07:00:26 <yamamoto> #startmeeting networking_midonet
07:00:26 <openstack> Meeting started Tue Feb 23 07:00:26 2016 UTC and is due to finish in 60 minutes.  The chair is yamamoto. Information about MeetBot at http://wiki.debian.org/MeetBot.
07:00:27 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
07:00:29 <kitsuneninetails> hello!
07:00:30 <openstack> The meeting name has been set to 'networking_midonet'
07:00:42 <yamamoto> #topic Agenda
07:00:44 <yamamoto> #link https://wiki.openstack.org/wiki/Meetings/NetworkingMidoNet
07:00:56 <yamamoto> #topic Announcements
07:01:07 <yamamoto> i have no announcements.  does anyone have any?
07:02:09 <yamamoto> #topic Bugs
07:02:42 <yamamoto> there's no bugs which is critical and new
07:02:50 <ryu25> nice!
07:03:04 <yamamoto> does anyone have any bug which needs wider attention?
07:03:29 <ryu25> i don't
07:04:33 <yamamoto> #topic make tempest jobs voting?
07:04:47 <yamamoto> i propose to make our tempest jobs voting
07:05:23 <ryu25> i'm ok with that
07:05:26 <yamamoto> it has been non-voting for reasons but we now can disable tests with devstackgaterc
07:05:35 <ryu25> right we can take fast action
07:06:05 <yamamoto> i'll submit project-config patch.
07:06:06 <kitsuneninetails> is this just a tempest job via devstack?
07:06:19 <kitsuneninetails> so, single-node all-in-one?
07:06:27 <kitsuneninetails> and just upstream tempest tests?
07:06:30 <yamamoto> kitsuneninetails: yes
07:08:25 <yamamoto> #topic what to do for stadium concept discussion?
07:08:45 <yamamoto> #link https://review.openstack.org/#/c/281628/
07:09:22 <yamamoto> ryu made a commet on the review.  thank you.
07:10:29 <ryu25> yeah, I am conflicted about this patch
07:10:42 <yamamoto> my understanding is ideally we want "midonet" as a big-tent project and make networking-midonet a part of it, right?
07:11:31 <ryu25> they don't necessary have to be the same, but i think it makes sense to put them together if they can be
07:11:41 <yamamoto> so being out of neutron staduim is not bad.
07:12:01 <ryu25> if networking-midonet can become a big tent project, then we can always add midonet there as well right?
07:12:27 <ryu25> i think that is more realistic
07:12:29 <yamamoto> i'm not sure.  i guess we should avoid playing tricks.
07:13:16 <ryu25> it may be considered trickery but i don't know if it really is
07:13:33 <ryu25> but i guess that's a conversation to have later
07:13:35 <yamamoto> ryu25: do you know what was a blocker of midonet as a big-tent project?
07:14:06 <ryu25> not in detail, but at one point, i believe Java/Scala got in the way
07:14:15 <ryu25> though apparently that was not really a true blocker
07:14:39 <ryu25> but it also demanded that the openstack infrastructure is updated to handle java projects
07:14:46 <yamamoto> then, just a matter of priority on our side?
07:15:20 <ryu25> that and also a question to whether such effort is really worthy at all
07:15:25 <ryu25> i can't really answer that part myself
07:16:01 <yamamoto> i see
07:16:30 <ryu25> but i would say that it's still something to consider
07:16:35 <ryu25> i think the conversation has stopped though
07:16:41 <ryu25> but not completely dead
07:17:00 <ryu25> so we can leave it off as a priority issue for now and that's not far from the truth
07:17:13 <yamamoto> ok
07:17:17 <yamamoto> thank you for explanation
07:17:40 <ryu25> there may be other reasons that I'm not aware of though
07:17:42 <ryu25> np!
07:17:50 <yamamoto> let's move on
07:17:50 <yamamoto> #topic Open Discussion
07:18:24 <ryu25> python-midonetclient repository will be removed (or at least deprecated)
07:18:54 <ryu25> today i tried doing pip install of networking-midonet and got stuck with midonetclient dependency missing.  And noticed that python-midonetclient package is not hosted in pypi
07:19:07 <ryu25> anyone know anything about that?
07:19:58 <ryu25> btw, I still think that we can move the python client code used in neutron to networking-midonet project and lose the dependency to python-midonetclient completely... but that is a separate discussion
07:20:30 <yamamoto> we haven't had it on pypi afaik
07:20:30 <ryu25> i'll be happy to discuss it offline if anyone is interested
07:20:49 <ryu25> really?  i thought Jaume put it up there at one point.  Maybe i'm forgetting something
07:20:54 <ryu25> ok well that explains it
07:21:18 <yamamoto> why the repository will be removed?
