22:03:20 <danwent> #startmeeting
22:03:21 <openstack> Meeting started Tue Jul  5 22:03:20 2011 UTC.  The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:03:22 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic.
22:03:32 <danwent> #topic melange
22:03:49 <danwent> hi sumit, good to see you're here...
22:04:07 <SumitNaiksatam> hey Dan
22:04:08 <danwent> we were waiting for someone from your team to joint to talk about quantum, so we're talking about melange first.
22:04:12 <danwent> troy?
22:04:12 <troytoman> Not much new to report. We're doing some work on IPv6 support and the ability to subnet blocks. Hoped to have the wiki updated by now but should get done this week.
22:04:15 <SumitNaiksatam> and everyone
22:04:41 <troytoman> Also looking at the network-refactoring discussion and starting to think about how to integrate with Nova.
22:04:58 <danwent> Ok, I have been going through the existing nova code and want to sync up with you about how the melange stuff will fit in.  can do it offline though.
22:05:03 <troytoman> we're also bringing some RAX nova core devs into the team to help us with that analysis
22:05:12 <danwent> very cool.
22:05:17 <markvoelker> o/
22:05:20 <troytoman> i think that's it unless there are questions
22:05:44 <danwent> #topic quantum
22:05:59 <danwent> whoops, forgot to send out the agenda: http://wiki.openstack.org/Network/Meetings
22:06:15 <danwent> for those just joining, we're switched the order to do melange first.
22:06:34 <danwent> nova refactoring.  ryu, want to give an update?
22:07:01 <ryu_ishimoto> sure
22:07:15 <danwent> wiki page is here: http://wiki.openstack.org/network-refactoring
22:07:21 <ryu_ishimoto> we've made progress ironing out the remaining issues offline with those involved
22:07:47 <ryu_ishimoto> in particular we are almost there with the libivrt driver refactoring
22:08:06 <ryu_ishimoto> sumit made a very nice proposal for that, and I will add that to the wiki
22:08:18 <salv-orlando> ryu_ishimoto: do you mean the network/linux_net driver?
22:09:02 <ryu_ishimoto> salv-orlando: I mean the VIF configuration for libvirt(the generation of interface XML section)
22:09:11 <salv-orlando> ok, I see.
22:09:23 <danwent> ryu_ishimoto: btw, I've downloaded your branch and will be playing with it after the meeting today.   What is your preferred method of feedback on the wiki-page content and code?
22:09:36 <danwent> still just email?
22:10:10 <ryu_ishimoto> danwent: Either works for me
22:10:12 <danwent> I think salvatore and I have found the problem with using launchpad answers for feedback: there doesn't seem to be a way to quote someone previous comments
22:10:37 <danwent> which is too bad because launchpad answers is nice and public, and has a good subscription mechanism :(
22:10:54 <salv-orlando> danwent: I don't understand actually why you can quote on comments for merge proposals and not answers...
22:10:54 <danwent> Ok, sumit anything to add?
22:11:05 <salv-orlando> ryu... one last thing
22:11:09 <danwent> salv-orlando:  yes, odd indeed :)
22:11:26 <ryu_ishimoto> salv-orlando: yup
22:11:29 <salv-orlando> I would like to play with your branch as well, but which one is exactly? I see 3 branches on the network-refactoring blueprint
22:11:33 <SumitNaiksatam> sorry i was sleeping :-)
22:11:49 <danwent> https://code.launchpad.net/~midokura/nova/network-refactoring
22:11:53 <danwent> at the top of the wiki page
22:11:55 <salv-orlando> ok, thanks
22:12:01 <SumitNaiksatam> so are we all in sync on the VIF driver stuff?
22:12:08 <ryu_ishimoto> danwent: yup that's it thanks Dan
22:12:13 <danwent> Sumit: I don't think so....
22:12:36 <danwent> I have an email out to Ryu on that, but I'll let him wake up before he needs to respond :)
22:12:40 <danwent> I know its early :)
22:12:54 <ryu_ishimoto> SumitNaiksatam: I have no problem with it, and i was JUST about to reply to that email
22:13:09 <SumitNaiksatam> ok great
22:13:17 <ryu_ishimoto> I'm really curious what Nova core guys feel about it too!
