17:02:30 <LouisF> #startmeeting service_chaining
17:02:31 <openstack> Meeting started Thu Sep 15 17:02:30 2016 UTC and is due to finish in 60 minutes.  The chair is LouisF. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:02:32 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:02:35 <openstack> The meeting name has been set to 'service_chaining'
17:02:40 <LouisF> hi all
17:02:51 <bcafarel> hello
17:03:07 <LouisF> bcafarel: hi
17:03:07 <yamahata> hello
17:03:09 <doonhammer> hello
17:03:11 <LouisF> yamahata: hi
17:03:16 <LouisF> doonhammer: hi
17:03:26 <LouisF> agenda https://wiki.openstack.org/wiki/Meetings/ServiceFunctionChainingMeeting#Agenda_for_the_Networking-SFC_Meeting_.289.2F15.2F2016.29
17:03:46 <LouisF> i want to go over status of work
17:04:11 <reedip-home> Hi
17:04:20 <LouisF> #topic work status
17:04:23 <janki> Hi
17:04:29 <LouisF> hi
17:04:56 <LouisF> https://review.openstack.org/#/c/351789/
17:05:02 <janki> This is my first meeting here. I am Janki.
17:05:14 <LouisF> janki: welcome
17:05:32 <bcafarel> janki: welcome
17:05:44 <janki> LouisF, bcafarel thank you
17:05:44 <LouisF> bcafarel: how is work on Replace OVSSfcAgent by a L2 agent extension progressing?
17:06:26 <bcafarel> I left it marked as "WIP" because I have to update unit and functional tests
17:06:40 <bcafarel> but code is OK and manual tests look fine
17:07:15 <bcafarel> (creating sample chains)
17:07:28 <LouisF> bcafarel: great when do expect the unit and functional test to be complete?
17:07:34 <reedip-home> Btw, it's my first meeting too :p
17:07:44 <LouisF> reedip-home: welcome
17:08:36 <bcafarel> LouisF: no more than a few days, I'd like to get this in review phase soon
17:08:49 <reedip-home> bcafarel: is it the link defined above???
17:08:57 <bcafarel> and I have igordcard waiting for it also :)
17:09:03 <bcafarel> reedip-home: indeed
17:09:11 <bcafarel> (and welcome also!)
17:09:16 <LouisF> bcafarel: excellent looking forward to the patch
17:09:19 <reedip-home> bcafarel: okay great :)
17:09:41 <LouisF> bcafarel: thanks for your work on this
17:10:23 <bcafarel> LouisF: you're welcome, it was a nice way to get into SFC code
17:10:36 <LouisF> Stephen Wong is not on so i'll skip to next item
17:10:38 <bcafarel> and it will make Neutron people happy too (and packagers)
17:10:47 <LouisF> bcafarel: definitely
17:11:05 <LouisF> s3wong: you arrived at exactly the right moment
17:11:08 <igordcard> great work so far
17:11:17 <s3wong> LouisF: hello
17:11:44 <LouisF> s3wong: hi the next item is the tacker drive for sfc
17:11:56 <LouisF> https://review.openstack.org/#/c/347568/
17:12:30 <LouisF> i noticed you had pushed a new patch
17:12:48 <s3wong> yes --- in the VNFFG plugin code, trozet uses the VIM driver directly to invoke chaining related calls
17:13:10 <s3wong> so there were two options for the n-sfc driver to conform
17:13:51 <s3wong> (a) register as a driver_type, and makes VNFFG plugin smart enough to know if vim=='openstack', load networking-sfc driver
17:14:08 <s3wong> (b) collapse the current networking-sfc driver into openstack vim driver
17:14:27 <s3wong> after some discussions, we decided to go with (b)
17:14:46 <LouisF> s3wong: sounds like the better approach
17:14:50 <s3wong> since in VNFFG plugin, trozet actually also collapsed that plugin into NFVO plugin
17:15:03 <s3wong> so now, for full end to end test, diga is working on it
17:15:23 <LouisF> s3wong: great sounds like good progress
17:15:53 <s3wong> our goal is to have at least one pass from TOSCA NFV VNFFG template to OVS networking-sfc driver by the end of the week
17:16:04 <LouisF> s3wong: thats great
17:16:05 <s3wong> keeping fingers crossed
17:16:29 <LouisF> s3wong: thanks for your work on this
17:16:35 <s3wong> LouisF: np
17:17:43 <LouisF> ok on to next item in agenda: driver interface pre-commit/post-commit change
17:17:46 <LouisF> https://review.openstack.org/#/c/363893/
17:18:09 <yamahata> That patch is updated.
