21:03:26 #startmeeting Networking 21:03:27 Meeting started Mon Jul 15 21:03:26 2013 UTC. The chair is markmcclain. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:03:28 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:03:30 The meeting name has been set to 'networking' 21:04:02 #link https://wiki.openstack.org/wiki/Network/Meetings#Bugs 21:04:03 markmcclain: hi! 21:04:37 Tomorrow's the H2 cut, so I'd like to spend most of the time talking about reviews we have outstanding 21:05:28 https://launchpad.net/neutron/+milestone/havana-2 21:05:54 we still have lot of items inflight 21:06:29 the good thing is that everything on here was proposed in time, but we don't have to cram it through just to make the h2 cut 21:07:41 if something needs another 1-2 days to clean up we'll still be ok for the larger Havana schedule 21:07:48 hi all, little late!! 21:08:03 * salv-orlando has already removed hist items which are unlikely to merge soon 21:08:32 I also want to say thanks to nati_uen_ for tracking down a nasty bug that was breaking the main devstack gate 21:08:55 #topic Bugs 21:08:59 markmcclain: Your welcome 21:09:08 https://bugs.launchpad.net/neutron/+bug/1194026 21:09:10 Launchpad bug 1194026 in neutron "check_public_network_connectivity fails with timeout" [Critical,In progress] 21:09:33 I have also found this one today: bug #1201478 21:09:35 Launchpad bug 1201478 in neutron "OverflowError: timeout is too large in gate-neutron-python26" [Undecided,In progress] https://launchpad.net/bugs/1201478 21:09:40 it's probably not as nasty 21:09:43 but still annoying 21:09:49 nachi has proposed a fix here: https://review.openstack.org/#/c/36890/ 21:10:11 salv-orlando: could you help me to improve commit message? 21:10:19 Looks like both amotoki and salv-orlando have been actively reviewing this one 21:10:23 I'll send you a draft after the meeting 21:10:38 yep, it's good for me. I am pestering nati_uen_ about the commit message 21:10:39 salv-orlando: Thanks! 21:10:54 because he's the lucky winner of today's pedantry prize 21:10:59 :) 21:11:09 Currently this test cause the infras team to turn off the neutron gate check. The longer this check is non-voting the greater the chance that bit rot sneaks in 21:11:33 armax: thanks for pointing out that bug 21:11:46 salv-orlando: haha I should still improve my english 21:12:29 markmclain: I think we should be able to merge it in the next hours 21:12:35 markmcclain: np 21:12:43 markmcclain: ^^ 21:13:04 salv-orlando: great 21:13:18 amotoki: would you mind looking at armax's fix? 21:13:42 markmcclain: i haven't checked it yet. 21:14:07 markmcclain: will look it soon. 21:14:14 amotoki: ok.. it's a small fix 21:14:30 Any critical bugs the team needs to know about? 21:15:02 #topic Docs 21:15:02 yeah…it's non critical but it'd be good to have a look at it as it may cause false positives during the review process 21:15:35 emagana has updated the docs for neutron rename 21:15:36 https://review.openstack.org/#/c/36322/ 21:15:37 on docs I am waiting for reviews on: https://review.openstack.org/#/c/36322/ 21:15:56 when folks have some time please review as quality docs are critical 21:16:10 thanks emagana for updating these 21:16:11 this one would also appreciate some love: https://review.openstack.org/#/c/32979/ 21:16:25 markmcclain: Thanks!! Please Folks, add your comments! 21:16:48 it's not strictly neutron…it documents how to pass network info to nova 21:17:22 armax: right but still impacts us 21:17:30 yup 21:17:43 nati_uen_: please look at my comments for https://review.openstack.org/#/c/36890/ 21:17:54 garyk: yes sir! 21:18:07 armax: I will review that. For some reason I can +2 it. 21:18:14 nice 21:18:15 how about renaming in other docs? Is anyone working on netconn-api or other guides? 