16:02:48 #startmeeting integration 16:02:48 Meeting started Thu Aug 4 16:02:48 2016 UTC. The chair is dfarrell07. Information about MeetBot at http://ci.openstack.org/meetbot.html. 16:02:48 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:02:48 The meeting name has been set to 'integration' 16:02:48 let me see 16:03:02 #chair jamoluhrsen LuisGomez anipbu zxiiro 16:03:02 Current chairs: LuisGomez anipbu dfarrell07 jamoluhrsen zxiiro 16:03:09 * dfarrell07 is called in 16:04:30 #topic packagin 16:04:36 #topic packaging 16:05:09 #info now building fedora packages (thanks to plaurin) 16:05:53 #info l2switch tutorial is now done, but sporadic behavior noticed 16:06:43 #info clustering tutorial coming next. 16:08:13 #info was a mention of tutorial using python cluster deployer, but suggestion is that we use the new cluster deployer scripts shipped in the distribution 16:09:50 #info new patch coming for cluster config scripts to make it slightly easier (don't need to give controller index) 16:10:26 #link https://opendaylight.readthedocs.io/en/latest/getting-started-guide/common-features/clustering.html?highlight=clustering <--- clustering config docs 16:13:16 #info the list of tutorials is now listed for the summit and hopefully we can leverage some of these l2switch/clustering/etc tutorials for the summit 16:17:25 #topic distribution 16:17:57 #info saw temp issue with karaf not starting. fixed now. 16:18:06 Peter Gubka proposed a change to integration/test: Add functional suites for bgpcep using exabgp https://git.opendaylight.org/gerrit/38667 16:18:25 #info distribution is WAY TOO BIG and now it will not upload to nexus, which breaks everything in CI 16:18:41 #info no system tests are running 16:19:19 #info autorelease can not produce anything for us. 16:19:58 #info LuisGomez thinks we cannot go much longer with this problem or we have to delay release or we have to change upload size limit on nexus 16:20:30 #info one solution is to revert patch to allow us to start offline, which brought in apache cxf 16:20:50 #info skitt notes that some of the biggest jars are coming from atrium 16:21:42 #info tykeal does not want to change upload limit, but if TSC chooses so we will change it. 16:23:15 #info atrium is just an example among many 16:23:37 public cloud appears to be working again 16:23:43 #info the distribution increased in size by nearly 80M just recently. 16:23:48 releng is starting to process csit 16:24:31 I'm entering a bug right now 16:24:33 #info we need a blocking bug for this distribution size issue. 16:24:41 #action skitt to file the dist size blocking bug 16:26:19 #info the bug is https://bugs.opendaylight.org/show_bug.cgi?id=6341 16:26:35 skitt make that a #link instead 16:26:41 #link https://bugs.opendaylight.org/show_bug.cgi?id=6341 16:27:18 #action LuisGomez, vratko to revert this patch https://git.opendaylight.org/gerrit/#/c/42842 16:27:45 #info LuisGomez suggests that we add some tests to check for distribution growth 16:28:08 revert is https://git.opendaylight.org/gerrit/43136 16:28:36 #link https://git.opendaylight.org/gerrit/43136 <--revert of 42842 16:31:49 #info we should change distribution to use odl-parent, then revert patch. 16:32:35 #action jamoluhrsen to reopen the offline distro bug and make it a blocker 16:33:17 #info basic growth test can be to check if distro size is > 500M to at least give us a warning that we are close to the nexus limit 16:35:06 #info it should now be safe to remove tcpmd5 from autorelease.... 16:35:27 #info zxiiro notes it's already been removed since july 22nd 16:35:44 #info beryllium SR3 distro ready to ship 16:36:02 #topic infra/builder 16:36:32 #info public cloud now seems to be working again now. close to 24hour outage 16:37:16 #info not sure how things happened, but rackspace was doing maintenance and maybe that was part of the reason 16:38:14 #info similar problems have happened in the past 16:39:41 #info the nexus timeout issue is still happening, even after LDAP mirror is in place. Feels like it's happening less often though. 16:42:04 #info jamoluhrsen points out that we are still seeing the ssh EOFError issues. 16:46:15 #info ssh EOFError started happening when we migrated to private cloud. are still happening now that we have moved test VMs to public with robot VM still in private 16:46:53 I think adding a retry on the ssh attempts will make the job more resilient to intermittent issues too 16:47:18 zxiiro: ack 16:47:34 #info no major infra changes coming now, so hopefully we can stabilize going forward 16:47:43 #topic misc 16:48:59 #info robot library organization was brought up last week.... we need better docs for these, and we better structure to be able to help new people coming to CSIT 16:49:30 #info also other projects are starting to use our robot tests. 16:50:53 #info overall questions around what Integration/* can/should do about helping fast and/or phased release strategy. 16:51:31 #info things like more automation around packaging, a working autorelease... 16:54:20 Luis Gomez proposed a change to integration/distribution: Pull karaf from odlparent https://git.opendaylight.org/gerrit/43139 16:54:31 #info dfarrell07 points out that ODL dev design forum is there in Sep to help plan some kind of pilot for fast/phased. 16:55:14 #info vratko notes that now is the deadline for us to decide if we are creating a stable distro, and it's pretty clear that we are NOT creating this. 16:56:11 #info odlparent needs to get the ball rolling in order for us to get any kind of stable release... first odlparent needs to become mature. skitt is driving this. 16:57:10 #info jamoluhrsen and LuisGomez on PTO next thurs. might cancel the meeting or dfarrell07 can run. will decide later 16:57:16 #endmeeting