21:00:34 #startmeeting networking 21:00:34 Meeting started Mon Aug 8 21:00:34 2016 UTC and is due to finish in 60 minutes. The chair is armax. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:35 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:37 The meeting name has been set to 'networking' 21:00:44 o/ 21:00:54 o/ 21:01:39 hello everyone 21:01:40 \o/ 21:01:42 o/ 21:02:00 #link https://wiki.openstack.org/wiki/Network/Meetings 21:02:08 #topic Announcements 21:02:27 late o/ 21:02:42 as some of you might have noticed from the ML, there’s a patch up for review to make Neutron the default networking for DevStack 21:03:01 during the review process a few things were spotted 21:03:31 like OVH, which is one of the cloud providers OpenStack infra is using that doesn’t exectly like the current patch 21:03:48 the dislike manifests itself by breaking one of the tempest tests for DVR 21:05:43 http://lists.openstack.org/pipermail/openstack-dev/2016-August/100888.html ? 21:06:04 hichihara: yes, links are provided on the meeting wiki page 21:06:28 armax: Ah, I didn't notice. 21:06:31 the course of action as we stand is to merge the default switch, skip the test, solve the issue and unskip the test 21:06:58 do we know why that test fails on only one cloud? 21:07:18 amuller: just ovh hardware seems to be affected 21:07:29 the root cause might change our strategy 21:07:34 amuller: it looks like it’s related on how the underlying networking box is configured 21:09:15 so we’ll have to see 21:09:25 perhaps a quick tempest test that tried to ping 10.x.y.199 would work - the assumption is something in their setup is responding 21:11:24 another reminder is the mid-cycle next week 21:12:05 we currently have these topics to discuss: 21:12:07 #link https://etherpad.openstack.org/p/newton-neutron-midcycle-workitems 21:12:43 we’ll provide a status update when we get back 21:13:01 armax : Is there any plan for remote setup to participate in mid-cycle? 21:13:21 Sukhdev: the usual IRC participation 21:13:31 Sukhdev: so, none in particular 21:13:42 armax : can we not set one up? 21:14:00 armax : I would love to participate remotely? 21:14:04 +1 21:14:33 Sukhdev: if someone wants to volunteer, by all means 21:15:13 armax: I am not familiar with the process to set it up - but. have partcipated in remote mid-cycle for Ironic 21:15:14 Sukhdev: though experience showed us that things don’t work very smoothly 21:15:22 it was very effective 21:15:39 OpenStack has phone lines that can be used by using laptops 21:16:13 We'll be grouped up according to the work items, so I suggest to find someone in a group to set up something on their laptop. 21:17:14 Volunteers can put their names on the etherpad? 21:18:56 HenryG: Sukhdev let’s add some pointers on the etherpad 21:19:30 Let me dig up the logistics to setup the conf call facilities offered by OpenStack and then we can go from there 21:19:36 there’s the timezone to factor in, but if folks are interested we should be able to engage them 21:20:55 ok let’s move on 21:21:03 #topic Blueprints 21:21:12 We’re getting closer to the end of N-3 21:21:36 some blueprints have still their whiteboards that are many months old 21:22:15 as blueprint assignee/approver have the the duty to provide a regular update on the progress of the effort 21:22:34 for the sake of people who are interested but cannot be involved on the day to day 21:23:11 I am reaching out the individual assignees/approvers to assess the latest status for the targeted efforts 21:23:48 it’ll be a huge favor to me, amongst others, if blueprints whiteboards were up to date 21:25:41 #topic Bugs and gate failures 21:25:58 electrocucaracha was our deputy for the week 21:26:05 :) 21:26:26 there were a couple o critical 21:26:47 1562878 21:27:02 bug 1562878 21:27:02 bug 1562878 in neutron "L3 HA: Unable to complete operation on subnet" [Critical,Confirmed] https://launchpad.net/bugs/1562878 - Assigned to Ann Taraday (akamyshnikova) 21:27:35 electrocucaracha: ack 21:27:42 I am looking at Grafana 21:27:58 and the unit test pipeline doesn’t look happy 21:28:00 #link http://grafana.openstack.org/dashboard/db/neutron-failure-rate?panelId=3&fullscreen 21:28:02 sorry that one is too old 21:28:07 anyone knows the root cause? 