15:00:07 <mlavalle> #startmeeting neutron_l3 15:00:08 <openstack> Meeting started Thu Oct 1 15:00:07 2015 UTC and is due to finish in 60 minutes. The chair is mlavalle. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:09 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:13 <openstack> The meeting name has been set to 'neutron_l3' 15:00:42 <mlavalle> Agenda: https://wiki.openstack.org/wiki/Meetings/Neutron-L3-Subteam#Agenda 15:00:53 <mlavalle> #chair carl_baldwin 15:00:54 <openstack> Current chairs: carl_baldwin mlavalle 15:01:18 <fitoduarte> hi 15:01:22 <mlavalle> #topic Announcements 15:01:43 <mlavalle> I don't have any annoucements. how about you carl_baldwin ? 15:01:49 <carl_baldwin> Nope 15:02:01 <haleyb> hi 15:02:07 <mlavalle> any other annoucements from the team? 15:02:36 <Swami> hi 15:02:39 <mlavalle> ok in that case, let's move on...... 15:02:39 <vikram> hi 15:02:46 <mlavalle> #topic Bugs 15:03:10 <mlavalle> so we don't have a lot of change in the list of bugs from last week, pretty much the same ones 15:03:23 <mlavalle> there was progress in most of them, though 15:03:46 <mlavalle> almost all of them have newer patchsets in their proposed fixes 15:04:14 <mlavalle> the one I am curious about is https://bugs.launchpad.net/neutron/+bug/1365473 15:04:14 <openstack> Launchpad bug 1365473 in neutron "Unable to create a router that's both HA and distributed" [High,In progress] - Assigned to Artur Korzeniewski (artur-korzeniewski) 15:04:51 <mlavalle> ownership was trnasferred from amuller to artur-korzeniewski just yesterday 15:05:16 <mlavalle> I haven't been able to ping either one of them, so I intend to follow up today 15:05:25 <Swami> mlavalle: I though amuller was done reviewing this patch. 15:05:44 <carl_baldwin> mlavalle: I reviewed that one yesterday. I can try to contact the new “author" 15:06:00 <mlavalle> carl_baldwin: thanks! 15:06:22 <mlavalle> https://bugs.launchpad.net/neutron/+bug/1494351 15:06:22 <openstack> Launchpad bug 1494351 in neutron "Observed StaleDataError in gate-neutron-dsvm-api tests if reference IPAM driver is used" [High,In progress] - Assigned to Pavel Bondar (pasha117) 15:06:31 <carl_baldwin> the review got me thinking of the coupling between the DVR classes in the L3 agent again. Still too much. I hope to have a patch set up soon. 15:06:50 <regXboi> carl_baldwin: ack - that's becoming painful 15:06:54 <pavel_bondar> working on that one (1494351) 15:07:06 <mlavalle> pavel_bondar pushed latest PS for this one. Waiting for review, I think 15:07:18 <carl_baldwin> pavel_bondar: I owe you a review on that one. Will bump it up on my list. 15:07:20 <mlavalle> pavel_bondar: am i correct? 15:07:33 <pavel_bondar> actually still have some failures 15:07:47 <pavel_bondar> but have an issue with reproducing it locally 15:08:05 <pavel_bondar> anyway reviewving there approach would be helpful 15:08:44 <pavel_bondar> working on this ticket at fullspeed now, so believe will get some results today/tomorrow 15:09:20 <mlavalle> pavel_bondar: thank you! 15:09:25 <carl_baldwin> pavel_bondar: ++ 15:09:37 <mlavalle> the next up is https://bugs.launchpad.net/neutron/+bug/1486795 15:09:37 <openstack> Launchpad bug 1486795 in neutron "DVR: create or update port by using notify specific host rather than fanout" [High,In progress] - Assigned to shihanzhang (shihanzhang) 15:10:16 <mlavalle> This one got a new patchset, it is waiting for cores reviews.... I believe regXboi gave it a +1 15:10:42 <regXboi> did i? 15:11:00 * regXboi goes and looks 15:11:00 <mlavalle> here's the patchset https://review.openstack.org/#/c/221209/ 15:11:11 <regXboi> I did 15:11:38 <regXboi> yeah, this has been through several revisions and I think others should look