18:32:08 <sc68cal> #startmeeting networking_fwaas
18:32:09 <openstack> Meeting started Wed Mar 16 18:32:08 2016 UTC and is due to finish in 60 minutes.  The chair is sc68cal. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:32:10 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:32:13 <openstack> The meeting name has been set to 'networking_fwaas'
18:32:22 <annp> hi sc68cal, all
18:32:31 <sc68cal> #chair SridarK
18:32:32 <openstack> Current chairs: SridarK sc68cal
18:32:34 <sc68cal> #chair xgerman
18:32:35 <openstack> Current chairs: SridarK sc68cal xgerman
18:32:44 <SridarK> Hi All
18:33:33 <sc68cal> Today I think is mostly just working on the actual release of mitaka
18:34:38 <SridarK> sc68cal: i think there is nothing pending w.r.t the release
18:34:47 <sc68cal> agreed
18:36:14 <sc68cal> so I think once the release happens, hopefully capacity will open back up and development can resume on the v2 api
18:36:46 <SridarK> sc68cal: yes hope so too
18:37:11 <SridarK> i guess the branch should open up for Newton real soon ?
18:39:35 <sc68cal> I would assume so. I think part of the problem is that none of the cores have done a release, it's always been handled for us, by either mestery or ihar
18:40:07 <SridarK> ok
18:40:40 <sc68cal> It appears we have a stable/mitaka branch cut
18:41:15 <sc68cal> so, I assume that we're open for newton based on that fact
18:41:18 <SridarK> ok cool
18:41:45 <SridarK> In terms of the N release,  we should get #link https://review.openstack.org/#/c/278863/ for observer hierarchy
18:42:11 <SridarK> the submitter it seems also got sucked into other things
18:42:36 <SridarK> but will be good to close it out early as a long standing thing
18:43:29 <sc68cal> yes, that patch is very important. SridarK or jwarendt - do you think you guys could push revisions if the submitter does not reappear?
18:43:55 <SridarK> sc68cal: yes will do so
18:44:53 <jwarendt> Yes
18:45:01 <xgerman> o/
18:45:36 <SridarK> jwarendt: i think u have been looking at the neutron (L3Agent) side of things
18:47:22 <jwarendt> Correct, but the agent.py on neutron side has direct calls right now into FWaaS class - instead of decoupled observer.  I have made the lookup dynamic, but observer would be much better.
18:48:08 <SridarK> jwarendt: yes agreed
18:48:35 <SridarK> i will reach out the submitter again to check on his availability and we can move forward accordingly
18:48:50 <SridarK> *reach out to the ..
18:49:00 <jwarendt> Plus for some reason there is a router_add but no router_delete on L3 agent side; hoping observer will properly listen to delete so fwaas can change state appropriately.
18:49:53 <SridarK> jwarendt: yes the router delete - earlier there was not much to do as the iptables will get cleaned out
18:50:14 <xgerman> well, we still need to remove it from the DB
18:50:17 <SridarK> jwarendt: now with router insertion things are a bit different - but yes that is a clean up
18:51:10 <SridarK> xgerman: yes i believe we have Foreign Key constraint
18:51:28 <SridarK> but let me look at that code again
18:51:31 <xgerman> ok, that would work
18:51:42 <SridarK> from the agent side of the world
18:52:16 <SridarK> Okay good i think we are all on the same page
18:53:35 <jwarendt> An update blocks if router in use, but can delete the router underneath the FWaaS without the FWaaS changing from ACTIVE state - even if the DB cleans up.
18:54:32 <jwarendt> Observers will help straighten that out.
18:54:42 <SridarK> jwarendt: yes agreed
18:54:49 <xgerman> +1
18:56:07 <sc68cal> cool
18:56:22 <SridarK> one other patch i wanted to discuss: #link https://review.openstack.org/#/c/291822/
18:56:22 <sc68cal> The other thing I wanted to bring up, is the state of 3rd party drivers
18:56:36 <SridarK> sc68cal: u read my mind :-)
18:56:38 <sc68cal> :)
18:56:55 <sc68cal> Now - macafee's driver, they actually have a CI job that passes reliably
18:57:09 <sc68cal> but the code itself needs a maintainer - to fix pylint issues
18:57:26 <SridarK> sc68cal: on this, i reached out to intel - there is some churn in ownership of McAfee - so things are in a bit of flux
18:57:55 <SridarK> I asked if they could take care of the pylint issues - which was one of ur main concerns
18:58:02 <sc68cal> yep
18:58:26 <SridarK> hopefully they can help with transtioning this to the new maintainers
18:58:27 <sc68cal> If they fix the issues, then that would be a positive step
18:59:04 <sc68cal> but what do we do about drivers like varmor and cisco?
18:59:11 <SridarK> we can consider giving some time until they sort out the maintainers - i will stay in touch with intel folks to help transition
18:59:14 <sc68cal> does the cisco driver have a 3rd party CI?
18:59:36 <SridarK> sc68cal: yes but i will be first to admit - the reliability is a bit suspect
18:59:53 <sc68cal> I haven't seen it report on a recent patch.
18:59:57 <sc68cal> or any, really.
19:00:05 <SridarK> sc68cal: i think we can start having vendors pull out to their own rep
19:00:08 <SridarK> *repo
19:00:47 <jwarendt> +1
19:01:03 <sc68cal> yeah, that would be good. The main thing is that we're supposed to vouch for the quality of these drivers, that are in tree - and to be honest the only one we can vouch for is the l3reference, since that gets tested at the openstack gate
19:01:15 <SridarK> sc68cal: yes earlier - we had some control on the ownership of the CI - now its in another part of Cisco
19:01:18 <SridarK> sc68cal: +1
19:01:49 <SridarK> and most vendors do have their networking-vendor repos
19:04:22 <SridarK> Would it be reasonable to ask if vendors can move their drivers out by N-2 ?
19:04:36 <sc68cal> that sounds good to me
19:05:07 <xgerman> +1
19:06:58 <SridarK> ok i will ping the vendor contacts as the first step
19:07:03 <xgerman> though in LBaaS we decided to leave them in
19:07:15 <xgerman> some vendors see it important to be part of the repo...
19:07:19 <SridarK> xgerman: hmm ok
19:07:39 <xgerman> we should ping them anyway and if they resist we can revisit
19:07:58 <SridarK> i am not religious on this, i thought in neutron in general the direction with vendor decomp - we want to ease vendors out
19:08:06 <SridarK> xgerman: yes sounds good to me
19:08:30 <sc68cal> I'm not religious either, but the pylint thing just highlighted a driver that needs an owner
19:09:24 <xgerman> agreed.
19:11:26 <SridarK> ok those were some things i wanted to discuss - did not have anything else - i can think of right now.
19:12:12 <sc68cal> that's all from me too. Anyone else?
19:12:26 <xgerman> well, we should mention that PTL proposals are open.. so if anyone wants to run...
19:12:52 <xgerman> for the PTL of Neutron...
19:13:27 <sc68cal> I wish anyone who wants to, all the best
19:13:49 * sc68cal leans back in his lazy boy chair and sips his drink
19:13:56 <SridarK> +1
19:15:56 <sc68cal> unless there is anything else, I think we can close this out and give everyone 15 back
19:16:01 <xgerman> +1
19:16:04 <SridarK> +1
19:16:29 <sc68cal> #endmeeting