17:00:05 <LouisF> #startmeeting service_chaining
17:00:10 <openstack> Meeting started Thu Feb 16 17:00:05 2017 UTC and is due to finish in 60 minutes.  The chair is LouisF. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:14 <openstack> The meeting name has been set to 'service_chaining'
17:00:16 <LouisF> hi all
17:00:19 <igordcard> hi LouisF
17:00:41 <LouisF> #topic agenda
17:00:46 <LouisF> https://wiki.openstack.org/wiki/Meetings/ServiceFunctionChainingMeeting#Agenda_for_the_Networking-SFC_Meeting_.282.2F16.2F2017.29
17:00:57 <bcafarel> hello
17:01:11 <igordcard> hi bcafarel
17:01:17 <LouisF> bcafarel: hi
17:01:21 <LouisF> igordcard: hi
17:02:03 <LouisF> wait a moment for others to join
17:03:31 <bcafarel> LouisF: thanks for creating the OVS db bug :)
17:03:39 <LouisF> bcafarel: np
17:03:48 <LouisF> ok lets start
17:04:13 <LouisF> Add SFC cli for OSC plugin https://review.openstack.org/#/c/409759/23
17:04:24 <LouisF> bcafarel: thanks for uploading patch
17:05:22 <LouisF> lets review and get this merged
17:05:23 <bcafarel> only differences there are: rebase fix, entry point for port pair group create, and merge a few recent changes on unit tests (from current repo)
17:05:35 <LouisF> bcafarel: thanks
17:05:44 <igordcard> hope to review soon
17:05:50 <LouisF> igordcard: thanks
17:05:59 <pcarver> hi
17:06:32 <bcafarel> np, I think the other remaining bugs are not blocking to get this merged (missing help page and unicode format on lists)
17:06:37 <bcafarel> pcarver: howdy
17:08:17 <LouisF> pcarver: hi
17:08:26 <igordcard> I will submit the sfc graph cli part later, and after this one merges
17:08:47 <igordcard> hi pcarver
17:09:16 <LouisF> #topic API reference
17:09:30 <LouisF> pcarver: what is status
17:10:25 <LouisF> for those who joined please review SFC cli for OSC plugin https://review.openstack.org/#/c/409759/23
17:10:47 <pcarver> not much change since last week, but after an tremendous amount of nitpicking and rebasing the BGPVPN API definition was finally merged and the BGPVPN API ref is pretty much perfect. So now I'm turning my attention to the SFC API def and ref
17:11:08 <LouisF> pcarver: great - thanks
17:11:23 <pcarver> I think the SFC API ref is in pretty good shape, but the def probably needs a lot of work.
17:11:40 <LouisF> pcarver: ok look forward to patches
17:11:47 <pcarver> I think reviewers are going to require it to be refactored before merging into neutron-lib
17:12:20 <LouisF> pcarver: was that done for bgpvpn?
17:13:02 <pcarver> it was less complicated that the sfc one, but I generalized the one bgpvpn specific validator
17:13:20 <LouisF> pcarver: ok
17:13:34 <LouisF> #topic ocata work
17:13:35 <pcarver> I think validate_list_of_allowed_values in sfc will probably need to move to a non-sfc specific validator
17:13:43 <pcarver> if there isn't an equivalent already available
17:13:55 <LouisF> pcarver: ok
17:14:10 <pcarver> And I need to figure out whether the expectation will be for each of the SFC elements to be split out
17:14:55 <pcarver> One of boden's comments on the API ref is " it appears there are actually 2 separate extensions providing these APIs"
17:15:24 <pcarver> I need to determine if they're going to expect the structure to be changed
17:15:50 <pcarver> e.g. for VLAN Trunk there are actually two api definition files
17:16:19 <pcarver> trunk.py and trunk_details.py
17:16:57 <LouisF> pcarver: you mean separate definition files: one for port-chains and another for flow-classifiers?
17:17:01 <pcarver> It's possible that they may want the SFC API definition split into port/pp/chain and flowclassifier or even into four separate pieces
17:17:05 <pcarver> I'm not sure yet
17:17:43 <pcarver> I'm going to do an initial patch set to factor out the validator, then see what sort of reviews I get
17:17:55 <igordcard> pcarver: pp/ppg/chain?
17:18:03 <pcarver> yes
17:18:07 <igordcard> o
17:18:10 <igordcard> ok
17:18:26 <LouisF> is that really needed?