22:13:27 <danwent> Ok, so let's just keep grinding on that discussion.
22:13:40 <danwent> ryu: +1
22:13:51 <salv-orlando> would you include me as well in this loop? I'm curious. I promise I will not disturb too much :-)
22:14:21 <danwent> salv: sorry, didn't realize you weren't.  I thought ryu CC'd everyone
22:14:23 <ryu_ishimoto> salv-orlando: Yeah of course.  The plan is to throw what we discussed to the entire community for more discussion as well
22:14:59 <ryu_ishimoto> salv-orlando: I kept adding people but I forgot to add you, sorry
22:15:04 <RamD> Could you include us all in that nova discussions
22:15:11 <salv-orlando> ryu_ishimoto: np
22:15:39 <salv-orlando> could we have a netstack mainling list ? :)
22:15:41 <ryu_ishimoto> sure i'll add the entire netstack group, the discussion started with something vey small and grew large
22:15:54 <RamD> salv-orlando:+1
22:16:26 <danwent> I'm in favor, I know there were concerns about this that others had mentioned to Rick, but let's revisit this.
22:16:38 <danwent> #action re-explore getting a netstack list
22:16:58 <danwent> #action danwent re-explore getting a netstack list
22:17:00 <salv-orlando> the fact is that we typically don't use the openstack mailing list to avoid spam
22:17:19 <danwent> salv: agreed.  a lot of these discussions are high volume
22:17:42 <danwent> so we don't want to be spamming the openstack list, but it makes sense that if someone wants the volume, they should be able to get it.
22:17:44 <RamD> Another suggestion: Create a design decision web page. and captute all the highlevel points there...
22:18:09 <danwent> RamD: should blueprints capture this?
22:18:17 <danwent> or are you talking about something else?
22:18:57 <RamD> danwent: I think a highlevel points that shows design decisions...and refer to BPs...that would be great
22:19:01 <danwent> btw, in the mean time, before we get a netstack list, please let ryu know if you want to be looped into the nova-refactoring discussion.
22:19:52 <danwent> RamD: do you mean kind of general design principles?
22:20:00 <danwent> that we've agreed on for quantum?
22:20:08 <RamD> yes
22:20:36 <danwent> I think that would be valuable.  I still have a half-written doc on that..... kind of lost momentum.  I will revive it and put it on a wiki.
22:20:58 <danwent> #action danwent, revive design principles portion of "what is quantum?" doc, post on wiki.
22:20:59 <RamD> Not to make it as big "documentation" effort...but to syncup any new people joining us
22:21:14 <RamD> danwent: Super...let me know if you need another kbd :-)
22:21:19 <danwent> RamD: agreed.
22:21:49 <danwent> OK, anything else on nova refactoring?
22:22:06 <ryu_ishimoto> that's it for now!
22:22:10 <danwent> thanks ryu.
22:22:19 <danwent> Ok, api auth.
22:22:23 <danwent> salvatore.
22:22:25 <salv-orlando> https://answers.launchpad.net/quantum/+question/163091
22:22:41 <danwent> salv: sorry for getting feedback to you so late.
22:23:00 <danwent> will respond to your most recent comments later today, I hope.
22:23:03 <salv-orlando> this is the address where we are discussing the proposed specification. I already received Dan's feedback and you are all warmly invited to give your feedback
22:23:29 <danwent> not being able to "quote" is a real downside here....
22:23:49 <salv-orlando> The interesting thing, is that, as Dan pointed out, authorization for Quantum kinds of goes back to the problem of the interaction with Nova wrt port creation and interface plugging
22:24:12 <salv-orlando> a problem we never fully fleshed out. Perhaps now the time is ripe to finalize that discussion
22:25:07 <danwent> Ok, is anyone else planning on commenting on the spec?
22:25:29 <danwent> Just trying to gauge how long salvatore should be waiting for feedback.
22:25:51 <danwent> ok, anything else on auth?
22:25:52 <markvoelker> I wouldn't mind having another read through….if I have anything I'll get it to you by tomorrow night.
22:26:06 <danwent> makrvoelker: great, thanks.
22:26:26 <danwent> ok, extensions discussion.
