22:00:48 <danwent> #startmeeting
22:00:49 <openstack> Meeting started Tue Jan 17 22:00:48 2012 UTC.  The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:00:50 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic.
22:00:53 <danwent> how've you been mark?
22:00:59 <salv-orlando> Hello!
22:01:10 <markvoelker> danwent: doing good thanks!
22:01:21 <somik> hi folks!
22:01:25 <danwent> #info agenda for today's meeting: http://wiki.openstack.org/Network/Meetings
22:01:29 <edgarmagana> ciao
22:01:34 <danwent> #topic General Topics
22:01:34 <mestery> Hi everyone!
22:01:50 <danwent> #info F-series release will be 'folsom'
22:02:11 <danwent> #info Folsom design summit: April 16-18, San Francisco, CA (http://www.openstack.org/conference/san-francisco-2012/)
22:02:20 <GheRivero_> night evryone
22:02:28 <danwent> sign-up is not yet open, but remember to block out the dates on your schedule.
22:02:50 <danwent> #topic Quantum Status
22:03:13 <danwent> #info next tuesday is quantum feature freeze for E-3: https://launchpad.net/quantum/+milestone/essex-3
22:03:46 <salv-orlando> danwent: feature freeze or branch point for E-3?
22:03:52 <danwent> we still have a LOT of issues not yet in review.  want to those things on gerrit by thursday/friday at the latest, if they are non-trivial reviews.
22:04:18 <danwent> salv-orlando: meant the same thing.. yeah, only referring to E-3 here.
22:04:29 <salv-orlando> k, cool
22:04:41 <danwent> we can pull bug fixes into E-3 branch if really needed, but definitely not features (and preferrably, no minor bugs)
22:05:15 <danwent> Thanks to all those who did reviews in the past week, our existing set of reviews is large, but doable, I think
22:05:24 <danwent> #info: current reviews: https://review.openstack.org/#q,status:open+project:openstack/quantum,n,z
22:05:47 <danwent> I wanted to highlight a couple of larger patches that we should start reviewing soon, as they may take some iteration.
22:06:14 <danwent> first, mtaylor will soon be pushing a review for removing the client code from the main repo.
22:06:44 <danwent> this obviously has the potential to be disruptive, so lots of review TLC is in order.
22:07:06 <danwent> Also, Salv's API work has two large reivews:  API filters and API error codes (seem ML for discussion)
22:07:21 <danwent> salv: wanted to get your thoughts on: Bug #803086
22:07:22 <uvirtbot> Launchpad bug 803086 in quantum "plugins.ini should be collapsed into quantum.conf to prevent configuration"sprawl"" [Medium,In progress] https://launchpad.net/bugs/803086
22:07:43 <danwent> seems like there is a lot of review feedback, I was trying to understand the current status
22:08:23 <danwent> salv-orlando?
22:08:37 <salv-orlando> I'm here
22:09:16 <danwent> thoughts on Bug #803086
22:09:17 <uvirtbot> Launchpad bug 803086 in quantum "plugins.ini should be collapsed into quantum.conf to prevent configuration"sprawl"" [Medium,In progress] https://launchpad.net/bugs/803086
22:09:17 <salv-orlando> I read your comment - last time I checked I was pretty sure the "old" behaviour (plugins.ini) still worked. Will have another crack
22:09:22 <salv-orlando> right today
22:09:38 <salv-orlando> Last time I checked was on Jan 10
22:09:42 <danwent> Ok, let me re-run my tests as well.  might have been on an old version or something.
22:09:55 <danwent> Ok, anyone from the Ryu controller team here today?
22:10:49 <danwent> I will respond on the ML in more detail, but my gut feeling is that this code should actually be a separate plugin, as it seems like primary overlap in functionality is just around the SQLalchemy DB model.
22:11:01 <salv-orlando> +1
22:11:18 <danwent> I'd like to create a base "sqlalchemy_plugin_base" class to make it easier for plugins to share DB code.
22:11:31 <danwent> #todo #danwent ML reply to Ryu controller team
22:11:49 <danwent> Ok, any other large reviews out there (or coming soon) that I missed?
22:12:14 <danwent> unfortunately, our core reveiwers will again be pretty swamped this release, so we need all cover devs to pitch in
22:12:15 <salv-orlando> I have a patch
22:12:36 <salv-orlando> which is going to be pushed soon which is around 200-300 hundred lines. It is for response pagination.
22:12:47 <salv-orlando> can't remember the bug id right now, sorry.
