16:02:32 <Sukhdev> #startmeeting ironic_neutron
16:02:33 <openstack> Meeting started Mon Nov 14 16:02:32 2016 UTC and is due to finish in 60 minutes.  The chair is Sukhdev. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:34 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:02:36 <openstack> The meeting name has been set to 'ironic_neutron'
16:02:54 <Sukhdev> #topic: Agenda
16:02:59 <Sukhdev> #link: https://wiki.openstack.org/wiki/Meetings/Ironic-neutron#Meeting_November_14.2C_2016
16:03:30 <Sukhdev> anybody wants anything added to the agenda?
16:03:50 <Sukhdev> #topic: Announcements
16:04:03 <Sukhdev> any announcements?
16:04:23 <sambetts> Attach/Detach VIF spec landed
16:04:33 <sambetts> so those patches need updating and reviewing
16:05:11 <Sukhdev> cool
16:05:15 <sambetts> vasyl is helping me update the client side patch because the API design changed at the last minute
16:05:41 <sambetts> I'm nearly done updating the Ironic side
16:06:45 <Sukhdev> sambetts : thanks
16:07:23 <Sukhdev> I spent some time last weekend organizing the wiki to reflect all the correct patches
16:07:58 <jroll> \o/
16:08:09 <Sukhdev> I should not call these patches POC code anymore :-)
16:08:34 <Sukhdev> make sure all the patches are listed there - it is easier to keep track of them
16:09:08 <Sukhdev> #topic: Security Groups
16:09:39 <jroll> this one is close
16:09:42 <Sukhdev> I noticed last week there were lot of reviews of the SG patches
16:09:45 <jroll> just missing some test cases
16:10:06 <Sukhdev> jroll : I am confused about the test cases that are needed
16:10:27 <jroll> Sukhdev: what's confusing?
16:10:58 <Sukhdev> jroll : about the neutron client setup for SG
16:11:14 <Sukhdev> for the verification
16:11:33 <jroll> not sure what you mean
16:11:44 <jroll> my comments say what is missing
16:12:32 <Sukhdev> jroll : not about what is missing - how to go about part - let me ping you offline
16:12:40 <jroll> Sukhdev: sure, sounds good
16:13:06 <Sukhdev> other than that this is pretty much good to go
16:13:54 <Sukhdev> I will try to add tests for verification part and upload it
16:14:16 <Sukhdev> any body else has any question about it other than test?
16:14:35 <Sukhdev> #topic: Port Groups
16:14:51 <Sukhdev> Port Groups is also looking good
16:15:03 <Sukhdev> Ironic patches are merged -
16:15:36 <jroll> there's still a few ironic patches https://review.openstack.org/#/q/topic:bug/1618754
16:16:20 <jroll> but yeah, we're getting there
16:16:30 <jroll> only one is code, the rest are docs/tests
16:16:43 <Sukhdev> right
16:17:32 <Sukhdev> nova side
16:18:01 <Sukhdev> nova side is sitting with -ve votes
16:18:10 <jroll> nova spec is looking good, hoping it'll get some eyes this week https://review.openstack.org/#/c/387534/
16:18:15 <jroll> nova spec freeze is thursday
16:19:18 <Sukhdev> we have few days - hopefully, this go
16:19:34 <jroll> ya :)
16:21:06 <Sukhdev> anything else on PG?
16:21:28 <Sukhdev> #topoic: Vlan Aware Servers
16:22:27 <sambetts> I don't think that worked :-p
16:22:41 <jroll> honestly I don't expect to see this make progress until late ocata or pike
16:22:55 <jroll> gotta finish attach/detach and portgroups first
16:23:45 <jroll> sambetts: agree?
