21:02:15 #startmeeting Networking 21:02:16 Meeting started Mon Feb 17 21:02:15 2014 UTC and is due to finish in 60 minutes. The chair is markmcclain. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:02:17 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:02:19 The meeting name has been set to 'networking' 21:02:21 hello 21:02:55 #link https://wiki.openstack.org/wiki/Network/Meetings 21:03:03 #topic Announcments 21:03:34 #info Oleg Bondarev (obandarev) is our newest core reviewer 21:03:40 Welcome to Oleg! 21:03:55 Congrats obandarev! 21:03:56 welcome! 21:04:00 hi! 21:04:01 Congratulation obondarev 21:04:16 #info Feature proposal deadline is the end of day tomorrow (Feb 18th) 21:04:18 Congrats obondarev 21:04:41 yea! 21:05:03 congrats oleg 21:05:35 #link https://launchpad.net/neutron/+milestone/icehouse-3 21:06:02 We have many blueprints that do not have any code associated with them 21:06:32 I will move any blueprints that do not have associated with them out of Icehouse 21:06:47 markmcclain: that will be tomorrow right? 21:07:08 just saying because I have one for which I'm finishing the code, and I was planning to push it tomorrow 21:07:14 salv-orlando: correct.. as long as it is Feb 17th somewhere in the world, code can be submitted 21:07:24 ah I thought 18th 21:07:34 if there oops you're right 18th 21:07:52 ok, that's fine then 21:08:08 oops? :( 21:08:27 yeah gongysh markmcclain had a kernel panic 21:08:27 gongysh: sorry forgot to delete 'if' 21:08:30 he's rebboting now 21:08:42 :) got it. 21:09:04 end of the day Feb 18th is the cutoff 21:09:53 if there is an important blueprint that needs an extension, please send me a PM and we can discuss the options 21:10:29 after the cutoff we'll spend time reviewing and revising 21:11:17 Note: this deadline only applies to features and not bugs 21:11:22 Questions? 21:11:42 markmcclain: so codes shouldn't be WIP? 21:12:05 nati_ueno: correct the proposal must be ready for review 21:12:13 markmcclain: gotcha 21:12:44 #topic bugs 21:13:02 The gate has been fairly stable since we got all of the fixes in place 21:13:09 we're currently working on new client release 21:13:16 Any other bugs the team should discuss? 21:13:59 #topic Docs 21:14:02 emagana: hi 21:14:08 markmcclain: Hi There! 21:14:41 In order to keep it short I have updated the wiki with the list of high priority bugs 21:15:06 most of them should be completed by this week (short fixes) 21:15:08 great thanks for updating the wiki 21:15:37 I have requested the entire team to create BP or bugs or any new feature to be follow in icehouse 21:16:08 So, just a kind reminder.. besides all new drivers in services I have not seen any major feature change 21:16:22 seems like we need to crosscheck any commit messages with DocImpact 21:16:38 to make sure there is the corresponding doc fix or bug 21:17:05 emagana: currently right now we have not merged any major changes 21:17:18 I expect we'll have a few land over the next two weeks 21:17:23 markmcclain: what about OVS and LinuxBridge code? 21:17:24 Anything else? 21:17:42 emagana: until we have a migration script we'll have to carry them 21:17:57 hopefully we'll get more info in the ML2 update section 21:18:01 markmcclain: Got it! 21:18:14 ML2 team, please point me to any BP describing this process. 21:18:14 #topic Nova Parity 21:18:19 emagana: thanks for the update 21:18:19 hi! 21:18:23 beagles: hi 21:18:44 not much movement last week, many of the reviews that were pending still are... maybe because of the sync up? 21:19:07 yeah a good chunk of the Nova team was in the meeting 21:19:27 I know I went through and looked at many of the Nova reviews 21:19:38 sweston and I were like ships in the night and missed each other, so we are going to sync up immediately after this meeting 21:19:55 ok 21:20:23 Anything else? 21:20:23 but that's pretty much it for now 21:20:33 beagles: thanks for updating 21:20:50 a few extra Nova bits 21:21:19 First, it was good to sit down with many of the Nova core team last week 21:21:38 we were able to talk about ways to improve the workflow a bit 21:21:54 arosen1 has been working on a few of these items 21:22:22 one of the items the Nova team did ask is that when changes are made the Neutron module within Nova that we get a Neutron core to look at the changes 21:22:52 a review looking at the changes from the Neutron perspective if very helpful for them 21:23:17 definitely 21:23:25 good idea markmcclain 21:24:06 was there discussion about interaction with the VIF plugin? 