whether the neutron team had re-take a look at this bug which has been there for quite some time? https://bugs.launchpad.net/neutron/+bug/1476527 17:13:27 <openstack> Launchpad bug 1476527 in neutron "[RFE] Add common classifier resource" [Wishlist,In progress] 17:13:37 <cathy_> ihrachys: Do you know? 17:14:23 <ihrachys> hm, it's approved so what is missing? 17:14:53 <ihrachys> note that as per armax, "but eventually this would end up being its own neutron stadium repo/LP project, and *not* the Neutron/Neutron-specs repos." 17:15:47 <ihrachys> if you want drivers to get back to it for some reason, we probably need to tag it as rfe again 17:15:53 <ihrachys> but I would better talk to armax 17:16:02 <ihrachys> so, is the latest resolution not in line with your vision? 17:17:20 <cathy_> ihrachys: per our discussion in last meeting, we would like it to be tagged as rfe-approved, not as a new stadium project 17:17:40 <ihrachys> cathy_: then I guess we need to run the RFE bug thru drivers team again 17:17:52 <cathy_> similar to the way L2 agent feature is handled 17:18:12 <ihrachys> I agree it's core to the project goals 17:18:39 <cathy_> ihrachys: Anything we need to do for "run the RFE bug thru drivers team again"? 17:18:55 <ihrachys> I think we should just replace rfe-approved tag with rfe 17:19:03 <ihrachys> lemme handle it 17:19:11 <cathy_> ihrachys: OK, thanks! 17:19:14 <igordcard> what about neutron-classifier? 17:19:27 <cathy_> This is the neutron-classifier feature 17:19:38 <cathy_> igordcard: This is the neutron-classifier feature 17:20:23 <cathy_> #action ihrachys will run the RFE bug thru drivers team again 17:20:28 <igordcard> neutron-classifier can still be core even if it is in its own repo 17:20:49 <igordcard> this common classifier seems like a new effort that doesn't use neutron-classifier at all 17:21:19 <ihrachys> not that there is much to use there, right? 17:21:22 <cathy_> igordcard: we will only have one classifier for the Neutron and the team will work together for that goal 17:21:32 <cathy_> ihrachys: I think so 17:21:51 <cathy_> ihrachys: but we will try to reuse existing implementation as much as possible 17:21:52 <sean-k-mooney> cathy_: there is a datamodle 17:22:17 <cathy_> sean-k-mooney: yes. 17:22:45 <igordcard> cathy_: but the model proposed in the CC wiki looks completely different from openstack/neutron-classifier 17:22:54 <sean-k-mooney> the implementation in the neutron-classifyer is type based so it is different form what exists in the security group api and networking-sfc 17:23:06 <cathy_> I think we need to first agree on the API, data model design, then we can evaluate existing cdoe 17:23:38 <sean-k-mooney> cathy_: yes i think that is a better first step 17:23:57 <LouisF> cathy_: agree 17:24:20 <s3wong> cathy_: true 17:24:32 <mohankumar_> cathy_ : +1 17:24:47 <sean-k-mooney> so in terms of design i see two different proposals for the api 17:25:22 <sean-k-mooney> form the security group api and networking sfc the l1-7 classifcaion exist withing a singel api resource 17:26:13 <cathy_> #topic API design discussion 17:26:20 <sean-k-mooney> in the neutron-classifier the classification is a compostion 17:27:01 <cathy_> sean-k-mooney: good point. 17:27:43 <LouisF> sean-k-mooney: can you elaborate 17:27:44 <igordcard> sean-k-mooney: cathy_ : this second approach is also described in more detail at https://bugs.launchpad.net/neutron/+bug/1476527/comments/26 17:27:44 <openstack> Launchpad bug 1476527 in neutron "[RFE] Add common classifier resource" [Wishlist,In progress] 17:28:00 <cathy_> how about someone evaluate the pros and cons of the two ways 17:28:52 <cathy_> anyone would like to take this evaluation work? 17:29:12 <sean-k-mooney> cathy_: i think it would be better to dicuss in a spec rather then one person evaluating 17:29:26 <ihrachys> + for a spec 17:29:32 <igordcard> I'll post a spec 17:29:51 <cathy_> igordcard: thanks! 