21:28:12 that bug is mixing two issues 21:28:23 the part that is specific to HA routers is low or medium priority 21:28:32 the gate issue is not specific to HA routers and is higher priority... 21:28:36 it should be split to two bugs 21:28:40 also it looks like greande is busted too in the gate queue 21:28:53 but I think that’s for nova-net 21:29:13 no, actually that’s gate-grenade-dsvm-neutron-multinode 21:29:26 basically, 1610960 1609693 and 1609540 were from this week 21:29:44 armax: I did not notice the py35 problem till now. I can look into it. 21:29:44 armax: that one we are about to solve, first patch in gate to merge. 21:29:49 oh boy 21:29:52 the gate is a disaster 21:29:58 armax: resolution https://review.openstack.org/#/c/352454/ 21:30:00 http://grafana.openstack.org/dashboard/db/neutron-failure-rate?panelId=8&fullscreen 21:30:09 and it's not us, yay. 21:30:11 :( 21:30:27 ihrachys: ack 21:30:45 armax: bug 1610960 gate-grenade-dsvm-neutron-multinode should be soon fixed 21:30:45 bug 1610960 in neutron "Invalid input for external_gateway_info. Reason: '' is not a valid UUID." [Critical,Confirmed] https://launchpad.net/bugs/1610960 21:31:00 should solve grenade failures. but I see dvr is not happy either. 21:31:11 ihrachys: ok, so I can look into the unit test failure 21:31:32 the failure on the lib periodic job is also a known issue 21:31:35 HenryG: any update? 21:32:16 Known since when? 21:32:23 since the last time we spoke? :) 21:32:35 Oh, I though that patch had merged. 21:32:41 um 21:32:49 then we’re not quite out of the woods yet 21:33:08 HenryG: the hacking rules one? nope, it's not 21:33:16 armax: the patch is https://review.openstack.org/350723 -- I'll sync with boden 21:33:36 HenryG, ihrachys ack 21:34:01 ok, ihrachys you’re looking into the dvr instability? 21:34:19 armax: no, why. I had plenty today to make me busy. 21:34:39 ihrachys: sorry I thought you said you were 21:34:45 armax: and I would be glad if someone else takes it, not me tomorrow 21:34:47 np 21:34:48 I will be looking at the unit test instability 21:35:05 anyone else willing to look into what the latest DVR instability is? 21:35:13 armax: i can look, thought we just had the os-client-config issue today 21:35:16 we;re about to peg at 100% 21:35:30 haleyb: occ was grenade only (for some reason) 21:35:30 haleyb: could be the same you reckon? 21:36:00 #link when looking at http://grafana.openstack.org/dashboard/db/neutron-failure-rate?panelId=8&fullscreen 21:36:07 #undo 21:36:08 Removing item from minutes: 21:36:09 #link http://grafana.openstack.org/dashboard/db/neutron-failure-rate?panelId=8&fullscreen 21:36:50 armax: i hadn't looked with the other one so bad, so i don't know 21:36:54 but the graphs for the gate queue look funny 21:36:59 http://grafana.openstack.org/dashboard/db/neutron-failure-rate?panelId=5&fullscreen 21:37:17 they seem to stop at yesterday 21:37:26 haleyb: I see https://review.openstack.org/352594 just got proposed. 