22:26:31 <salv-orlando> Is anyone planning to have a discussion around whether plugging an interface into a network is an operation that should be explictly be performed by tenants or by nova on behalf of the tenant (ie: should be an operation of an 'admin API')?
22:27:02 <danwent> salv-orlando: I assume you mean logical plugging?
22:27:14 <salv-orlando> yeah, logical of course :-)
22:27:33 <danwent> I believe the plan is that it will be possible for tenants to plug directly, or something else can "orchestrate" on behalf of the tenants.
22:27:42 <danwent> that something else could be donabe, or it even could be nova
22:27:58 <SumitNaiksatam> that was my understanding
22:28:01 <danwent> nova already does a far amount of orchestration, though I'm not sure thats the long-term goal.
22:28:11 <danwent> any other opinions out there?
22:28:29 <salv-orlando> Ok, so can we be safely sure we don't need a discussion on this point?
22:28:35 <carlp> sounds reasonable to me
22:28:46 <troytoman> we are thinking that we should limit orchestration in Nova as a general rule
22:29:01 <danwent> salv-orlando: one of the things I am thinking about when reviewing ryu's branch is what it would mean for nova to orchestrate quantum
22:29:05 <troytoman> that said plugging in a vif within Nova is something that we would probably do
22:29:52 <danwent> ok, extensions.
22:29:55 <salv-orlando> danwent: That's why I want to play with the refactoring branch, as this might also affect the way in which we do authorization in Quantum.
22:30:10 <danwent> salv-orlando: I agree.
22:30:20 <danwent> is james here?
22:30:22 <salv-orlando> for extensions, is James on line?
22:30:41 <danwent> or ying?
22:30:58 <danwent> James sent an email about discussing extensions.
22:30:59 <RamD> danwent: What is the q ..sorry
22:31:07 <danwent> it will probably happen later this week.
22:31:19 <danwent> holler if you want to be involved and we'll make sure you get the invite.
22:31:25 <troytoman> I don't think we've found a time yet.
22:31:40 <danwent> yeah, last email in the thread was from troy
22:31:41 <RamD> danwent: Yes. Will followup with James/Ying on the extensions discussions
22:31:48 <danwent> RamD: thx.
22:32:07 <danwent> and last but not least for quantum: Mark
22:32:14 <danwent> any updates on the GUI?
22:32:21 <troytoman> i think we'll have some updates to the extension merge prop that handles plug-ins etc.
22:32:22 <danwent> + client lib stuff?
22:32:29 <markvoelker> Actually, it might be "least" for this week. =)
22:32:35 <danwent> :)
22:32:51 <danwent> troy: great.
22:32:56 <markvoelker> Due to the holidays and folks being on PTO for the US Independence Day holiday and medical reasons there hasn't been a huge amount of progress.
22:33:06 <markvoelker> However, Arvind did get a prelim mockup branch posted
22:33:10 <markvoelker> It's linked here:
22:33:17 <markvoelker> https://code.launchpad.net/~asomya/+junk/dashboard-quantum-mockup
22:33:51 <danwent> cool.
22:33:51 <markvoelker> There are some known issues yet as it's what he had time for before leaving the country (he'll be out all this week as well).  Those are in the whiteboard on the BP:
22:33:56 <markvoelker> https://blueprints.launchpad.net/quantum/+spec/quantum-client-gui
22:34:21 <markvoelker> Tyler is beginning work on the refactor but has been out today and yesterday.  I'll have him get a branch linked as soon as he returns.
22:34:34 <danwent> great, thanks for the update.
22:34:41 <danwent> anything else on quantum?
22:34:43 <RamD> test-framework for Qunatum..Is Santosh or any one else doing that?
22:34:51 <danwent> unit tests?
22:34:53 <danwent> or system test?
22:35:00 <danwent> unit test stuff should be in
22:35:04 <danwent> salv, somik?
22:35:14 <danwent> brad?
22:35:18 <bhall> savlatore and I have gotten some pieces for unit testing in and are currently working on more tests
22:35:20 <salv-orlando> Thankfully, there's no independence day in the UK
22:35:21 <bhall> filling out the other pieces
22:35:27 <markvoelker> RamD: were you referring to: https://blueprints.launchpad.net/quantum/+spec/quantum-system-test ?