17:18:20 <yamahata> The remaining stuff is to answer LouisF 's question.
17:18:32 <yamahata> It seems like you want to know details, not high level description.
17:18:48 <LouisF> yamahata: some detail would be useful
17:18:54 <yamahata> I put detailed commit message.
17:19:20 <yamahata> Maybe you want something different...
17:20:09 <LouisF> yamahata: can you describe how existing drivers, eg ovs, onos would not be affected by this change
17:20:44 <LouisF> i would like everyone to review this
17:20:53 <yamahata> It's described in commit message line 19-
17:21:49 <yamahata> Probably it's not what you want to see, though. Can you elaborate what you want to see?
17:22:13 <LouisF> yamahata: thanks that is fine
17:23:17 <LouisF> mohan and vikram should review this as they work with the onos driver
17:23:48 <LouisF> they are not on  the call  - i will contact them
17:24:41 <LouisF> ok moving on;
17:24:55 <LouisF> fix tempest tests https://review.openstack.org/#/c/363413/
17:25:21 <LouisF> bcafarel: you had a comment regarding the migrations
17:25:53 * bcafarel looks
17:25:57 <bcafarel> ah yes
17:26:30 <bcafarel> I guess they got accidentally in this patch, these are the migrations from the tenant->project revert
17:27:08 <LouisF> bcafarel: those migration changes should be removed from this patch - right?
17:27:09 <bcafarel> there is https://review.openstack.org/#/c/367770/ for just the revert (revert revert actually)
17:27:47 <bcafarel> LouisF: it'd be better as they are not really related
17:28:21 <LouisF> bcafarel: ok i will contact patch author make corrections
17:28:26 <LouisF> bcafarel: thanks
17:28:50 <bcafarel> LouisF: np
17:29:32 <LouisF> ok next on to symmetric chain api https://review.openstack.org/#/c/308274/
17:30:10 <LouisF> Pavel was working on this but he is not on
17:30:33 <LouisF> i will contact him on that
17:31:45 <LouisF> ovs driver/agent spec for symmetric chains https://review.openstack.org/#/c/368155/
17:32:24 <LouisF> please review this
17:32:54 <LouisF> patch for work on the code will follow in a few weeks
17:34:34 <LouisF> any other comments before we llok at bugs
17:34:37 <LouisF> look
17:35:40 <LouisF> #topic bug scrub
17:36:06 <LouisF> https://bugs.launchpad.net/networking-sfc/+bug/1599037
17:36:08 <openstack> Launchpad bug 1599037 in networking-sfc "make get_port_mask more efficient by avoiding some unnecessarily mask checking" [Undecided,In progress] - Assigned to ding bo (longddropt)
17:36:31 <LouisF> this has been around for a while but needs to be  rebased
17:36:54 <LouisF> i will contact ding bo
17:37:48 <LouisF> https://bugs.launchpad.net/networking-sfc/+bug/1582238
17:37:49 <openstack> Launchpad bug 1582238 in networking-sfc "Add "priority” field in flow-classifier" [Undecided,In progress]
17:38:20 <LouisF> mohan has been working on this - i will ping him on status
17:39:12 <LouisF> https://bugs.launchpad.net/networking-sfc/+bug/1617904
17:39:13 <openstack> Launchpad bug 1617904 in networking-sfc "set [flowclassifier] section default drivers to ovs " [Undecided,In progress] - Assigned to xiaodongwang (xiaodongwang991481)
17:39:40 <LouisF> i will check with xiaodong on this
17:40:21 <LouisF> https://bugs.launchpad.net/networking-sfc/+bug/1587486
17:40:22 <openstack> Launchpad bug 1587486 in networking-sfc "Support SFC Encapsulation" [Undecided,In progress] - Assigned to Igor Duarte Cardoso (igordcard)
17:40:46 <LouisF> igordcard: what is the status of this?
