21:00:25 #startmeeting quantum 21:00:26 Meeting started Mon Jan 28 21:00:25 2013 UTC. The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:27 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:29 nati_ueno: everything ok? 21:00:30 The meeting name has been set to 'quantum' 21:00:44 #info agenda: http://wiki.openstack.org/Network/Meetings 21:00:44 jo 21:00:44 hi all! 21:00:49 hi 21:00:55 hi all 21:01:14 danwent Thanks going better 21:01:17 #info we will be trying a slightly different format for the meeting, trying to get the quantum sub-team leads more involved. we'll probably have to iterate a coupel times to get it right. 21:01:43 #info Updated the Quantum Community Project page: http://wiki.openstack.org/QuantumStarterBugs 21:01:52 any other announcements before we get things started? 21:02:14 #topic quantum documentation 21:02:41 do we believe we have filed doc bugs for all G-1 and G-2 items by now? 21:03:25 #info we had some questions around what plugins where compatible with which nova virt layer configs, so I added this issue: https://bugs.launchpad.net/openstack-manuals/+bug/1104828 21:03:27 Launchpad bug 1104828 in openstack-manuals "Q-admin: compatibility matrix for q-plugin vs. nova-virt " [High,Confirmed] 21:03:38 I will likely be contacting you about this if you are a plugin maintainer 21:03:50 danwent: this bug is too low https://bugs.launchpad.net/openstack-manuals/+bug/1099574 21:03:54 Launchpad bug 1099574 in openstack-manuals "Quantum DB migration" [Medium,Confirmed] 21:04:18 markmcclain: ok, bumped to high 21:04:43 markmcclain: i also got a question a few days ago about configuring metadata with grizzly, is there a doc issue tracking that? 21:05:17 danwent: yes, https://bugs.launchpad.net/openstack-manuals/+bug/1099573 21:05:18 Launchpad bug 1099573 in openstack-manuals "Quantum Metadata Proxy should be documented" [Medium,Confirmed] 21:05:46 markmcclain: ok, i'm marking everything we definitely need to do for grizzly as a high, but it seems like the docs people are marking them as medium :) 21:06:09 markmcclain: you can email annegentle to get permissions to set the priorities of doc issues yourself if you need to do it frequently. 21:06:18 i'll bump that one 21:06:27 https://blueprints.launchpad.net/quantum/+spec/metadata-non-routed 21:06:51 there is another BP for metadata, which should be implemented. 21:07:02 salv-orlando: i saw you added the api docs bugs you promoised, thanks. any idea why they don't show up here? https://bugs.launchpad.net/openstack-api-site/+bugs?field.tag=netconn-api 21:07:12 gongysh: included a comment about that further down the agenda 21:07:13 missing tag 21:07:15 gongysh: i think mark is talking about that later on the agenda 21:07:40 salv-orlando: a, perfect. 21:07:51 danwent: fixed 21:08:18 #info in the next three weeks, as we wrap up G-3, please make sure that as features merge, we add doc issues for them. If its a community-wide feature, doc issue should be 'high'. 21:08:31 salv-orlando: thanks, seeing them now. 21:08:39 Ok, any other doc issues before we move on to G-3? 21:09:06 #topic grizzly-3 status 21:09:46 #info we're going to try and have a slightly different format here. We'll go through the 5 community-wide 'high' issues, then give the sub-team leads a chance to comment on the status of any other blueprints. 21:10:03 but first off, I wanted to give me regular warning, just so people aren't surprised. 21:10:28 #info g-3 branch date is Feb 19th, 3 weeks away. 21:11:24 #info all features must be in code review a week before the branch date (and even that isn't a guarantee of merging). The goal here is to avoid a final rush of last minute reviews, were core devs are pressured to quickly review a ton of code, as this tends to lead to low standards 21:11:53 #info for G-3, the only feature-freeze exceptions would be for community-wide 'high' priorities (no matter how important a feature is to you!) 21:12:09 any questions concerns on these before we dive into the current 'high' issues? 21:12:38 Multiple L3 + DHCP agents: 21:12:39 dev: yong 21:12:40 sub-team L3/IPAM/DHCP (markmcclain) 21:12:42 spec: https://blueprints.launchpad.net/quantum/+spec/quantum-scheduler 21:12:43 review: https://review.openstack.org/#/c/18216/ 21:12:44 active core devs: garyk, salv-orlando, danwent 21:12:45 status: ? 21:12:49 gongysh, markmcclain status? 