16:04:08 #startmeeting Integration Weekly Meeting 16:04:08 Meeting started Thu Jan 22 16:04:08 2015 UTC. The chair is dfarrell07. Information about MeetBot at http://ci.openstack.org/meetbot.html. 16:04:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:04:08 The meeting name has been set to 'integration_weekly_meeting' 16:04:24 #chair LuisGomez tykeal 16:04:24 Current chairs: LuisGomez dfarrell07 tykeal 16:04:35 oh you silly person ;) 16:04:38 lol 16:05:15 think something is wrong with my audio 16:05:32 zxiiro: I don't see you in the call... so you may have dropped? 16:05:44 correction, I see you but you don't list as having audio 16:06:56 * dfarrell07 will not be able to attend gathering (due to location on East coast of US) 16:07:09 ^^but will be around on IRC/Hangouts/whatever 16:11:46 #info LuisGomez goes over highlights of next week's Integration Team gathering 16:13:48 #info Discussion about how to include community members that can't make meeting F2F 16:16:01 #topic Updates on Trello tasks 16:18:00 #info JJB very critical, CASP3R not in call and haven't seen him on IRC much recently, but folks think he's making progress. Need to sync up with him. 16:19:30 #info dfarrell07 gives updates on RPM, Docker and Puppet packing/distribution work 16:19:55 #info RPM work is being consumed and tested by various things, including Puppet mod, so pretty sure it's gtg 16:20:48 #info Docker packaging has all the basics done, still need to loop back to building examples on top of Karaf image to show how images with more complex features should work 16:21:23 #info ODL Puppet mod is going very well, that's what dfarrell07 has been spending all his time on. All the basics working, see GitHub Issues for details of what needs to be done. 16:21:31 #link https://github.com/dfarrell07/dfarrell07-opendaylight ODL Puppet mod GitHub repo 16:21:43 #undo 16:21:43 Removing item from minutes: 16:22:00 #link https://github.com/dfarrell07/puppet-opendaylight ODL Puppet mod GitHub repo 16:24:47 zxiiro: sounds like you were just volunteered ;) 16:24:51 #chair catohornet 16:24:51 Current chairs: LuisGomez catohornet dfarrell07 tykeal 16:26:49 tykeal: what was i volunteered for? (had a emergency to take care of so missed some of the conversation) 16:27:09 zxiiro: talking about / training on JJB at the meetup next week 16:27:18 ah, yeah i'm fine with that :) 16:29:33 infinite build loop sounds *super intense*, like a normal inf loop but super computationally heavy lol 16:39:29 tykeal: I feel like it might be doable as long as the loop ends at 1 integration job and we don't have integration jobs triggering more jobs. 16:46:00 zxiiro: the thing is, I think with what is being suggested that downstream projects would have to watch not just the merge job(s) but the integration jobs as well 16:47:11 tykeal: true, and if we make integration trigger merge, we could cause a build loop 16:47:20 yes 21:45:06 tykeal and zxiiro, downstream projects just need to listen to upstream merge as they are doing now, no need to create a chain of integration after integration 21:56:33 LuisGomez: I think i found a solution that will satisfy both you and tykael's requirements 21:57:07 LuisGomez: I'm need to add some reverse dependency support to our JJB scripts to support it though hoping to get a patch tonight 21:57:35 LuisGomez: you can see a POC I built here which I think works https://jenkins.opendaylight.org/sandbox/view/integration/ 21:58:10 basically builds will look like this a-merge -> b-integration -> b-merge -> STOP 21:58:32 in my POC i purposely made a build loop so project A depends on project B and project B depends on project A 21:59:13 but using a INTEGRATION variable I'm able to pass to the merge job to tell the job to not fire off any jobs if an integration job kicked it off 21:59:54 this stops the loop, and we can depend on "merge" only jobs for deploying artifacts 22:02:33 hi zxiiro, maybe i am missing somehting here but i do not see the need of b-integration -> b-merge 22:03:20 b-integration builds the project code and as a result you can upload artifacts without invoking merge job 22:04:07 LuisGomez: yes but as tykael pointed out if we use both that means projects would have to know to watch both merge and integration jobs 22:04:30 LuisGomez: he had an issue with that so if we get the integration job to launch the merge job that won't have to change 22:05:28 zxiiro, when you say watch, what does it mean? what is the impact? i think i miss this part 22:05:36 LuisGomez: by projects i mean downstream projects would have to keep an eye on both their upstream's merge and integration jobs 22:06:17 why? 22:06:32 downstream project only looks at upstream merge 22:06:37 LuisGomez: i'm not sure what the impact is either but tykael had an issue with it, I guess you should ping