21:00:42 #startmeeting quantum 21:00:43 Meeting started Mon Jan 7 21:00:42 2013 UTC. The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:44 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:46 hi 21:00:46 Greetings to all for the new year! 21:00:46 The meeting name has been set to 'quantum' 21:00:53 hi 21:00:56 #link agenda: http://wiki.openstack.org/Network/Meetings 21:01:06 #topic announcements 21:01:45 #info many thanks to gongysh for all of this work answering questions on https://answers.launchpad.net/quantum/ . I'd encourage everyone to subscribe to notifications and help out 21:02:08 Hi! 21:02:09 #info grizzly-2 branch point is tomorrow 21:02:36 all feature code for G-2 must be merged at that point. beyond that time, only 'critical' or 'high' bugs will be backported for the G-2 release 21:02:45 hi 21:02:57 #info heads up that the next folsom stable release will be end of january 21:03:33 #info devstack gate for quantum is currently broken due to webob version conflict with other projects. we believe this patch (already approved) will fix it: https://review.openstack.org/#/c/19138/ 21:03:52 (we are still waiting on gating to merge it, and confirm that things are fixed) 21:03:58 danwent: Any "new features"/plugins will NOT be accepted after that? 21:04:07 any other annoncements? 21:04:21 emagana: this is only with respect to going into the "G-2 release" 21:04:29 not grizzly as a whole (if that is your question) 21:04:39 G-2 is frozen tuesday, and released on thursday 21:04:53 once G-2 is frozen and branch, G-3 opens up. 21:05:02 danwent: got it! I misunderstood you. Thanks 21:05:11 danwent: I did some devstack testing this weekend and it failed with webob 1.2.3 21:05:26 mlavalle: yes, that is what the above patch is supposed to address 21:05:34 #topic grizzly-2 milestone release 21:05:36 ok 21:06:06 ok, main goal for this meeting is to make sure we can close out the several patches we still have in review for G-2 by tomorrow 21:06:16 markmcclain: https://blueprints.launchpad.net/quantum/+spec/quantum-db-upgrades 21:06:28 are we waiting on merge of quota change first? https://review.openstack.org/#/c/19039/ 21:06:28 yes… just pushed a bunch of cleanup changes 21:06:42 main review is here: https://review.openstack.org/18341 21:06:42 no.. we can do it after 21:07:02 the quota ext makes a slight change that we can handle as a migration 21:07:11 markmcclain: ok. 21:07:36 I approved 19039… did I do wrong? 21:07:41 markmcclain: so my understanding is that we are good to merge on the patch, thoughs salv-orlando is going to finish his testing before we merge. 21:07:53 salv-orlando: 19039 21:07:57 is fine to merge 21:08:02 ok, thanks 21:08:03 we just should backport it stable 21:08:11 sorry *should not* 21:08:14 If everything goes according to plan we can merge db upgrades in a few hours 21:08:24 I am rebasing my patch with db changes on top of it already 21:08:31 salv-orlando: and we know how everything tends to go according to plan :P 21:08:50 I also updated instructions how to init migrations with a folsom db 21:08:58 *existing 21:09:03 No this time it's true :) - thanks to what mark said above ^ ^ 21:09:10 one question about db migration, why not use the tech the same as nova? 21:09:40 gongysh: i believe this is discussed in the spec 21:09:50 gongysh: we need the ability to control the order of migrations a little more than what sqlalchemy allows 21:10:04 sorry sqlalchemy-migrations 21:10:13 ok, got it. 21:10:18 thanks 21:10:58 next up. nati_ueno https://blueprints.launchpad.net/quantum/+spec/quantum-security-groups-iptables 21:11:18 danwent: I got +1 from gary, akihiro, yong 21:11:22 nati_ueno: are we still dealing with the issue of the noopfirewalldirver on nova? 21:11:43 danwent: It is not related with linuxbridge one 21:12:01 nati_ueno: ok 21:12:07 danwent: I'm not sure why, but it occurs with ovs only 21:12:10 here is reivew: https://review.openstack.org/#/c/16210/ 21:12:22 why? 21:12:33 i think its because existing bridge with quantum does not use type="bridge" ? 21:12:45 (unless that changed recently) 21:12:57 I can enable the nova firewall driver with linux bridge vif driver. 