at it 15:11:38 <mlavalle> so anyway, we need core reviewers for this one 15:12:14 <carl_baldwin> Will look today 15:12:34 <mlavalle> finally in our list https://bugs.launchpad.net/neutron/+bug/1486828 15:12:34 <openstack> Launchpad bug 1486828 in neutron "L3: Notify specific agent rather than fanout when associating floatingip" [High,In progress] - Assigned to changzhi (changzhi) 15:12:36 <haleyb> i believe there is a similar change for FIP as well 15:12:43 <haleyb> that one :) 15:13:03 <haleyb> https://review.openstack.org/#/c/215136/ 15:13:23 <mlavalle> yeah, he is asking help mocking RPC in unit tests 15:13:38 <regXboi> yes - I pointed amuller at this patch 15:13:54 <regXboi> but the submitter isn't getting the recommendations 15:14:09 <regXboi> this patch still needs test work 15:14:42 <mlavalle> so, should we wait for amuller to get back to the submitter with further help? 15:14:56 <amuller> sorry, what patch? 15:15:16 <mlavalle> amuller: https://review.openstack.org/#/c/215136/ 15:15:39 <mlavalle> the submitter is asking for help mocking RPC 15:15:41 <amuller> oh... 15:16:03 <amuller> I don't think that's the only thing he's asking for =D 15:16:04 * carl_baldwin knows amuller loves helping people mock stuff 15:16:10 <amuller> heh 15:16:28 <amuller> I was trying to get him to write proper tests without mocking the entire world 15:17:02 <carl_baldwin> :) 15:17:08 <mlavalle> ++ 15:17:17 <amuller> and I was hoping he could do that without me doing it for him 15:17:21 <amuller> but it doesn't seem to be the case 15:18:11 <amuller> I'll try to find a test that's similar enough that he could go on 15:18:31 <mlavalle> amuller: thanks! 15:19:17 <mlavalle> so that's all the bugs I have today.... regXboi do you have anything? or anyone in the team? 15:20:07 <mlavalle> ok, let's move on..... 15:20:17 <mlavalle> #topic Routed network segments 15:20:34 <carl_baldwin> Hi 15:20:43 <carl_baldwin> Still collecting reviews. 15:20:49 <regXboi> mlavalle: sorry no - you've hit my list 15:21:27 <carl_baldwin> I would like amuller and kevinbenton to review soon as they both had a lot of feedback before. 15:21:32 <carl_baldwin> #link https://review.openstack.org/#/c/225384/ 15:22:13 <carl_baldwin> I need to start doing the work breakdown for this and get started on the first steps. 15:22:20 <carl_baldwin> That is all from me. 15:22:30 <mlavalle> thanks! 15:22:41 <mlavalle> #topic DVR 15:22:51 <Swami> mlavalle:hi 15:22:58 <mlavalle> regXboi, haleyb, Swami you are up 15:23:14 <Swami> I have been working with haleyb on some of the bugs that we have been seeing in upstream 15:23:36 <Swami> Right now there are two bugs that might be the reason for the inconsistent behavior in the gate. 15:23:45 <regXboi> references? 15:24:06 <Swami> regXboi: hold on I will paste the patches. 15:24:23 <Swami> #link https://review.openstack.org/#/c/228582/ 15:24:42 <Swami> #link https://review.openstack.org/#/c/229561/ 15:25:15 <Swami> The second patch that I posted, still does not have a bug associated with it. I will file a bug today. Right now I am testing it in the gate to see if I am seeing the floatingip race condition. 15:25:21 <mlavalle> I will add them to the wiki so we can track them easilly next week 15:25:32 <Swami> Also the second patch needs some more work, from the server side to clean up the namespace. 15:25:52 <Swami> Myself and haleyb were looking at the floatingip race condition yesterday from the log. 15:26:23 <haleyb> Swami: ok, that's the one where the namespace gets deleted from under us with a command in-flight 15:26:36 <Swami> So in order to avoid this race, I am suggesting that we don't cleanup the fip namespace after every iteration and only clean up when the network is removed. 