21:00:22 <danwent> #startmeeting quantum
21:00:23 <openstack> Meeting started Mon Feb 25 21:00:22 2013 UTC.  The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:24 <rkukura> hi
21:00:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:27 <openstack> The meeting name has been set to 'quantum'
21:01:07 <garyk> hi
21:01:15 <danwent> #info agenda: https://wiki.openstack.org/wiki/Network/Meetings
21:01:16 <mestery> hi
21:01:19 <danwent> #topic announcements
21:01:28 <SumitNaiksatam> hi all!
21:01:50 <danwent> #info great work everyone on getting G-3 out the door.  That span of time is the most stressful and difficult part of the 6-month release cycle
21:02:15 <danwent> now we just need to make sure we limit code churn moving forward, focus on fixing bugs, and improving docs.
21:02:45 <danwent> #info as i mentioned on the core list, we'll be following a slightly different meeting agenda for the remainder of the release cycle
21:03:12 <danwent> #info we are now in the RC1 phase.  https://launchpad.net/quantum/+milestone/grizzly-rc1
21:03:35 <danwent> #info there are no official dates for RC1 release, each project decides when their RC1 is.
21:04:05 <danwent> #info RC1 should be what we consider as the grizzly release, unless a release blocker is found in an RC build.
21:04:32 <danwent> #info current plan it to branch for RC1 at March 12th (in two weeks)
21:04:47 <danwent> #info In one week (March 5th) Bar for bug fixes will increase to high priority fixes only (similar to what would be required for getting into stable branch)
21:05:24 <danwent> #info and as a team, we need to be putting a heavy clamp on any new bugs being filed that are really "mini-features".
21:05:38 <danwent> ok, any other announcements?
21:05:59 <mlavalle> danwent: I proposed a presentation for the summit. Please vote for it http://www.openstack.org/summit/portland-2013/vote-for-speakers/presentation/559
21:06:16 <danwent> mlavalle: please leave this kind of discussion for open discussion
21:06:20 <mlavalle> danwent: it's an intro to Quantum
21:06:32 <danwent> mlavalle: sure… just don't want a flood of such requests during the meeting :)
21:06:42 <mlavalle> sorry ;-)
21:06:54 <danwent> #topic RC1 feature freeze exceptions
21:07:16 <danwent> ok, we have two FFEs outstanding for quantum, the multi-agent work, and the lbaas work.
21:07:33 <danwent> #info FFE issues need to be merged by end-of-day tuesday.
21:07:33 <garyk> danwent: multi looks good. still testing
21:07:39 <danwent> garyk: great
21:07:55 <danwent> testing is key on that one, given that it is specific to multiple hosts, etc.
21:08:03 <danwent> first, off, multi-agent
21:08:08 <danwent> https://blueprints.launchpad.net/quantum/+spec/quantum-scheduler
21:08:13 <danwent> https://review.openstack.org/#/c/21175/
21:08:36 <danwent> is yong here?
21:08:40 <emagana> danwent: any guidelines on best practices for  testing it
21:09:21 <garyk> emagana: need to run multiple agents on multiple hosts.
21:09:21 <danwent> emagana: probably best to talk to yong about that…. i was still unclear on how to enable some aspects of the patch last time i review it
21:09:29 <garyk> danwent: this depened on a nova patch
21:09:44 <danwent> garyk: is this the host binding patch?
21:09:46 <garyk> danwent: https://review.openstack.org/#/c/21141/
21:09:55 <garyk> danwent: yup.
21:10:24 <garyk> danwent: i can try and update it tomorrow if there is no activity
21:10:55 <danwent> yeah, i think that would be good
21:11:06 <danwent> the current commit message seems way out of date "This is a one line change in nova/network/quantumv2/api.py "
21:11:32 <danwent> and the reason we need the patch for grizzly isn't clear, as that review was originally targeting something that did not make it into grizzly
21:11:51 <danwent> #info multi-agent depends on change to nova to send port binding information: https://review.openstack.org/#/c/21141/
21:12:03 <garyk> danwent: understood. 'll update tomorrow
21:12:31 <danwent> thx
21:12:44 <danwent> ok, garyk is there another core dev testing the patch, or am I that other core dev?
21:12:56 <garyk> danwent: amotoki is also testing
21:13:00 <amotoki> danwent: which patch?
21:13:02 <danwent> great.
21:13:06 <amotoki> multi-agent one?
21:13:09 <danwent> yes
21:13:14 <garyk> amotoki: yongs multihost
21:13:23 <garyk> sorry multi -agent
21:13:31 <amotoki> yes. i am testing it.