21:13:05 i finally reviewed this last night, sorry for the LONG delay :) 21:13:13 i am reviewing. still have to take a look at thelatest chnages. hopefully tomorrow 21:13:42 XML patches should be ok 21:13:50 gongysh: yes, that's up next 21:14:13 I need to re-review 21:14:16 quantum scheduler has something to deal with. 21:14:25 nothing additional to add on quantum-scheduler bp 21:14:44 ok. where do we expect to be on this blueprint by next week? 21:14:49 host, collection action, extension name. 21:15:09 gongysh: sorry, don't understand 21:15:23 I have respond your comments in review. 21:15:34 I conclude three point from your comments. 21:15:41 gongysh: ok, will take a look later today 21:16:23 ok, so I expect we still think we'll be in review next week? 21:16:37 yes. 21:16:40 what are our goals to achieve this week? get key design questions cleared up? 21:17:02 I am updating the patches when I got comments. 21:17:27 #info multi-agent blueprint is in feedback mode in review. Expected to still be in review next week. 21:17:28 I hope so, that depends on reviewers. 21:17:48 #info goal is to clean up design questions and be to detailed code review and testing by next week 21:17:58 (may be optimistic, but i'm an optimist :) ) 21:18:13 the code can be tested now. 21:18:14 XML for Quantum v2 API: 21:18:15 dev: yong 21:18:16 sub-team: API (salv-orlando) 21:18:18 review: https://review.openstack.org/#/c/19998/ 21:18:19 active core devs: markmcclain, garyk 21:18:19 status: ? 21:18:33 Garyk approved. 21:18:37 I need finish reviewing 21:18:39 it has passed garyk's test. 21:18:55 Yup. tested and it worked. Yong solved all problems found 21:19:00 I have worked with him some time last week. 21:19:21 salv-orlando: we need an extra +2 for approval. 21:19:22 ok, sounds great. I am not aware of any design issues still outstanding here, so it sounds like we expect this to merge fairly soon? 21:19:28 I just need to double check collections and handling of null values. Should be ok though since Gary has spent an awful lot of time on it. 21:19:58 This should be implemented for the next meeting 21:20:12 #info xml API patch is further along in review, no major design issues. Should be merged by next week. 21:20:17 salv-orlando: we have some test cases for it, nul, collections, etc. 21:20:26 Security Groups (OVS plugin): 21:20:26 dev: nachi 21:20:28 sub-team: security groups / fw (arosen) 21:20:29 review: https://review.openstack.org/#/c/19436/14 21:20:30 active core devs: arosen, amotoki, rkukura 21:20:31 status: ? 21:21:05 nati_ueno is at the hostital, may be slow to respond 21:21:15 arosen, can you comment? 21:21:17 On review also blocked by nova bug 21:21:58 nati_ueno: can you paste the nova bug that it is blocked on? does that bug have active reviewers or do we need to identify them? 21:22:03 Akihiro looks almost ok for the patch 21:22:27 nati_ueno: patch looks pretty close to being good for me. 21:22:32 https://bugs.launchpad.net/nova/+bug/1050433 21:22:35 Launchpad bug 1050433 in nova "LibvirtBridgeDriver crashes when spawning an instance with NoopFirewallDriver" [High,In progress] 21:22:35 I'll check it out again today 21:23:22 nati_ueno: looks like daniel b has nova bug assigned, but i don't see a patch 21:23:27 may be good to ping him 21:23:33 https://review.openstack.org/#/c/19126/ 21:23:51 This is the review 21:24:05 ah, its just not tagged with the bug-id 21:24:26 Yes I comented it on review 21:24:37 I will update the bug to indicate that this is the fix. That patch is stalled on some of the vif-plugging discussions we've been having on the ML, but i'm optimistic that that log jam may be opening up soon 21:24:58 I agree 21:25:14 ok, any other concerns around the quantum side fo the patch? 21:25:22 What are our goals for moving this forward in the next week? 21:25:45 problem is why this bug is blocking our patch? 21:25:48 I hope it merged in this week 21:26:07 #info security group patch is blocked by quantum vif-plugging issue: https://bugs.launchpad.net/nova/+bug/1050433 21:26:08 Launchpad bug 1050433 in nova "LibvirtBridgeDriver crashes when spawning an instance with NoopFirewallDriver" [High,In progress] 21:26:32 Yong we can't disable nova security group 21:26:41 One question on this review - are we trying to avoid sending notifications within transactions? 