tykael to clarify why he's against it 22:06:50 yes, i have to 22:07:03 LuisGomez: he's in #opendaylight-releng if you want to join that 22:07:06 LuisGomez: to discuss 22:07:17 yes 23:39:22 LuisGomez: FYI Your request was merged. Integration jobs should now deploy artifacts 23:39:46 thanks zxiiro 01:28:15 Venkatrangan Govindarajan proposed a change to integration: Fix VTN Coordinator Tests for change in response Code https://git.opendaylight.org/gerrit/14410 01:28:40 Venkatrangan Govindarajan proposed a change to integration: Fix VTN Coordinator Tests for change in response Code https://git.opendaylight.org/gerrit/14411 11:17:30 Good morning 15:38:22 morning Jamie 16:58:28 mlemay: Howdy 16:58:37 mlemay: still clearing out from snow here :S 00:29:24 Vamsi Krishna Devaki proposed a change to integration: Fix - Bug 2638 - Integration templates have incorrect directory name https://git.opendaylight.org/gerrit/14466 23:58:51 Can I get link to integration Trello site 00:00:30 I got it. 01:49:36 Has anyone tested this ova in Fusion? 01:49:38 https://wiki.opendaylight.org/view/CrossProject:Integration_Group:Test_VMs 23:02:39 Hello 12:37:44 Good morning 21:17:14 Ed Warnicke proposed a change to integration: Add neutron feature to integration. https://git.opendaylight.org/gerrit/14524 01:05:24 Vamsi Krishna Devaki proposed a change to integration: Fix - Bug 2639 - karaf feature repo not available on HELIUM-SR1.1 when deployed using cluster deploy script https://git.opendaylight.org/gerrit/14526 13:26:00 Good morning 19:36:57 zxiiro here is the JJB yamb 19:36:57 https://pad.opendaylight.org/p/Integration-jjb 19:37:15 https://jenkins.opendaylight.org/sandbox/job/integration-master-csit-l2switch-only/configure is the job in the sandbox 19:37:57 seem like we're not setting the $PROJECT anywhere 19:40:40 CASP3R: oh we need to add that in. It was changed not too long ago. Let me get you the setting 19:41:14 Ok thanks 19:43:56 CASP3R: I bolded the changes I made. Basicaly in the "- project:" section I added a project: 'integration' and in the job-template i added a parameters for the project 19:47:13 Thanks zxiiro, that got me pass that error, let me find some more errors in my job setting. If i find something i can't fix then i bug you. 19:47:51 CASP3R: sounds good. Sorry I should let you know about that build change. 19:47:56 should have* 19:55:41 Daniel Farrell proposed a change to integration: Move to new Karaf-based RPM https://git.opendaylight.org/gerrit/14552 22:11:42 zxiiro https://jenkins.opendaylight.org/integration/view/All/job/integration-master-performance-plugin/plot/ 22:12:39 CASP3R: it's supported, search for plot on this page http://ci.openstack.org/jenkins-job-builder/publishers.html 22:54:22 CASP3R, did we have pass/fail criteria in the past with these cbench tests in CI? 22:54:38 Nope 22:54:50 If the test completed it passed 22:54:59 we just plotted the results 22:55:22 jamoluhrsen: Those need attention, imho 22:55:46 They use a subset of (very old) WCBench code 22:55:58 +1 to that 22:56:01 yeah, ok. 22:56:27 I think the main value at this point would just be having something running to catch any regression, even if it's a narrowly focused case. 22:58:57 jamoluhrsen: Yeah, no doubt. We (Integration team) have been pretty clear with wanting (better) CI for perf tests. It's pretty non-trivial though. I have a lot of thoughts on it, would be happy to share them 22:59:44 One of the reasons I went on this packaging/deployment binge was to support future perf CI 23:00:12 realized that any future solution needed a more reliable way to stand up/consume ODL 23:00:37 Also, good to see you CASP3R :D 23:00:47 Wish I could join in CA 23:00:49 I been a busy man 23:01:43 CASP3R: Know the feeling, lol. No worries implied :) 12:09:23 Good morning 20:50:35 Hope the Integration gathering is going well, @all. :) 20:50:57 We are having lunch at eh Integration Gathering 20:52:16 catohornet_: Cool :) feel free to ping me if there's anything I can contribute to, as always 20:52:33 Sorry again that I can't be there 21:54:53 There’s an integration party? 21:55:02 and its not being live streamed? 21:55:29 dfarrell07: lets have our own party, with blackjack and booze 21:55:45 ;) 00:06:09 Vamsi Krishna Devaki proposed a change to integration: Fix - Bug 2639 - karaf feature repo not available on HELIUM-SR1.1 when deployed using cluster deploy script https://git.opendaylight.org/gerrit/14526 11:39:19 Good morning 16:06:58 dfarrell07: Error: Can't start another meeting, one is in progress. Use #endmeeting first. 16:07:08 there's a meeting running? 16:07:52 dfarrell07: according to the bot there is 16:08:03 * dfarrell07 is looking at history 16:08:19 yeah, we didn't end the last one 16:08:21 looks like the meeting from last week didn't get closed 16:08:24 #endmeeting