16:03:52 #startmeeting OPNFV BGS daily release readiness check-in 16:03:52 Meeting started Tue Apr 28 16:03:52 2015 UTC. The chair is frankbrockners. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:03:52 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:03:52 The meeting name has been set to 'opnfv_bgs_daily_release_readiness_check_in' 16:04:02 #info Frank Brockners 16:04:03 <[1]JonasB> #info Jonas Bjurel 16:04:04 #info Daniel Smith 16:04:13 #info Peter Bandzi 16:04:25 #info Fatih Degirmenci 16:04:30 #info Tim Rozet 16:04:41 #info rprakash 16:04:55 #info Morgan Richomme 16:05:19 looks like we have a quorum 16:06:02 let's do the standard quick roundtable 16:06:26 #topic daily status report on progress and plans 16:06:46 Tim, Jonas - could you give a brief update? 16:06:58 <[1]JonasB> #info We're working hard on the autodeploy, hope to have it finished soon, going into the LF lab soon 16:07:30 <[1]JonasB> #info Agreed network setup with Peter 16:07:37 #info Heather Kirksey 16:07:56 #info Are there any open questions on the LF HW config for now? 16:08:19 <[1]JonasB> #Info, not now, doesnt mean there will be no 16:08:39 ok - understand 16:08:55 Tim - quick update from your end? 16:09:04 #info we were able to do automatic deployment on POD2 16:09:26 <[1]JonasB> Congrats ! 16:09:30 #info unable to access horizon due LF firewall I think 16:09:34 thanks :) 16:09:46 congrats as well! 16:09:51 +1 16:10:00 <[1]JonasB> Is the maintenance window over and done? 16:10:00 #info Bryan Sullivan 16:10:15 #info all the nodes show up as properly configured and I checked pacemaker to make sure openstack-services were up 16:10:25 #info great so its amatter of port opening for Horizon but you can do cli now 16:10:28 #info I would like to try creating networks and tenants and pinging next 16:10:43 #info Are you in touch with Aric/Konstantin on the FW issue for Horizon? 16:11:02 #info sent out an email earlier today to Konstantin with lf-leads on it 16:11:27 trozet: when do you think you will be ready to start doing stuff from Jenkins? 16:11:55 [1]JonasB: I did change with vlan this morning... But in case there will be need something reconfigured on POD1 we can do most of configs without affecting POD2 16:11:59 fdegir: I think as soon as I can verify a few things on the current deployment then we can link to jenkins 16:12:00 * frankbrockners Fatih reads my mind,... 16:12:13 I feel like it is this week 16:12:27 yeah 16:12:28 really glad to hear! 16:12:28 definetly 16:12:51 good stuff.. 16:13:30 Peter - Can you give us a quick update on the LF HW config (and the maintenance work done earlier today)? 16:13:35 #info other than that I'm cleaning up some of the deploy stuff, and making a clean for jenkins 16:13:42 end status 16:14:00 #info default plan for public changed to 411 16:14:08 plan = vlan 16:14:15 #undo 16:14:15 Removing item from minutes: 16:14:18 <[1]JonasB> Is that on Wiki now? 16:14:33 <[1]JonasB> pbandzi: Is that on Wiki now? 16:14:39 #info LF HW update: default vlan for public changed to 411 (per pbandzi) 16:14:47 not in wiki yet. but these are transparent to nodes 16:14:55 #info pbandzi: did you test to see if that tag 0 is gone, and there is no tag coming back to the UCS host? 16:14:55 I will put it to wiki anyway 16:15:12 #info plan 0 tag still remains however 16:15:22 #undo 16:15:32 sory :) plan=vlan 16:15:39 hmm weird 16:16:08 #info vlan 0 tag on host interface still remains however (per pbandzi) 16:16:36 that's all 16:16:37 frankbrockners: can we get info from Cisco on why? 16:16:53 sounds like a bug? 16:16:55 #info research for the reason of vlan tga 0 ongoing 16:17:05 trozet: I am in process of getting this info 16:17:10 ah ok 16:17:11 thanks 16:17:17 trozet: yes - as soon as your Cisco reps have the info :-) 16:18:03 anything else? 16:18:07 <[1]JonasB> Nope 16:18:46 morgan_orange and trozet: Planning to run functest smoketests on pod2 deployment? 16:19:03 yes I should have access to POD 2 16:19:12 so can try start doing that tomorrow 16:19:28 morgan_orange: ok. Let me know what time you plan on trying 16:19:49 just wait for redhat green light 16:19:51 will it be installed on the jumphost as baremetal, vm, container? 16:20:26 we install the tools on the jumphost and perform Rally bench suite and tempest smoke test suite 16:20:52 #info I'll send out an email signaling ready for functest/jenkins when setup is verified as OK 16:21:00 ok 16:21:03 question morgain - does this means we will have two different sets of setups 16:21:06 morgan_orange: tools are installed natively on the jumphost - correct? (i.e. no VM, not containerized) 16:21:08 for each type of testing / FUEL and FORMEAN? 