21:39:06 i do see the strange "mountains" with grafana now, but see a linuxbridge spike 21:39:27 ok, let’s move for now, but before we do, we have njohnston deputy for the week 21:39:30 thanks njohnston 21:39:36 happy to assist 21:39:42 anyone with spare cycles for next week? 21:40:44 no-one? 21:41:10 ok 21:41:22 #topic Docs 21:42:03 let me remind you that getting closer to feature freeze means we should also start looking at providing documentation for what has been built in the Newton timeframe 21:42:30 I would invite folks to look at the list 21:42:31 #link https://bugs.launchpad.net/neutron/+bugs?field.tag=doc 21:43:24 and make sure we close as many bugs as we can by writing the appropriate documentation material 21:43:54 It seems we need volunteer for doc 21:44:27 Many bugs without assignee 21:44:50 * asettle pokes in head. You should ask in the docs meeting if anyone would like to volunteer. We have a few networking people interested in working on that stuff. 21:45:48 I took for from the end since those are mine 21:46:00 once we go into feature freeze, I’ll be also going over this list and chase down folks 21:46:32 *four 21:46:50 ok, let’s move on 21:46:55 #topic Transition to OSC 21:46:58 And we also have new another tag for api-ref now, https://bugs.launchpad.net/neutron/+bugs?field.tag=api-ref 21:47:08 hichihara: ack 21:47:23 amotoki is offline, rtheis perhaps you have anything to share? 21:48:03 armax: I will prepare a transition status update for mid-cycle next week. 21:48:28 progress continues with osc and osc plugin patches 21:48:36 rtheis: ack 21:48:44 nothing else from me 21:48:47 Is there still a plan for an API reference clean up sprint, or was that folded into the mid-cycle? 21:48:47 rtheis: thanks 21:49:22 njohnston: I guess that amotoki is preparing 21:49:22 njohnston: I’d defer the answer to amotoki, but I hope we’ll be pushing at the mid-cycle too 21:49:37 #topic Keystone v3 21:49:53 the schema migration patch merged 21:50:25 dasm, HenryG: what are you baking next for us? 21:50:41 was anyone complaining from 3party? all smooth? 21:50:55 armax: couple things need to be cleaned up, like temporary fixes from subprojects and missing deprecation warnings. 21:51:04 dasm: ok 21:51:26 ihrachys: noticed just two problems: nova had mocked calls to neutronclient, and because of this, couldn't follow changes 21:51:31 the next two big steps would be to fix the API and sweep the codebase? 21:51:47 dasm: links? 21:52:10 armax: digging 21:52:28 armax: https://review.openstack.org/#/c/349299/ 21:52:37 https://bugs.launchpad.net/nova/+bug/1608258 21:52:37 Launchpad bug 1608258 in OpenStack Compute (nova) "test_neutronv2 unit tests fail with python-neutronclient 5.0.0" [High,Fix released] - Assigned to Matt Riedemann (mriedem) 21:52:47 ihrachys: thanks 21:53:11 and also this one: https://bugs.launchpad.net/bugs/1610905 21:53:11 Launchpad bug 1610905 in neutron "Updating agent with non admin context returns 500 error." [Medium,Confirmed] - Assigned to Kengo Hobo (hobo-kengo) 21:53:12 ihrachys: thanks 21:53:51 but second one seems to be not connected to latest change, according to HenryG 21:53:56 dasm: how will plugin API work? will we receive both project and tenant IDs and will need to reconcile ourselves? or you will do it in api layer and provide a single value for us? 21:54:44 ihrachys: we shouldn't break backward comp, so idea was to do another extension 21:55:33 ihrachys: everything (hopefully) will be done on api site. 21:55:41 *side 21:55:45 dasm: my understanding is that we’d be supporting both for sometime 21:56:10 ihrachys: and that the programmatic API would continue to see/handle tenant_id 21:56:25 armax: yes, both will be supported. 21:56:27 cool with me, less work on objects side 21:56:40 #topic neutron-lib 21:56:47 HenryG: anything worth raising for awareness? 21:59:10 looks like we lost HenryG 21:59:12 1min 21:59:15 oh well 21:59:18 let’s get it back then 21:59:33 thanks everyone for partecipating 21:59:37 #endmeeting