13:57:16 <jonasbj> #startmeeting Weekly Fuel@OPNFV sync 13:57:16 <collabot> Meeting started Thu Aug 13 13:57:16 2015 UTC. The chair is jonasbj. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:57:16 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic. 13:57:16 <collabot> The meeting name has been set to 'weekly_fuel_opnfv_sync' 13:57:33 <jonasbj> #info Jonas Bjurel 13:57:34 <szilard> #info Szilard Cserey 13:57:40 <szilard> hi Jonas 13:57:47 <mskalski> #info Michal Skalski 13:57:51 <jonasbj> Hi Szilard! 13:58:00 <jonasbj> Hi Michal 13:58:07 <szilard> Hi Michal :) 13:58:14 <mskalski> Hi guys:) 14:01:10 <jonasbj> I guess itt will be a short meeting - people are still on vaccation 14:01:27 <jonasbj> But let's start anyway. 14:01:45 <jonasbj> Lets start with stable arno 14:01:48 <szilard> yes 14:01:56 <jonasbj> #topic stable arno 14:02:12 <jonasbj> #info The build issues have been resolved 14:02:32 <szilard> super great, good job Jonas! :) 14:02:55 <jonasbj> #info They were due to upstream Fuel changes that is not under fuel repo change control 14:03:23 <jonasbj> mskalski: Are you looking into why this happend and how we can avid it in the future? 14:03:48 <szilard> yes, could be great if you can give a brief info 14:04:26 <mskalski> jonasbj: It happen because devops repository which contain only packages for building iso is no divided by fuel releases 14:04:31 <ruijing> don't understand what issue. 14:04:53 <mskalski> so when they change something in this ubuntu repo 14:04:58 <mskalski> it affect all versions 14:05:24 <mskalski> I write about it to devops team but don't have response from them yet 14:05:38 <jonasbj> ruijing: The issue is that even if we nail down our upstream repos to a specific commit - the build is not reproducable 14:05:39 <ruijing> you mean released artifect is not immutable. 14:06:20 <mskalski> I assume that should not happen often but still it is a problem 14:06:25 <jonasbj> mskalski: Can we keep an own mirror? 14:06:37 <ruijing> yes. it is big issue for us. we need to have repeatable build tool and repeatable source code. 14:07:59 <jonasbj> mskalski: Can you look into this issue and come back with a proposal? 14:08:15 <mskalski> jonasbj: I think yes but the we just need to change this line in prepare-build-env.sh https://github.com/stackforge/fuel-main/blob/master/prepare-build-env.sh#L44 14:09:14 <jonasbj> Everyone - should we try with local repo mirrors for the B-release 14:10:19 <jonasbj> +1 14:11:20 <szilard> good idea 14:11:24 <jonasbj> Any oppinions? 14:12:23 <jonasbj> #agree We need to look into ways to accomplish reproducable builds, using local mirrors or otherwise. 14:12:59 <jonasbj> Szilard - status on the POD move and needed deploy changes for that. 14:13:22 <ruijing> use vagrant and save box in our repo? 14:13:29 <szilard> #info I pushed the DEA/DHA config files for POD2 14:13:52 <jonasbj> Does it work yet? 14:13:53 <szilard> #info I'm updating releng/jjb to run the Production Autodeployer for master 14:14:07 <szilard> I tested for multinode it works 14:14:14 <szilard> now I'm testing for HA 14:14:25 <jonasbj> szilard: Great! 14:14:28 <szilard> waiting to be finished 14:14:37 <ruijing> Great! 14:14:51 <mskalski> When do you guys plan to switch to 6.1 ? 14:15:04 <jonasbj> Szilard: What about the clean up script - Tim has started working on a common one? 14:15:28 <ruijing> I already switch to 6.1 since odl plugin is based on 6.1 14:15:30 <szilard> Yes he did the cleanup script, and created a patch 14:15:40 <jonasbj> mskalski: I think we should switch master next week when Stefan is back! 14:15:41 <szilard> but now we have to integrate into jenkins 14:15:43 <szilard> somehow 14:16:01 <jonasbj> szilard: That's good! 14:16:20 <ruijing> hi, szilard. jenkins is up? can you tell me URL? 14:16:21 <szilard> so we have to invoke that script in a separate job afte deploy finished 14:16:43 <jonasbj> Should we move to SR1 14:16:57 <jonasbj> #topic SR1 14:17:07 <szilard> https://build.opnfv.org/ci/ 14:17:17 <jonasbj> So just to refresh us on status 14:17:35 <jonasbj> Michal - status on the plug-in, anything left? 14:18:48 <mskalski> In branch with juno/helium (https://github.com/stackforge/fuel-plugin-opendaylight/tree/juno/helium) there is nothing happen, so this what I show on the demo 14:19:17 <jonasbj> mskalski: Great, good status there. 