21:13:13 I think at least we have to doc it. 21:13:21 gongysh: Yes I agree 21:13:55 basically, this happens if you use LibvirtBridgeDriver or any driver that derives from it, including OVS hybrid driver 21:14:18 but will not happen with QuantumLinuxBridgeVIFDriver( 21:14:25 anyway, this is a detail, so back to the main patch 21:14:39 nati_ueno: so you +1's, but not +2s? 21:15:00 danwent: Akihiro was +2 21:15:07 LibvirtBridgeDriver expects libvirt Iptables Driver, that is the reason of the issue. 21:15:13 danwent, I can change my score to +2, if gate test passes. 21:15:47 gongysh: Gating failure is because of webob version, so it is not related this patch. 21:15:58 ok… hopefully by end of meeting the gate will have run to merge the webob change, at which point we would need to rebase 21:16:09 nati_ueno told me not to approve patches if ... 21:16:09 danwent: secgroup patch conflicts with aaron's patch. 21:16:38 gongysh: Yes you are right. lets wait. sorry 21:16:47 danwent: we need to determine which is merged first 21:16:51 amotoki: I agree 21:17:09 nati_ueno: linux bridge patch is ranked higher, so I don't want to jeopardize that merging in G-2 21:17:32 amotoki: how significant are the conflicts? 21:17:48 either one can be merged first. 21:18:01 #info expect db-migration patch to merge today 21:18:15 arosen1: are you oke with merging linux bridge stuff first? 21:19:07 is there a dependency between db-migration patch and security groups? I imagine security groups patch does not already include db migration code? 21:19:33 Ah That's could be 21:19:34 I think most challenging is db-migration patch to nati_ueno's patch. 21:19:35 danwent: sure it's an easy fix to update it. 21:19:46 there is, but I'll SG migration 21:19:52 arosen1: thanks. ok, we will merge linux birdge first. 21:19:58 markmcclain: in main patch, or afterward? 21:20:03 I'll rebase on nachi's patch in that case. 21:20:11 arosen1: Thanks! 21:20:11 depending on which lands first 21:20:16 markmcclain: ok 21:20:56 ok, so, sounds liek there are no outstanding concerns with SG patch? 21:21:21 gongysh: are you still planning on waiting until gate passes? that will require a base on webob patch, i believe 21:21:38 so db-migration first or sg patch first? 21:21:55 no, I don't need to wait. 21:22:10 gongysh: sounds like either is OK. mark will add the DB migrations to his patch or as a separate patch later 21:22:16 ok, so can we just merge it now then? 21:22:25 ok. 21:22:30 fire away :) 21:22:47 Ah.. that was long way.. 21:23:00 #info plan is to merge bridge security groups patch now. will rebase nvp security groups patch on top. markmcclain will add db migration for SG to his patch. 21:23:01 Thanks! 21:23:09 phew :) 21:23:17 ok, next up, salv-orlando https://blueprints.launchpad.net/quantum/+spec/quantum-service-type 21:23:18 yay! 21:23:27 salv-orlando: i think we're good to go here, once you add the migrations? 21:23:37 I think there's no other reason for concern 21:23:45 garyk and I have both been +2 21:24:02 danwent: salv-orlando : i can +2 now. it lloks good to me 21:24:07 salv-orlando: ETA on that repush? 21:24:24 I need a couple of hours, so it will be in garyk's REM phase 21:24:29 salv-orlando: i'm just worried, as garyk goes offline 21:24:30 yeah 21:24:37 but perhaps I can get a +2 from him tomorrow morning 21:24:45 it's still in time for making it into G-2 21:24:46 ok, so plan on pushing later today, and garyk +2's in teh morning? 21:24:52 salv-orlando: i will check it first thing in the morning 21:24:55 salv-orlando: yes, things can merge on tuesday 21:25:01 ok, thanks 21:25:02 FYI jeblairReminder: jenkins is offline for an emergency security-related upgrade, ETA in a few hours. 21:25:08 salv-orlando: I just got back from vacations but I can do the review today. 