16:01:44 <Sukhdev> #startmeeting: ironic_neutron
16:01:45 <hshiina> o/
16:01:45 <openstack> Meeting started Mon Aug 22 16:01:44 2016 UTC and is due to finish in 60 minutes.  The chair is Sukhdev. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:46 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:48 <openstack> The meeting name has been set to '__ironic_neutron'
16:02:07 <Sukhdev> #topic: Agenda
16:02:12 <Sukhdev> #link: https://wiki.openstack.org/wiki/Meetings/Ironic-neutron#Meeting_August_22.2C_2016
16:02:35 <Sukhdev> todays agenda is short and sweet - should be done quick
16:02:35 <devananda> o/
16:02:39 * devananda lurks from a cafe in NYC
16:02:41 <Sukhdev> feel free to add
16:03:14 <Sukhdev> #topic: Announcements
16:03:30 <sambetts> o/
16:03:37 <Sukhdev> I will be out next week - gone to VmWorld in Las Vegas
16:03:45 <mjturek1> enjoy!
16:03:50 <Sukhdev> Need a volunteer to chair the meeting
16:04:06 <jroll> I could do that
16:04:19 <Sukhdev> jroll : perfect - thanks
16:04:24 <jroll> np
16:05:27 <Sukhdev> BTW, I will be presenting Bare Metal deployments in multi-tenant data centers in Future:Net next week -
16:06:03 <Sukhdev> Future:net is running in parallel to VMWorld
16:06:27 <Sukhdev> Anybody has any other announcement?
16:06:39 <jroll> awesome :)
16:07:21 <Sukhdev> it seems it is a very hot topic - there is another speaker on Bare metals topic as well
16:07:38 <mjturek1> Sukhdev: any chance it'll be recorded?
16:08:24 <Sukhdev> mjturek1 : I have no idea - but, they have sent me instructions as if it will recorded
16:08:45 <mjturek1> ahhh, would love to hear it if it does get recorded :)
16:09:01 <Sukhdev> mjturek1 : http://www.vmware.com/futurenet
16:09:35 <mjturek1> very cool - thx
16:09:46 <Sukhdev> moving on -
16:09:57 <Sukhdev> today's agenda is short
16:10:05 <Sukhdev> Two items to cover -
16:10:15 <Sukhdev> 1) PortGroups
16:10:25 <Sukhdev> 2) Security Groups
16:10:33 <mjturek1> for portgroups - in response to last week's meeting I drafted up this nova patch https://review.openstack.org/#/c/357222 (WIP because of a failing test case)
16:10:51 <mjturek1> would like to know if people think this is sufficient
16:11:01 <Sukhdev> #link: https://review.openstack.org/#/c/357222
16:12:20 <sambetts> seems like a resonable way to go from a quick scan
16:12:26 <jroll> mjturek1: I saw that this morning, I think it's okay... checking one thing
16:12:34 <mjturek1> cool cool
16:12:54 <jroll> ok yeah, I think that's fine
16:13:01 <jroll> the other thing we could do is just not present resources for it
16:13:14 <jroll> which might be more confusing for ops, but less annoying for nova users
16:13:30 <sambetts> as in don't return the ports that have port group ids on the port list?
16:13:50 <mjturek1> jroll: think it'd be as likely to land?
16:13:54 <jroll> er, sorry, don't present resources for a node with portgroups
16:14:03 <jroll> e.g. here https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L259
16:14:25 <jroll> mjturek1: idk, it's an extra api call in a pretty hot path, it's worth the conversation though
16:14:50 <sambetts> I think we could do in on the Ironic side using the API versions if we hide ports that have port_group_ids from API versions prior to supporting that field right?
16:15:44 <jroll> ehhhhhh, seems super hacky
16:16:27 <sambetts> yeah :/ just throwing it out as an alternative if we didn't want to put that logic in Nova
16:17:57 <sambetts> the fact we're having to bump the API version so nova understands portgroups just enough to say hey I don't understand those, seems to defeat the point of having the API version there
16:18:30 <jroll> yeah, it's a bit odd
16:19:07 <mjturek1> sambetts: yeah it is a bit strange.
16:19:20 <sambetts> another alternative is to prevent vif_port_id being set on a port that has a portgroup_id on the Ironic side
16:19:48 <sambetts> then Ironic will throw a you can do this error at that point
16:20:13 <sambetts> or a 400 you should be using my portgroup... error
16:20:56 <sambetts> then no version bumb is required, it'll just error at the port.update call
16:21:09 <sambetts> bump*
16:21:29 <sambetts> thats ^ less hacking and probably useful anyway
16:21:40 <sambetts> outside this scenario
16:22:03 <jroll> yeah, maybe we leave a comment on that patch?
16:22:16 <jroll> that does seem useful
16:23:08 <sambetts> leaving a comment now
16:23:19 <mjturek1> thanks sambetts, I think I understand
16:23:46 <jroll> sambetts: the portgroups API patch, to be clear :)
16:24:25 <Sukhdev> the other two patches on port groups require reviews as well folks
16:24:40 <sambetts> jroll: thanks :)
16:24:40 <Sukhdev> #link: https://review.openstack.org/#/c/347549
16:24:54 <Sukhdev> #link: https://review.openstack.org/#/c/332177/
16:25:54 <Sukhdev> anything else on port groups?
