22:03:28 <danwent> #startmeeting
22:03:29 <openstack> Meeting started Tue Feb  7 22:03:28 2012 UTC.  The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:03:30 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic.
22:03:46 <danwent> #link  agenda: http://wiki.openstack.org/Network/Meetings
22:04:10 <salvorlando> hello netstackers!
22:04:30 <danwent> #info just a reminding to everyone that feature freeze/branch point for E-4 will be one week earlier than normal.  Feb 21st.
22:04:50 <danwent> we need to avoid last minute merges that take our time away from testing.
22:05:12 <danwent> #info: curent items for E-4: https://launchpad.net/quantum/+milestone/essex-4
22:05:36 <danwent> biggest thing that stands out is that v1.1 client is unassigned
22:05:48 <salvorlando> agreed
22:06:03 <danwent> Salv, I know you mentioned that you might be able to do that, though at the cost of other API work.  Is that still the case?
22:06:11 <salvorlando> I'd be happy to take it,  but I don't feel sure about ability of deliver... not a lot of spare dev cycles.
22:06:29 <salvorlando> I have two outstanding work items on API.
22:06:43 <danwent> salvorlando: can either of those items drop?
22:07:07 <danwent> seems like having a client that can leverage 1.1 features is important.
22:07:08 <salvorlando> maybe resource links...it does not appear to be a huge amount of requests
22:07:12 <salvorlando> of request for it.
22:07:16 <salvorlando> agreed.
22:07:33 <salvorlando> I can definitely drop resource links (already dropped rate limiting)
22:07:48 <danwent> yeah, I'm all for pruning to the most important stuff early
22:08:10 <salvorlando> as concerns pagination for responses, I'd try to merge it. The code is already there, altough I need to rebase and repush as the review expired
22:08:29 <danwent> yeah, i remember the code was really close.
22:08:56 <salvorlando> we just decided to defer it to make reviews easier.
22:09:04 <danwent> ah, ok.
22:09:21 <danwent> so it sounds like i have a soft yes on the client work?
22:09:51 <salvorlando> It is a soft yes... right :)
22:10:00 <danwent> ok, thanks :)
22:10:19 <salvorlando> will make it either a "hard yes" or a  "no" in the next 36 hours...
22:10:25 <danwent> ok, sounds good.
22:10:44 <danwent> debo-os update on merge of system test stuff into devstack?
22:11:00 <danwent> do we know who to talk to in order to try and get it running automatically on builds?
22:11:33 <danwent> carlp: do you still have that infrastructure available?
22:12:13 <danwent> hrmm…  we'll try and come back to that topic when the owners aren't asleep :)
22:12:14 <carlp> Yes
22:12:21 <danwent> ah, just in time
22:12:29 <carlp> Sorry, got two things going at once :)
22:12:41 <carlp> Yes, the hardware is available and ready for use!
22:13:03 <danwent> #action #debo-os  contact #carlp with info on how to run quantum excercise.sh within devstack.
22:13:14 <danwent> that would be really great to have that going and emailing us if things break
22:13:23 <carlp> yes, that was my plan
22:13:37 <danwent> sweet.  we should have enough to get the whole system up and running end-to-end
22:14:14 <danwent> I have some additions to the excercise.sh script, but for now I'd just focus on getting the env setup with the initial version of the script.
22:14:23 <danwent> anything else on system test?
22:15:14 <danwent> Ok, next up in the linux bridge plugin, which is long overdue for merging.  Salv and I have promised reviews today + tomorrow, so hopefully we can iterate quickly and have it in by end of week.
22:15:25 <danwent> sumit, anything to add?
22:15:32 <SumitNaiksatam> great...thanks.
22:15:35 <SumitNaiksatam> no nothing to add
22:15:39 <SumitNaiksatam> we did some testing
22:15:43 <SumitNaiksatam> edgar and shweta
22:15:48 <SumitNaiksatam> so far so good
22:16:09 <SumitNaiksatam> Shweta found another issue with creating networks for non-admin users
22:16:18 <SumitNaiksatam> not linux bridge plugin specific
22:16:19 <danwent> great.  we'll have to figure out the questions of how to handle shared code, but I think merging it and getting it in for E-4 is the key thing to focus on.
22:16:24 <SumitNaiksatam> we can take that up later
22:16:33 <SumitNaiksatam> yeah, good point on the shared code
22:16:33 <danwent> sumit: is this the email she sent out last night?
22:16:35 <SumitNaiksatam> saw the emails
22:16:47 <SumitNaiksatam> yeah, the one she sent yesterday night
22:17:37 <danwent> yeah, would be good to get an exact env. and set of commands to run.  once a bug like that is reproducable, they are usually quite easy to fix.  I  saw a bug like that when working on the unit tests, but it was actually the result of a misconfig.  Would be good to see how to repro.