15:26:39 <Swami> haleyb: yes 15:27:13 <Swami> I did a similar change in the server side a while back but now I will use the same context to issue a namespace cleanup in the agent as well. 15:27:36 <Swami> Once I have the server side up I will open it up for reviews. 15:27:50 <regXboi> from my point I'm working to accumulate some more coding resources for DVR tasks and next week I'll be looking at DVR performance 15:28:06 <Swami> regXboi: thanks 15:28:26 <Swami> regXboi: what do you mean by coding resources for DVR, from community or internal. 15:28:35 <regXboi> Swami: I mean internal 15:28:40 <Swami> regXboi: ok 15:29:04 <haleyb> regXboi: let me know if i can help 15:29:16 <regXboi> haleyb: ack 15:29:17 <Swami> also there was issue with the snat static routes and there is patch that addresses this issue. 15:29:37 <Swami> #link https://review.openstack.org/#/c/228026/ 15:30:02 <Swami> regXboi: please also count me in for any DVR issues that you are targetting upstream. 15:30:26 <regXboi> Swami: ack 15:30:27 <mlavalle> i'll track that patchset in the wiki as well 15:30:33 <carl_baldwin> Swami: ack 15:30:41 <carl_baldwin> I’ll have some review time today. 15:30:43 <Swami> carl_baldwin: regXboi: This patch needs your attention, I think I have addressed all comments. #link https://review.openstack.org/#/c/225319/ 15:31:04 <regXboi> Swami: ack - will look sometime today 15:31:20 <Swami> I forgot to all one other patch that haleyb is working on 15:31:37 <Swami> ip_lib patch to check for existence of a device before returns the handle for the device. 15:31:59 <Swami> s/all/add 15:32:23 <Swami> #link https://review.openstack.org/#/c/229010/ 15:32:33 <Swami> that's all I have now for DVR 15:32:52 <mlavalle> Swami: thanks! 15:32:56 <Swami> hopefully by next week we can fix all the instability issues. Keeping my fingers crossed. 15:33:06 <mlavalle> haleyb, regXboi: anything else? 15:33:19 <regXboi> mlavalle: nope - I expect to have more next week 15:33:28 <haleyb> nope 15:33:34 <regXboi> once I have the net node baselines in place to compare to 15:34:07 <mlavalle> thanks, moving on 15:34:22 <mlavalle> #topic BGP dynamic routing 15:34:27 <tidwellr> Hi 15:34:33 <mlavalle> hi there ! 15:35:41 <tidwellr> I'm working on decomposing https://review.openstack.org/#/c/201621/, it's really big 15:36:00 <devvesa> thanks tidwellr :) 15:36:05 <carl_baldwin> Really Big (tm) 15:36:12 <tidwellr> carl_baldwin made a brave attempt at a review and ran out of gas 15:36:21 <mlavalle> LOL 15:36:23 <tidwellr> devvesa too :) 15:36:23 <vikram> great 15:36:45 <carl_baldwin> tidwellr: Overall, it is looking very god. 15:36:48 <carl_baldwin> good 15:37:05 <tidwellr> carl_baldwin: well, you know ;) 15:37:14 <devvesa> yes it does 15:37:27 <tidwellr> vikram, how are the agent patches looking? 15:37:31 <vikram> devvesa: glad to see you ;) 15:37:55 <devvesa> :) 15:37:56 <tidwellr> I still need to make route discovery address scope aware 15:38:09 <vikram> tidwellr: looks good to me ;)... Scheduler needs some attention 15:38:46 <carl_baldwin> tidwellr: I’m going to hit address scopes hard as soon as I wrap up this DVR thing for DVR + HA. 15:39:04 <vikram> carl_baldwin: it would be helpful if you have a glance over other patches 15:39:21 <carl_baldwin> vikram: ack 15:39:27 <tidwellr> yes, the agent patches are in better shape than https://review.openstack.org/#/c/201621/ 15:39:30 <vikram> carl_baldwin: thanks 15:39:47 <tidwellr> you won't run out of gas with those :) 15:39:50 <vikram> devvesa: also promised to review those ;) 15:40:10 <devvesa> I did. And I will do :) 15:40:41 <vikram> devvesa, carl_baldwin, tidwellr: thanks 15:40:41 <tidwellr> I have been able to take the code we have up for review, deploy it, and actually route to tenant networks with it 15:40:51 <mlavalle> ++ 15:40:58 <carl_baldwin> tidwellr: awesome! 