22:03:26 <dendrobates> #startmeeting
22:03:27 <openstack> Meeting started Wed Jun  1 22:03:26 2011 UTC.  The chair is dendrobates. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:03:28 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic.
22:03:51 <dendrobates> no one updated the agenda, so we will just wing it this week
22:04:14 <dendrobates> #topic old business
22:04:41 <dendrobates> did the data extension discussion conclude last week?
22:04:54 <dendrobates> or do we still need to discuss it?
22:05:17 <dendrobates> hi dan
22:05:24 <dendrobates> did the data extension discussion conclude last week?
22:06:16 <dendrobates> ying: you were part of that, did it conclude to your satisfaction?
22:07:01 <dendrobates> is anyone else here
22:07:02 <ying> Yes, I think our conclusion is that we will keep core simple and do PortProfile in extension now.
22:07:06 <dendrobates> good
22:07:28 <dendrobates> ying: is that acceptable for your needs
22:07:56 <ying> yes
22:08:01 <dendrobates> great
22:08:22 <dendrobates> was there anything else from that meeting that we need to come back to?
22:08:42 <ying> I have a question, we will have a some code skeleton available for core api soon, right?
22:08:56 <dendrobates> ying: for quantum?
22:09:06 <danwent> hi folks
22:09:06 <danwent> am i online
22:09:06 <danwent> I am wondering around the airport trying to get a decent signal
22:09:06 <danwent> anyone?
22:09:06 <ying> yes
22:09:06 <danwent> can you see me?
22:09:07 <danwent> hello?
22:09:16 <dendrobates> wow dan
22:09:20 <ying> dan: yes
22:09:28 <dendrobates> danwent: all came through at once
22:09:53 <danwent> yes, quantum poc code will be soon.
22:10:01 <dendrobates> ying: I think the midokura/NTT work is the basis of that
22:10:05 <danwent> sorry, probably my lossy connection :)
22:10:30 <ecarlin> you need infiniband wifi :-)
22:10:32 <danwent> ying: salvatore just got us API code, so we're going to be able to have an end-to-end system available by next week, i suspect
22:10:38 <ewanmellor> Salvatore said that he's pushed code already.
22:10:46 <danwent> :)
22:11:04 <danwent> yes, salvatore's code looks good, so i think we're in good shape
22:11:06 <dendrobates> #topic quantum updates
22:11:17 <dendrobates> danwent: any other updates
22:11:19 <ying> danwent: thanks
22:11:23 <danwent> we're going to throw together some more unit tests, the cli utility to drive the API
22:11:30 <danwent> in a couple days things should be ready for an initial review.
22:11:45 <ying> danwent: is Salvatore's code available online?
22:12:05 <danwent> ying: yes, let me grab URL
22:12:54 <ewanmellor> danwent: What's the plan for system testing (if any)?
22:13:12 <danwent> somik, brad, are you there?  my connection is slow and pages are not loading.  can you send out salvatore's latest link?
22:13:12 <danwent> ying: otherwise, we can email it to you tonight.
22:13:12 <danwent> I am propping my laptop up on a southwest ticket counter right now :)
22:13:12 <danwent> trying to get a decent signal
22:13:19 <bhall> ying: https://code.launchpad.net/~netstack/network-service/quantum-api-workinprogress
22:13:24 <ying> danwent: thanks
22:13:59 <dendrobates> any other quantum updates before we move on?
22:14:19 <dendrobates> #topic  donabe updates
22:14:41 <dendrobates> I am working on the Donabe draft API
22:15:00 <dendrobates> If  I can get approval, I will send it out for review.
22:15:19 <dendrobates> It is based on existing work at Cisco, so there is hurdle to jump
22:15:27 <dendrobates> speaking of which...
22:15:45 <dendrobates> We have made progress on opening the cisco floodgates
22:16:17 <danwent> sorry guys, dropped off signal again.  Anyone seeing this?
22:16:17 <danwent> ewan: let's chat later.  I can't keep a decent signal here.  short answer: we'll be doing some, but help is definitely appreciated.
22:16:20 <dendrobates> we should soon be clear for devs like ying to start pushing their work
22:16:45 <ewanmellor_> danwent: OK
22:17:01 <ewanmellor_> I just dropped off myself (coffee shop in SF)
22:17:48 <dendrobates> is anyone here to talk about melange?
22:18:07 <troytoman> i can provide some updates
22:18:16 <dendrobates> #topic melange update
22:18:23 <ewanmellor_> ying: https://code.launchpad.net/~netstack/network-service/quantum-api-workinprogress is the code you asked for.
22:18:32 <troytoman> we have some base functionality for setting up blocks/allocating IPs etc.
22:18:47 <ying> ewanmellor_: thanks
22:18:55 <troytoman> it is currently in a branch on launchpad, i'll try to get an API draft and branch location out soon
22:19:10 <dendrobates> troytoman: awesome, thanks
22:19:23 <troytoman> our next steps are to look at the network refactoring and muti-nic and how we start to integrate with Nova
22:20:04 <dendrobates> #topic open discussion
22:20:05 <danwent> dendrobates: looking forward to draft.
22:20:10 <danwent> webex?
22:20:27 <dendrobates> So, I can setup any webex meetings that anyone wants.
22:20:32 <ryu_ishimoto> I started a branch for refactoring:  http://code.launchpad.net/~midokura/nova/network-refactoring
22:20:43 <dendrobates> I can will try to get a netstack account that I can share with thte group
22:21:02 <danwent> If others are ok with it, I would like to at least give webex a try.