21:18:33 amotoki: I will do the netconn-api renaming 21:18:34 salv-orlando: you the man :) 21:18:47 salv-orlando: thanks 21:18:55 i will do it for ops-guide 21:18:57 amotoki: I don't think that anyone has doubled checked the netconn-api docs. 21:19:15 just one more update, will be adding one more commits for the rest of the manuals (nova, installation, etc..) please, also review them 21:19:26 markmcclain: we mostly already used the generic term 'network' there. But something might have been missed. 21:19:32 markmcclain: i am talking about renaming to neutron in netconn-api 21:19:54 amotoki: there are also bugs https://bugs.launchpad.net/openstack-api-site/+bugs?field.tag=netconn-api but a few in progress, which is good 21:20:13 #action salv-orlando to ensure the q-word is not anymore anywhere in netconn-api 21:20:24 thanks salv-orlando 21:20:26 salv-orlando: thanks 21:20:30 q-word lol 21:20:52 Any other docs related items? 21:21:21 salv-orlando: https://bugs.launchpad.net/openstack-api-site/+bug/1201596 21:21:23 Launchpad bug 1201596 in openstack-api-site "Project rename for netconn-api" [High,Confirmed] 21:21:24 just for you 21:21:44 annegentle_: that's very kind of you :) 21:22:49 #topic API 21:23:03 salv-orlando: hi 21:23:59 since today is mostly about open reviews 21:24:12 the status is that firewall and VPN API patches are very close to merge 21:24:33 snaiksatam added something to the firewall patch over the weekend and this triggered another round of review 21:25:16 nati_uen_ has just lifted the wip put on the API patch 21:25:30 salv-orlando: thanks for the review, this was a pending item per your comment earlier i realized i hadn't added 21:25:33 for VPN… so we are ready for another and hopefully last round of review 21:25:46 np 21:26:06 salv-orlando: I have fixes the code based on Salvatore and Akihiro's comment, and pushed it 21:26:11 I think service provider patch is slipping to H-3, unless I hear otherwise from enikanorov_ 21:26:21 enikanorov: ^^^ 21:26:30 well 21:26:44 I'd like to hear something from markmcclain since he's still -2 on this patch 21:26:54 I have moved the authZ patch to h-3 as the last bit I need to are some changes into the agent mgmt patch and gongysh has just pushed another patch there, so I do not want to interfere 21:27:08 salv-orlando: enikanorov pinged me about it.. I've got comments to publish on service_types 21:27:14 I'm ready to resolve any comments (Have updated the wiki with description already) 21:27:37 it's also a medium feature and I'd like to ensure that VPN and FW get reviewed first with everyone's limited cycles 21:28:02 #topic VPN 21:28:04 salv-orlando: we can put my patch to h-3 21:28:12 yep, i think we're in no hurry for service types but i would like to see it merged in early H-3 21:28:14 OK we are working on review comment. 21:28:22 enikanorov: same here 21:28:27 salv-orlando: akihiro: and the others thenk you for your reviews 21:28:48 Cleanup for API side is finished. Swami is working on Client side. 21:29:00 I'll fix the driver side in today 21:29:22 One good news is heat support will be come up after we merged the vpn. 21:29:25 That's all 21:29:41 markmcclain: I have discuss with you namespace wrapper 21:29:56 nati_uen_: ok.. we can discuss offline 21:30:09 markmcclain: OK I put my thought on the gerrit 21:30:13 OK thanks 21:30:39 nati_uen_: who's going to write the api docs for VPN? 21:30:45 and regular docs? 21:31:16 markmcclain: Satich was raised hand 21:31:30 markmcclain: But If he don't do that, I'll take it 21:31:30 nati_uen_: I can help you on that! 21:31:34 emagana: Thanks! 21:31:45 nati_uen_: let's touch based offline! 