14:19:50 <mskalski> jonasbj: what is a date of SR1 release? 14:19:55 <jonasbj> szilard: I know you been working on the deployer for 6.1 and plugin installation - status? 14:20:04 <jonasbj> mskalski: Sometime Sep 14:20:06 <szilard> end of september is the release 14:20:11 <szilard> mid september is code freeze 14:20:18 <mskalski> ok thanks 14:20:27 <szilard> as agreed with Frank Brockners 14:20:57 <szilard> #info I have adapted Autodeployer for 6.1 and plugin installation and ODL 14:21:02 <jonasbj> mskalski, could you info the status you just gave 14:21:42 <mskalski> #info In branch with juno/helium (https://github.com/stackforge/fuel-plugin-opendaylight/tree/juno/helium) nothing happen from month, so this what I show on the demo 14:21:48 <szilard> #info ODL is not very stable, yesterday Michal and I found out that puppet manifest fails 14:22:33 <szilard> #info because ovsdb plugin in ODL is not started first time 14:22:41 <jonasbj> are those our puppet manifests? 14:22:48 <szilard> #info we restarted ODL, and second time it worked 14:23:03 <mskalski> jonasbj: no, its OpenDaylight controller behaviour 14:23:09 <szilard> yes 14:23:24 <jonasbj> Ok 14:24:16 <szilard> #info I also added Daniel's pre-deploy functionality in Autodeployer, with values 14:24:26 <mskalski> I thought how to resolve this from puppet for example always restart ODL but it will be nasty piece of code 14:24:29 <szilard> #info in Arno it was just an empty dictionary 14:25:03 <szilard> Michal can you still put that code there, and you can take it out later when ODL becomes stable 14:25:23 <jonasbj> Okay, it seems that much of the pieces for SR1 are in place, we just need to rebase the build system. 14:25:53 <mskalski> I also compare my approach with devstack script and I do the same thing as there 14:26:18 <jonasbj> #info Next week we will start rebasing the build system for Fuel 6.1 and turn any reminding patches/mods into Fuel plugins. 14:26:19 <szilard> It's enough to restart once or twice only 14:26:31 <lmcdasm> jonas - question for you? 14:26:41 <jonasbj> lmcdasm: Go 14:26:51 <lmcdasm> should we be looking to add support for the SFC-ODL modules / build (rather than just lithium)? 14:27:03 <lmcdasm> or you want to just have lithium karaf up for SR1 14:27:05 <szilard> well it would be nice :) 14:27:10 <lmcdasm> and then SFC-ODL with chain for B-rel? 14:27:30 <jonasbj> SFC is for B in my oppinion 14:27:30 <lmcdasm> since its just (thx to Michal's work) simplying a different rpm build 14:27:38 <lmcdasm> ok - just wanted to see your plan :) 14:27:51 <jonasbj> Whatabout Lithium, should we go for that in SR1 14:28:04 <mskalski> lmcdasm: what is difference between SFC-ODL and standard ODL? I assume that different features enabled but something more? 14:28:30 <lmcdasm> ya.. they have hacked it a bit 14:28:45 <lmcdasm> there is the modules and a lib you need to get 14:28:57 <lmcdasm> not a huge deal easy to add in a .spec file for a rpm 14:29:10 <lmcdasm> but its a question of our "track" and what we want to say we have support for and when 14:29:21 <lmcdasm> and the sfc stuff (with gbp/nsh) isnt quite there yet 14:29:41 <lmcdasm> (i think :) ) 14:29:56 <lmcdasm> yes jonas - i think lithium for SR1 is the idea no? 14:30:10 <mskalski> for me adding support for SFC is a interesting thing to do 14:30:19 <lmcdasm> my understanding would be that in the build environment we are able to generate that rpm during the build 14:30:20 <jonasbj> mskalski: What do you think for SR1 Helium or Lithium? 