21:26:45 #info seems possible to merge quantum OVS security groups this week, but relies on a nova change, so its unclear 21:27:30 if we use LibvirtOpenVswitchDriver, we can use that no firewall 21:28:02 gongysh: nati_ueno's change requires IPtables, which requires hybrid driver for OVs, which is where the problem is 21:28:03 Yong it not working by the bug 21:28:17 ok. 21:28:26 rkukura: not sure, someone more familiar with the change would need to comment 21:28:55 but I think ovs bridge is depending on open flow. 21:29:15 we can discard the hybrid driver now. 21:29:27 Rkukura I looks it should be. I will check the code 21:29:44 LBaaS Agent-based Plugin: 21:29:45 dev: Ilya + enikanorov 21:29:46 sub-team: lbaas (danwent) 21:29:47 spec: http://wiki.openstack.org/Quantum/LBaaS/Agent 21:29:48 reviews: https://review.openstack.org/#/c/20225/ and https://review.openstack.org/#/c/20579/ 21:29:49 Agree with gongysh here. 21:29:49 active core reviewers: garyk and ?? 21:30:26 went on review today and received a couple of comments 21:30:27 I won't be able to review before Friday 21:30:36 there is a ton of code duplication here. an extra set of eyes would be very helpful 21:30:37 enikanorov: yup, saw that 21:30:51 garyk: interesting, dup with what parts of the code? 21:30:54 existing code? 21:31:08 (or within the patch itself?) 21:31:19 danwent: there are the lbaas classes. yup within the pacth set 21:31:31 garyk: got it 21:31:34 i think that i added comments about this 21:31:46 ok, so salv-orlando said he can be the second core dev 21:31:51 I will update the agenda for next time 21:31:55 maybe i am missing something (i am not the sharpest tool in the shed at the moment) 21:32:16 I hope we discuss some lbaas stuff at the end of the meeting (related to one of these reviews) 21:32:44 enikanorov: ok, sounds good. we can actually discuss it in a few mins during the "sub-teams" section 21:33:21 well. I've wrote an email about the concern of having conf-based dev management 21:33:25 #info lbaas agent driver posted for review ( https://review.openstack.org/#/c/20225/ and https://review.openstack.org/#/c/20579/ ). garyk and salv-orlando will be core devs (more welcomed!) 21:33:40 since it's incompatible with typical haproxy use case 21:34:00 youcef has proposed a solution, we're evaluating it now 21:34:03 enikanorov: ok 21:34:14 HAProxy driver for Agent-based Plugin: 21:34:14 dev: oleg 21:34:16 sub-team: lbaas (danwent) 21:34:17 spec: http://wiki.openstack.org/Quantum/LBaaS/HAProxyDriver 21:34:18 review: (none yet. expected by next week?) 21:34:19 anticipated core reviewers? 21:34:23 enikanorov: are you contact for this as well? 21:34:41 haproxy driver is expected on review at the end of the week 21:35:03 #info lbaas HA proxy driver review expected by end of this week 21:35:28 ok, before we move on to to the sub-team reports, there's one other comment that I meant to make above but forgot. 21:36:02 as of today, we have 28 unmerged blueprints, which needs to be at 0 in three weeks through some combination of merging blueprints and dropping blueprints from the milestone. 21:36:26 so many! 21:36:28 by tomorrow, I want to be under unmerged 25 blueprints, and by a week from now, under 15 unmerged blueprints. 21:37:06 by merging, u mean change two into one magically? 21:37:08 so, my current target is that there are several blueprints that are 'not started', that will be booted out of the G-3 release in the next few days (i've already updated the whiteboards to warn the developers) 21:37:25 gongysh: haha, no i meant, by merging the code into the repo (i.e., implementing it). 21:37:38 got it. 21:37:45 gongysh: though that's a clever trick :) 21:38:28 Ok, so sub-team leads, please chime in at this point with additional items you think need team updates (none required). 21:38:47 also, if there are active technical discussions going on the ML around items in your sub-team, please highlight those 21:38:57 let's start with API. anything salv-orlando ? 21:39:14 I see: 21:39:14 I am the approver for 4 blueprints. 21:39:15 - https://blueprints.launchpad.net/quantum/+spec/quantum-api-wadl 21:39:16 - https://blueprints.launchpad.net/quantum/+spec/support-pagination-in-api-v2 21:39:17 - https://blueprints.launchpad.net/quantum/+spec/load-plugin-supported-extensions 21:39:23 beyond what we talked about above 21:39:25 thanks for pasting the list before me. 21:39:35 salv-orlando: sorry to step on your toes! 21:39:36 We already said XML is in good shape 21:39:55 Pagination is too in good shape, but I need another core there. 