21:03:43 <ttx> #startmeeting 21:03:44 <openstack> Meeting started Tue Jun 12 21:03:43 2012 UTC. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:03:44 <bcwaldon> ttx: fire him! 21:03:46 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:03:53 <ttx> bcwaldon: can't, he is my boss. 21:04:03 <ttx> Agenda @ http://wiki.openstack.org/Meetings/ProjectMeeting 21:04:14 <ttx> #info 3 weeks left until the milestone-proposed cut for Folsom-2 (July 3) 21:04:26 <ttx> #topic Actions from previous meeting 21:04:33 <ttx> * ttx to point Swift distro packagers to required config changes in Swift 1.5.0: DONE 21:04:52 <ttx> For completeness: see https://lists.launchpad.net/openstack/msg12754.html 21:04:58 <ttx> * bcwaldon to set assignees for glance-client-parity and streaming-server specs: DONE 21:05:02 <ttx> * jog0 to update status for general-host-aggregates bp: DONE 21:05:06 <ttx> * ttx to fix jgriffith access to bp updating: DONE 21:05:12 <ttx> * lzyeval to update status of ec2-id-compatibilty bp: DONE 21:05:21 <ttx> * vishy to contact assignees for config-drive-v2, delete-in-any-state and user-configurable-rbac and confirm F3 targeting 21:05:39 <ttx> looks DONE now 21:05:44 <ttx> #topic Keystone statu 21:05:47 <ttx> s 21:05:49 <ttx> :) 21:05:53 <ttx> heckj: o/ 21:05:53 * heckj waves 21:06:03 <ttx> welcome back :) 21:06:04 <ttx> #link https://launchpad.net/keystone/+milestone/folsom-2 21:06:16 <ttx> Got a few questions for you 21:06:20 <heckj> shoot 21:06:24 <ttx> * https://blueprints.launchpad.net/keystone/+spec/stop-ids-in-uris (Guang Yee) 21:06:36 <ttx> How is that going ? termie said last week he didn't understand why that was needed, or "essential"... 21:07:10 <termie> ttx: the best description i have run into was that people are worried about token ids leaking in logs 21:07:10 <heckj> That blueprint will be dealt with in the V3 API, shouldn't be listed as essential. Certainly won't be done by this milestone at the rate we're going. 21:07:20 <heckj> termie: correct 21:07:31 <ttx> heckj: downgrade to "High" ? 21:07:46 <heckj> ttx: done 21:07:49 <ttx> cool 21:07:52 <termie> while i feel like that is probably nice to think about, i also don't think our tokens are long-lived enough for that to matter much 21:08:00 <ttx> * https://blueprints.launchpad.net/keystone/+spec/implement-v3-core-api (heckj) 21:08:11 <ttx> how is it going so far ? 21:08:20 <ttx> ...and how likely is it to be completed within the next 3 weeks ? 21:08:38 <heckj> ^ still working through draft updates, so no initial work has started here. Given the timing, I suspect it's not going to be done this milestone, and we'll need to delay it 21:08:38 <uvirtbot> heckj: Error: "still" is not a valid command. 21:08:54 <heckj> will retarget the BP 21:08:54 <bcwaldon> uvirtbot: thanks for the update 21:08:55 <uvirtbot> bcwaldon: Error: "thanks" is not a valid command. 21:09:21 <heckj> one of these days I'm going to have figure out what that thing does 21:09:34 <bcwaldon> heckj: it's a bender 21:09:50 <markwash> insert girder? 21:09:59 <bcwaldon> 31 degrees, 32 degrees 21:10:01 <ttx> * https://blueprints.launchpad.net/keystone/+spec/keystone-ipv6-support (ayoung) 21:10:19 <ttx> Not started either ? Any chance /that one/ will be in F2 ? 21:10:35 <ayoung> ttx, my vote is no 21:10:37 <heckj> ttx: no progress as yet, expected as documentation on a deployment option in documentation. Quite possibly. 