16:23:58 <sambetts> Yeah, they are certainly higher priority and we aren't going to get the nova spec for the configdrive in before freeze
16:25:03 <sambetts> other than that part everything else should be Ironic side
16:25:23 <jroll> yep
16:26:03 <Sukhdev> that naturally leads to the next topic -
16:26:25 <Sukhdev> #topic: Attach/detach API
16:26:45 <Sukhdev> as sambetts pointed out - the spec is now approved
16:27:14 <jroll> as well as the nova blueprint
16:27:20 <jroll> so time to get moving on code :D
16:27:35 <Sukhdev> sambetts : I listed three patches on the wiki (nova, ironic, client)
16:27:38 <jroll> we also decided nova portgroups depends on this, it will make it much simpler to implement
16:28:34 <sambetts> Yeah and that was a question I had, do you want my attach/detach spec API patch to Ironic to include the portgroups mapping logic and the port mapping logic or just the port part with portgroups in a follow up
16:28:59 <sambetts> s/spec//
16:29:38 <jroll> sambetts: not sure, I suspect the latter, just for smaller patches
16:29:50 <sambetts> yeah thats what I was thinking
16:31:18 <Sukhdev> we need it for both though
16:31:34 <jroll> nobody disagrees :)
16:32:13 <sambetts> Cool, the status of the patches are: Ironic one is nearly finished being updated
16:32:27 <sambetts> Ironic Client one Vasyl was going to update for me
16:32:46 <jroll> nice
16:33:05 <sambetts> and Nova one will need updating once client functions are laided out
16:35:13 <Sukhdev> sounds reasonable - anything else on this?
16:35:42 <sambetts> Nothing else from me
16:35:48 <sambetts> just pushing on with it
16:36:07 <jroll> ditto
16:36:27 <Sukhdev> with my change in job, I am trying to set up the BM environment so that I can test all of this
16:36:37 <Sukhdev> I lost my setup when I left Arista :-)
16:36:59 <Sukhdev> #topic: Open Discussion
16:37:14 * sambetts is also working to build a new test environment after mine was aquired by another team :(
16:37:59 <Sukhdev> sambetts : ha ha - it can be painful
16:38:25 <sambetts> yeah
16:38:38 <Sukhdev> anybody wants to discuss anything else networking related?
16:38:48 <sambetts> btw what are the plans/goals for 3rd party CI of the network drivers + ml2 supprt?
16:39:04 <sambetts> have we laid out any rules for that?
16:39:21 <sambetts> s/rules/goals
16:39:48 <jroll> we haven't
16:39:56 <jroll> it really shouldn't differ based on driver, right?
16:40:29 <sambetts> no, jsut different local_link_connection info to be passed for different vendors
16:41:12 <Sukhdev> sambetts : that is part of the CI test setup
16:41:31 <sambetts> the unit tests?
16:41:55 <jroll> sambetts: hrm, those vendors don't have code in ironic, so seems like it'd be weird to ask them for CI
16:42:04 <Sukhdev> I can give you example of how we did at Arista
16:42:36 <sambetts> jroll: yeah :/ its more of a cross project ml2 driver repo X ironic CI I guess, like how nova CI's ironic
16:43:07 <sambetts> so networking-cisco should proably CI Ironic or the other way around
16:43:53 <jroll> ya
16:43:56 <Sukhdev> sambetts : you do not need ironic specific CI - unless you want to test the internals of ironic through CI
16:43:57 <jroll> the first, I think
16:44:33 <sambetts> yeah that seems to make the most sense, to make sure nothing that merges into net-cisco breaks the intergration with ironic
16:45:00 <Sukhdev> yup - so, there are two parts to it sambetts
16:45:28 <Sukhdev> 1) the merges into networking-cisco - those are your local unit tests
16:45:56 <Sukhdev> 2) the CI to catch the issues with upstream - i.e at the API integration level
16:46:38 <Sukhdev> that is the way I had setup for Arista - it works very effectively
16:46:47 <Sukhdev> unless, you want an overkill :-):-)
16:47:50 <sambetts> yeah, I'll have to play around and see what extra equipment I can squeeze into my CI environment :)
16:50:11 <sambetts> thats all I had today
16:50:21 <Sukhdev> anything else - anybody?
16:50:33 <jroll> nope :)
16:50:43 <Sukhdev> cool - we are done
16:50:47 <Sukhdev> thanks
16:50:51 <jroll> thanks Sukhdev
16:50:59 <Sukhdev> #endmeeting