22:12:55 <danwent> Ok, that's been targeted for E-3 for a while, right?
22:13:18 <salv-orlando> Yeah. Code is basically ready, but i still need to sort out XML serialization for the link to the next page.
22:13:37 <danwent> cool.  just wanted to confirm it wasn't something brand new.  sounds good.
22:13:43 <SumitNaiksatam> dan: I am planning to submit the linux bridge plugin for review
22:13:44 <salv-orlando> will either push tomorrow or defer
22:13:56 <salv-orlando> odds are 4/11 to push tomorrow 3/1 to defer
22:14:05 <danwent> Yeah, any big patches should be in by thursday.
22:14:44 <danwent> Sumit:  ok, just make sure we get it in early, and be sure to help out with others reviews as a thank you :)
22:15:04 <SumitNaiksatam> dan: sure, edgar is my proxy :-)
22:15:11 <edgarmagana> :-)
22:15:25 <danwent> indeed, but we need all the help we can get reviewing.
22:15:35 <SumitNaiksatam> dan: yeah, kidding :-)
22:15:41 <danwent> good good :)
22:15:51 <danwent> ok, is debo here?
22:16:04 <danwent> i injected a VPN item into the agenda last minute
22:16:17 <debo-os> yeah
22:16:42 <danwent> Ok, cloudpipe thoughts?
22:16:46 <debo-os> Thx a lot for the feedback
22:17:07 <debo-os> I think I summarized the feedback and the action items for cloudpipe on the ML
22:17:22 <debo-os> we will aim to do a no frills version for E
22:17:37 <debo-os> assuming that a lot of will change with the real VPN service in F
22:17:46 <danwent> debo-os:  I think that makes sense.
22:18:14 <danwent> my main concern is whether we can get anything into Nova during E-4…. they will be in a pretty hard feature freeze based on concerns over stabalization
22:19:01 <danwent> I would say that if you don't think you can do something in the next few days, and have a shot for E-3, we should ping vish to get his thoughts on this.
22:19:19 <debo-os> I will try to make it to E3 but lets ping Vish anyway
22:19:21 <danwent> would help to have it well scoped what changes you would be making to nova, as I think they will be pretty targeted.
22:19:26 <danwent> ok, good idea.
22:19:49 <debo-os> I think I have the changes scoped out for Nova and I had run them past Soren and Vish
22:19:56 <danwent> Ok, anything else in terms of e-3?
22:20:34 <danwent> Ok, one topic I wanted to bring up with Nova + Quantum Integration testing
22:21:10 <danwent> we get a fair amount of breakage in QuantumManager due to nova changes, and presumably the fact that the unit tests + smoke tests run before nova commits don't test QuantumManager well.
22:21:35 <danwent> jaypipes sent out a cool preso on Tempest, if you haven't seen it yet: https://docs.google.com/a/nicira.com/presentation/d/1M3XhAco_0u7NZQn3Gz53z9VOHHrkQBzEs5gt43ZvhOc/edit#slide=id.p
22:21:59 <debo-os> on that note ... Dan, would it make sense to focus on devstack stability w. Quantum + Nova
22:22:11 <debo-os> ultimately a naive user would use something like devstack
22:22:17 <bhall> danwent: do you have another link?
22:22:32 <danwent> I think we've made good progress on Nova-parity with Essex, but not very much progess with system test.
22:22:38 <danwent> bhall: that one not working?
22:22:42 <jaypipes> bhall: http://joinfu.com/slides
22:22:47 <bhall> jaypipes: thanks
22:23:04 <jaypipes> np
22:23:05 <danwent> whoops, yeah, not public
22:23:34 <danwent> debo-os: yes… devstack stability is a start
22:23:45 <danwent> and things like tempest build on devstack
22:24:32 <danwent> I'd really like to see us making more progress on this front.  I'm going to start poking on it in my "free" time, but I'd really love if someone decided to take some more ownership here.
22:24:41 <danwent> I'm pretty swamped with a lot of project management stuff
22:24:51 <danwent> and haven't had as much time to code as I would like.
22:25:23 <danwent> #info On a related note, we have our bug-squashing day on Feb 2nd.
22:25:27 <GheRivero> is that test integration for essex or folsom release?
22:26:02 <danwent> I think we need to get the infrastructure in place ASAP.  I'd love to see work done in E-4, but if not, then it will need to happen in F
22:26:14 <danwent> http://wiki.openstack.org/BugSquashingDay/20120202
22:26:34 <bhall> danwent: I agree.. we definitely need it
22:27:02 <danwent> since we were thinking of focusing on code quality + coverage for bug squashing, one thing we could do was help people setup devstack + tempest and write integration tests.