21:10:51 <heckj> Or not, if ayoung says... 21:10:55 <ttx> heh 21:11:06 <ayoung> ttx, IPv6 support is not there yet in upstream Eventlet. I suggest that if someone wants it, they use HTTPD for the near term 21:11:19 <heckj> will retarget that too 21:11:23 <ttx> note that you're preparing yourself a large F3 plate, which is not really the best idea in the world :) 21:11:40 <heckj> didn't say I'd retarget that to F3 21:11:51 <ttx> ayoung: if the implementation of that bp depends on ipv6 support in eventlet... sounds a bit far away to me 21:11:59 <ttx> heckj: ack 21:12:06 <heckj> removed from any milestones or series for now 21:12:14 <ttx> * https://blueprints.launchpad.net/keystone/+spec/rbac-keystone (dolphm) 21:12:22 <ttx> Slow progress, does that mean it might miss F2 too ? 21:13:12 <ttx> ..or that there are still 3 weeks left after all ? 21:13:17 <heckj> That will likely split per this mornings keystone meeting 21:13:46 <ttx> cool. Split is good, couldn't actually figure what work was involved in this one 21:13:52 <ttx> heckj: anything else ? 21:13:54 <heckj> The implementation of an API supporting RBAC will be tied to V3 API, but we'll also be documenting and setting some defaults, suggested values for policy.json and implementing a policy engine in keystone itself 21:14:10 <ttx> so APIv3 is the big blocker IIUC 21:14:31 <heckj> Quite a bit of the desired feature work is dependent on it 21:14:52 <ttx> heckj: any other comment ? 21:15:05 <heckj> nope 21:15:08 <ttx> Anyone else, questions about Keystone ? 21:15:19 <ayoung> ttx, BTW, that blueprint should really be for IPv6 support across all Openstack projects that are based on Eventlet 21:16:07 <ttx> ayoung: should probably be filed as separate blueprints... since there is no need to roll it all out at the exact same moment 21:16:25 <ttx> #topic Swift status 21:16:32 * chmouel waves 21:16:33 <ttx> chmouel: yo 21:16:47 <ttx> I just had a question about the python-swiftclient split, how far is it from completion ? 21:17:09 <chmouel> it's done since bcwaldon merged the review on glance 21:17:25 <chmouel> there seem to be some problem with the functional tests 21:17:30 <chmouel> which I am going to work on 21:17:34 <ttx> chmouel: was the client lib removed from swift proper ? 21:17:38 <chmouel> yes 21:18:03 <ttx> devcamcar: does that affect horizon ? 21:18:09 <chmouel> I have logged a bug on horizon to have a look at it so they can switch from cloudfiles 21:18:11 <ttx> or did you anticipate it ? 21:18:19 <ttx> ok. 21:18:20 <chmouel> they are not using swift.common.client 21:18:22 <chmouel> but python-cloudfiles 21:18:23 <devcamcar> ttx: no, we've been using python-cloudfiles 21:18:35 <ttx> oh, so that's actually good news for you :) 21:18:40 <devcamcar> we do plan to move to python-swiftclient when this all settles 21:18:54 <chmouel> well python-cloudfiles is somewhat slower than swiftclient :) 21:19:25 <ttx> chmouel: could you post to the ML when the final bits land ? This will affect distros 21:19:45 <chmouel> ttx: yes I was planning to do that just after this meeting 21:19:45 <ttx> they should anticipate the work, before 1.5.1 is released 21:19:49 <ttx> awesome 21:19:58 <chmouel> so another thing 21:20:02 <chmouel> There has been some discussion to have devstack gating on swift commits 21:20:03 <ttx> #action chmouel to post about python-swiftclient separation to the ML 21:20:26 <chmouel> and we still need to look over how's that going to work 21:20:57 <chmouel> so I am going to get in touch with the CI to understand how's that going to work with the swift specifics 21:21:06 <Ravikumar_hp> question: any documentation on python-swiftclient 21:21:25 <chmouel> Ravikumar_hp: there is none 21:21:26 <Ravikumar_hp> list of APIs and how to call 21:21:39 <chmouel> is there any documentation for other *clients library btw: ? 