15:41:00 <devvesa> cool! 15:41:09 <vikram> great work! 15:41:18 <Swami> ++ 15:42:44 <tidwellr> unless vikram has anything else, I think that's it on the BGP front 15:43:08 <vikram> I don't have anything to highlight 15:43:18 <vikram> tidwellr: We are in sync ;) 15:43:27 <mlavalle> tidwellr, vikram thanks for the update 15:43:37 <mlavalle> next one.... 15:45:56 <mlavalle> #topic DNS 15:45:56 <mlavalle> #topic DNS 15:45:56 <mlavalle> so, functionality in https://review.openstack.org/#/c/212213/ for floating ips is complete 15:45:57 <mlavalle> I have thoroughly tested it in my sandbox and seems to be bug free 15:45:57 <mlavalle> I have also incorporated the designate team feedback as far as creating the zones for reverse lookups from and admin user context 15:46:21 <carl_baldwin> mlavalle: Great! 15:46:31 <mlavalle> I just realized yesterday that I forgot to make PTR pointers optional based on a config parameter. But that is a minor change that will be taken care of today 15:46:48 <mlavalle> that was feedback from the designate team 15:47:18 <mlavalle> so I am starting to work on the nova side to create instace ports with dns_name 15:47:32 <mlavalle> and that will close the whole circle 15:47:58 <carl_baldwin> mlavalle: Any chance the Jenkins tests can be cleaned up? Those tend to scare reviewers away. 15:48:09 <mlavalle> I will also start bugging johnthetubaguy to make sure dns makes it to nova priorities in this cycle 15:48:42 <mlavalle> carl_baldwin: I haven't paid attention to that because i think the problem is that we don't have designate in the gate 15:49:21 <mlavalle> but over the next few days i plan to run the whoile test suite in my sandbox 15:49:36 <mlavalle> so we have a better picture of that 15:49:57 <carl_baldwin> mlavalle: ok 15:49:57 <mlavalle> i will also ping anteaya to start working with the infra team to get designate + neutron in the gate 15:50:11 <mlavalle> makes sense? 15:50:27 <anteaya> mlavalle: all you need to do is write the tests and jobs 15:50:33 <carl_baldwin> Regarding Nova. I think we should have the code up. Otherwise, I think it will go like it has before. They haven’t come through for us. 15:50:37 <anteaya> mlavalle: I can help you find samples if you need them 15:50:48 <anteaya> mlavalle: find me in infra after the meeting 15:50:51 <mlavalle> carl_baldwin: agree 15:51:09 <mlavalle> anteaya: i'll ping you today or tomorrow, thanks! 15:51:14 <anteaya> sure 15:51:28 <mlavalle> that's all I have today 15:51:56 <mlavalle> #topic Address Scopes 15:52:12 <mlavalle> carl_baldwin: since you mentioned this above, you want to discuss any further? 15:52:22 <carl_baldwin> I kind of gave an update here. Will start hitting it hard soon. 15:52:31 <mlavalle> cool 15:52:36 <mlavalle> in that case 15:52:49 <mlavalle> #topic Open agenda 15:53:15 <mlavalle> I have a personal announcement to share with the team 15:53:24 <mlavalle> tomorrow is my last day at Rackspace 15:53:48 <mlavalle> this coming Monday I start at IBM as an OpenStack developer focused in Neutron 15:54:07 <mlavalle> uopstream, so you guys are not getting rid of me..... :-) 15:54:07 <carl_baldwin> You’ll still be the same mlavalle to us! We’re happy you’ll have time to continue working with us. 15:54:15 <vikram> mlavalle: Best of luck! 15:54:24 <pc_m> mlavalle: good to see you're still involved in OS. 15:54:27 <mlavalle> thanks! 15:54:29 <Swami> mlavalle: congrats! and best of luck. 15:55:14 <mlavalle> any other open topics? 15:56:01 <mlavalle> ok, in that case, let's give back 5 minutes back to everybody's agenda 15:56:07 <mlavalle> #endmeeting