13:02:09 #startmeeting openstack-salt 13:02:10 Meeting started Tue Apr 5 13:02:09 2016 UTC and is due to finish in 60 minutes. The chair is jasondotstar. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:02:11 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:02:14 The meeting name has been set to 'openstack_salt' 13:02:21 #topic roll call 13:02:23 o/ 13:02:27 o/ 13:02:29 hello 13:02:34 hi cznewt 13:03:14 what's new? :-) 13:03:30 we have good news :) 13:03:38 oh nice 13:03:50 dvr support in review process 13:03:56 \o/ 13:04:15 who else we got? 13:04:21 and bird of feather has been accepted 13:04:26 niiiiiice 13:05:06 yes :) 13:05:32 ok I just pinged the rest of the team, hopefully they will show 13:06:05 #topic Introduction 13:06:14 this meeting is for the openstack-salt team 13:06:25 if you're interested in contributing to the discussion, please join #openstack-salt 13:06:39 meetings are Weekly on Tuesdays at 1300UTC 13:06:46 #link http://eavesdrop.openstack.org/#OpenStack_Salt_Team_Meeting 13:06:53 #link https://wiki.openstack.org/wiki/Meetings/openstack-salt 13:07:13 #topic Review past action items 13:07:30 is genunix around today? 13:07:36 o/ 13:08:04 ohai liqw 13:09:33 so let's move to agenda from past? 13:09:44 we don't have genunix yet, but is there anyone who can speak to this: 13:09:45 genunix to start the RPM packaging spec 13:09:51 if not, we'll move on 13:10:00 jasondotstar to reach out to dirk (RPM-packaging team) to discuss RPM packaging 13:10:01 he's away today 13:10:06 cznewt: ack. 13:10:37 so I reached out to dirk and Igor from the RPM-packaging team 13:11:06 we are going to have a short meeting today to discuss RPM packaging. this should give us the info we need to write up a spec 13:24:43 well, let's step back a sec 13:24:43 i think you're saying that we need to use salt to stand up multiple interconnected services 13:24:49 and do end-to-end tests 13:24:54 yes 13:24:58 to ensure that all our formulas work in concert, right? 13:25:11 yes, pretty much 13:25:14 right, ok 13:25:45 so I think there's a facility for that where nodepool is involved 13:25:50 basically we have salt orchestration which shlould bring the services up in right order 13:25:58 right 13:26:13 i can follow up on this 13:26:39 but the actual scenario testing and result evaluation is not clear on how to do 13:26:54 basically we need more than "services are running" 13:27:05 cznewt: agreed 13:27:06 but to be able to create net/server/volume 13:27:12 spawn infra 13:27:23 and if this work, we can mark it success 13:27:34 spawn resources i should say 13:27:36 I gotcha 13:28:22 #action jasondotstar to follow up on how we run integration/functional tests that incorporate multi-node services 13:28:33 +1 13:28:51 ok let's move on 13:29:09 cznewt to follow up on DVR networking and the neutron salt formula with OVS support 13:29:38 yes this has been contributed by wondra 13:29:46 excellent! 13:30:07 wondra: can you tell some more things about the implementation? 13:30:34 Frankly? The base is probably yours from Icehouse time. 13:30:46 I just updated it, tested, added DVR parameters. 13:31:37 There are a few bugfixes and I renamed bridge and switch to compute and network, because it was highly confising. 13:32:17 yes, I'm thinking on how to create new test-lab with this network setup 13:32:30 About the reviews.. I don't really know what the package fuel-utils does. I just copied it. I actually think it is obsolete. 13:33:05 is there a review to look at on this? 13:33:22 And about the testing. How exactly does the test infrastructure work now? I read in the docs that it spins up some VM, but what does it do then? 13:33:27 https://review.openstack.org/#/c/300725/2 13:33:43 this package is not required 13:33:52 OK, will delete. 13:33:58 and generally no packages in formulas 13:34:36 #link https://review.openstack.org/#/c/300725/2 13:35:11 we can d 13:35:30 o review comments at the end of session 13:35:44 there are 2 points I'd like to go over first 13:36:09 regarding the summit and propagation of our effort 13:36:14 ok 13:38:21 1. bird of feather talk https://www.openstack.org/summit/austin-2016/summit-schedule/events/9449?goback=1 13:39:12 it will be about real life experience, but i'd like to empasise on how to join us 13:39:37 the benefits of our solution over the competition etc :) 13:40:07 I'd make the presentation shareable to the rest of the team 13:40:16 yes 13:40:19 that's awesome 13:40:21 +1 13:40:30 #link https://www.openstack.org/summit/austin-2016/summit-schedule/events/9449?goback=1 13:41:28 we certainly want to bring up what makes us different 13:42:20 Ok can you give me action to prepare the presentation doc 13:42:47 sure thing 13:43:05 #action cznewt to prepare the BoF session presentation slide for review with the team 13:43:25 #action cznewt to prepare the BoF session presentation slides for review with the team 13:44:27 cznewt: there was a second item? 13:45:35 yes it's the vendor involvement 13:45:54 if to gather them to BoF presentation 13:46:05 to give them some reason to come/contribute 13:46:15 to this presentation 13:46:50 are there specific vendors you want to target? 13:48:03 sdn/disk 13:48:24 vendors, to integrate their backends 13:48:57 if they were to be summoned to BoF talk, there should be part in how to integrate new things 13:49:15 +1 13:49:28 this is the main question whether to direct them to BoF or make some informal meeting on integration 13:49:41 or make the integration part of BoF 13:49:51 the BoF is like a kickoff 13:52:46 i'd think 13:52:49 we invite them there 13:53:01 and have them hear what we have to offer 13:53:26 but an informal meeting might work too.... something to think about 13:53:57 something we can talk about more in chan 13:54:03 ok, agreed, we can tell them more at end BoF and if there's demand we'll make informal meeting 13:54:12 +1 13:54:18 ok final 5 minutes 13:54:24 #topic Open Discussion 13:54:29 now to the dvr stuff 13:54:31 any other items to bring up 13:54:37 yes, let's finish up with that 13:54:39 yes and the test labs 13:54:47 go for it 13:54:57 we have ubu/rh dvr/contrail single/cluster 13:55:07 this is 2x2x2 - 8 possible solutions 13:55:10 setups 13:55:31 how to manage this 13:55:51 if we add any more it goes to 16...32 :) 13:55:56 possible setups 13:56:12 how to choose what setups to support 13:57:01 hmmmmmmmm 13:57:06 good point 13:57:58 Crazy idea - do it in serial on a btrfs filesystem with snapshots. In testing the setups, you always change 1 option, so the snapshots will form a binary tree :-) 13:58:13 :) 13:58:15 :-) 13:59:54 time is running up, we can discuss in our channel this, I'd like to get to some resolution as I want to setup at least the dvr/ubuntu single/cluster for start 13:59:57 something for us to think about 14:00:01 +1 14:00:06 thanks guys 14:00:08 #endmeeting