16:26:22 <Sukhdev> jroll : BTW, how much time do we have for the cut-off of Neutron release?
16:26:33 <jroll> Sukhdev: Newton?
16:26:54 <Sukhdev> yes - sorry.... I always mess up neutron newton :-)
16:27:11 <jroll> I want to cut off portgroups at feature freeze, I think
16:27:16 <Sukhdev> fingers are so tuned to typing neutron
16:27:17 <jroll> which is the week ending sept 2
16:27:19 <jroll> it's a huge patch
16:27:31 <jroll> *maybe* a week after that, but idk
16:27:39 <jroll> #link http://releases.openstack.org/newton/schedule.html
16:27:42 <jroll> is the newton schedule
16:28:02 <sambetts> Looking at the portgroups patches I'm confused by the commit messages, both seem to say they are adding the portgroups API
16:28:56 <mjturek1> sambetts: sorry, the first patch is just adding the portgroups API controller
16:29:02 <mjturek1> no interaction with ports or nodes
16:29:22 <Sukhdev> jroll : thanks - we do have some
16:29:39 <Sukhdev> time
16:29:46 <mjturek1> sambetts: I'll clear up the messages today
16:30:12 <sambetts> mjturek1: thanks :) it seems like they need a rebase too
16:30:22 <mjturek1> will do
16:30:43 <jroll> Sukhdev: this week and next week
16:31:44 <Sukhdev> jroll : puts pressure on me to get the security groups patch completed this week - as I am out next week - but, one week is sufficient :-):-)
16:32:01 <Sukhdev> anything on port groups?
16:32:06 <jroll> Sukhdev: assuming reviewers aren't insanely busy, it's sufficient
16:32:09 <jroll> (hint: they are)
16:32:52 <Sukhdev> :-)
16:33:13 <jroll> we have priorities and lots of other things to do, please understand if it doesn't make it
16:33:35 <Sukhdev> Understood
16:33:57 <Sukhdev> one the second topic - Security Groups - I finally started to work on it - got delayed because of critical customer delivery
16:34:12 <Sukhdev> it is my goal to complete it this week
16:35:10 <Sukhdev> I will be pushing a patch within a day or two for early assessment from you guys
16:35:58 <jroll> cool
16:36:27 <Sukhdev> other than that I do not have any other update
16:36:46 <Sukhdev> #topic: Open Discussion
16:37:13 <Sukhdev> Two weeks to the goal line - as jroll mentioned
16:37:19 <Sukhdev> lets give it a final push
16:37:32 <mjturek1> \o/
16:37:47 <sambetts> Sukhdev: inspector plugin documentation is here BTW http://docs.openstack.org/developer/ironic-inspector/usage.html#plugins
16:38:04 <sambetts> Sukhdev: see the local_link_connection plugin part :)
16:39:35 <Sukhdev> sambetts : so, I use the ramdisk image created by devstack for my testing - If I modify to pass the lldp parameter - I should be able to test it?
16:40:09 <Sukhdev> #link: http://docs.openstack.org/developer/ironic-inspector/usage.html#plugins
16:40:35 <sambetts> Sukhdev: you have to modify the inspector DHCP configuration to pass the new kernel parameter, and add local_link_connection to the list of enabled plugins in the inspector config
16:42:04 <sambetts> Sukhdev: devstack should be using IPA for inspector already
16:42:38 <sambetts> just need to enable the LLDP collection
16:43:16 <sambetts> and the plugin to handle the new data
16:43:19 <Sukhdev> sambetts : I pulled a fresh install on Friday with devstack - this means I have everything in my setup -
16:44:03 <sambetts> awesome then you only need the two changes to the configs and it should work
16:45:43 <Sukhdev> sambetts : presently my test script sets up the local_link_information by doing manual inspection of physical connectivity (kind of similar to what LLDP would do)
16:45:54 <Sukhdev> at the time of node creation
16:47:22 <Sukhdev> sambetts : ipa-collect-lldp=1 does this go in the ironic config file?
16:47:37 <sambetts> no, the inspector DHCP configuration file
16:48:12 <Sukhdev> sambetts : I will ping you off-line to understand it - this will be really quick for me to test it out
16:48:28 <Sukhdev> sambetts: and will simplify my test scripts
16:48:33 <sambetts> Ok sure :D
16:48:51 <Sukhdev> anything else
16:49:25 <Sukhdev> devananda : enjoy your trip in NYC
16:49:43 <devananda> Sukhdev: thanks! here for conferences this week
16:50:00 <Sukhdev> good weather to be in NYC
16:50:14 <Sukhdev> If nothing else - we are done
16:50:35 <Sukhdev> Reminder jroll will chair this meeting next week
16:50:40 <Sukhdev> Thanks folks
16:50:47 <jroll> thank you Sukhdev
16:50:53 <Sukhdev> bye
16:51:01 <Sukhdev> #endmeeting