21:02:29 #startmeeting Networking 21:02:30 Meeting started Mon Mar 3 21:02:29 2014 UTC and is due to finish in 60 minutes. The chair is markmcclain. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:02:31 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:02:33 The meeting name has been set to 'networking' 21:02:49 #link https://wiki.openstack.org/wiki/Network/Meetings 21:03:01 #topic Announcements 21:03:11 markmcclain: :P 21:03:27 The Icehouse-3 deadline is tomorrow 21:03:52 #link https://launchpad.net/neutron/+milestone/icehouse-3 21:04:57 We've still got 51 blueprints that could land tomorrow 21:05:08 this number is significantly higher than it should be 21:05:37 markmcclain: 51! Yikes. 21:06:00 markmcclain: unless everybody takes large quantities of methamphetamins right now 21:06:35 well I do have a few 5hr energies handy 21:07:03 In general, many patches have gone through more cycles than expected 21:07:11 hi sorry late 21:07:32 additionally, we've got several drivers/plugins that don't have testing systems available 21:09:00 If the testing reports are not available with the driver configured/running, we cannot merge the code 21:09:04 markmcclain: on that topic, is linuxbridge tested in the gate? 21:09:16 marun: no 21:09:17 markmcclain: For drivers whose third party testing is --><-- close, if it's enabled by tomorrow are they good? 21:09:25 markmcclain: should it be? 21:09:26 marun: I don't think so no,. 21:10:07 marun: the QA team is looking at changing our testing matrix a bit 21:10:10 It think we should be holding ourselves to the same standards we are enforcing for 3rd party. 21:10:26 * marun gets off the soapbox. 21:10:28 according to the policy we should note the linux bridge plugin is not tested, and possibly earmarked for removal. 21:10:37 I also noticed that there are a plugins that are voting without running the posted patch sets it seems. 21:10:43 so we can discuss over the next few weeks what the correct matrix should look like 21:10:44 and technically I think we're talking about the bridge mech driver 21:10:52 salv-orlando: agreed 21:11:10 is it time to panic for those of us who have existing drivers without voting tests? (tail-f ncs in my case) 21:11:21 salv-orlando: although, we probably want to test the linuxbridge and openvswitch plugins until they are removed 21:11:28 arosen: yes it seems that a few systems are not running the tests properly 21:11:30 salv-orlando: mech driver and agent 21:11:33 marun: but I think it's just a matter of testing matrix as markmcclain says. We can enforce linux bridge testing in the gate 21:11:50 salv-orlando: absolutely. glad to hear it's in-process. 21:11:55 lukego: I think you have ~6 months before panicking 21:12:17 (I can't wait that long for my panic fix. I'll have to get it elsewhere...) 21:12:39 lukego: so existing drivers that don't have testing by March 18th will be marked as pending removal in Juno. 21:13:07 and the release notes will state the plugin is untested 21:13:29 markmcclain: thanks for clearing that up. will try to have testing working by then. (birth of first son is on critical path but he's probably going to have the decency to wait until his deadline :)) 21:13:47 lol 21:13:59 markmcclain: If a plugin is in review pending a clean tempest run posted, when that happens can it merge? 21:14:14 lukego: keep in mind that you can test on the soapbox repo anytime and markmcclain has stated he wants to see you stable there before you vote 21:14:43 mestery: yes in a few instances I've removed the -2 when tempest is functioning 21:14:43 is 3/6 the deadline for approval of plugins to be merged? 21:14:45 anteaya: right. I need to work out how to do that. I only recently agreed to take back maintenance of the driver for Icehouse from my buddies at tail-f 21:14:58 hemanthravi: the deadline is end of day March 4th 21:15:07 markmcclain: thank you! 21:15:15 lukego: okay glad to be informed of the change, we can talk after the meeting to help you understand how to test the sandbox repo 21:15:42 anteaya: (oops - I updated my name on the wiki, didn't think to make a further notification) 21:16:06 lukego: np, you telling me on irc was what works best for me 21:16:10 we have a plugin with review commens fixed and third part test setup running. Is there anything else on my part to get this approved 21:16:35 hemanthravi: it is just the reviewers working through their queues 21:16:54 thx 21:18:07 The other item I wanted to mention is we're now getting the time of year where the government messes with your clock 21:18:31 Different parts of the world will be transitioning to/from summer time over the next few weeks 21:18:49 yay summer!!! 21:19:01 * mestery thinks it must be time to start planting the spring crops. 21:19:17 salv-orlando: you have to wait a bit ;-) 21:20:19 there is also a last minute issue regarding migration file names 21:21:01 mestery: :) 21:21:15 shivharis: we can discuss in bugs 21:21:28 #topic Bugs 21:21:37 markmcclain: k 21:22:27 so the full tempest job is still failing at times 21:22:45 I think we have the same bugs as the last week 21:22:57 unless I've missed some patches 21:23:19 salv-orlando: I don't think you've missed anything. The team have mainly be focused on reviews 21:23:28 Any other bugs? 21:23:40 shivharis: what issue did you want to raise? 