21:40:03 I'm the only one reviewing it so far. 21:40:11 quantum-api-wadl expected to send review by this week 21:40:17 I'll take a look at pagination 21:40:20 #help need another core dev on https://blueprints.launchpad.net/quantum/+spec/support-pagination-in-api-v2 21:40:23 zyluo: was going to ping you about that. 21:40:29 #info markmcclain will help on pagination 21:40:46 Once we see the code for quantum-api-wadl (some is already on review) we can make a call there 21:40:55 the last one, it's borderline at the moment. 21:41:01 I need an update from the assignee. 21:41:18 salv-orlando: agreed. I'll update the BP suggesting we bump it. its not critical 21:41:19 Without it, it's probably going to be booted out next week. (also because it's not really critcial) 21:41:24 :) 21:41:34 ok, anything else on API? 21:41:37 Holizon support also needed for pagenation 21:42:01 nati_ueno: is there a BP filed? 21:42:04 nati_ueno: that is desirable, but not mandatory 21:42:07 i assume this is not required? 21:42:09 yeah, great 21:42:14 API works as before 21:42:33 if you don't ask for pagination. Otherwise we would have broken bw compatibility :) 21:42:47 There wa bug when glance support pagenation 21:42:47 ok, time is getting late, so let's keep moving to updates from mark's subteam(s) 21:43:19 markmcclain: probalby don't need to go through all of them, just highlight anything you see as important for the community to have an update on. including active technical discussions / design reviews. 21:43:39 metadata on non-routed networks will be available this week 21:43:58 btw, please use info tags for announcements so they are capture in notes 21:44:07 #info metadata on non-routed networks will be available this week 21:44:20 The defaut limit was 20 and holizon didn't support pagenation. Then some image disapered from horizon 21:44:46 oops… any rate moved a few BPs out of Grizzly and will finish moving more out that aren't started or need design decisions 21:45:05 markmcclain: ok, thanks 21:45:34 nati_ueno, we disable pagination by default, so it will be ok 21:45:37 also see l3plugin mentioned on the agenda: Migrate L3 from mixin to plugin: https://blueprints.launchpad.net/quantum/+spec/quantum-l3-routing-plugin 21:46:03 Alexxu it sounds good! Thans 21:46:04 yeah.. folks should look at the mailing for the discussion 21:46:09 danwent: Bob should be posting that review tomorrow, as far as I know. 21:46:09 nati_ueno: you're welcome to look at the patch on gerrit. This will allow you to see whether the problem occurs or not 21:46:16 i think there is a design here to review. I think people generally agree this is the right direction, but what is unclear is whether this can be done non-disruptively for G-3 21:46:25 I have code for that one. Planned to upload for review this week 21:46:40 mestery: great. I think seeing the code change is probably the only way to answer the "how disruptive" question, so yeah, I think that's the next step 21:46:46 Salvatore: ok I will review it 21:47:00 #info expect review for l3-plugin change later this week. 21:47:01 danwent: Yes, I asked Bob to post the review. 21:47:28 ok, we're running late, any other important sub-team updates about core quantum? 21:47:40 (we'll still have our regular agenda items after this) 21:47:58 I wanted to make sure people saw this post by the Arista folks: https://blueprints.launchpad.net/quantum/+spec/ovsplugin-hardware-devices 21:48:01 bobmel design was not very disruptive. (assuming we're talking about the one I reviewed) 21:48:07 There is an active ML discussion on it. 21:48:49 #info I also think we are wrapping up our long-thread on vif-plugging on the ML. Would be good to have people read the conclusions, as it will likely result in mandatory plugin changes in H to simplify nova. 21:48:50 No, I've tried to make it very incremental 21:48:56 are there plans to or need to do port security extension for linuxbridge and openvswitch? 21:49:05 danwent: I saw that thread, seems as if there may be some opportunity for the Arista stuff to fit in nicely with thee ML2 plugin. 21:49:36 mestery: yes, that is rkukura's impression as well, so hopefully it will work out. 21:49:51 danwent rkukura: great! 