21:31:52 emagana: gotcha 21:32:03 we have good wiki, so IMO it is easy to write docs 21:32:17 Ok.. anything about VPNs? 21:32:34 markmcclain: That's all. 21:32:45 #action emagana create bp on docs for VPNaaS 21:32:54 nati_uen_: thanks for the update 21:33:05 #topic FWaaS 21:33:08 SumitNaiksatam: hi 21:33:12 hi 21:33:17 All fwaas patches are getting reviews, and we have two or more cores on each. The driver patch needs a little more core review attention. 21:33:27 I got great feedback on the API patch. 21:33:34 I believe all reviewers' comments have been addressed and I don't think there are outstanding issues. 21:33:36 I think I reviewed it but id not left comment 21:33:41 the driver patch 21:33:51 I will have another look at it. 21:33:54 yeah, big thanks to Salvatore, Akihiro, Aaron, Eugene, Oleg and Paul for the continued reviews. Hopefully we can wrap this patch soon. 21:34:03 i mean the API patch 21:34:13 great glad to see we've got good core coverage 21:34:18 and hopefully the agent and driver patches as well 21:34:30 Who's writing api and regular docs for this? 21:34:31 Devstack patch (Ravi) is also close to getting pushed, and if it lands soon it might help in reviews. 21:35:23 If you don't mind, I want to help write doc since I am not familiar with fwass. 21:35:32 i will start working on the docs with the rest of the team 21:35:38 gongysh: yeah sure 21:35:46 :#action SumitNaiksatam create bp on docs for FWaaS 21:35:58 SridarK (agent patch) has a question on the RPC version 21:36:06 SridarK: you there? 21:36:11 #action SumitNaiksatam create bp on docs for FWaaS 21:36:24 y On RPC - adding a method for Agent -> Plugin - since it is a new method - backward compatible - so we bump up the minor version on both ends ? 21:36:45 the question is more in the context of the L3 agent 21:36:57 got some diff opinions, follwing comments in rpc/dispatcher.py 21:37:08 Yes I have some concern 21:37:11 since we are using the l3 agent footprint for the firewall agent 21:37:30 Current code changes l3-agent directly. 21:37:31 ping me if you want me to look at a review WRT rpc versioning ... i added the support for it in the rpc code originally 21:37:58 thanks 21:38:56 nati_uen_: can you clarify by what you mean by changing code directly? 21:39:33 SumitNaiksatam: current code is adding mix in for l3-agent class. IMO, we agreed we don't touch l3-agent in this experimental version of services 21:39:35 the fwaas agent functionality kicks in only if the fwaas driver is configured 21:40:19 so fw and vpn will conflict for l3 agent? 21:40:24 i thought that this is the understanding and we try to look for a common framework across services in H3 21:40:28 nati_uen_: i don't recall that we had this agreement, in fact the agreement was to leverage the footprint of the l3 agent using the mixing approach 21:40:44 enikanorov: it will not be conflict 21:41:04 nati_uen_: are you talking about the issue we have discussed today? 21:41:10 about l3 agent? 21:41:25 enikanorov: yes 21:41:42 ah IMO, Mark said we should use inherit approach at first. 21:41:52 by conflict I meant that it would not be possible to configure both vpn and fw at the same time 21:41:54 If it is my misunderstandings, it is OK 21:42:19 inheritance was my preferred path for early development 21:42:37 enikanorov: it is not tested, but it looks like we can use both of it. (i'm little bit worrying about iptables though) 21:42:54 nachi's fix for the gate issues to the l3 agent should land first 21:42:57 markmcclain, nati_uen_: we are using inheritance 21:44:04 SumitNaiksatam: it looks adding mixin to l3-agent 21:45:01 ok.. can we chat about this after the meeting? 21:45:49 markmcclain: SumitNaiksatam 21:45:52 markmcclain: sure 21:46:14 any others questions/thoughts on FWaaS? 