17:19:04 <pcarver> LouisF: I don't know if it's really needed, but there were a lot of iterations on bgpvpn and it was simpler
17:19:27 <LouisF> pcarver: ok
17:19:32 <pcarver> I'm going to start with just making the validator generic rather than sfc specific
17:20:07 <pcarver> but I'm not going to be surprised if I get a ton of nitpicking
17:20:40 <LouisF> :(
17:21:03 <pcarver> I probably won't get it done before next week, so maybe I'll be able to talk directly with people at the PTG
17:21:16 <LouisF> ok thanks
17:21:20 <pcarver> I'm planning to be in Atlanta Wednesday-Friday
17:21:42 <LouisF> ocata work : Symmetric Chain Support for OVS driver and agent https://review.openstack.org/#/c/410482/
17:22:00 <LouisF> i think we are close to approval
17:22:57 <LouisF> please review
17:23:49 <LouisF> ocata work: Support MPLS correlation without SFC Proxy https://review.openstack.org/#/c/420339/
17:24:04 <LouisF> igordcard: i had some comments on this
17:24:35 <igordcard> LouisF: I've hopefully completed that one now... will address the comments too and likely fix what's making the tempest tests fail
17:25:01 <igordcard> LouisF: with that, sfc graphs is now very close... hoping to finish by friday/monday
17:25:37 <LouisF> igordcard: good - look forward to patch updates
17:25:41 <igordcard> LouisF: I still have to start adding OSC support and api def/ref though - for the new sfc graph resource
17:25:54 <LouisF> igordcard: thanks
17:27:45 <LouisF> #topic stable/ocata branch
17:28:29 <LouisF> i would like to get this done by end  of month
17:28:49 <LouisF> to align with the official ocata release
17:29:32 <LouisF> hopefully we can get the items discussed today complete by then
17:30:03 <LouisF> thoughts?
17:30:22 <igordcard> LouisF: end of the month including everything reviewed?
17:30:47 <LouisF> igordcard: that would be the goal
17:30:57 <bcafarel> so symmetric chain support and MPLS correlation in networking-sfc itself, API ref and OSC plugin in other repos?
17:31:12 <LouisF> bcafarel: yes
17:31:32 <LouisF> and sfc graph
17:31:48 <igordcard> OK that gives me about a week
17:32:14 <bcafarel> igordcard: better drink a lot of coffee in Atlanta then!
17:32:27 <igordcard> bcafarel: yeah......
17:32:57 <LouisF> igordcard: is not we can move this work to pike
17:33:06 <LouisF> if not...
17:33:27 <igordcard> LouisF: right right, deadliest line is 10th of march right? so we need some days to make sure everything is perfect
17:33:53 <LouisF> or we can delay the stable/ocata branch for sfc for a week...
17:34:51 <igordcard> LouisF: does that have to be made in advance, or can the 10th of march be considered "tentative"?
17:35:46 <LouisF> igordcard: is can be tentative and we can re-evaluate at end of month
17:36:37 <LouisF> #topic PTG meeting
17:36:54 <LouisF> i will not be attending
17:37:07 <LouisF> bcafarel: you are going?
17:37:13 <LouisF> igordcard: you also?
17:37:27 <igordcard> LouisF: yes
17:37:35 <LouisF> anyone else?
17:37:53 <bcafarel> yes, I will be there all week, mostly QA the first days, then neutron and octavia rooms
17:38:21 <igordcard> I'll be wed to fri, so I still have ~2 days to work
17:39:08 <LouisF> we have the regular irc meeting next week if scheduling is ok for you two?
17:39:09 <bcafarel> pcarver too for these days right?
17:39:29 <LouisF> and paul?
17:39:52 <LouisF> if not we can skip next week's irc meeting
17:39:54 <igordcard> LouisF: I may join but not sure how much I'll contribute
17:40:17 <bcafarel> if I have my hours correctly it will be during lunch time in Atlanta
17:40:28 <igordcard> bcafarel: yeah... noon I think
17:40:39 <pcarver> I'll be arriving late Tuesday night
17:41:20 <LouisF> bcafarel: right - ok let have next meeting on march 2?
17:41:57 <igordcard> good with me
17:41:58 <bcafarel> LouisF: yes, sounds easier, no need to come in if we are all busy in rooms or having lunch :)
17:42:08 <LouisF> bcafarel: ok
17:42:15 <pcarver> ok
17:42:45 <bcafarel> any particular topic to watch for regarding sfc?
17:43:17 <LouisF> igordcard: you are working on the common classifier?
17:43:28 <igordcard> LouisF: yes
17:43:41 <igordcard> we have a PoC almost done
17:43:49 <LouisF> will that be discussed at the ptg?
17:43:58 <igordcard> LouisF: yes we'll try to bring that up
17:44:34 <bcafarel> current agenda items: https://etherpad.openstack.org/p/neutron-ptg-pike
17:45:52 <igordcard> it would be interesting to see an integration PoC with networking-sfc in the future, but first the ccf needs to go beyond PoC
17:46:13 <igordcard> by the end of pike or during queens would be the plan
17:46:16 <LouisF> integration of sfc with ironic may worthwhile looking at
17:46:42 <LouisF> igordcard: yes indeed
17:47:41 <LouisF> a port-chain that includes both vnfs and pnfs using ironic
17:48:24 <igordcard> nice
17:49:33 <bcafarel> noted, it sounds nice indeed
17:49:58 <igordcard> you tested with  containers in the past right? kuryr?
17:50:11 <bcafarel> there is the usual stadium topic too, and maybe also the "use tempest stable api" item
17:51:10 <LouisF> igordcard: some testing with containers
17:51:32 <LouisF> #topic other items
17:52:15 <LouisF> ok next meeting on March 2
17:52:21 <LouisF> safe travels
17:52:47 <LouisF> bye all
17:52:52 <igordcard> bye
17:52:54 <LouisF> #endmeeting