21:24:27 beagles: a little bit 21:25:05 we discussed a few ways to make Nova wait until Neutron has notified that the port is ready before booting 21:25:26 Any other Nova items? 21:26:15 #topic Tempest 21:26:17 mlavalle: hi 21:26:20 hi 21:26:40 The group contributing to api tests has been chugging along 21:26:52 You sent out an update via email earlier today: http://lists.openstack.org/pipermail/openstack-dev/2014-February/027362.html 21:26:54 #link http://lists.openstack.org/pipermail/openstack-dev/2014-February/027362.html 21:26:59 earlier today I sent a message to the ML 21:27:20 with all the patch sets classified to make it easy to cores help us review the code 21:27:40 as you can see in the message, we are covering the complete api 21:27:41 that email was very helpful, thank you for sending it out 21:28:05 we are in a good position to be done by the end of ichouse 21:28:15 icehouse^^^ 21:28:22 that's great news 21:28:56 I am also planning to create a view in gerrit, to make it even easier to review 21:29:06 awesome 21:29:08 that was an idea from the qa irc meeting of last week 21:29:26 that's a good one 21:29:27 that's all I have, thanks 21:29:43 any questions on Tempest? 21:30:21 i have one, but i'd like to leave it to open discussion 21:30:27 it mostly technical 21:30:30 enikanorov_: ok 21:30:39 we'll move for now then 21:30:50 #topic L3 21:31:04 Doesn't look like carl_baldwin is around today 21:31:43 There was a meeting of the folks interested in DVR last week 21:32:48 #link https://docs.google.com/presentation/d/1XJY30ZM0K3xz1U4CVWaQuaqBotGadAFM1xl1UuLtbrk/ 21:33:06 Those are the latest slides for more info please be sure to join their IRC meeting 21:33:11 #topic IPv6 21:33:14 sc68cal: hi 21:33:49 hello 21:33:56 There is nothing to really report since last week 21:34:07 ok 21:34:22 just need to get some review attention, right? 21:34:41 yes 21:35:22 ok 21:35:30 Any questions on IPv6? 21:35:48 #topic ML2 21:35:56 mestery, rkukura: hi 21:36:35 markmcclain: hi! 21:37:08 this was raised earlier.. what is the status on the migration blueprint? 21:37:09 So, the main item here is that the Migration tool BP still doesn't have any code pushed out yet. 21:37:10 #link https://blueprints.launchpad.net/neutron/+spec/ml2-deprecated-plugin-migration Migration Tool BP 21:37:20 Without that, we can't start deprecating Linuxbridge and OVS plugins 21:37:23 correct 21:37:36 rkukura and I got ahold of the assignee, but no code has been proposed yet. 21:38:04 So, it's kind of stuck right now I think. 21:38:21 Would this be a candidate for an extension if someone else picks it up and does something as simple as possible? 21:38:59 markmcclain: I'm willing to take it on. 21:39:12 marun: That would be awesome! 21:39:18 marun: thank you for stepping up 21:39:27 marun: +1 21:39:57 rkukura: yes this is definitely a candidate for an extension 21:40:54 Besides a more reviews anything else to highlight? 21:41:14 nothing else new I can think of 21:41:16 Reviews for all the MDs proposed would be great to get from cores markmcclain 21:41:58 all the generic stuff needed for SR-IOV is in review 21:42:05 ok and I've cleared any -2s due to testing requirements? 21:42:09 on the neutron side, but not on nova 21:42:41 rkukura: I thought the PCI stuff on nova was moved out of Icehouse now by Russell? 21:42:48 I thought I saw an email about that. 21:43:13 I think his reasons were review cycles this late in Icehouse, let me find that and send it to you rkukura. 21:43:35 mestery: Most likely some nova pieces won't make icehouse, but we should get the generic support into neutron if possible 21:43:47 rkukura: Agreed on that point. 21:44:36 ok.. if there are any bits that are not generic let's temporarily move them to WIP 21:46:14 That's all on ML2 from me this week. 21:46:16 looking at https://review.openstack.org/#/c/73500/ 21:46:31 that's the generic one we need to merge early right? 21:46:45 markmcclain: one of them 21:47:24 markmcclain: others are https://review.openstack.org/#/c/72452/ and https://review.openstack.org/#/c/72334/ 21:48:10 ok 21:48:21 I'll work on getting reviewers for those 21:48:25 Any other ML2 items? 