21:21:41 <bcwaldon> chmouel: isn't it the same as the old swift.client docs? 21:21:50 <bcwaldon> chmouel: yes, current glance client 21:22:18 <ttx> #action chmouel to discuss with CI on enabling swift in devstack-gate 21:22:40 <chmouel> Ravikumar_hp: oh yeah actually they should get auto generated soon we have just merged a commit but I am not sure where this is going to be exposed 21:22:42 <ttx> chmouel: Anything else ? 21:22:51 <chmouel> noip 21:23:08 <ttx> Other questions on Swift ? 21:23:39 <ttx> #topic Glance status 21:23:43 <ttx> bcwaldon: o/ 21:23:46 <ttx> #link https://launchpad.net/glance/+milestone/folsom-2 21:24:01 <ttx> Looking at the road to api-v2, looks like continuous progress... 21:24:04 <bcwaldon> indeed 21:24:14 <bcwaldon> we've come up with some more bps in the past week or two 21:24:19 <ttx> api-v2-links depends on api-v2-refactor-schemas, which in unassigned / unknown status ? 21:24:19 <bcwaldon> but we are making steady progress 21:24:32 <bcwaldon> ttx: you should really stop caching that info before the meetings 21:24:49 <bcwaldon> ttx: I've gotten efficient at updating it just as the meeting starts :) 21:24:56 <ttx> sigh 21:25:03 <bcwaldon> le sigh indeed 21:25:23 <ttx> Actually it's http://wiki.openstack.org/releasestatus/ that is not refreshed at the right time :) 21:25:27 <bcwaldon> aha 21:25:49 <bcwaldon> I don't have anything to report 21:26:16 <ttx> bcwaldon: you still have an essential depending on a medium :P 21:26:25 <bcwaldon> why yes I do 21:26:38 <bcwaldon> I've created a paradox! 21:26:39 <ttx> well, that makes it essential as well :) 21:26:53 <bcwaldon> fixing it right now 21:26:54 <ttx> or the other just pretends to be essential. 21:27:12 <ttx> Did you hear from dprince about swift-tenant-specific-storage ? 21:27:20 <bcwaldon> negative 21:27:23 <bcwaldon> or I forgot 21:27:24 <ttx> How essential is this ? 21:27:32 <bcwaldon> not as essential as I thought 21:27:41 <bcwaldon> could be Med or High 21:27:47 <ttx> make it high :) 21:28:04 <bcwaldon> k 21:28:05 <ttx> Also had a question about essential stuff being targeted to folsom-3... 21:28:12 <chmouel> I can help if dprince want 21:28:12 <ttx> I understand why you absolutely need separate-client and glance-client-parity... 21:28:26 <bcwaldon> chmouel: I'll keep that in mind, thanks :) 21:28:26 <ttx> But shouldn't https://blueprints.launchpad.net/glance/+spec/streaming-server be "High" priority instead ? 21:28:36 <ttx> Sounds like something you'd really like to have, rather than something we can't release without. 21:28:42 <bcwaldon> ttx: true 21:28:46 <bcwaldon> done 21:28:49 <ttx> thx 21:28:54 <ttx> bcwaldon: Anything else you wanted to mention ? 21:29:00 <bcwaldon> no sir 21:29:09 <ttx> Cool. Questions on Glance ? 21:29:32 <Ravikumar_hp> Python-glanceclient - Folsom-3? 21:29:47 <bcwaldon> Ravikumar_hp: yes 21:30:54 <ttx> #topic Quantum status 21:30:59 <ttx> danwent: hey 21:31:06 <ttx> #link https://launchpad.net/quantum/+milestone/folsom-2 21:31:06 <danwent> ttx: hello 21:31:21 <danwent> you'll noticed i flagged a couple things with slow progress to highlight them :) 21:31:36 <danwent> the good news is that we got the bulk of the API stuff merged 21:31:42 <ttx> ooh. 21:31:49 <ttx> Good thing I refreshed 21:32:04 <ttx> So now that's done, should we consider that Melange as a separate project is dead ? 