07:21:33 <yamamoto> in favor of the one in midonet repo?
07:22:19 <ryu25> yes, that is the preferred approach (not mine) and it will not be changed
07:23:23 <yamamoto> #link https://review.openstack.org/#/c/283026/
07:23:32 <yamamoto> is this a part of the move?
07:24:09 <ryu25> yes i believe so
07:25:01 <ryu25> this is one reason why I prefer to move the neutron API code out of midonet and into networking-midonet
07:25:12 <ryu25> the dependency is very small
07:26:12 <yamamoto> are there folks to prefer to keep it in midonet?
07:27:02 <ryu25> yes, at least I know the main midonet contributors prefer it that way
07:27:12 <ryu25> oh wait, are you referring to the neutron client code?
07:27:19 <ryu25> that one, i don't think anyone cares too strongly
07:27:28 <ryu25> python-midonetclient project, people have strogn preferences
07:28:04 <yamamoto> yes, my question was about neutronclient
07:28:17 <yamamoto> maybe mdts or something uses it?
07:28:20 <ryu25> yes there are strong supporters.
07:28:27 <ryu25> yup mdts does
07:30:34 <yamamoto> i tend to think mdts and python-midonetclient should be out of midonet repo but i guess this meeting is not right place to discuss.
07:31:12 <ryu25> i agree with that though i certainly see their view with regards to simplifying maintenance
07:32:06 <yamamoto> kitsuneninetails: do you have any idea when zephyr gets ready to consume for networking-midonet gate jobs?
07:32:58 <kitsuneninetails> would it need to run in a devstack env?
07:33:29 <kitsuneninetails> if it can just run by itself and execute tests, whenever we want to create the jenkins jobs for it should be fine
07:33:45 <kitsuneninetails> if it has to run with devstack, it'll be a little bit
07:34:08 <ryu25> kitsuneninetails: that would be very cool to be able to run with devstack
07:34:33 <kitsuneninetails> It's planned, but it'll take some dev work
07:35:35 <ryu25> understood, and i'll be more than happy to help out on that effort
07:35:37 <yamamoto> kitsuneninetails: it doesn't have a way to run against an existing deployment?
07:36:56 <kitsuneninetails> it's planned, but no, not currently.  Currently it uses the running neutron deployment, but will bring up and tear down the midonet environment (including tearing down any neutron data along with the midonet-specific data)
07:38:10 <kitsuneninetails> I am trying to get it running against an existing env without creation/teardown, but devstack is hard-locked to using "screen", and I am trying to come up with a way to send a ctrl-C to a screen session so I can restart midolman when necessary
07:38:55 <yamamoto> do you need to restart neutron server as well?
07:40:15 <joejoe> kitsuneninetails: I wonder if killing/restarting the screen session itself would work
07:40:59 <kitsuneninetails> I will try that out and play with it a bit.  I might not even need to restart, now that I think about it some more
07:41:10 <yamamoto> i tend to think networking-midonet jobs should use midonet from packages rather than via devmido in the first place.
07:41:12 <kitsuneninetails> I will have to devote some time to getting it to work with devstack
07:41:55 <ryu25> yamamoto: i agree on the CI jobs.  devmido was created only if you want to do parallel development
07:42:39 <ryu25> kitsuneninetails: do all tests require agents to be restarted?
07:42:51 <ryu25> is that part of the setup or something?
07:43:27 <kitsuneninetails> just a part of the physical layer startup, not the test system
07:43:39 <kitsuneninetails> test system just assumes everything is running
07:43:57 <ryu25> ah ok so with some tweaks, we may just be able to run these tests with a devstack environment
07:45:14 <kitsuneninetails> basically, we just need a PTM manager which just sets itself up to an already running physical layer, rather than setting itself up.
07:47:10 <yamamoto> kitsuneninetails: thank you for explanation
07:47:24 <yamamoto> does anyone have any other topic to discuss?
07:47:36 <ryu25> nope
07:47:52 <yamamoto> let me go back to bugs
07:47:58 <yamamoto> #topic Bugs
07:48:13 <yamamoto> any volunteer for bug deputy for this week?
07:49:01 <ryu25> i haven't done it in a while, so i can take it if no one volunteers
07:49:46 <yamamoto> it might be time to revisit if the bug deputy process is worth for us though.
07:50:13 <yamamoto> ryu25: thank you
07:52:36 <yamamoto> let me give you back a several minutes
07:52:41 <yamamoto> thank you for attending
07:52:44 <yamamoto> #endmeeting