21:01:31 #startmeeting Networking 21:01:32 Meeting started Mon Aug 5 21:01:31 2013 UTC and is due to finish in 60 minutes. The chair is markmcclain. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:01:33 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:01:36 The meeting name has been set to 'networking' 21:01:57 #link https://wiki.openstack.org/wiki/Network/Meetings 21:02:16 hello all! 21:02:22 #topic Announcements 21:02:39 hi 21:02:40 #info Bug Day August 8th 21:02:42 hi 21:03:37 The number of open unresolved open bugs is currently hovering around 200 21:04:24 Thursday will be a good time to work on closing some 21:04:49 so more if a Bug Squashing Day than Bug Generating Day 21:05:02 s/if a/of a/ 21:05:10 are all of the 200 unassigned? 21:05:13 med_: If we find add new bugs that ok too 21:05:30 armax: it is a mix 21:05:45 k 21:05:46 some of the unassigned bugs need to be reviewed to make sure they still apply 21:06:08 The other quickly approaching deadline is August 23rd 21:07:11 That is the last day to propose code for Havana 21:07:15 #link https://wiki.openstack.org/wiki/FeatureProposalFreeze 21:07:30 * salv-orlando I fainted as I read "this is" instead of "that is" 21:07:52 haha… we still have a bit of work to do 21:08:44 #info End of day August 23rd is Feature Proposal Freeze 21:08:54 https://launchpad.net/neutron/+milestone/havana-3 21:09:05 Overall for H3 we're looking ok 21:09:28 9 implemented and another 20 in some state of review 21:10:03 we've got 15 or so that need code 21:10:50 #topic Bugs 21:11:15 Thanks to nati_ueno for working on the gate problem test 21:11:51 it is start vorting again 21:11:53 The infra team changed the neutron devstack gate to voting 21:11:59 Great work nati_ueno! 21:12:06 mestery: Thanks 21:12:13 which helps to prevent bit rot 21:12:34 markmcclain: speaking of deadlines, what about dkehn's bare metal support patches ? 21:13:03 lifeless: those patches have been set of a high priority item for H3 21:13:37 last I checked there were outstanding comment on the server patchset 21:13:44 markmcclain: cool! dkehn keeps running into hiccups and issues; -really- keen to get it landed ;) 21:13:52 same here 21:13:54 ah! dkehn ^ 21:14:24 We currently don 21:14:34 't have any critical bugs to talk about 21:14:47 any bugs that need to brought up to the team? 21:15:14 markmcclain: I have a non-critical that I need to ping you about; not sure if the meeting is the right place. 21:15:14 #topic Docs 21:15:35 2nd round of name changes has been completed and successfully merged 21:15:37 lifeless: we chat offline about it 21:15:40 markmcclain: ack 21:16:05 emagana_: great job! 21:16:06 I am now chasing individuals bugs and also reading the whole guide making sure everything make sense 21:16:20 I may bother people asking for links update 21:16:48 There are plenty of externals links in the manuals pointing to un-updated wikis 21:17:07 thanks for looking for those 21:17:37 VPNaaS and FWaaS can be reported by nati_ueno and SumitNaiksatam 21:17:43 nati_ueno: Thanks! 21:17:55 Thats all from my side! 21:18:15 nati_ueno: how are the VPN docs progressing? 21:18:26 markmcclain: no update for docs yet. 21:18:53 markmcclain: I'm working on API for VPN 21:19:11 pcm_: cool 21:19:17 pcm_: Thanks 21:19:22 np 21:19:26 SumitNaiksatam: Any updates on the FWaaS docs? 21:19:30 nati_ueno: I can help if you need it! 21:19:40 markmcclain: i am just getting started 21:19:55 ok 21:20:00 emagana_: Thanks. I'm waiting the driver patch is merged. Because it may be changed 21:20:14 Any other questions on docs? 21:20:23 #topic API 21:20:24 we need to get ml2 info into the docs at some point soon 21:20:31 #undo 21:20:32 Removing item from minutes: 21:20:46 rkukura: correct 21:21:00 rkukura: I can help with that. 21:21:03 I'll add that as a major doc item we'll be tracking the rest of this cycle 21:21:24 sounds good - thanks mestery and markmcclain 21:22:16 #topic API 21:22:23 salv-orlando: hi 21:22:27 hi 21:22:56 Mostly quiet today. We've merged FW and VPN extensions. 21:23:05 We still have a few extensions outstanding 21:23:19 but all the API-related items are already under review 21:23:40 any question? 21:23:50 great to hear 21:24:04 # topic VPN 21:24:15 OK Thanks. DB part is merged. 21:24:24 so waiting ipsec driver is meged 21:24:41 I got +2 from Akihiro. 21:24:51 markmcclain: Do you have some concerns fro the patch? 