21:25:32 #info service-id is done. salv-orlando is adding db migration. danwent and garyk to re-review. should merge tuesday morning 21:25:46 emagana: things, though I think review on this is pretty much done expect for new db migration 21:26:03 next up, lbaas crud 21:26:06 sthakkar: https://blueprints.launchpad.net/quantum/+spec/lbaas-plugin-api-crud 21:26:09 danwent: ok 21:26:12 sthakkar is proxying for leon 21:26:31 looks like leon responded to mark's comment: https://review.openstack.org/#/c/16919/ 21:26:36 so i think there are a couple small issues 21:26:47 regarding unit tests that need to be cleaned up 21:26:48 there's a -1, but its more of a style think from avishay 21:26:59 commit message definitely needs to be updated 21:27:09 one more thing: Leon needs to make folder names consistent 21:27:16 other than that we are looking okay. two things to do: clean up style & tests 21:27:17 there are lb and loadbalancer 21:27:39 enikanorov: ok will mention that as well 21:27:43 enikanorov: yeah, i agree 21:27:50 i've already made this comment 21:27:55 ok, any other outstanding concerns? 21:27:57 home Leon will fix that 21:28:02 *hope 21:28:06 sthakkar: when is leon online (he's based on china, right?) 21:28:20 danwent: yep, around 4pm pst 21:28:45 leon is based on chain? 21:28:49 sthakkar: ok, can we have him update those style issues first thing his time, so people here can ideally approve by end of day monday? 21:29:10 gongysh: yeah, sthakkar should know exactly where 21:29:11 If he is, I will work closely with him. 21:29:18 danwent: yes, i think that should be fine. we'll make that happen 21:29:38 sthakkar: ok, great. and the two core devs on this are salv-orlando and markmcclain ? 21:29:50 I am core dev on Leon's patch, yes 21:30:01 with lots of great reviews from the marantis folks as well :) 21:30:10 sthakkar: can u tell me something about leon? 21:30:25 gongysh: perfect, that'd be great. he's based out of beijing 21:30:26 ]yes 21:30:38 ok, so those are the four critical or high blueprints for G-2 21:30:45 that should be the focus for us in the next two days. 21:31:19 #info lbaas crud expected to merge late monday or early tuesday (US) 21:31:34 #info currently there are no 'critical' or 'high' bugs targeted for G-2 21:31:47 if something arises, please bring it to the attention of the entire core team 21:32:02 we also have some untriaged bugs that I am working through: https://bugs.launchpad.net/quantum/+bugs?field.status=NEW&field.importance=UNDECIDED 21:32:16 if you see something on this list that corresponds to your sub-team, please jump in and triage it yourself. 21:32:31 Anything else for G-2 before we move on to talking about stable? 21:33:12 #topic quantum stable 21:33:33 garyk, doing a fantastic job on this as always, thanks for the email to the core team 21:33:42 anything to add here? 21:33:46 danwent: thanks 21:33:56 here is the list of current stable reviews garyk sent out: 21:33:57 https://review.openstack.org/#/c/19079/ - validity checks for gateway on subnet 21:33:57 https://review.openstack.org/#/c/19051/ - exception cleanup 21:33:59 https://review.openstack.org/#/c/18818/ - ensure allocation pools are deleted 21:34:00 https://review.openstack.org/#/c/18819/ - fix exception output (I need to write a test case on master - i do not think this should prevent it) 21:34:00 https://review.openstack.org/#/c/18651/ - atomic access to databases 21:34:15 danwent: a bit of karma to those will help 21:34:23 garyk: have they finalized date for folsom/stable release at end of jan? 21:34:37 danwent: https://review.openstack.org/#/c/19096/ 21:34:50 danwent: it is at the end of the month as far as i know 21:35:18 garyk: something specific about that last review, or just adding it to the list? 21:35:44 everyone on the core team should make sure they knock of a few stable reviews during their review day 21:35:46 danwent: just adding it to the list. the backport here was a bit challenging due to major code changes 21:36:17 garyk: yeah, agreed. my feeling is that the further away we get from a release, the higher the bar is for something to be backported, since the cost goes up. 21:36:31 +1 to that 21:36:35 agreed 21:36:50 ok, anything else on stable? 21:36:54 nope 21:37:16 #topic quantum tempest 21:37:21 mlavalle: any updates? 