22:27:19 <debo-os> I can do devstack and tempest setup
22:27:29 <danwent> There is a local meetup in SF, which might serve as a central location (sorry salv!) or we could meet in the southbay.
22:27:29 <debo-os> and maybe some tests too
22:27:35 <debo-os> southbay :)
22:27:38 <danwent> debo-os: that would be awesome.
22:28:12 <salv-orlando> my thoughts will be with you :)
22:28:34 <salv-orlando> anyway, what are your targets for the bug squashing day... what will be considered as a success?
22:28:38 <danwent> debo-os:  Ok, I will sync up with you directly on this.  Perhaps Nicira or Cisco could play host for an in person meetup.  I will send an email to the list to see if there are enough people who would like to meet up in person for the bug squashing day to make organizing an event worth it.
22:29:13 <danwent> salv-orlando:  my goal was to get a functioning system test environment up and running, with multiple people knowledgeable enough to add tests.
22:29:16 <debo-os> danwent: awesome ...
22:29:35 <danwent> carlp says he has infrastructure to run these tests automatically on each commit, which is ultimately what we want.
22:30:03 <danwent> Great, thanks for volunteering debo.
22:30:18 <danwent> #todo #debo-os and #danwent sync up on tempest + system testing
22:30:31 <danwent> I think this stuff is realy important for the success of the project.
22:30:40 <danwent> OK, one last topic to bring-up.
22:31:23 <danwent> Since we are getting close to the main essex release, we want to make sure that distros that package quantum have correct packaging, and are aware of recent changes to our packaging (e.g., splitting client repo out).
22:31:58 <danwent> there are a few listed here: http://wiki.openstack.org/QuantumPackages
22:32:01 <cdub> danwent: if that's a question, fedora is aware
22:32:13 <danwent> I'll also be contacting folks at Stackops, etc.
22:32:23 <danwent> cdub: great, will assume you have that taken care of.
22:32:47 <cdub> yup, rkukura will holler if things break
22:32:47 <danwent> if there are other distros that any of you think should package quantum, but don't, let's add them to the page and reach out to them.
22:33:12 <danwent> Ok, any other topics to bring up on Quantum? or open discussion?
22:33:23 <GheRivero> i'm taking care of debian packages and also stackosp
22:33:48 <danwent> awesome, thanks GehRivero.  Does stackops package quantum already, or will it be new with the final Essex release?
22:33:54 <danwent> Geh -> Ghe
22:34:01 <bhall> danwent: did you and monty chat about how plugins are going to work with the repo split?
22:34:30 <GheRivero> it's a work in progress
22:34:34 <danwent> based on his reply, I think his comment about plugins wasn't related to the repo split
22:34:38 <danwent> GheRivero: thx
22:34:41 <anotherjesse1> danwent: love the work you guys are doing for devstack integration
22:34:42 <bhall> oh ok
22:34:54 <anotherjesse1> thanks :)
22:35:12 <danwent> anotherjesse1:  thanks!  was just thinking the same thing about all the cool stuff you folks have been doing with devstack!
22:35:22 <danwent> makes a devs life so much easier
22:35:40 <danwent> Ok, last call, any questions/comments?
22:35:50 <cdub> any folks going to FOSDEM?
22:36:24 <danwent> not I… wish I could
22:36:28 <anotherjesse1> danwent: we do multi-node devstack by setting ENABLED_SERVICES=n-cpu,n-net,n-api (and MYSQL_HOST, RABBIT_HOST, GLANCE_HOSTPORT) - is it possible to do multi-node quantum?
22:36:38 <anotherjesse1> perhaps an email?
22:37:03 <danwent> anotherjesse1:  yes, quantum can do multi-node
22:37:28 <danwent> was planning on looking at how to do multi-node in devstack.  please send email with thoughts.
22:37:36 <bhall> anotherjesse1: I have it working on our end.. I have some changes to propose around that fairly soon
22:37:47 <danwent> last i checked it was "coming soon" on the webpage, or is it already supported?
22:37:53 <danwent> ah, got it
22:38:13 <danwent> yeah, multi-node will be very important to us using devstack for serious system test.
22:38:39 <danwent> thanks folks.  please remember to focus on reviews this week, and if you have non-trivial patches, to get them done by thursday.
22:38:51 <danwent> #endmeeting