21:13:52 <danwent> ok, i'll still recheck the patch to make sure that my feedback was addressed, but with the lbaas stuff I won't have time to do sufficient testing of multi-agent, so its good to have two other cores testing.
21:14:08 <danwent> #info garyk and amotoki are the two core devs working on reviewing + testing multi-agent.
21:14:23 <danwent> ok, anything else on multi-agent?  are there devstack changes that are needed for more people to test?
21:14:37 <garyk> danwent: i think a few things:
21:15:02 <garyk> 1. at the moment it supports ovs - which needs few patches for devstack - state_pacth and firewall driver
21:15:17 <danwent> documenting this will also be a big effort, so please make sure those items are getting filed.
21:15:29 <garyk> 2. it requires a number of hosts - for dhcp/l3 agents. so it is a bit challenging
21:15:48 <garyk> danwent: i have on my todo to update the wiki for devstack and ovs
21:16:02 <danwent> sorry, update it for what exactly?
21:16:22 <danwent> is there a wiki page for the multi-agent stuff?
21:16:31 <danwent> sorry, i'm not sure what you're saying
21:16:44 <garyk> danwent: not yet. yong has it documented. we should make a wiki. i'll try and create one tomorrow
21:16:54 <danwent> ok.
21:16:56 <danwent> thx
21:17:02 <danwent> ok, moving on to lbaas
21:17:16 <danwent> https://blueprints.launchpad.net/quantum/+spec/lbaas-namespace-agent
21:17:21 <danwent> https://review.openstack.org/#/c/22794/
21:17:39 <danwent> as we talked about last week, markmcclain has volunteered to help get the lbaas stuff in for Grizzly
21:18:06 <danwent> he is simplifying the model significantly, going with an approach that should be familiar to people since it closely aligns with what we do for l3 + dhcp
21:18:26 <danwent> there will still be a driver interface, so other models can be supported out of tree with grizzly, and in tree with havana
21:18:45 <danwent> I've been reviewing + testing this, and will be one of the core devs
21:18:56 <danwent> expect a WIP devstack patch for this later today.
21:19:12 <enikanorov> what's "wip" ?
21:19:15 <danwent> I think we still need a second core dev
21:19:18 <danwent> "work in progress"
21:20:05 <garyk> danwent: i can try
21:20:24 <danwent> garyk: thanks, though you're already quite loaded :)
21:20:39 <danwent> if anyone else can help out as well, that would be great
21:20:40 <enikanorov> one concern about proposed patch: seems that it lack some error recovering. say if  for some reason haproxy is left running, then next deployments may fail
21:20:57 <enikanorov> also, user will not know about it since objects status will not be updated accordingly
21:21:10 <enikanorov> danwent: we're also testing it.
21:21:13 <danwent> enikanorov: ok, let's discuss this as part of the review.
21:21:31 <danwent> enikanorov: yup, great.
21:21:59 <danwent> ok, i will hunt around for another core.  If we don't get one by end-of-day today, probably no point in pushign for the tuesday night deadline.
21:22:13 <SumitNaiksatam> danwent/amotoki: I believe the FFE applies to KC's Horizon patch as well- https://review.openstack.org/#/c/21486/? I saw amotoki's comment to that effect...
21:22:14 <danwent> markmcclain: sorry to keep talking… you should be leading this :)
21:22:44 <markmcclain> no worries.. you've covered main points
21:22:46 <danwent> amotoki: did gabriel grant an FFE in horizon?
21:22:52 <danwent> for the lbaas stuff?
21:23:12 <danwent> or rather, did gabrial + ttx :)
21:23:14 <amotoki> danwent: i have not heard about it from gabriel yet.
21:23:38 <danwent> SumitNaiksatam: ok, then it sounds like maybe amotoki was just talking about the main quantum FFE.
21:23:54 <SumitNaiksatam> amotoki: I noticed you mentioned on the review that this has FFE, so just wanted to confirm
21:23:55 <danwent> SumitNaiksatam: there was a thread with Gabriel that you included me on, right?  I don't think he responded though.
21:24:08 <SumitNaiksatam> danwent: yeah
21:24:25 <SumitNaiksatam> maybe KC can just keep the review updated
21:24:34 <danwent> #TODO amotoki talk with gabriel about whether LBaaS work in Horizion has an FFE or not.  Needs to be additional FFE beyond quantum FFE.
21:24:38 <amotoki> SumitNaiksatam: yes. i am talking that KC's patch depends on the status lbaas itself.