17:41:35 <igordcard> LouisF: let me get the link
17:41:47 <igordcard> this is patch 1 out of 3 https://review.openstack.org/#/c/346175/
17:41:53 <igordcard> I'll bring updates to it soon
17:42:09 <igordcard> patch 2 out of 3 is dependent on https://review.openstack.org/#/c/351789/
17:42:26 <igordcard> (enable nsh encap/decap, not yet submitted)
17:42:45 <LouisF> igordcard: thanks
17:42:53 <igordcard> I will bring the new driver interface to patch 1/3 too, it is important to have the correct way of updating a port chain
17:43:06 <igordcard> patch 3/3 is the sfc graph, will submit later
17:44:25 <igordcard> that's all
17:44:51 <LouisF> igordcard: thanks  for your work
17:45:08 <LouisF> https://bugs.launchpad.net/networking-sfc/+bug/1605675
17:45:10 <openstack> Launchpad bug 1605675 in networking-sfc "Models and migrations are out of sync, not tested" [Undecided,In progress] - Assigned to Henry Gessau (gessau)
17:45:40 <LouisF> i will contact henry gessau on this
17:47:31 <LouisF> ok thats all i have
17:47:52 <reedip-home> LouisF : quick query.
17:48:03 <LouisF> reedip-home: go ahead
17:48:49 <reedip-home> This was a part of the query shared on the dev list
17:50:12 <reedip-home> As I am currently restarting my work on TaaS, wanted to know about this
17:50:36 <LouisF> ok i saw that
17:51:13 <LouisF> from Subrahmanyam Ongole  - right?
17:51:19 <reedip-home> LouisF : TaaS team had an initial discussion at the Austin Summit
17:51:23 <LouisF> is that you?
17:51:24 <reedip-home> Yes
17:51:41 <reedip-home> But the topic never got the momentum
17:51:52 <reedip-home> Umm, no :)
17:52:41 <reedip-home> We can discuss and share our requirements with Networking SFC
17:52:45 <LouisF> are you thinking of adding taas to sfc?
17:52:58 <reedip-home> Kind of
17:53:31 <reedip-home> I do not have the complete picture but I wanted to know if that is possible
17:53:49 <LouisF> do you have the taas requirements?
17:53:53 <reedip-home> In the O/P release??
17:54:39 <s3wong> reedip-home: don't know much about TaaS; is it a standalone service (in VM, for example, with Neutron ports), or is it just OVS redirecting traffic to a collector?
17:55:21 <LouisF> reedip-home: is there a description or spec?
17:55:27 <reedip-home> s3wong : taas is being deployed ad a service
17:55:55 <reedip-home> LouisF, s3wong: hold on
17:56:00 <s3wong> reedip-home: with Neutron ports associated?
17:56:42 <reedip-home> #link : https://github.com/openstack/tap-as-a-service
17:56:56 <reedip-home> s3wong: yes, a neutron network
17:57:44 <LouisF> reedip-home: we can investigate and see how sfc and taas can potentially integrate
17:58:21 <reedip-home> LouisF: lets have a discussion on this , if possible on the ML and the weekly Meetings
17:58:24 <s3wong> reedip-home: so your intent is that instead of having users explicitly place the TaaS Neutron ports as part of the chain, you want to have networking-sfc APIs/resources to automatically insert tap into the chain?
17:58:44 <reedip-home> s3wong : I believe so
17:59:15 <reedip-home> The time is short so therefore we can move this discussion for other forums.
17:59:37 <LouisF> reedip-home: we can discuss offline
17:59:45 <LouisF> ok thanks all
17:59:52 <s3wong> reedip-home, LouisF: ML, I guess?
17:59:53 <LouisF> bye
18:00:05 <LouisF> s3wong: yes
18:00:11 <bcafarel> bye
18:00:18 <s3wong> bye
18:00:21 <LouisF> #endmeeting