22:21:05 <dendrobates> that way anyone can call a webex meeting.
22:21:37 <jamesurquhart> dendrobates: You should talk to Chris Hoff. He uses WebEx for CloudAudit calls.
22:21:41 <danwent> ryu: thanks.  how does this compare to the original branch?
22:21:51 <dendrobates> danwent: I think it will be good for somethings, but it can be difficult for people for whom english is a 2nd or 3rd language
22:21:53 <danwent> (original ntt/midokura branch)
22:22:02 <danwent> dendrobates: true
22:22:14 <dendrobates> jamesurquhart: great, I'll ping beaker
22:22:22 <ryu_ishimoto> danwent:  It's a lot less disruptive after discussing with Vish and Trey how we should approach the refactoring
22:22:32 <danwent> on webex, I am in favor of at least giving it a try.
22:22:45 <ryu_ishimoto> still need to discuss further with quantum/ipam/multi-nic to finalize the approach, however
22:22:47 * ewanmellor_ starts WebEx vs GoToMeeting smackdown
22:22:50 <danwent> ryu: ok, i will take a look.
22:23:04 <danwent> ewan :)
22:23:18 <ryu_ishimoto> danwent: thanks
22:23:38 <dendrobates> ewanmellor: :)
22:23:50 <jamesurquhart> ewanmellor_: We can draft a custody agreement. Alternate weeks and two months each summer. ;)
22:24:04 <danwent> james: :)
22:24:11 <dendrobates> btw, I am in the valley this week if anyone wants to have a netstack meetup
22:25:04 <danwent> dendrobates:  I am back tomorrow.  will send email.
22:25:13 <dendrobates> wow we can have a meeting in record time, if no one can stay connected.
22:25:17 <SumitNaiksatam> I would like to bring up the topic of semantics of "attach" in quantum for discussion
22:25:23 <danwent> :)
22:25:34 <dendrobates> SumitNaiksatam: go ahead
22:25:34 <danwent> sumit: yes, that is a great email thread.
22:25:35 <SumitNaiksatam> we had a discussion Dan and myself over emails
22:25:48 <SumitNaiksatam> Dan: did you get a chance to read my last email wherein refered to the 802.1Qbh configuration?
22:26:12 <SumitNaiksatam> hoping that others got a chance to think about this as well?
22:26:55 <danwent> sumit: i have been meaning to respond via email, but have been in meetings the past two days with travel.
22:26:55 <danwent> sumit: i have some follow-up questions for you.  but short answer is we definitely need a way for vntag/vepa to work.
22:26:59 <danwent> sumit: I actually think it fits into the model in general, though I am not sure about some libvirt specifics.
22:27:10 <danwent> particularly that "profile" attribute was not in the main libvirt docs.
22:27:21 <danwent> I can't load web pages right now, so I will have to provide more details later :)
22:27:26 <SumitNaiksatam> ok
22:27:59 <SumitNaiksatam> the point being that there is some information which needs to go into the libvirt "interface" configuration which, we think, should come from quantum
22:28:09 <danwent> sumit: generally, i am curious how the profile attribute compares to the XML that I sent out (with the instanceid, etc.)
22:28:18 <SumitNaiksatam> in this case it happens to be the "port-profile"
22:28:46 <ewanmellor> Why is libvirt relevant, in a discussion about vntag?  The VMware nova backend uses the VIM API directly.
22:28:52 <SumitNaiksatam> dan: profileid is a separate attribute
22:29:21 <SumitNaiksatam> ewan: we are not referring to ESX
22:29:26 <danwent> sumit: sometimes libvirt can be a bit of a pain as it tries to perform both compute + network orchestration.  If we can't work around it, we'll have to figure out someway for nova to get that info from quantum, but I think that breaks the general model that quantum can change the network connectivity without nova's help.
22:29:54 <SumitNaiksatam> dan: yes, thats precisely the reason I wanted to have this discussion
22:30:13 <ewanmellor> SumitNaiksatam: Are you talking about using libvirt for network-plane operations?
22:30:18 <dendrobates> danwent: SumitNaiksatam : maybe this is a good topic for a webex discussion
22:30:20 <danwent> ewan: I think sumit was talking about vntag/vepa on linux
22:30:36 <SumitNaiksatam> ewan: dan: yes
22:30:40 <danwent> sumit: great.  I think we're on the same page about what needs to be possible.  We just need to work out the best way for it to happen so that we can have as clean of a separation of quantum + nova as possible.
22:30:42 <SumitNaiksatam> mean to clarify
22:30:48 <danwent> dendrobates: :)
22:31:08 <SumitNaiksatam> dendorobates: agree, this discussion is more involved
22:31:19 <SumitNaiksatam> we might want to carry it on over emails until the next meeting
22:31:35 <danwent> sumit: sounds good, or over webex if that is faster.
22:31:41 <danwent> (yet still open :) )
22:31:55 <danwent> at the least though, we'll want to send the conclusion to the list to wrap up the thread.
22:32:16 <SumitNaiksatam> as far as the libvirt "interface" configuration goes, there are pieces of information which nova would know, and there are others which quantum would
22:32:39 <SumitNaiksatam> just reiterating
22:32:50 <dendrobates> ok, anything else?
22:32:58 <danwent> nope
22:33:13 <ewanmellor> Can we make sure that Quantum operations aren't defined in terms of libvirt semantics?
22:33:21 <danwent> +1 :)
22:33:50 <dendrobates> #endmeeting