21:25:16 nati_ueno: I need to circle back on the driver now that ctypes was replaced with mount 21:25:36 markmcclain: gotcha Thanks 21:25:40 That's all from my side 21:25:49 Thanks for updating 21:26:07 #topic FWaaS 21:26:40 hi 21:26:45 Driver patch is ready, agent patch needs to add one more response to an outstanding comment, CLI (needs one more +2) and Horizon are also ready 21:27:27 we have SridarK and RajeshMohan today (back from India) 21:27:31 SumitNaiksatam: great looks like we have plenty of core coverage on the open reviews 21:27:57 markmcclain: yes, in my opinion the CLI is ready to right away 21:28:14 agent should be ready by tomorrow or so 21:28:24 driver is already ready but will need to wait for agent 21:28:37 i sent out the proposal on the explicit commit on firewall 21:28:43 got positive feedback 21:28:55 are we all fine in making this change? 21:29:09 that is changing from the current implicit commit to explicit commit? 21:29:22 SumitNaiksatam: The explicit commit model is fine with me. 21:29:32 mestery: thanks 21:29:45 we're adding a bit of state management in restful APIS 21:29:49 which should be stateless 21:29:52 but that said 21:30:02 we've violated principles so many times 21:30:13 that I just risk sounding pedant 21:30:20 haha 21:30:21 and also I don't have a viable alternative 21:30:29 so I should shut up 21:30:32 salv-orlando: :-) 21:30:32 * salv-orlando doing that now 21:30:45 any thoughts from others? 21:31:01 should we submit a patch? 21:31:14 SumitNaiksatam: I wanna know implementation plan 21:31:22 the details are always the big hangup 21:31:29 WTF, the agent restareted with uncomit change? 21:31:45 nati_ueno: WTF? :-) 21:31:57 aaa sorry typo 21:32:00 What's if 21:32:01 nati_ueno: that can be handled 21:32:12 i will put this in a spec 21:32:26 so server side keeps previous state? 21:32:27 SumitNaiksatam: I think writing a BP is a logical next step 21:32:29 can I assign to the driver api implantation? 21:32:30 i will add it to the existing google doc 21:32:46 nati_ueno: yes, it has both pending and committed changes 21:32:51 it -> server 21:33:02 given the late stage in the cycle.. I'd rather track this separately 21:33:28 markmcclain: separately as in a different blueprint? 21:33:32 yes 21:33:36 markmcclain: yeah sure 21:33:56 can I? 21:34:14 i am fine not putting this change for H, however i think its something that is required, so better have it before people really start using this 21:34:26 hence this suggestion 21:34:34 i will put the blueprint out 21:34:54 SridarK, RajeshMohan anything else to add? 21:35:05 Nothing from side... 21:35:15 will push out review comments responses in the next day 21:35:19 for Agent 21:35:24 SumitNaiksatam: agreed… timing is my biggest concern… three weeks is a real short timeline 21:35:33 SumitNaiksatam: If I get the proposal right, this would change the API in a backward incompatible way. So better have a blueprint soon. 21:35:35 markmcclain: i am fine either way 21:35:46 salv-orlando: yes, sure 21:35:51 as we'd rather avoid a bw incompatible change after Havana is released 21:36:00 +1 21:36:07 salv-orlando: thats right, hence the suggestion to do now, versus later 21:36:17 that is assuming we want to do it this way 21:37:00 #action SumitNaiksatam to file BP on rule commit changes 21:37:10 Any other questions for FWaaS team/ 21:37:53 #topic LBaaS 21:38:04 hi 21:38:44 Looks like most of the changes have just minor clean up items 21:38:44 there is one major patch on review - integration of lbaas plugin with service types. 21:38:57 Akihiro did initial review 21:39:08 I'd like salv-orlando and nati_ueno to take a look also 21:39:08 https://review.openstack.org/#/c/38120/ 21:39:20 as they participated in service types discussion a lot 21:39:29 salv-orlando: do you have cycles to look at this? 21:39:34 I will find cycles 21:39:42 thanks 21:40:04 enikanorov_: gotcha 21:40:27 Anything else? 21:40:35 there are two outstanding items - some driver/agent infrastructure refactoring made by obondarev to simplify plugin-driver implementation for vendors 21:40:50 to be pushed on review this week 21:40:59 and delegating port creation to driver 21:41:31 that one is a part of 'multiple vips per pool' BP which full impl could be postponed to I 21:42:14 that's all from lbaas side 21:43:10 thanks for the update 21:43:16 #topic ML2 21:43:22 mestery, rkukura: hi 21:43:46 My only update here is that support for ML2 with OVS/LB agents is now in devstack. 21:43:56 great work 21:44:00 It would be great to have people try this out and report any issues (ML2 related). 