21:37:30 looks like we're now focusing on the two blueprint 21:37:35 https://blueprints.launchpad.net/tempest/+spec/quantum-basic-api 21:37:35 https://blueprints.launchpad.net/tempest/+spec/quantum-extended-api 21:37:50 any update from the tempest team meeting last week? 21:37:56 danwent: I sent a message to the ML with the inventory of all the BP's 21:38:00 (only needed if its relevant to the quantum team) 21:38:06 mlavalle: yup, thanks. 21:38:18 Jay cleanups some bps. 21:38:18 Jay approved my two BP's 21:38:25 working on them now 21:38:43 ok, anything else on tempest? 21:38:48 nope 21:38:49 mlavalle: My team guy is working on refactoring quantum tempest code. Let's talk later 21:38:59 nachi, i +2'd your patch to get temptest + devstack gating 21:39:05 danwent: Thanks 21:39:06 nati-ueno 21:39:08 ok 21:39:18 we've waited long enough to hear about any "style" concerns in my opinion :) 21:39:31 danwent: ha ha Yes true 21:39:31 https://review.openstack.org/#/c/18665/ 21:39:40 #topic quantum horizon 21:39:54 amotoki + nati_ueno, looks like you both made good progress here. 21:40:05 #info https://blueprints.launchpad.net/horizon/+spec/improve-quantum-summary-table (in review, expected for G-2: https://review.openstack.org/#/c/19037/) 21:40:07 garyk: Could you take a look 18665. if this one is merged gating will works 21:40:16 #info https://blueprints.launchpad.net/horizon/+spec/quantum-floating-ip (moved to G-3) 21:40:18 nati_ueno: sure 21:40:24 garyk: Thanks! 21:40:25 #info https://blueprints.launchpad.net/horizon/+spec/quantum-l3-support (in review, promising for G-2: https://review.openstack.org/#/c/19028/) 21:40:41 sounds like gabriel thinks both fo the G-2 items will still make it, which is great 21:40:52 anything else on horizon + quantum? 21:41:05 I have addressed concerns from Akihiro and Gabriel in latest patch 21:41:25 nati_ueno: thanks. will check soon. 21:41:30 #topic open discussion 21:41:31 amotoki: Thanks! 21:41:50 Can anyone tell me why the xcp+ovs patch is languishing? 21:41:51 https://review.openstack.org/#/c/15022/ 21:41:52 ok, i skipped the docs section before, in the interest of time. only open issue that is 'high' is https://bugs.launchpad.net/openstack-manuals/+bug/1088304 21:41:53 Launchpad bug 1088304 in openstack-manuals "quantum security group doc is confusing when it comes to use nova security group features" [High,In progress] 21:42:18 mnewby: i owe you a review on that as a thank you for the tempest work. 21:42:47 mnewby: but in general, when there is plugin work for a particular platform that people aren't all that familiar with and probably don't have a test setup for, getting reviews is much harder 21:43:04 mnewby: i find barter to be the best approach :) 21:43:05 danwent: My concern is that this is testable in two ways 21:43:16 mnewby: verifying that libvirt support still works - easy 21:43:20 oops, haha 21:43:47 danwent: verifying xcp is supported - easy, too, since people like John Garbutt of citrix have signed off on the changes as working 21:43:54 anyone else want to volunteer? 21:44:07 danwent: I even went as far as documenting the steps taken to verify xcp support, and still no +2s from core. 21:44:13 ok, if our main goal is making sure the code looks reasonable, and that it doesn't break existing stuff, that seems pretty easy. 21:44:25 danwent: I understand the horse-trading aspect, but it's been almost 2 months. 21:45:07 ok, on a related note, I expect a lot of people to be pushing new plugins for new platforms soon 21:45:40 mnewby: where is the document, can I have a look? 21:45:41 i have been encouraging people to add a link to their code on the main quantum wiki, where it talks about plugins that are out of tree 21:46:01 gongysh: http://wiki.openstack.org/QuantumDevstackOvsXcp 21:46:07 as a first step, so the quantum team can look at the code and provide high-level suggestions. 21:46:18 then the trick will be to find someone on the core team willing to maintain that plugin. 21:47:03 to ease the maintenance burden, i'm encouraging people to keep their plugins small and simple to start, until they have someone from their team contributing actively to the community. 