21:48:29 markmcclain: thanks! 21:48:36 none from me 21:48:41 None here 21:49:05 #topic VPN 21:49:17 nati_ueno: do we have the agent code yet? 21:49:18 or is it likely to miss? 21:49:53 markmcclain: sorry disconnected 21:49:56 so topic is vpn? 21:50:06 nati_ueno: no worries 21:50:19 yes was wondering that status of the agent 21:50:36 We have WIP code for agent, and me and Rajesh is working on it. I think we can make it ready for review until tommorow 21:50:55 nati_ueno: ok 21:50:57 Anything else? 21:51:03 markmcclain: FYI, rebased VPNaaS Service Type Framework server code and updated review. 21:51:14 pcm_: thanks 21:51:16 I'll update wiki for how to use it 21:51:20 pcm_: Thanks. 21:51:25 That's all from me 21:51:28 nati_ueno: np 21:51:34 need reviews of it. 21:52:04 it is great if we can have Service Type Framework and SSL-VPN in icehouse 21:52:18 ok 21:52:30 i have vendor plugin out now and it depends on STF 21:52:35 nati_ueno: Could you also update docs besides the wiki? 21:52:43 emagana: sure 21:52:51 nati_ueno: Thanks! 21:52:52 the important thing is that we get them right 21:53:02 #topic FWaaS 21:53:09 i've been running STF for months now 21:53:23 Looks like we have several items needing review 21:53:28 hi 21:53:35 yeah, things are ripe for review 21:53:50 as with VPNaaS, STF is waiting for another core 21:54:16 once, that gets in we have the firewall insertion patch 21:54:17 I added a comment about the relative lack of tests for that change 21:54:19 <10% of code 21:54:27 STF? 21:54:51 yes 21:54:55 ok, i see that 21:54:59 Anything else on FWaaS? 21:55:11 turning fwaas on in the gate 21:55:40 running FWaaS tempest tests will require that 21:55:55 I think we're probably ready to add it 21:56:04 ok 21:56:06 let's wait until Thursday 21:56:14 sure, whenever you decide 21:56:30 I'll follow up with on Thursday 21:56:33 #topic LBaaS 21:56:36 sure 21:56:58 enikanorov_ and I have been discussing models in the main channel, so I know the status of the subteam 21:57:04 hi, so we have to put on hold l7 feature until the isntance approach is sorted out 21:57:10 #topic Open Discussion 21:57:21 enikanorov_: you had an open discussion topic 21:57:35 yeah, that's a question about lbaas tempest scenario 21:57:47 we only have 3 minutes, so may have to jump over to -neutron to finish discussing 21:58:02 may be someone culd tell if running http backends on the host (instead of tenant network) is a good idea 21:58:13 that could save a VM spinup 21:58:29 but maybe it's not a real-life case 21:58:53 i'm not of the opinion that we need vm's for all network testing 21:59:04 unless there is vm-specific config required 21:59:21 so this would be spinning up a server process in a namespace? 21:59:29 marun: agree. but then loadbalancer will access members which are out of tenant network 21:59:34 and then having it load balanced? 21:59:38 markmcclain: no 21:59:52 we can't do so because in that case test will not be generic 22:00:03 so other solutions (other than haproxy) will not work 22:00:25 yeah I think we should spin up the members as normal 22:00:29 even if it is slower 22:00:45 otherwise we're testing a case that isn't representative of deployment 22:01:00 ok, good to know your opinion 22:01:10 so in fact we have 2 patches (with both variants) 22:01:19 ok.. I'll look for them 22:01:22 so far i tend to go with VM approach as well 22:01:32 Ok we're over our allotted time 22:01:34 the only item I had is that we're adding parallel testing in non voting mode: https://review.openstack.org/#/c/74131/ … I've run some parallel tests already with 'check experimental', and I'm quite positive we can switch the test to voting before I-3 deadline. The only observed issue so far is the 'famous' SSH protocol banner error, which however seems related to floating IP reassignment. I will look into that, I'm not even sure whethe 22:01:35 it's a neutron or testing issue. 22:01:39 especially when other approach not working well :) 22:02:01 salv-orlando: +1 to enabling parallel testing 22:02:30 as long as it works quite reliably I'm happy too. But we should triage it as non voting for 1-2 weeks 22:02:37 at least 22:02:42 agreed 22:03:34 Thanks to everyone for stopping by this week. We have a lot of exciting items in review. 22:03:49 #endmeeting