17:30:32 <igordcard> sorry I need to leave the meeting now - will catchup afterwards with the logs 17:30:33 <cathy_> #action igordcard will post a spec on the pros and cons of the two classification ways 17:30:38 <igordcard> cya, thanks 17:31:00 <davidsha> bye 17:31:17 <sean-k-mooney> well should the spec not capture the implentation approch rather then the pros and cons 17:31:31 <sean-k-mooney> ie discribe the datamodels and api for both solutions 17:32:08 <cathy_> sean-k-mooney: yes, that should be included otherwise people can not evaluate whether the pros and cons make sense 17:32:10 <sean-k-mooney> as the spec is reviewed we can then converge on one implenttion that meets the contraitns 17:32:52 <cathy_> sean-k-mooney: I would think before we decide on one approach, no need to dive into very detail 17:33:18 <cathy_> after the team reaches consensus on one way, we can have another spec detailing the API, data model etc. 17:33:58 <cathy_> I think we first need to agree on the high level which way to go (with enough detail for the team to make decision) 17:34:54 <sean-k-mooney> sure though at least a preliminay defintion of the api and data model will be needed but i will not neet to be complete 17:35:17 <cathy_> sean-k-mooney: yes, agree. 17:35:47 <sean-k-mooney> something akin to https://review.openstack.org/#/c/318317/5/doc/source/devref/openvswitch_agent.rst i think would be a good guide 17:35:48 <cathy_> I think we have completed the second topic on the agenda:-) 17:36:53 <LouisF> sean-k-mooney: yes that is a good example for the spec 17:37:34 <cathy_> sean-k-mooney: This is "implementation detail", which should be the spec. But for evaluation, we probably do not need this detail. 17:37:59 <ihrachys> you can always start with a less detailed draft and then fill it in with details 17:38:53 <LouisF> there should be sufficient to be able to identify the pros/cons of each approach 17:39:11 <sean-k-mooney> cathy_: yes that was the implentation but i was using it to illistrate how both options are compared rather then present one as the solution and the other as an alternitive 17:39:15 <LouisF> sufficient detail 17:39:32 <cathy_> sean-k-mooney: OK, Sure 17:40:07 <cathy_> sean-k-mooney: I think it is important not to present one as the solution and the other as alternative 17:40:41 <cathy_> I will work with igordcard on this comparison spec 17:40:57 <cathy_> anyone else would like to contribute to this spec? 17:41:23 <cathy_> ihrachys: yes 17:41:57 <cathy_> We are done with the topics on the agenda. Any other topic you would like to discuss? 17:42:28 <davidsha> cathy_: I've made some PoC code for the Flow manager if people would like to give their opinions 17:42:55 <davidsha> #link https://review.openstack.org/#/c/323963/ 17:43:19 <cathy_> davidsha: that is related to L2 agent extension, not quite the flow classifier, right? 17:43:50 <jlibosva> davidsha: is it implementing the draft that ajo proposed? 17:44:01 <LouisF> cathy_: I can help with spec also 17:44:11 <cathy_> LouisF: great, thanks! 17:44:34 <cathy_> #action Louis and Cathy will work with igordcard on the spec 17:44:48 <davidsha> cathy_, jlibosva : kinda, it's a bit out of date now but I'm working on it. 17:45:08 <ihrachys> davidsha: not related to the meeting, but looks less scary than I thought :) 17:45:30 <davidsha> ihrachys: You haven't seen the prioritization yet @.@ 17:45:54 <cathy_> davidsha: yes not related to the meeting since this meeting is specific to FC:-) 17:46:18 <cathy_> davidsha: but I will take a look after you update it. Could you add me as reviewer? 17:46:19 <ihrachys> ok folks, I need to run. ciao. 17:46:34 <cathy_> ihrachys: ok, thanks for joining 17:46:40 <davidsha> cathy_: kk, I'll do that now. 17:46:49 <cathy_> davidsha: thanks. 17:47:08 <cathy_> I will end this meeting if no other topic on FC. 17:47:15 <cathy_> 5 17:47:16 <cathy_> 4 17:47:19 <cathy_> 3 17:47:21 <cathy_> 2 17:47:23 <cathy_> 1 17:47:36 <cathy_> thanks everyone 17:47:40 <davidsha> thanks 17:47:42 <mohankumar__> bye 17:47:45 <cathy_> bye for now 17:47:52 <cathy_> we will start the next meeting in 3 weeks 17:48:04 <cathy_> #endmeeting