15:59:20 #startmeeting Infra Work Group Weekly 15:59:20 Meeting started Wed Jan 11 15:59:20 2017 UTC. The chair is jmorgan1. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:59:20 Useful Commands: #action #agreed #help #info #idea #link #topic. 15:59:20 The meeting name has been set to 'infra_work_group_weekly' 15:59:27 #topic Roll Call 15:59:31 #info Jack Morgan 16:00:18 #info Fatih Degirmenci 16:00:31 #info Hongbo 16:00:34 info Tianwei Wu 16:01:15 hw_wutianwei: need "#info" 16:01:32 #info Tianwei Wu 16:01:36 sorry 16:01:49 can you hear me? 16:02:03 #info David McBride 16:02:08 #info Bryan Sullivan 16:02:13 no, hongbo2122 16:02:19 #info Trevor Bramwell 16:02:38 could someone please give me the GTM link? 16:02:43 I will try again logining 16:02:59 I can not join the GTM 16:03:03 https://global.gotomeeting.com/join/819733085 16:03:03 dmcbride: https://global.gotomeeting.com/join/819733085 16:03:12 #topic Dovetail Support 16:03:35 go ahead 16:04:46 #info MathewLi presents slides on Dovetail requirements 16:05:03 I am back 16:05:33 #info Uli Kleber 16:05:40 #info details on Dovetail wiki page 16:05:50 * uli-k sorry took me a while to connect..... 16:07:46 #info Dovetail needs stable version of OPNFV deployed in order to run tests 16:07:47 #info rprakash 16:08:11 #info Due to OPNFV turning off CI against stable branches (colorado), Dovetail will not have possibility to run against it 16:08:42 #info In order to run Dovetail, we need Colorado deployed on baremetal PODs 16:09:06 #info But we still lack dynamic allocation of resources so Dovetail probably needs 4 baremetal PODs, 1 per installer 16:09:17 #undo 16:12:01 #info An additional ARM POD is needed for running Dovetail against ARM on Fuel 16:12:08 #info Which makes 5 PODs in total 16:13:13 #info Apex is working on ARM support for E release 16:13:51 #info Fatih asks how oftern we need to run Dovetail against stable branches 16:16:12 #info Dovetail will be part of release testing process 16:18:27 #info had some discussion on dedicated hardware vs dovetails jobs on current hardware 16:18:37 While we are developing Dovetail for Danube it should be run as a job against Danube during the R&D cycle. This is just another test job. 16:18:54 After release we would want to validate dovetail against stable. 16:19:01 When Dovetail is "ready" 16:19:31 ChrisPriceAB: but then this makes me ask the same question 16:19:48 ChrisPriceAB: we need test strategy to come up with how to test the platform itself 16:19:57 sounds like this ask is gated until Dynamic CI is ready, no? 16:20:05 ChrisPriceAB: we can't have everything running as part of CI pipeline against the master 16:20:15 ChrisPriceAB: or one single run will take days to complete 16:20:18 Yes fdegir but that is a question for the Danube project rather than DoveTail 16:20:38 ChrisPriceAB: yes, it is a general quesiton and valid for other test proejcts that are not part of the CI master runs 16:20:54 as well^ 16:20:56 not as part of CI, maybe weekly or on change fdegir 16:21:03 #info points out that the amount of testing is short - a coupls of tests / a few hours 16:21:48 #undo 16:21:48 Removing item from minutes: 16:21:53 ChrisPriceAB: weekly could be done 16:21:59 #info Hongbo points out that the amount of testing is short - a coupls of tests / a few hours 16:23:04 The difference is after an OPNFV release when the dovetail suite is to be made generally available it needs to be validated against stable. 16:24:19 #action fdegir to work with Dovetail in Releng to coordiate a test plan in the short 16:24:44 #info fdegir mentions that this will be difficult as we get closer to release time as release testing increases 16:26:25 #action MatthewLi to work with releng project to coordiate planning 16:27:52 #topic CI resource redundancy 16:28:59 #info due to recent ourtage and maintenance work jenkins jobs failed and projects suffered 16:29:23 #info need to rebalance jenkins jobs to different labs to add redundancy 16:30:32 #info Huawei has POD available for dynamic ci or similar efforts 16:32:35 #action tianwei_wu to return to once Huawei POd as available to discuss allocation 16:32:37 #link https://wiki.opnfv.org/display/pharos/Huawei%27s+munich+Lab 16:33:41 #linkhttps://wiki.opnfv.org/display/pharos/Huawei+Xi%27an+Lab 16:33:56 #action Fatih and Jack to propose a solution to rebalancing jenkins jobs in the labs 16:34:04 :) 16:36:40 #info the Infra PTL need to consolidate slides to present to TSC then request time at TSC to present 16:37:52 #action aricg to follow up with Rudy on UCS support and present update next week 16:38:12 bramwelt ducking and weaving nicely! 16:38:50 ;) 16:45:26 +1 to move 16:47:53 The first step toward healing jmorgan1... maybe it's me 16:49:21 #action jmorgan1 Move Infra WG meeting minutes to general meeting page 16:51:35 #info move Infra meeting minutes to Meeting space for reporting and send out update to mailing list 16:51:45 #topic MS4 for Danube 16:52:30 #info David presents his slides on improving definition of MS4 16:53:37 #info MS4 needs better definition of what it means and how to measure it 16:54:22 #info MS4 is "infra updates completed" currently for projects/senarios, CI capacity, and Infra WG improvements 16:56:03 #info issue #1 timing - when should we have this milestone (currently its 3+ months before release) 16:56:55 #info issue #2 release manager - what questions should be asked to provide status update on completion of this milestone 16:57:50 #info issue #3 derive capacity - how to collect requirements and what is the process for this 16:58:36 #info issue #4 Infra WG plans / improvements - how to coordiante with release manager and how to approve the plan 17:03:06 #info David to upload slides to Software release wiki space and send link to Jack to include in meeting minutes 17:03:43 to end the meeting? 17:03:56 to jmorgan1 17:03:58 #info We will discuss this topic again next week to propose solutions to the problem statemnts 17:04:08 #endmeeting