21:32:32 <danwent> from a folsom release perspective, yes. there may still be people that use the essex version 21:32:39 <ttx> sure. 21:33:01 <danwent> ok, so the two big issues are DHCP and the Quantum/Nova integration 21:33:18 <danwent> DHCP: the folks working on this say they are still confident with the F-2 date 21:34:05 <danwent> but I haven't seen much yet. Will be trying to kick this work into high gear this week. Otherwise, we potentially need to revisit what we think we'll be able to do with the F-2 deliverable (my goals was that basic uses cases would be functionally complete by F-2) 21:34:14 <ttx> that's how I understand the "Slow progress" you have there 21:34:32 <ttx> In other news, can authorization-support-for-quantum be considered 'Implemented' ? 21:34:39 <ttx> I saw the change landed 21:34:48 <danwent> ttx: one patch landed, but there's one more coming 21:34:53 <ttx> oh, ok 21:35:04 <danwent> kevin's making good progress on that though, so i'm not worried about it. 21:35:29 <danwent> i am worried about the nova/quantum integration, as it looks like tr3buchet may no longer be able to work on that in F-2 21:35:32 <ttx> Could you look into and set the priority and series goal for https://blueprints.launchpad.net/quantum/+spec/cisco-plugin-v2-api-support (proposed for F3) ? 21:35:48 <danwent> ttx: ah, i think that's a new one. hadn't seen it yet. 21:35:51 <danwent> will do. 21:36:13 <danwent> done 21:36:18 <ttx> danwent: could you explain what that nova/quantum integration blueprint will cover exactly ? 21:36:35 <ttx> Nova QuantumManager for quantum v2 ? 21:36:38 <danwent> ttx: nova will use the v2 quantum apis when VMs are created and destroyed 21:36:49 <danwent> meaning that quantum will take care of IPAM as well. 21:37:04 <danwent> and that you can now create networks directly using the quantum API, rather than using nova-manage 21:37:06 <ttx> ok, yes, would have been good to have that in F2 21:37:17 <danwent> ttx: indeed, its quite central. 21:37:18 <ttx> so that it sees some mileage 21:37:35 <danwent> so if I can't get someone to work on it, i'll probably drop my other F-2 stuff and do it myself :( 21:38:11 <ttx> #help volunteer needed to help with improved-nova-quantum-integration 21:38:24 <danwent> other than that, things are going wel though. lot of good code being written an reviewed. a few new core devs. 21:38:30 <ttx> danwent: I think it might be worth it 21:38:41 <danwent> ttx: agreed. 21:38:42 <ttx> (drop the other things to get that into F2) 21:38:53 <ttx> danwent: Anything else ? 21:38:59 <danwent> don't think so. 21:39:03 <ttx> Questions on Quantum ? 21:39:34 <ttx> #topic Nova status 21:39:52 <ttx> So we don't have vishy, i'll just ask a few questions just in case the assignees can answer 21:39:59 <ttx> #link https://launchpad.net/nova/+milestone/folsom-2 21:40:07 <ttx> Plan & progress looks good to me 21:40:21 <ttx> Quick review of the status on the essential stuff: 21:40:34 <ttx> * https://blueprints.launchpad.net/nova/+spec/general-host-aggregates (jog0) 21:40:44 <ttx> * https://blueprints.launchpad.net/nova/+spec/finish-uuid-conversion (mikal) 21:41:06 <ttx> jog0, mikal: if you're around, a quick update on status for those BPs ^ 21:41:15 <ttx> jgriffith: around ? 21:41:32 <jgriffith> Yep 21:41:49 <jog0> ttx: making good progress. Working with Citrix guys to keep xen hv_pools working. 21:41:50 <jgriffith> ttx: Things are looking good on cinder 21:41:58 <ttx> jgriffith: do you know how volume-decoupling goes ? 21:42:11 <ttx> jog0: thx, still on track for F2, then ? 21:42:24 <jgriffith> ttx: Yeah, alot of what I'm doing this morning is part of that: 21:42:38 <jgriffith> Between sleepsonthefloor and myself we've about got it done 21:42:45 <ttx> cool 21:42:47 <jgriffith> I'm just finishing some ec2 compat issues 21:42:47 <ttx> jgriffith: What's left to do before Cinder becomes a first-class citizen ? 