21:24:48 before the merge, the migration file needs updates. 21:25:26 i have tempest etc., but will have to update the final migration file 21:25:58 shivharis: which review? 21:26:25 brocade md review 21:27:03 https://review.openstack.org/#/c/60129 21:27:22 shivharis: thanks 21:27:30 Any other bugs the team should discuss? 21:28:00 i have two but can wait till v6 section 21:28:26 sc68cal: noted 21:29:29 Looks like emagana is not online, so we'll skip docs for this week 21:29:34 #topic Nova Parity 21:29:36 beagles: hi 21:29:43 hi... 21:30:31 so it looks like the biggest item is getting a few more eyes on that callback 21:30:37 not much news there, I encourage those who are interested to give the admin-event-callback-api stuff a look over 21:30:41 yes 21:30:47 +1 on getting that callback in this cycle 21:30:54 that's the most important thing at the moment 21:31:06 beagles: Link? 21:31:08 well.. maybe not :) 21:31:10 https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/admin-event-callback-api,n,z 21:31:19 one of the most important things 21:31:23 beagles: Thanks! 21:31:35 https://review.openstack.org/#/c/75253/ 21:31:37 arosen has a lot of relevant irons in the fire as well 21:31:48 and... 21:31:54 Yup it would be awesome to get some reviews from everyone on those. 21:32:03 there is one thing I forgot about that needs highlighting... 21:32:11 beagles: what is that? 21:32:22 because of vif driver changes on the nova side, security groups are not working with ovs 21:32:41 so nati_ueno has been working on those 21:32:50 * beagles digs for links 21:33:02 eeek, is that still unresolved? 21:33:16 yep 21:33:19 last I checked... 21:33:36 so any cycles anyone can spare would be uber welcome 21:33:46 I'll re-review https://review.openstack.org/#/c/21946/ 21:33:46 please review https://review.openstack.org/#/c/21946/ 21:33:46 for review, etc 21:33:55 nati_ueno: I see you have posted new patches, thanks, I'll also review! 21:34:09 ALSO 21:34:10 mestery: Thanks! 21:34:11 The question freequently raised in vif security reviews is why we need to split spoofing attributes. now we have 4 spoofing related fields. 21:34:24 feedback from everyone is welcome. 21:34:40 amotoki: Sounds good. 21:34:43 amotoki: so I wanna discuss with this in the bug report which has history of the discussion 21:35:06 I think the bug is a good place to discuss this 21:35:23 nati_ueno: i understand it. it was discussed about a year ago and many folks are not aware of it. 21:35:25 Any other nova parity issues? 21:35:56 worth mentioning but not high priority is that I've had folks (sweston, HenryG) approach me about parity related work... once the dust settles we should reconvene and figure if it makes more sense to simply prioritize stability and interoperability before closing any remaining gaps, but that's for later 21:35:56 amotoki: yeah,,, I should do happy birthday for the bug... 21:36:32 beagles: we need to be able to both 21:37:41 yes, I agree. And I am available to work with beagles to create a plan for this, if necessary 21:38:49 beagles: thanks for the update and thanks for the call to get more reviews on the callback series that Dan &Aaron worked on since it does ensure ports are wired before booting 21:38:55 #topic Tempest 21:39:25 mlavelle is unable to attend today 21:39:52 #link https://wiki.openstack.org/wiki/Network/Meetings#Neutron_Tempest_.28mlavalle.29 Update Open Tempest Reviews 21:40:04 #topic L3 21:40:07 carl_baldwin: hi 21:40:12 markmcclain: hi 21:40:19 #link https://wiki.openstack.org/wiki/Meetings/Neutron-L3-Subteam 21:40:34 Meetings on Thursdays at 1500. The first one was good but short. 21:40:44 l3-high-availability still in active review. New patches will be up today. 21:40:57 neutron-ovs-dvr is progressing well. 21:41:08 - documentation about changes to L2 (ML2/OVS) agent is available. Link is on the sub team page. 21:41:15 great 21:41:22 - documentation about changes to L3 will be up soon. 21:41:37 Other than that, the sub team is starting to look at L3 blueprints for Juno (e.g. Dynamic Routing, DNS improvements) 21:41:42 I don't think we have cores dedicated to the HA reviews 21:41:57 carl_baldwin: May I? 21:42:18 nati_ueno: You're welcome 21:42:26 carl_baldwin: sure. which one? 21:42:27 nati_ueno: thanks 21:42:30 https://review.openstack.org/#/c/71586/ 21:42:35 markmcclain: gotcha 21:42:39 Looks like it will miss the deadline tomorrow. Any feeling about the chances for an exception? 21:43:10 carl_baldwin: we'll have to do a risk assessment on it 21:43:21 we can chat offline about it 21:43:29 Ok, we'll take that offline. That is all I have. 21:43:33 I think we need one more core for this, if we are trying it to the icehouse 21:44:02 nati_ueno: it's my queue too 21:44:09 carl_baldwin: thanks for the update 21:44:11 carl_baldwin: l3-high-availability BP has several reviews. are there any priorities and dependencies among them? 21:44:37 markmcclain: sure 21:44:44 amotoki: I'll outline that on the sub team page today. 