21:47:38 there isn't really a decision we need to make today, this is more of just a heads up. I expect a flood of additional plugin in G-3, and not enough cycles from the core team to handle them all. 21:47:55 danwent: That suggests that such plugins should not be maintained by core. 21:48:09 danwent: A stable api might be preferable. 21:48:14 so if you're planning on contributing one, your best bet is to keep it simple, and to contribute to the community, so someone on the core team will decide its worth maintaining. 21:48:16 (plugin api that is) 21:48:17 I am feeling the plugins give us too burden to develop new features 21:48:31 mnewby: yes, that is the default position for all such plugins 21:48:39 danwent: cool 21:49:22 ok, any other open discussion? 21:49:33 Hi guys, talking about plugins 21:49:51 mnewby: +1 21:49:51 garyk helped a lot in reviewing the initial HyperV plugin 21:50:16 gongysh: yes, i understand. its a balance we need to draw as a team. 21:50:17 I was wondering if there is any chance to get it approved in the next days 21:50:39 gongysh: but adding new plugins can add new horsepower to the community as well, if it coaxes people to be more involved in community aspects. 21:51:01 alexpilotti: top reviewing focus for the next two days is in G-2 slotted items. 21:51:13 alexpilotti: is there a particular reason why merging is urgent? 21:51:16 alexpilotti: Is a wiki explaining the process to test it? 21:51:42 alexpilotti: or rather, more urgent than anyone else's urgent requests for reviews :) 21:52:01 danwent: the only reason is that we would like to send in also the NVGRE tunnel support afterwards 21:52:11 Initial patchset for the ML2 plugin should show up in the next week or two, so we can eventually look at much smaller/simpler drivers (with stable API) rather than full-fledged plugins 21:52:23 danwent: of course it's not importat to be G2 21:52:35 rkukura: What's ML2 plugin? 21:53:06 alexpilotti: you can stack patches on top of each other, no need for waiting for us sloths to review your patch :) 21:53:06 alexpilotti: ok, i would focus on identifying the core devs that will help you with the review and merge, in which case, it should merge quickly. 21:53:27 I can be one core - I think we're on the same time zone 21:53:32 I hope we can have a wait for rkukura's ML2 plugin, which allows coreplugin to driver multiple drivers at the same time. 21:53:50 nati_ueno: https://blueprints.launchpad.net/quantum/+spec/modular-l2 21:53:59 alexpilotti: remember that before this can merge, you need to get a core maintainer for the plugin. 21:53:59 rkukura: I got it 21:54:22 so one quantum server can servers multiple L2 agents. 21:54:44 alexpilotti: and that maintainer should send a note to the core team explaining why they think adding the plugin to the core repo is a good idea. 21:54:51 danwent: ok, from what I remember in a conversation we had 21:55:15 that somebody from our team could do that task 21:55:27 or is this something that happens afterwards? 21:55:51 alexpilotti: Provide instructions for testing will be great as well 21:56:02 emagana: will do! 21:56:16 alexpilotti: at some point in the future, hopefully someone from your team can do that. 21:56:22 danwent: not belonging to the Quantum core reviewers of course! 21:56:45 alexpilotti: in the mean time, you can hopefully convince an existing core dev that they can work with this person before they are a core. 21:56:51 alexpilotti: As Dan mentioned before, adding a link to the specific plugin wiki in the main quantum wiki 21:57:14 emagana: ok, tx. Do you have a link at hand? 21:57:29 alexpilotti: if you keep the plugin simple to start, and provide obvious value (i would argue hyper-v does), then convincing a core dev to do this shouldn't be too hard. 21:57:39 ok, let's take this discussion offline, so we can wrap up ontime. 21:57:50 anything else, before we close? 21:57:52 danwent: ok tx! 21:58:01 Brocade has a q-plugin and currently located at an external git site http://github/brocade/brocade - seeking help from code devs to review in order to push into quantum. Please send me email at sharis@brocade.com. 21:58:01 http://wiki.openstack.org/Quantum 21:58:25 ok, thanks folks. let's close up G-2! 21:58:28 #endmeeting