16:09:07 #startmeeting networking_cisco 16:09:08 Meeting started Tue Nov 8 16:09:07 2016 UTC and is due to finish in 60 minutes. The chair is sambetts. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:09:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:09:12 The meeting name has been set to 'networking_cisco' 16:09:51 Hey team sorry for the late start, clocks changing etc has screwed up everyones calendars 16:10:00 sambetts: Hi Sam, anything on the agenda? 16:10:02 As always the agenda is here https://wiki.openstack.org/wiki/Meetings/networking-cisco 16:10:07 o/ 16:10:21 yeah my calendar showed meeting in another hour so I guess I have some things to change :) 16:11:17 bradjones: Have you set the calendar item for the meeting to UTC? Then it should show it correcly 16:12:15 bobmel: I have now :) 16:13:31 Ok so first off 16:13:37 #topic Announcements 16:14:00 #action Clocks have gone back please ensure that your calendars are setup with 16:00 UTC for this meeting 16:15:12 RDO team have unpinned our newton and mitaka releases, so both RDO releases should be using networking-cisco 4.0.0 16:16:31 sambetts: maybe use info keyword for such announcements 16:17:50 oh yeah that make more sense 16:17:52 #undo 16:17:53 Removing item from minutes: 16:18:06 #info Clocks have gone back please ensure that your calendars are setup with 16:00 UTC for this meeting 16:18:12 #info RDO team have unpinned our newton and mitaka releases, so both RDO releases should be using networking-cisco 4.0.0 16:18:26 any other annoucements? 16:20:59 nothing from me 16:21:15 me neither 16:22:28 cool moving on then 16:22:32 me neither 16:22:39 #topic Driver support going forward 16:23:12 Basically this topic is about the quality of networking-cisco and the drivers we support 16:23:46 we have a large number of different drivers in the networking-cisco package, managed by different teams 16:24:24 currently there is some inconsistncy with the test coverage provided for some of the drivers in the repo 16:24:31 sambetts: indeed 16:24:47 I can provide some more info here 16:24:51 and also there are some drivers that havn't got an active developer to support them 16:25:32 I tried to calculate the coverage for each of the different drivers results were: 16:25:34 SAF ~= 56% 16:25:39 ASR1K / CSR1KV ~= 73% 16:25:43 CPNR ~= 53% 16:25:48 ML2 Apic ~= 81% 16:25:53 ML2 N1kv ~= 88% 16:25:57 ML2 Nexus ~= 91% (WOOOO! :D) 16:26:03 ML2 UCSM ~= 74% 16:26:24 ^ as a rough guide for the plugins currently letting us down 16:26:32 Do we have a set bar? 16:26:51 we're aiming for 80% across the project 16:27:09 we need to come up with a timeline for everyone to hit that 80% 16:27:33 I think at the moment we should just aim to hit that by the end of this openstack release cycle 16:27:40 so got a few months to sort it out 16:28:23 caboucha: nice job with the nexus plug :) 16:28:29 *plugin 16:28:45 Some of the files makes sense to exclude from coverage calcs. I've been told there is a way to do that but I have not tested it. 16:29:13 yeah if you can let me know what files then I can sort it out for you 16:30:12 bradjones: Ok, I'll get back to you with some examples. 16:30:35 we're currently @ about 62% coverage across the project, but that might be a little higher if we're currently including files which don't make sense 16:30:59 The problem we're going to run into is that some of the drivers listed here under 80% coverage aren't getting activly developed, and so that puts us in a difficult position if we want to push for 80% coverage across the project 16:31:45 so we need to reach out to the teams that own the drivers, and the suggestion we have is that they either step up to the mark to increase their coverage or we move them out of networking-cisco 16:31:51 For the ASR1k/CSR1kv we're a bit low on some parts of the config agent code and then we have some files which do not make sense to test but which reduce overall coverage 16:33:03 sambetts: I think that makes sense. 16:33:23 my 2 cents would be let's keep networking-cisco for actively developed drivers & if we have drivers that are no longer being developed they should be moved to a staging repo or something 16:34:20 yup, we also want to push for CI for all in-tree drivers 16:34:28 bradjones: Is there such a staging repo? 16:34:42 bobmel: not yet but we could create one if needed 16:34:43 or where you thinking of the repos in CiscoSystems on github` 16:34:46 ? 16:34:57 we havn't created it yet, we were thinking networking-cisco-staging or something like that in git.openstack.org 16:35:10 ok 16:37:34 Given that this is a short release, those affected need to get going soon 16:38:34 For the ASR1k/CSR1kv I think it should not be difficult to meet the requirement in time. 16:39:42 bobmel: excellent :D 16:39:57 yeah, we will need to have at least one release with the drivers deprecated before we can remove them from tree and when we do we will have to cut a major version number release, standard deprecation period in OS is 3 months 16:40:03 we'll have to reach out to SAF folks offline 16:41:17 bradjones: I get the feeling you interpreted my statement as if I was volunteering for the ASR1k/CSR1kv work :-) 16:42:19 bobmel: you mean you're not volunteering ;-) 16:42:21 haha 16:42:50 CPNR is another plugin that I'm unsure of activity, we'll need to reach out to all the teams offline to ensure everyone is aware 16:43:03 yup 16:44:07 it might be worth us maintaining a list of driver contacts in the project repo README or something, just so that we know who to call if something breaks too 16:44:31 sambetts: Good idea 16:46:00 third party CI is the other peice of the puzzle here too 16:46:18 so ideally we should have third party CI covering every driver we have in the project 16:47:01 currently we only have nexus, tailf, ucsm and cpnr covered 16:47:22 I'm poorly updated on the CI side of things 16:48:50 I thought we once had something for ASR1k too 16:49:54 our third party CI is currently down, I see an old job that isn't running any longer for csr1k 16:49:58 I don't see one for asr 16:50:51 Ah, yes, wrong of me, if was for CSR1kv we had it 16:51:22 the third party CI is currently down due to lab issues which doesn't help the scenario which means none of our plugins are being CI'd right now 16:52:22 I think we'll need further downstream discussions regarding third party CI and equipment etc 16:53:17 sambetts: Yes, agree. 16:53:44 ok as we only have 10 mins left I'm going to move on to the next topid 16:53:47 topic 16:53:52 #topic CI Status 16:54:31 #info CI is down due to a problem with the node running jenkins/zuul etc we are working on fixing the problem 16:54:53 sambetts: For the old CSR1kv CI I think it is Nikolay who did most of the work. I helped out in the beginning. 16:55:51 ok cool, yeah the equipment might still be there for running that job, we might just need to fix up the code that runs the tests 16:56:47 the neutron PTL is working on creating a list of tempests tests that all neutron plugins should meet, we should switch our testing over to use that suite of tests once it is confirmed 16:57:44 #topic Sub/Driver team feedback 16:58:01 Sub-teams any driver specifc updates? 16:58:25 I'll take silence as a no 16:58:30 #topic Review Priorities 16:59:37 I would like to see https://review.openstack.org/#/c/363835/ merged before anything else so we can start adding release notes for any further commits to networking-cisco 16:59:46 including the driver deprecations 17:00:04 sambetts: i'll take a look at the release notes one again 17:00:16 thanks :) 17:00:35 following that caboucha you might be interested in this patch https://review.openstack.org/#/c/366261/ 17:00:47 but folks we're out of time 17:01:14 so thanks again for attending and see you in 2 weeks time 17:01:21 #endmeeting