22:17:54 <SumitNaiksatam> ok, will get back to you on that
22:18:02 <danwent> We also need to get reviews in on the Ryu plugin, as that too was deferred from E-3
22:18:07 <SumitNaiksatam> we just wanted to quickly check if you were also seeing it
22:18:30 <danwent> Sumit:  smart thing to do until we get better test coverage in Nova for Quantum Manager.
22:19:06 <danwent> on that note:  i sent an email about a patch to improve quantum manger test coverage.
22:19:09 <danwent> here's the review: https://review.openstack.org/3858
22:19:26 <danwent> I'd encourage people to review it, and think about additional testing that could be added.
22:19:46 <danwent> One huge gap is to improve the melange test coverage.
22:19:57 <danwent> my patch does not address it.
22:20:12 <danwent> (this is melange client code within Quantum Manager that calls melange, not the melange service itself)
22:20:39 <danwent> is debo-os here?
22:21:12 <danwent> #action #debo-os check with vishy to confirm that nova work on VPN means that there is a decent VPN + Quantum optioin for Essex
22:21:40 <danwent> assuming that is the case, we'll defer further VPN parity work to Folsom (or a separate service… TBD)
22:22:03 <danwent> anyone here working on the horizon + quantum stuff?
22:22:46 <danwent> I saw that vishy fixed the vif API bug, which was the main thing breaking quantum + horizon.  But we still need to shift horizon more to a quantum-manager style model, rather than a manual vif-plugging model.
22:23:10 <danwent> #action #danwent poke ML to see if we need any BPs/bugs targeted for quantum + horizon work.
22:23:55 <danwent> Last topic on my list is the authz middleware.   I'm hoping someone from Nicira can take this one on in E-4, but its definitely at risk, so if its important to you, I'd speak up.
22:24:40 <danwent> btw, got an email update from debo-os via phone
22:25:03 <salvorlando> excuse? email update via phone??
22:25:07 <danwent> he's working on modularizing the excersie.sh code, so we can have multiple tests that have a shared "library" of quantum setup logic
22:25:23 <danwent> yes, he says his phone IRC client isn't working.
22:25:47 <danwent> we'll have to figure out an appropriate disciplinary action :)
22:25:47 <salvorlando> :)
22:26:23 <danwent> Ok, so E-4 is branching Feb 21st, don't forget.
22:26:24 * markvoelker votes that debo must bring us all donuts
22:26:37 <danwent> donuts via IRC?? delicious
22:26:48 <bhall> mmm.. donuts
22:27:13 <danwent> and please take some time to contribute to one of the testing efforts, quantum-manager unit tests, quantum unit tests, or quantum + nova system/integration test.
22:27:39 <danwent> as the system becomes more complex, it will be brittle is we don't have great test coverage, so we need to suck it up and write some tests :)
22:27:58 <danwent> Ok, anything else specific to E-4?
22:28:31 <danwent> #topic open discussion
22:28:43 <danwent> I have a couple things to mention
22:29:08 <danwent> as discussed last week, and in the main openstack meeting, we're planning on applying to become a core project for the Folsom release
22:29:31 <danwent> some folks from melange have also been talking about whether it makes sense to merge melange into quantum
22:29:59 <danwent> separate code for the most part, but a shared API layer, docs, project org, meetings, etc.
22:30:13 <danwent> probably a good thing to discuss in detail at the summit (April 16th, save the date!)
22:30:25 <danwent> but wanted to let it stew a bit before then.
22:30:39 <salvorlando> I think it makes sense to finally have a project called NetStack, as it was meant to be in the beginning!
22:30:53 <danwent> :)
22:31:17 <danwent> salvatore's an originalist.  the one thing people will never agree on is naming :)  probably best discussed over beer
22:32:00 <danwent> only other topic was to mention a bay area meeting we were will be talking about quantum: http://www.meetup.com/openstack/events/41485502/
22:32:02 <mestery> Note: Most things are best discussed over beer. :)
22:32:02 <salvorlando> jokes apart, I'm for the merger.
22:32:12 <danwent> this thursday
22:32:27 <salvorlando> mestery: or over alcoholic beverages in general
22:32:58 <danwent> ok, any other items for open discussion before we adjorn for beverages?
22:33:38 <danwent> btw, it was realy exciting to have so many new people participating in the bug squashing
22:33:45 <rkukura> anyone object to dropping setup_*.py files?
22:33:45 <danwent> 15+ people, very cool
22:34:06 <bhall> rkukura: I'm all for it.. nuke 'em!
22:34:25 <GheRivero> shoot them all rkukura
22:34:31 <danwent> rkukura: yeah, really appreciate the work you've been doing
22:35:18 <danwent> Ok folks.  several threads on the ML to follow-up on.  Reviews for linux bridge + ryu plugins, and general dev for E-4.  Sounds like a plan!
22:35:26 <danwent> have a good week all
22:35:32 <salvorlando> you too
22:35:35 <danwent> #endmeeting