21:24:56 <enikanorov> btw, we will also need python-quantumclient patch for lbaas-namespace-agent
21:25:13 <markmcclain> enikanorov: I've got one build.. I'll post into review
21:25:21 <enikanorov> ok
21:25:37 <danwent> markmcclain: ok, is there a bug tracking that?
21:25:51 <garyk> markmcclain: will this be able to work if the distribution does not support namespaces?
21:25:54 <markmcclain> no… I'll create one for it
21:26:07 <markmcclain> #todo markmcclain create cli bug for lbaas changes
21:26:20 <danwent> ok, anything else on lbaas?
21:26:43 <danwent> #topic quantum bugs (markmcclain)
21:26:58 <markmcclain> garyk: the current WIP does not support namespaces….I'll look into no ns support
21:27:10 <markmcclain> we have 29 bugs still open for RC1
21:27:12 <garyk> markmcclain: ok, thanks
21:27:22 <danwent> https://launchpad.net/quantum/+milestone/grizzly-rc1
21:27:52 <danwent> ok, i've been going through and pruning bugs that seem more like features
21:28:01 <danwent> or rather "mini-features"
21:28:25 <garyk> danwent: markmcclain a lot of the bus in progress are either abandoned or need to be rebased...
21:28:49 <danwent> basically, we need to shift toward having only bugs that are real fixes to issues that will make the existing quantum feature-set easier for people to use
21:29:35 <danwent> garyk: ok, i have not yet looked at bugs where the title implies it is a real bug
21:30:17 <danwent> #info if you are the proud owner of an RC1 bug, please update that bug sometime today with current status, whether you expect to hav ea review posted by end of this week, etc.  and if you already have a review, please make sure it is rebased.
21:30:45 <danwent> #info and if you have a bug that you don't think you can get to, please unassign yourself, so someone else can pick it off.
21:31:07 <danwent> #info I'd like anything that is borderline "mini-feature" to be merged by EOD tomorrow.
21:31:22 <danwent> one such item that I wanted to bring up was external network support for L3
21:31:32 <danwent> https://bugs.launchpad.net/bugs/1056437
21:31:33 <uvirtbot> Launchpad bug 1056437 in quantum "L3 agent should support provider external networks" [High,In progress]
21:31:56 <danwent> or rather, provider netowrk support for l3
21:32:05 <danwent> and the related bug for NVP l3: https://bugs.launchpad.net/bugs/1130211
21:32:06 <uvirtbot> Launchpad bug 1130211 in quantum "provider network for L3 gateways (NVP)" [Medium,In progress]
21:32:20 <rkukura> I'll update that by tomorrow morning - need to test whethr it already works for linuxbridge, and fix for openvswitch
21:32:30 <danwent> this is valuable, but also means code changes to some pretty critical functionality.
21:32:44 <salv-orlando> If I'm not able to make my patch of a decent size in two hours, I will postpone
21:33:16 <danwent> if people think we should extend the deadline past tuesday for this, let's discuss, but otherwise i'm going to enforce tuesday deadline.
21:33:49 <danwent> #info in general, please keep an eye out to prevent new bugs being targeted for RC1 unless they truly fix something for a grizzly user.
21:34:12 <danwent> there are a few other bugs I want to talk about, but we're a bit behind schedule, so i'll send email.
21:34:20 <danwent> #todo danwent send out email on quantum bugs.
21:34:27 <danwent> anything else on bugs?
21:34:53 <danwent> (basically, my email will be about a few bugs that I consider important for grizzly, but seems stalled waiting on design discussions or feedback)
21:35:10 <danwent> #topic python-quantumclient
21:35:18 <danwent> (markmcclain + yong)
21:35:31 <markmcclain> it doesn't appear yong is online
21:35:41 <danwent> indeed.
21:35:50 <danwent> i think he was working on getting a new milestone setup
21:36:02 <danwent> ah, looks like that is done
21:36:02 <markmcclain> yeah.. we decided to setup at 2.2 milestone
21:36:26 <markmcclain> will retarget the bp and bugs to that milestone
21:36:40 <danwent> perfect, was just going to suggest that
21:36:57 <markmcclain> decided to stay with 2.2 since the changes are still compatible with 2.1
21:37:03 <danwent> and are we expecting roughly the same timelines for 2.2 as grizzly (i.e., expect a freeze mid-march?)
21:37:08 <danwent> k
21:37:16 <markmcclain> yes.. plan to keep things simple
21:37:29 <danwent> great
21:37:35 <amotoki> needs to be releate before rc1 i think.