21:43:01 <jgriffith> ttx: It already is in my book :) 21:43:04 <jog0> ttx: yes. Not sure if every part will be in by F2 (such as reming duplicate metadata concepts). but most will be ready 21:43:06 <jgriffith> ttx: Fix the ec2 stuff 21:43:19 <jgriffith> ttx: add some tests to devstack 21:43:35 <ttx> jgriffith: Is there anything that will get broken in F2 by the completion of this ? 21:43:37 <jgriffith> We're just about fully functional... 21:43:43 <ttx> Like horizon, which has switch-to-cinder-client targeted to F3 now ? 21:43:53 <jgriffith> That's up to everybody else (ie vish) 21:44:09 <jgriffith> We have it now so everything is turned on/off by flags 21:44:15 <jgriffith> So you can go nova-volume or cinder 21:44:23 <jgriffith> this includes devstack as well as "nova" 21:44:43 <ttx> jgriffith: we'll probably make the decision post-F2 to do the full switch or not 21:44:45 <jgriffith> I have a couple of challenges to get this to work correctly in nova api's but I should get there 21:44:51 <ttx> together with the core confirmation 21:44:55 <jgriffith> ttx: The only other concern is TESTS 21:45:04 <ttx> jgriffith: TESTS are good 21:45:13 <jgriffith> ttx: the majority of the existing tests aren't so good though 21:45:20 <jgriffith> ttx: Most are all using fakes for everything!! 21:45:30 <ttx> heh 21:45:39 <ttx> nothing like a test that can't fail. 21:45:43 <jgriffith> :) 21:45:49 <ttx> jgriffith: anything else ? 21:46:14 <jgriffith> ttx: Not unless anybody is interested, or better yet wants to help out :) 21:46:26 <jgriffith> It's been me and sleepsonthefloor and that's about it 21:46:32 <annegentle> jgriffith: what updates to the Compute docs do you expect you'll need? 21:46:35 <ttx> #help Cinder needs more useful test coverage 21:46:50 <ttx> Questions on Nova/Cinder ? 21:46:53 <jgriffith> ttx: bahhh 21:46:54 <jgriffith> docs 21:46:56 <annegentle> #help Cinder needs documentation 21:47:12 <jgriffith> I'll have to tackle that in force after this week 21:47:22 <jgriffith> annegentle: I'll make a note to link up with you 21:47:24 <ttx> ok, great 21:47:30 <ttx> #topic Horizon status 21:47:31 <annegentle> jgriffith: maybe you and sleepsonthefloor can tell me more about it 21:47:35 <devcamcar> o/ 21:47:35 <annegentle> jgriffith: sounds good 21:47:36 <ttx> devcamcar: hey 21:47:41 <devcamcar> ttx: howdy 21:47:47 <ttx> devcamcar: we'll talk about a stable/essex point release after the quick folsom-2 status 21:47:51 <jgriffith> perhaps I'll send a note out to the ML just to update folks??? 21:47:52 <devcamcar> ok 21:47:54 <ttx> #link https://launchpad.net/horizon/+milestone/folsom-2 21:48:01 <ttx> jgriffith: good idea 21:48:07 <jgriffith> ttx: KO 21:48:08 <devcamcar> F2 blueprints are generally progressing well 21:48:13 <ttx> #action jgriffith to update the ML with Cinder progress 21:48:19 <devcamcar> https://blueprints.launchpad.net/horizon/+spec/nova-volume-optional may get bumped to F3 21:48:26 <devcamcar> when we make the rest of the cinder client changes 21:48:39 <devcamcar> in fact, i'm going to do that now, been meaning to move it 21:48:44 <ttx> devcamcar: Two blueprints were added to F2 without having their series goal confirmed, would be good to look into those and set Priority and Series goal accordingly: 21:48:48 <ttx> * https://blueprints.launchpad.net/horizon/+spec/use-common-jsonutils 21:48:51 <ttx> * https://blueprints.launchpad.net/horizon/+spec/automatic-secure-key-generation 21:49:30 <devcamcar> ttx: will fix the series goal now 21:49:36 <ttx> Last question, about the "Essential" blueprints in Folsom-3 (I don't really like to have essential stuff left on the last milestone): 21:49:45 <ttx> * https://blueprints.launchpad.net/horizon/+spec/switch-to-cinder-client 21:49:50 <ttx> I understand why this one is essential, and work is started, so we should be alright there... 