14:30:45 <lmcdasm> pointing at a source we want (i.e helilum / lithium - you can do either source build or gz fetch method) 14:30:51 <lmcdasm> i think that would be doable no Michal? 14:30:51 <ruijing> SFC has 2 dependency 14:30:56 <lmcdasm> go rui! 14:31:26 <ruijing> 1. OVS with NSH (ETA: commit to upstream: Oct) 14:31:41 <mskalski> jonasbj: lithium is already implemented in juno/lithium branch so it can be Lithium for SR1 14:31:55 <ruijing> 2. enable SFC in karaf (maybe GBP/SFC or netvirt/SFC) 14:32:00 <lmcdasm> <nice info Michal> 14:32:13 <lmcdasm> and rui - thanks for that - this is why i think we are good with Lith for SR1 14:32:15 <lmcdasm> since its done 14:32:26 <lmcdasm> and then target SFC support with NSH when its due for B release 14:32:28 <jonasbj> mskalski: What do you need to do with the plugin to go with Lithium? 14:32:51 <lmcdasm> (or maybe if the time slips b-sr1 with frameowrk for it in B).. its up to Jonas who knows the velocity of work for fuel really 14:32:57 <jonasbj> ruijing: Those dependencies will be for the B-release 14:33:30 <mskalski> jonasbj: nothing because OpenStack Juno not allow us to for example use L3 managed fully by ODL 14:33:51 <mskalski> jonasbj: so current implementation is the same as for helium 14:33:55 <ruijing> I am looking at how to install ovs and enable sfc. 14:34:09 <jonasbj> #agree We will go with ODL Lithium for SR1 14:34:22 <lmcdasm> we arent looking for the L3 stuff anyway 14:34:42 <lmcdasm> and by b-rel we can see how far along they are - rui - getting the OVS to make a SFF rigth now doesnt work 14:34:43 <jonasbj> lmcdasm, and others: Correct - no L3 for SR1 14:34:56 <lmcdasm> that ODL (tim and I were talking as well) is really dumb and you have to send it anything 14:35:01 <lmcdasm> (we can chat offline if you like) 14:35:12 <ruijing> Right. We need to wait for pushing NSH to ovs 14:35:19 <jonasbj> Anything more on SR1 plans? 14:35:30 <mskalski> Im curently investigating L3 with lithium and Kilo which will be a part of Fuel 7.0 14:36:04 <ruijing> when Fuel 6.1 release? 14:36:12 <jonasbj> mskalski: That could be a B-project milestone/sprint to have that working! 14:36:18 <mskalski> ruijing: is already released 14:37:31 <jonasbj> For the sake of formality - lets say something on B-release 14:37:42 <jonasbj> #topic B-release 14:38:21 <jonasbj> #info A new fuel@OPNFV gerrit repo is set-up 14:39:17 <jonasbj> #info can all fuel@OPNFV committers send me a +2 on the following patch just so we know the access policies are correct 14:39:29 <jonasbj> #link https://gerrit.opnfv.org/gerrit/#/c/1164/ 14:39:31 <lmcdasm> patch id please (sorry lots of mail clutter) 14:39:32 <lmcdasm> haha 14:40:20 <jonasbj> #info we also have a own Jira area which we should start using 14:40:33 <lmcdasm> are we going 6.1 for B-rel? 14:40:49 <jonasbj> lmcdasm: No - 8.0 14:41:12 <jonasbj> As it stands today - provided release in Mar 14:41:52 <jonasbj> mskalski: Thanks for your +2 :-) 14:42:21 <jonasbj> But B-release is a little vague to mee - yet 14:43:19 <jonasbj> Anything more this week? 14:43:39 <ruijing> just question, why B-Release use Fuel 8.0 instead of Fuel 7.0? 14:44:05 <szilard> good question :) 14:44:12 <jonasbj> Can we have the formal sync and casual meeting swaped next week - I will travel on Thu 14:44:37 <mskalski> In Fuel 8.0 we should have OpenStack Liberty maybe this is one of the requirments 14:44:43 <szilard> That's okay, I will be on holiday on Thu (national) 14:44:51 <jonasbj> ruijing: At current TSC is setting Liberty as a targer 14:45:49 <ruijing> yes, understood 14:46:35 <jonasbj> Shall we end the call? 14:46:46 <jonasbj> 3 14:46:51 <ruijing> 2 14:46:56 <jonasbj> 1 14:46:58 <mskalski> one more thing from me I will be on holiday from 24.08 to 4.09 14:47:01 <jonasbj> Thanks everyone 14:47:07 <jonasbj> #endmeeting