21:49:59 rkukura: my impression was that libvirt spoof prevention was sufficient 21:50:40 enikanorov: is there more discussion needed on device management and LBaaS, or is current ML thread sufficient? 21:51:06 (as in, do we need to talk about this at the team meeting, or is ML discussion sufficient?) 21:51:09 well, I'd like to hear some more opinions or more detailed opinions 21:51:24 if we have time now, then it'd be good 21:51:27 Ok. I'm working through my backlog of design reviews for quantum. 21:51:41 enikanorov: unfortunately we don't, so we'll just have to live with encouraging people to chime in the ML 21:51:47 #topic python-quantumclient 21:52:03 danwent: I'm just wondering if these plugins need to implement the API extension? 21:52:05 thanks to markmcclain for creating the 3.0.0 release target on launchpad: https://launchpad.net/python-quantumclient/+milestone/3.0.0 21:52:43 rkukura: ah, ok. 21:53:01 #info thanks to markmcclain for creating the 3.0.0 release target on launchpad: https://launchpad.net/python-quantumclient/+milestone/3.0.0 21:53:26 markmcclain, gongysh, please make sure priorities are set for those items. 21:53:45 if anyone else is expecting a client change to make grizzly, please make sure it is targeted at this release. 21:53:58 #topic quantum stable 21:54:14 garyk: how are we looking for openstack stable folsom release in a few days? 21:54:27 Sorry I should go thanks bye 21:54:33 danwent: we are looking good. i am backporting the port satus fixes (unless people object). 21:54:36 nati_ueno: bye! 21:54:42 nati_ueno: fo well and feel better 21:55:01 ok, please ping the core team directly if additional urgent fixes are needed to make it in the stable drop. 21:55:06 anything else on stable? 21:55:33 #topic quantum system test 21:56:02 two blueprints for system test, but neither assigned to grizzly. are they active? https://blueprints.launchpad.net/tempest/+spec/quantum-basic-api 21:56:02 https://blueprints.launchpad.net/tempest/+spec/quantum-extended-api 21:56:11 danwent: I am developing the code for the first BP 21:56:22 achieving good progress 21:56:40 mlavalle: ok, great. do tempest blueprints not need to be assigned to a release? perhaps not, since it is just system test. 21:57:10 #info making good progress on tempest + quantum work: https://blueprints.launchpad.net/tempest/+spec/quantum-basic-api 21:57:22 danwent: my impression they don't have to. But I will check with Jay on the meeting on Thursday and report back to you 21:57:28 mlavalle: thanks. 21:57:35 #topic Quantum + Horizon 21:57:52 amotoki and nachi are both out, so i'll just paste amotoki's summary 21:58:01 https://blueprints.launchpad.net/horizon/+spec/nova-net-quantum-abstraction (I will upload a patch this Tuesday) 21:58:01 https://blueprints.launchpad.net/horizon/+spec/quantum-floating-ip (I will upload a patch this Tuesday) 21:58:03 https://blueprints.launchpad.net/horizon/+spec/quantum-network-topology 21:58:04 https://blueprints.launchpad.net/horizon/+spec/quantum-vnic-ordering (Nachi's colleagues has a good progres) 21:58:05 https://blueprints.launchpad.net/horizon/+spec/quantum-security-group (I started) 21:58:06 https://blueprints.launchpad.net/horizon/+spec/quantum-lbaas 21:58:23 SumitNaiksatam: any update on lbaas work? 21:58:55 #topic open discussion 21:59:03 ok, two minutes to spare (phew!) 21:59:08 any open discussion topics? 21:59:11 danwent: you mentioned last week the need to improve the developers documentation. If we are not achieving any progress there, I want to take ownership of it. Here's what I propose: 21:59:28 mlavalle: by all means! 21:59:30 danwent: I want to give a presentation during the Summit on Quantum internals and how to hack it. I will develop the presentation and the documentation together 21:59:46 danwent: Two caveats though. You won'y see much progress from today until G-3. I want to deliver as much of the tempest tests as possible 22:00:12 danwent: regarding LBaaS, KC is making progress 22:00:16 for G3. After that, I will devote time the the Summit presentation and the documentation 22:00:28 mlavalle: that would be great from my perspective. thanks. 22:00:38 SumitNaiksatam: ok, i figured, but always good to ask 22:00:51 ok, that's all folks. have a good afternoon/evening/morning! 22:00:55 #endmeeting