21:49:59 <ttx> * https://blueprints.launchpad.net/horizon/+spec/quantum-workflow-integration 21:50:11 <ttx> This one has no assignee, should it just be assigned to "Nebula" team ? 21:50:41 <ttx> This sounds like a risky thing to consider "Essential"... Could you explain why we can't release Folsom without that feature ? 21:51:26 <devcamcar> ttx: we can probably downgrade that to "high" 21:51:35 <ttx> sounds like a plan. 21:51:37 <devcamcar> it will be a poor user experience without it though 21:51:57 <ttx> well, "High" should still definitely get done :) 21:52:08 <devcamcar> ttx: true enough ;) 21:52:12 <ttx> So, second subject, Horizon 2012.1.1 21:52:25 <ttx> The process is usually as simple as designating a commit ID as your candidate 21:52:40 <ttx> But the tarballs are currently not generated for stable/essex and we need to clean that up first 21:52:46 <devcamcar> ttx: yes, we've gotten a number of back ports into stable/essex that are helpful, so as soon as one more lands, we can tag it 21:52:54 <devcamcar> ttx: ok 21:53:02 <ttx> I'll discuss with markmc and see how we'll proceed here exactly 21:53:20 <ttx> (markmc is the release team member that handles stable release updates) 21:53:43 <devcamcar> for the record, a few stable/essex reviews here would help: https://review.openstack.org/#/c/7730/ 21:53:58 <ttx> fwiw, here is how he handled the 2011.3.1 releases: http://wiki.openstack.org/StableBranchRelease 21:54:01 <devcamcar> #help need reviews on https://review.openstack.org/#/c/7730/ 21:54:14 <ttx> devcamcar: so, i'l discuss with him, fix CI and come back to you 21:54:20 <devcamcar> ttx: sounds good 21:54:37 <ttx> #action ttx to clarify Horizon 2012.1.1 release process and fix CI to match 21:54:47 <ttx> devcamcar: Anything else ? 21:54:53 <devcamcar> ttx: nope 21:54:57 <ttx> Questions for Horizon ? 21:55:39 <ttx> #topic Other Team reports 21:55:46 <ttx> annegentle, jaypipes, mtaylor: ? 21:56:48 <annegentle> doc team met yesterday - here are notes: 21:57:00 <annegentle> #link http://eavesdrop.openstack.org/meetings/openstack-meeting/2012/openstack-meeting.2012-06-11-20.03.html 21:57:21 <annegentle> I didn't hire an intern for the summer, sorry folks. 4 final candidates took other positions. 21:57:34 <annegentle> Need to start in March not April next year. 21:57:44 <ttx> intern drought in Texas 21:57:52 <ttx> Any other team lead with a status report ? 21:57:58 <annegentle> I'm testing chinese language output through the maven plugin 21:58:04 <annegentle> Will report back on progress. 21:58:21 <annegentle> Looking for writers for http://etherpad.openstack.org/EssexOperationsGuide 21:58:39 <annegentle> And reviewers for the high availability document draft https://review.openstack.org/#/c/8139/ 21:59:03 <ttx> #help <annegentle> Looking for writers for http://etherpad.openstack.org/EssexOperationsGuide 21:59:05 <annegentle> Also, the CI team is working on a doc-related consistency project 21:59:10 <ttx> #help <annegentle> And reviewers for the high availability document draft https://review.openstack.org/#/c/8139/ 21:59:31 <annegentle> maybe mtaylor or jeblair or clarkb can explain better 21:59:43 <ttx> #topic Open discussion 21:59:51 <ttx> Anything else, anyone ? 22:00:36 <ttx> Quick glance at the effect of the BugTriage day on Nova: http://webnumbr.com/untouched-nova-bugs 22:01:05 <ttx> Now we need to discuss how to make that more permanent :) 22:01:09 <ttx> #endmeeting