21:44:09 this include gre and vxlan support 21:44:17 Yes, VLAN, VXLAN, and GRE support. 21:44:30 nice 21:44:34 rkukura: mestery: awesome work! 21:44:38 woot. 21:44:42 On the MechanismDriver front, we're starting to see some MechanismDrivers show up now in BPs and code. 21:44:44 mestery: do we have wiki page about how to use devstack for ml2? 21:45:01 I wanna know localrc 21:45:06 gongysh: Yes: https://wiki.openstack.org/wiki/Neutron/ML2#Using_ML2_in_Devstack 21:45:28 thanks 21:45:31 When we make ML2 as default in devstack? 21:45:33 We're working on an OpenDaylight MechanismDriver now as well, and hope to get that in for H3 yet. 21:45:49 nati_ueno: Soon, I hope. I need to talk to Dean and Troy this week about that. 21:46:00 ^^^ Dean and Sean 21:46:02 mestery: Error: "^^" is not a valid command. 21:46:03 mestery: Thanks 21:46:44 Any other questions for ML2 team? 21:46:49 At this point, I suggest we require all new merges to include support for ml2 for anything they do to openvswitch or linuxbridge 21:46:50 mestery: may be, if neutron-team agrees to let ML2 as default, 21:47:00 mestery: we can do it just changing default plugin on devstack 21:47:16 nati_ueno: Yes, but I'd still like to get their input and blessing. :) 21:47:26 mestery: gotcha :) 21:47:26 rkukura: I agree, and I've been noting that in review already too. 21:47:44 rkukura: I'm fine enforcing that policy 21:47:59 I noticed we missed ext-gw-mode extension for ml2, so will file bug for that 21:48:17 rkukura: Good catch, thanks! 21:48:19 rkukura: It was merged before ML2 21:48:34 it should be super-trivail because 21:48:37 it's L3 21:48:39 rkukura: thanks for filing it 21:48:39 agreed 21:49:08 #topic Stable Branch 21:49:25 garyk couldn't make it today 21:49:40 The stable branch is currently frozen 21:50:41 If something is missing please let me and garyk know. 21:51:18 #topic Open Discussion 21:51:38 I wanted to give you an heads up on Bug 1207402 21:51:40 Launchpad bug 1207402 in neutron "neutron should automatically stamp the database version when it is deployed" [High,In progress] https://launchpad.net/bugs/1207402 21:52:11 Basically now that we have migrations, the mechanism that auto-generates the schema is superfluous if not dangerous 21:52:24 agreed 21:52:26 dangerous because it does not stamp to any version 21:52:59 and also because it does not perform table deletion or update (like when you add an attribute) 21:53:02 so we're removing it 21:53:09 or at least… we're proposing to remove it! 21:53:21 if you have a reason for not removing it, please speak up! 21:53:27 We have 3 patches 21:53:55 the devstack patch is the first that should go in: https://review.openstack.org/#/c/40243/ 21:53:57 salv-orlando, I haven't seen the test_extension_ext_gw_mode.py tearDown fix merge yet, nay news 21:54:31 s/nay/any 21:54:43 dkehn: I put it in WIP as the problem is a bit more general with test_config. I'm trying to do something which is not a workaround if I can find some cycles for it 21:55:04 The other two patches are neutron, and are currently in draft as they cannot be merged if the devstack patch does not go in 21:55:05 salv-orlando, k, I'll leave it in my pathc then for now 21:55:15 After Sumit and Salvatore's advice I've drastically cut the size of the embrane's neutron plugin patch, furthermore we now have a core-dev sponsor (Thanks SumitNaiksatam). Please give a look to it if you have some cycles :) 21:55:17 #link https://review.openstack.org/#/c/38222/ 21:56:04 salv-orlando: +1 removing the auto generated db schema 21:56:06 patch 1: https://review.openstack.org/#/c/40241/ and patch 2 (the real deal is here): https://review.openstack.org/#/c/40296/ 21:56:57 salv-orlando: is that draft mode? 21:57:09 draft mode for both patches 21:57:14 do you get access denied? 21:57:18 salv-orlando: yes 21:57:19 salv-orlando: Permission denied. 21:57:20 yes 21:57:28 On both patches 21:57:49 I always forget about it. Instead of adding all of you as reviewers, I'll put them in wip 21:58:01 salv-orlando: Thanks 21:58:12 done 21:58:59 salv-orlando: Thanks. 21:59:04 tick-tock. 22:00:09 salv-orlando: thanks for working on this change 22:00:31 ivar-lazzaro: thanks for reducing the size of the patch 22:00:46 Any of other items for open discussion? 22:01:00 It'd be great if the blueprint could be accepted in H3, given that the code seem easy enough for you yo review ofc... please consider that 22:01:14 yo/to 22:01:45 Ok we're out of time… 22:02:16 Remember Thursday is a bug day 22:02:19 #endmeeting