17:01:30 <LouisF> #startmeeting service_chaining 17:01:31 <openstack> Meeting started Thu Oct 26 17:01:30 2017 UTC and is due to finish in 60 minutes. The chair is LouisF. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:32 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:01:34 <openstack> The meeting name has been set to 'service_chaining' 17:01:41 <LouisF> hi all 17:01:48 <LouisF> #topic Agenda 17:02:10 <LouisF> https://wiki.openstack.org/wiki/Meetings/ServiceFunctionChainingMeeting#Agenda_for_the_Networking-SFC_Meeting_.2810.2F26.2F2017.29 17:03:27 <LouisF> Good progress this week - several merges 17:04:06 <LouisF> #topic Zuul v3 migration 17:04:29 <vks1> hi LouisF 17:04:33 <vks1> sorry I am late 17:04:57 <LouisF> vks1: np, you are first 17:05:02 <vks1> bcafarel amy not be joining today 17:05:10 <LouisF> Bernard and Igor cannot make it 17:05:13 <vks1> I talked to him 17:05:17 <LouisF> vks1: yes 17:05:31 <LouisF> got email from Bernard on status 17:05:34 <LouisF> Milestone 1 release was finally published (some last zuulv3 problems, but not from us): https://review.openstack.org/#/c/512432/ 17:05:50 <LouisF> zuul v3: current status is stable (green everywhere, including python-neutronclient), next week I will resume the work to move job definition in networking-sfc repo (and clean the legacy ones) 17:06:20 <vks1> ok 17:06:22 <LouisF> #topic Tempest work 17:06:40 <LouisF> Tempest plugin: no feedback on http://lists.openstack.org/pipermail/openstack-dev/2017-October/123814.html 17:06:54 <vks1> I saw that 17:07:06 <LouisF> Bernard is going to check with Armando on status 17:07:33 <vks1> ok 17:07:37 <LouisF> #topic Tap SF 17:08:06 <LouisF> API and driver changes merged! 17:08:14 <vks1> yeah thanks :) 17:08:23 <vks1> neutronclient is pending to be merged 17:08:26 <LouisF> https://review.openstack.org/#/c/449954/ 17:08:43 <LouisF> https://review.openstack.org/#/c/465057/ 17:08:49 <LouisF> thanks for your help 17:09:01 <LouisF> what is link? 17:09:18 <vks1> https://review.openstack.org/#/c/491185/ 17:10:08 <LouisF> thanks 17:10:15 <LouisF> #topic SFC graph 17:10:36 <LouisF> amotoki had some feedback on the CLI patch, mostly minor items 17:10:55 <LouisF> python-neutronclient patches: amotoki has reviewed them, and will cut a new release when they get in https://review.openstack.org/#/q/project:openstack/python-neutronclient+%22sfc+plugin%22 (bug fixes) https://review.openstack.org/#/c/491185/ (tap CLI) https://review.openstack.org/#/c/487190/ (graphs CLI) 17:11:32 <LouisF> #topic NSH dataplane 17:11:45 <LouisF> https://review.openstack.org/#/c/373465/ needs rebase after the tap merge 17:12:26 <LouisF> #topic Non-transparent SF 17:12:42 <LouisF> new patch coming soon 17:12:51 <LouisF> #Other items 17:13:06 <LouisF> vks1: anything else? 17:13:10 <vks1> me and bcafarel were discussing on dead SF detection 17:13:26 <LouisF> this will be a quick meeting 17:13:35 <vks1> dead peer detection 17:13:52 <LouisF> can you elaborate? 17:14:16 <vks1> detection of SH which are not reachable in chain 17:14:32 <vks1> I think that's what you also suggested in last to last meetings 17:14:52 <vks1> I saw a bug also in launchpad related to that 17:14:59 <vks1> not sure that was the same bug 17:15:11 <LouisF> vks1: yes we need to look at that 17:15:34 <vks1> I think I can pursue that 17:15:51 <vks1> bcafarel also said he will also contribute 17:16:07 <vks1> do you have any thoughts on to start with 17:16:18 <vks1> probably starting some discussion doc 17:16:25 <vks1> on something on etherpad 17:16:37 <LouisF> there a couple of options: 1. use another service to monitor SF liveness, 2. monitor it directly from upstream SFF 17:16:40 <vks1> or pulling up a bp 17:16:57 <LouisF> vks1: yes etherpad is good 17:17:53 <vks1> how should we start then ? 17:18:55 <LouisF> vks1: describe possible options, say, one using ceilometer 17:19:27 <LouisF> we need to add SFC metrics also, which can be retrieved using ceilometer 17:19:45 <vks1> but how come ceilometer periodically check SFs 17:19:47 <vks1> ? 17:20:01 <LouisF> the VMs 17:20:06 <vks1> yeah 17:20:45 <LouisF> anyway lots of work to do 17:20:46 <vks1> the VMs failure detection will require some sort n/w probe 17:21:07 <LouisF> that must exist already? 17:21:19 <vks1> oh let me check that 17:21:27 <LouisF> ok 17:21:42 <LouisF> needs investigation 17:21:58 <LouisF> thats all i have today 17:22:27 <LouisF> i will be at the Summit for next two weeks and Bernard has offered to run the meetings 17:22:27 <vks1> ceilometer has health mon support 17:22:32 <vks1> OK 17:23:13 <LouisF> vks1: yes so we can use that and use that to control the SF weights in the PPG 17:23:23 <vks1> yeah 17:23:35 <LouisF> is a SF VM fails set its weight to 0 17:24:25 <LouisF> vks1: thanks 17:24:26 <LouisF> bye 17:24:34 <vks1> LouisF: bYE 17:24:45 <LouisF> #endmeeting