16:00:15 <frankbrockners> #startmeeting FDS synch
16:00:15 <collabot> Meeting started Thu Oct 13 16:00:15 2016 UTC.  The chair is frankbrockners. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:15 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:00:15 <collabot> The meeting name has been set to 'fds_synch'
16:00:28 <frankbrockners> #info Frank Brockners
16:00:42 <jlinkes> #info Juraj Linkes
16:01:27 <andy_vanko_> #info Andrej Vanko
16:01:33 <mmarsale> #info Maros Marsalek
16:01:52 <frankbrockners> #info agenda for today: quick check in on key open issues: (a) QEMU / vhost-user reconnect issue (b) HA setup & testing (c) security groups implementation (GBP/HC/VPP)
16:02:21 <frankbrockners> lets start with (a)
16:02:57 <frankbrockners> does anyone have the latest status? Michal got info on a possible resolution from Damjan but needed to patch 16.09
16:03:21 <frankbrockners> andy_vanko_ tomas_c jlinkes?
16:03:43 <frankbrockners> ah -- and another item for the agenda
16:03:45 <trozet> #info Tim Rozet
16:04:13 <frankbrockners> #info (d) status of nosdn-fdio-noha scenario (is the bug that requires neutron restart resolved)?
16:04:16 <jlinkes> #info we tried a bunch of different things for (a), but so far, none worked
16:04:36 <frankbrockners> jlinkes - did we build a new VPP with Damjan's patch included?
16:05:28 <jlinkes> #info we tried vpp 16.12 with Damjan's patch, but that didn't work with HC - mmarsale already solved the issue and we'll try that again tomorrow
16:05:54 <andy_vanko_> <frankbrockners> rechecking with tomas now for any update
16:06:09 <frankbrockners> ok - thanks andy_vanko_
16:06:10 <jlinkes> #info we also tried to build 16.09 vpp with the same code changes, but that didn't work, even when we built qemu 2.7.50
16:06:16 <frankbrockners> let's move to (b) in that case
16:07:15 <frankbrockners> #info marcel deployed nosdn-fdio with HA successfully (though not sure whether any testing was done) and is now trying odl_l2-fdio-ha (with OS HA only - i.e. no ODL clustering)
16:07:34 <frankbrockners> jlinkes, andy_vanko_ - do we know the status?
16:08:01 <tomas_c> #info Tomas Cechvala, sorry for later notice
16:08:13 <frankbrockners> jlinkes - could you also update us on the deployment with HA on the Cisco UCSB lab?
16:09:47 <jlinkes> #info trozet advised to try nofeature scenario to check that have the correct setup - if that deploys correctly we'll investigate what's causing issues in the fdio scenario; if it doesn't deploy correctly, there's most likely an issue with our setup
16:11:37 <frankbrockners> jlinkes - how is the fdio deployment with HA going?
16:12:08 <frankbrockners> (on both CENGN/supermicro and Cisco UCSB)
16:12:41 <jlinkes> the deployment on Cisco UCSB ends with a false positive with improperly setup openstack and database on controllers
16:13:03 <frankbrockners> let's info that...
16:13:07 <frankbrockners> #info the deployment on Cisco UCSB ends with a false positive with improperly setup openstack and database on controllers
16:13:22 <frankbrockners> jlinkes - did you already touch base with trozet on this one?
16:13:55 <trozet> frankbrockners: ^^^^^^^
16:14:03 <jlinkes> frankbrockners: yes, I mentioned that trozet advised to try the nofeature scenario - that's related to this
16:14:15 <frankbrockners> ah - ok
16:14:48 <frankbrockners> got it - but based on the fact that at least on CENGN the nofeature scenario deployed - we should be ok, or?
16:15:47 <jlinkes> that also could've been a false positive
16:16:08 <frankbrockners> ah ok - let's check that and circle back with tim
16:16:25 <frankbrockners> let's briefly touch base on (c) security groups implementation
16:16:31 <andy_vanko_> #info frankbrockners - ref "odl_l2-fdio-ha (with OS HA only - i.e. no ODL clustering)" - no update on this
16:17:03 <jlinkes> that's from Marcel's side, right?
16:17:13 <jlinkes> i.e. Supermicro
16:17:41 <frankbrockners> #info Andrew will provide details on how VPP can be configured to do basic stateless filtering later today on https://wiki.fd.io/view/VPP/SecurityGroups
16:18:35 <frankbrockners> #info those examples will cover basic security groups settings as are used in Functest today (i.e. rules like "allow icmp and ssh for v4 and v6 and block anything else - both on ingress and egress"
16:18:56 <frankbrockners> #info examples will also highlight the associated java APIs.
16:19:22 <frankbrockners> this should be good enough for HC to start testing - as well as GBP to then work via HC
16:19:56 <frankbrockners> tomas_c and mmarsale - any brief update on where we are with implementations on HC and GBP?
16:20:06 <frankbrockners> when could we start end to end testing? Monday?
16:20:10 <mmarsale> #info HC SecGroups phase0: Outstanding issues:
16:20:10 <mmarsale> #info 1. Port ranges - InReview/InTest
16:20:10 <mmarsale> #info 2. Egress ACLs implementation (based on APIs by Ole, no impl though) - ETA: tomorrow, but APIs are already provided by HC.
16:20:10 <mmarsale> #info The rest (mix permit/deny, mix L2/L3) is already done
16:20:30 <mmarsale> #info so Monday looks promising
16:20:43 <frankbrockners> mmarsale - thanks - well prepared as always!
16:21:00 <tomas_c> frankbrockners: still working on it
16:21:11 <tomas_c> just basic tests done
16:21:17 <frankbrockners> tomas_c - does Monday sound feasible?
16:21:34 <tomas_c> yes, it does, i'd like to have it by tomorrow
16:21:38 <tomas_c> tested
16:21:42 <tomas_c> with HC
16:21:44 <frankbrockners> very good - thanks tomas_c
16:21:53 <tomas_c> frankbrockners: np
16:21:58 <frankbrockners> let's move to (d) - nosdn-fdio scenario real quick
16:22:32 <frankbrockners> seanatcisco, fpan - any updates on resolving the open bug which blocks successful testing in OPNFV?
16:23:13 <fpan> yes, there is effort upstream in networking-vpp to resolve this issue, patches in review and testing now
16:23:26 <frankbrockners> let's info that
16:23:58 <frankbrockners> #info nosdn-fdio scenario: there is effort upstream in networking-vpp to resolve the key issue which blocks successful testing in OPNFV, patches in review and testing now
16:24:04 <frankbrockners> thanks fpan
16:24:11 <frankbrockners> is there anything else for today?
16:24:46 <andy_vanko_> frankbrockners - not from me
16:24:55 <frankbrockners> if that's not the case...
16:25:08 * frankbrockners thanks everyone!
16:25:08 <frankbrockners> #endmeeting