21:37:44 <amotoki> s/releate/released/
21:38:01 <markmcclain> amotoki: ok
21:38:03 <garyk> amotoki: good point. we should aim a few days before
21:38:10 <danwent> amotoki: yeah, so nova + horizon can pull it in?
21:38:22 <amotoki> danwent: i think so
21:38:29 <danwent> amotoki: agreed, good point
21:38:46 <danwent> we'll also need bugs against horizon + nova to update their pip-requires for this new version
21:38:59 <danwent> amotoki + garyk can you handle that for horizon + nova respectively?
21:39:13 <markmcclain> horizon is not pinned to a particular version
21:39:14 <garyk> danwent: sure
21:39:25 <danwent> markmcclain: but we want to enforce a minimum, right?
21:39:25 <amotoki> danwent: sure.
21:39:36 <markmcclain> yes
21:40:12 <amotoki> i plan to update pip-require to >= 2.2
21:40:29 <markmcclain> amotoki: perfect
21:40:32 <danwent> ok, so as far as an exact date for 2.2, let's look at a calendar.  what seems reasonable?
21:41:07 <markmcclain> Mar 1?
21:41:17 <danwent> remember that nova + horizon can choose their own RC1 dates.  nova's is likely to be later than ours.  i'm not sure about horizon
21:41:35 <danwent> markmcclain: ok, if you think that is feasible based on the set of reviews outstanding
21:41:46 * markmcclain double checks reviews
21:41:56 <danwent> I was going to say the 6th :)
21:42:09 <markmcclain> I'll fine with a later date
21:42:25 <danwent> hehe, ok, let's say the 6th.  you can always get it done earlier :)
21:42:49 <danwent> #info expect 2.2 release of python-quantumclient, compatible with grizzly features, near March 6th,
21:43:00 <danwent> ok, anything else on client/cli?
21:43:12 <markmcclain> that's all I've got
21:43:18 <danwent> thx
21:43:24 <danwent> #topic documentation
21:44:41 <danwent> ok, so I was going to go through each sub-team and try to identify key doc issues, but I think that would take too long.  Instead, could each sub-team lead sometime this week put together an email to the team on doc issues for grizzly, making sure to file doc issues for any features that are not covered?
21:44:54 <danwent> we'll need issues both for admin guide + api guide in some cases.
21:45:16 <danwent> by next week's meeting, I'd like to be able to go over where our major gaps are in terms of needing people to put together content for docs.
21:45:31 <arosen> sure thing
21:46:01 <danwent> #topic horizon + quantum (amotoki)
21:46:44 <danwent> amotoki: hello
21:46:51 <amotoki> hi
21:47:04 <amotoki> According to horizon G-rc1 page and ttx page, LBaaS quantum horizon is granted as FFE.
21:47:13 <danwent> amotoki: sweet
21:47:28 <amotoki> i will mail Gabriel (cc-ed to dan)
21:47:33 <danwent> any other FFEs or major bugs we're tracking that we need the team to be aware of
21:47:41 <danwent> amotoki: please also include SumitNaiksatam + KC
21:47:52 <amotoki> danwent: of cource.
21:48:07 <amotoki> horizon quantum bugs are https://bugs.launchpad.net/horizon/+bugs?field.tag=quantum
21:48:21 <amotoki> they are not targeted to rc1 yet, but it is worth fixed.
21:48:48 <amotoki> i will take care of them.
21:49:13 <danwent> amotoki: ok, great.  would be good to make sure gabriel knows which ones you consider important and to have them targeted at RC1.  otherwise they risk falling off of his radar.
21:49:38 <danwent> amotoki: great.  i admit i haven't been using the horizon stuff in my testing lately, but I really need to give it a shot now to see all of the exciting new features :)
21:49:41 <amotoki> will cover them in the mail i send later.
21:49:42 <SumitNaiksatam> amotoki: thanks for the confirmation!
21:49:44 <danwent> anything else on horizon :)
21:49:48 <danwent> ?
21:49:51 <amotoki> that's all from me.
21:50:03 <danwent> thx
21:50:05 <garyk> sleep is on the horizon :)
21:50:09 <danwent> #topic open discuss
21:50:12 <danwent> ion
21:50:14 <danwent> garyk: :)
21:50:25 <danwent> garyk: any stable release on the horizon?
21:50:31 <danwent> or none planed?
21:50:50 <garyk> not that i am aware of. i'll get back to you about that at next weeks meeting
21:51:09 <danwent> garyk: yeah, my guess is that people will want to get the grizzly RCs out… makes sense
21:51:12 <danwent> anything else?
21:51:31 <danwent> ok, getting a meeting done on time :)
21:51:38 <danwent> #endmeeting