22:35:39 <salv-orlando> it would be a bit of a nonsense, after all
22:35:51 <danwent> bhall: it would be great if we had docs on how to run the unit tests... I think you were going to do that?
22:36:04 <danwent> salv: yes... wouldn't expect the UK to celebrate that :)
22:36:06 <RamD> markvoelkar: yes..wanted to find if any other piecees need to be done
22:36:11 <bhall> sh run_tests.sh .. but yeah, I can create a document :)
22:36:11 <somik> salv-orlando: lol
22:36:32 <salv-orlando> I said because I used last week to make some progress starting from Brad's branch
22:36:35 <markvoelker> bhall: what about getting things into Jenkins/Smokestack?
22:36:58 <danwent> mark: I think that is waiting on incubation...
22:37:09 <danwent> really need to chase down dendrobates and get an update on that
22:37:27 <danwent> #action danwent: find out current status of incubation from dendrobates
22:37:30 <markvoelker> danwent: Ah, ok.  RamD or myself can harass him. =)
22:37:37 <carlp> I'm building a Jenkins for Openstack/Ceph, and was planning on running some networking stuff on it too
22:37:38 <danwent> mark: the more the better :)
22:38:04 <danwent> carlp: great.  what's Ceph?
22:38:31 <carlp> Storage System that we are developing.  Unrelated to networking, but I have interest in both :)
22:38:45 <danwent> carlp: ah, cool.
22:38:46 <salv-orlando> bhall: updated branch for unit tests is in lp:~netstack/quantum/quantum-unit-tests
22:38:58 <carlp> I'm hoping to get Jenkins running this week, I'll be happy to give logins to those we deem worthy :)
22:39:17 <danwent> carlp: very cool.  please keep me in the loop.
22:39:23 <carlp> will do
22:39:29 <danwent> (if you deem me worthy) :P
22:39:39 <salv-orlando> carlp: just a curiosity... why is the project call Ceph?
22:40:02 <danwent> anything else on quantum?
22:40:15 <salv-orlando> on unit tests...
22:40:34 <salv-orlando> I have implemented several test cases for the API, will complete them by tomorrow.
22:41:05 <salv-orlando> FakePlugin has ben enhanced to use in-memory sqlite db with models provided in quantum.db. This will ensure high code coverage.
22:41:22 <bhall> awesome.. soon it will be time to look at code coverage numbers :)
22:41:24 <salv-orlando> We need unit tests for db, cli, wsgi and plugin-interface as well
22:41:25 <somik> that sounds like a good improvement
22:41:49 <salv-orlando> the branch is on netstack ownership, so feel free to add stuff!
22:42:08 <danwent> salv: great work.  is there a blueprint were we are tracking the areas you think need work?
22:42:09 <salv-orlando> In the meanwhile I did a self "cease and desist" on proposing other branches for merge...
22:42:21 <danwent> is it: https://blueprints.launchpad.net/quantum/+spec/api-spec-unit-tests
22:42:24 <danwent> or something else?
22:42:38 <salv-orlando> yes, that one
22:42:45 <danwent> k, thanks.
22:42:51 <danwent> last call on quantum...
22:43:08 <danwent> anyone here to talk about donabe?
22:43:33 <danwent> #topic open discussion
22:43:50 <salv-orlando> I have a question on the nestack-core team
22:44:02 <danwent> sure.
22:44:25 <salv-orlando> I sent an email to Rick, as we decided a few weeks ago, proposing myself as a core member, but I'm not on the list.
22:44:32 <danwent> really?
22:44:43 <salv-orlando> I know it is not nice to self-candidate but that's what we agreed :-)
22:44:55 <danwent> #action dendrobates make salv-orlando a core member
22:45:09 <salv-orlando> k, thanks
22:45:15 <danwent> you've contributed much more code than most people in that group... not sure how that happened.
22:45:32 <danwent> thanks for bringing it up.
22:45:35 <danwent> anything else?
22:45:53 <danwent> (not to mention done reviews, fixed bugs, etc.)
22:46:04 <danwent> ok, folks.  have a good one!
22:46:11 <danwent> #endmeeting