21:46:14 ok.. I want to cover the other teams before we run out of time 21:46:20 gotcha 21:46:28 #topic ml2 21:46:35 rkukura, mestery: hi 21:46:49 hi 21:47:09 looks like ML2 changes are moving through 21:47:10 the multiple tunnel type patch is in, but needs a quick fix 21:47:27 https://review.openstack.org/#/c/37020/ 21:47:35 this fix needs one more core 21:47:58 the GRE and VXLAN type drivers are very close - should get in tomorrow 21:48:11 I'll take a look at 37020 21:48:23 markmcclain: thanks 21:48:26 thanks 21:48:48 yeah I know I've reviewed a few iterations of the GRE and VXLAN drivers 21:49:02 the team has been been super quick at responding to feedback 21:49:38 anything else for ML2? 21:49:40 the ml2-portbinding BP is moved to h-3, and arosen has been raising questions about the port binding extension and GenericVifDriver stuff it depends on 21:50:17 #topic LBaaS 21:50:24 also, the MechanismDriver BP is merged, but is not showing this in the BP 21:50:24 agent scheduling patch is waiting for 3rd week for cores attention. 21:50:50 I think it has chance to get in for H-2 21:50:57 enikanorov: I will go on with the agent scheduling review. 21:51:06 gongysh: thanks. 21:51:15 https://review.openstack.org/#/c/32137/ 21:51:59 that's the most important item for lbaas for the moment. 21:52:20 enikanorov: don't hesitate push me if I am lazy. 21:52:44 ok.. after gongysh gets done with his review we'll get 2nd core from the others who already looked at it 21:52:49 enikanorov: thanks for the update 21:52:52 gongysh: ok =) 21:53:25 #open discussion 21:53:40 Any other major H2 items I've skipped? 21:53:40 markmcclain: today a new bug on lbaas came in 21:53:53 it was about no namespace support 21:53:53 enikanorov: which bug 21:53:58 let me find 21:54:08 https://bugs.launchpad.net/neutron/+bug/1201249 21:54:09 Launchpad bug 1201249 in neutron "quantum-lbaas-agent ignores use_namespaces = False" [Low,In progress] 21:54:27 https://review.openstack.org/24739 preserve iptables counters is getting close, any chance for H-2? 21:54:31 so the question would be... if we plan to support such option ever? 21:54:41 my take was 'no' 21:55:23 enikanorov: that's correct originally we decided the feature needed namespaces to work 21:56:28 garyk: can you take a second look at https://review.openstack.org/24739? 21:56:30 I struggle to understand how the comment in the review is related to the specific issue reported. 21:56:52 I am referring to the review for bug 1201249 21:56:55 Launchpad bug 1201249 in neutron "quantum-lbaas-agent ignores use_namespaces = False" [Low,In progress] https://launchpad.net/bugs/1201249 21:57:07 salv-orlando: reported wants this feature 21:57:16 that's it, probably 21:57:43 but i think it only practically solvable via vm or hw appliances 21:58:05 reported=reporter 21:58:11 yeah.. the lbaas agent was never designed to work without namespaces 21:58:19 markmcclain: sure 21:58:27 garyk: thanks 21:58:39 markmcclain, thx for the reviews on https://review.openstack.org/#/c/30441/ which are completed, focusing on the front-end https://review.openstack.org/#/c/30447/ 21:59:24 garyk: yes, thanks 21:59:45 dkehn: last time I checked, there was no test coverage for db-related code 22:00:33 enikanorov, markmcclain indicated taht we usually don't do db related items like cascade, etc 22:01:03 enikanorov, if thats whatt you are refferring to 22:01:33 i'm refering to unit tests for api of extra dhcp options mixin 22:03:17 enikanorov, will look in it, take we talk off line 22:03:30 enikanorov, can we talk off line 22:03:41 ok.. anything else? 22:03:55 dkehn: I'll look at your new patch soon. 22:03:57 dkehn: sure, in about 8 hours. 22:04:17 enikanorov, lol, ok, I'm back in USA tz will ping you tomoroow 22:04:29 ok 22:04:55 Thanks to everyone for their hard work on H2.. remember the cut is end of day tomorrow 22:05:02 #endmeeting