16:21:21 would it not be better to make a portable container or something 16:21:34 so that we dont have a bunch of variants running around (disucssion with Jose). 16:22:24 otherwise we set outselves up to do the same "setup" work again and again - seems a bit wasteful in terms of "manual" work and error prone 16:23:26 Jose told me that, no objection for a container 16:23:45 but support will be needed... 16:23:49 i would perfer that since then we have a static package (that can eventually also be rolled out automatically) 16:24:01 i can support Jose (have been ) as needed with that 16:24:22 or others with the setup.. he has a host and all the tools so its really a question of just setting up the tools inside a container 16:24:28 and i can do that with you guys 16:24:53 ok cool one more thing to learn..good project :) 16:24:56 for tomorrow testing if you want manual for time thats fine, but im thinking its worth the effort to make this a "box" we can then mass produce 16:24:58 sounds like a good plan - if you could jointly work on a dockerfile 16:25:00 aye! 16:25:00 morgan_orange: is there an scripted installer available that installs from the same source to the jumphost right now? or is it manually installed/configed? 16:25:26 set an AP on me to follow up and get that done by EOW for testing runs starting next monday? 16:25:30 does that work for all? 16:25:49 one ask: if we containerize things - can we make sure we create one container with "everything functest" - i.e. rally/tempest *and* robot? 16:26:09 that was my idea Frank - i dont know enough of whats in there, but i would like all testing in one spot yup 16:26:12 * fdegir likes this container thingy 16:26:19 I think so, at the moment Robot and rally have been deployed on intel POD 2 and taht is fine 16:26:33 #action lmcdasm (Daniel) to work with morgan_orange and pbandzi to containerize the current functest environment 16:26:40 and then we can look at having two different "startups" (or more) to make whatever environment is testing against (FUEL, FORMEAN, OSCAR) to align networking , etc) 16:26:42 we have a procedure to set up the tools then script to perform the test and simplify integration into CI 16:27:09 Khaleesi can auto-install and run tempest...in case that is helpful 16:27:35 via ansible 16:27:37 trozet: I believe things are driven through Rally currently 16:27:40 With Rally it also one line to setup Tempest and run it 16:27:45 ah ok 16:28:41 anything else to discuss? 16:28:45 so I suggets to start manually once green light received, and based on the procedure we may think to the container scenario with jose_lausuch and pbandzi 16:29:11 we will probably have work on tests 16:29:24 the smoke test suite on E/// POD lead to more than 70 errors now 16:29:34 we understand some (issue with cinder) but not all.. 16:30:44 morgan_orange: Just a heads up - I would expect some network test cases to fail in the HA setup, there are bugs with opendaylight that I haven't filed/fully investigated yet 16:32:58 ok folks - are we done? Anything else to chat through? 16:33:21 Can i ask tim do you want me to test Jenkins job for LF POD2? 16:33:44 sure. should we do jenkins first or let functest team go first? 16:33:44 or wait like Morgan for your green signal for doing anythin on LF POD2 16:34:27 OK then if you want me to contribute send me the Error report for tests and I can analyse and help clear them 16:34:31 rprakash: you don't need to wait for Hello World to get yourself familiar with stuff 16:34:46 you can start doing experiments while waiting trozet 16:35:07 why dont we let functest go first 16:35:09 Ok let me do with hellow world for now 16:35:12 and then incorporate real stuff 16:35:19 ill send out an email when its ready, and morgan_orange can run tests 16:35:27 then when hes done we can re-deploy wiht jenkins 16:35:29 sound good? 16:35:35 ok 16:35:51 ok rprakash: I'll loop you in on the emails so you know when its ready to go 16:37:40 #info morgan_orange will work with trozet to get functest to run on POD2 16:38:01 #info rprakash will get ready for Jenkins integration in parallel 16:38:33 #info once functest on POD2 is operational, rprakash will do Jenkins integration 16:38:46 does this capture it? 16:38:51 yep 16:38:57 yep 16:39:04 done for today? 16:39:41 thanks everyone... see you tomorrow! 16:39:47 #endmeeting