21:45:18 carl_baldwin: thanks. it will really help reviewers. 21:45:18 carl_baldwin: perfect 21:45:30 #topic IPv6 21:45:33 sc68cal: hi 21:45:36 carl_baldwin: Could you add how can I test these patches on that page? 21:45:36 hello 21:46:14 So, in the process off fixing 1262759 we registered another bug identified by Robert Li 21:46:21 https://launchpad.net/bugs/1262759 21:46:26 https://bugs.launchpad.net/neutron/+bug/1284518 21:46:40 ah too hasty, didn't do the link 21:46:57 #link https://bugs.launchpad.net/neutron/+bug/1284518 Link Local Address should be allowed to be set as router interface for IPv6 subnet 21:47:38 #link https://review.openstack.org/#/c/76125/ Fix for LLA address on routers 21:47:50 yes I think that both me and Akihiro have looked at it 21:48:17 yes it is in my radar. 21:48:37 sc68cal: what about the current status for ipv6-two-attributes? 21:48:38 https://review.openstack.org/#/q/status:open+project:openstack/neutron+branch:master+topic:bp/ipv6-two-attributes,n,z 21:49:00 salv-orlando: I believe I just need to bump the alembic script, and make sure I addressed all reviewer comments 21:49:21 I will be working with Shixiong tomorrow for his patchset 21:49:49 ok thanks. What about stateless IPv6? 21:50:09 the patch was failing jenkins, including unit tests as of this morning (GMT) 21:50:36 Yes - I believe there is still some work on that patch - I am working with Shixiong to help him with Git and running the unit tests 21:51:13 sorry I asked because I though Shixion was working on DHCP and Dazhao on stateless IPv6 21:51:25 No problem - I can clarify that 21:51:41 cool. 21:51:54 Shixiong's patch does the work to make dnsmasq work properly for all the modes, while Dazhao's patch teaches neutron how to calculate EUI-64 addresses 21:52:11 At Comcast, we will be using non-openstack routers to announce IPv6 routes and prefixes 21:52:33 so we will be running with address_mode = slaac and ra_mode not set 21:52:45 ok.. so let's see what get can done tomorrow 21:52:58 this might be a candidate for an FFE 21:53:03 FFE? 21:53:12 feature freeze exception 21:53:31 maybe - although I'm not quite confident in the patches just yet 21:53:48 Note that Horizon team is considering IPv6-two-attributes supports as potential FFE because basic IPv6 features are one of important community features. I hope it will land soon (even if it doesn't land for I-3) 21:53:58 I'll check back in with both sc68cal and salv-orlando tomorrow and we'll reassess where we are 21:54:31 Agreed - please feel free to attend tomorrow's meeting or we can discuss on the ML 21:55:35 amotoki: agreed.. if we do an FFE it would have to land this week otherwise the horizon team feels the burden the late merge 21:55:35 Any other questions on IPv6? 21:55:50 sc68cal: thanks for the update 21:55:52 #topic ML2 21:55:59 mestery, rkukura: hi 21:56:08 Top priority is to get marun's migration tool merged 21:56:12 markmcclain: Hi! 21:56:24 https://review.openstack.org/#/c/76533/ 21:56:32 rkukura: I'll be submitting an updated patch with lb support shortly. 21:56:34 Thanks marun for that review 21:56:43 marun: great! 21:56:46 That's our highest priority issue on ML2 right now. 21:57:02 I've already secured an FFE for the migration tool 21:57:06 The one thing I still need is to know how to determine that the db is at a given state. 21:57:09 Thanks markmcclain! 21:57:10 We've got several drivers close to merging, some still needing 3rd party testing 21:57:19 so we don't have press to get merged tomorrow 21:57:24 state => alembic version 21:57:25 markmcclain: great 21:57:31 markmcclain: great! 21:57:59 The generic stuff for SR-IOV is all done 21:58:14 rkukura: the lack of 3rd party testing for the drivers is the biggest blocker for several 21:58:46 markmcclain: Hoping at least some of those will come through in time 21:58:54 great news the sr-iov is all merged 21:59:24 So we need to focus on reviewing the drivers that do have 3rd party testing in place 21:59:24 rkukura: I'm concerned that <24 hrs of reporting gives us no baseline in how accurate the systems report 21:59:34 ++ 22:00:15 Thanks for the update on ML2 22:00:20 Given drivers have little risk, will FFEs be considered? 22:00:52 rkukura: my opinion is that "yes in theory" but there is also an upper limit on the number of FFEs the core team can handle 22:01:03 rkukura: while they are low risk they are burden on the release team 22:01:40 Ok we've run out of time for today 22:02:01 markmcclain: could you share your concerns for service type framework implementation in VPN STF review? we would like to confirm it again. 22:02:05 I know the statuses of each of the adv service teams 22:02:42 amotoki: we can jump over to #openstack-neutron since we've overrun our window here 22:02:54 * pcm_ interested too 22:03:24 For those with patches in the queue. Please be very responsive to review comments so that they are ready for reviewers to take another look 22:03:39 markmcclain: or we can contineu in the review. 22:03:55 Thanks for everyone who has made a push during the last week to reviews completed 22:03:58 #endmeeting