16:00:56 <rgoulding> #startmeeting kernel projects
16:00:56 <odl_meetbot> Meeting started Tue Jul 24 16:00:56 2018 UTC.  The chair is rgoulding. Information about MeetBot at http://ci.openstack.org/meetbot.html.
16:00:56 <odl_meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:56 <odl_meetbot> The meeting name has been set to 'kernel_projects'
16:01:01 <rgoulding> #topic agenda bashing
16:01:09 <rgoulding> #info Move odl-jolokia from controller to aaa
16:01:16 <rgoulding> #info odlparent 3.1.3/yangtools-2.0.9
16:01:24 <rgoulding> #info netconf/netconf-it
16:01:33 <vpickard> #info clustering
16:01:50 <vpickard> pantelis_: hey Tom, you going to be able to join the kernel call?
16:02:29 <pantelis_> oh yeah - internal stuff going on
16:04:13 <rgoulding> #topic clustering update
16:04:56 <rgoulding> #link https://jira.opendaylight.org/browse/CONTROLLER-1853
16:05:04 <rgoulding> #info in controller csit, ODL is restarted a number of times
16:05:13 <rgoulding> #info on the fourth time it is restarted it failes
16:05:19 <rgoulding> #info due to bundles failing to start
16:05:36 <rgoulding> #info the problem appears to be that the seed node failed to start and join the cluster
16:05:52 <rgoulding> #info jamoluhrsen also saw this
16:06:07 <rgoulding> #info appears to be a duplicate of CONTROLLER-1849
16:06:28 <vpickard> #info 3 mins for datastore to converge
16:06:36 <rgoulding> #info trying to converge before saying “hey im ready"
16:07:29 <rgoulding> #info anything using clustered-app-config will wait indefinitely for the READ to complete
16:07:38 <rgoulding> #info anything depending on DS is going to fail
16:13:05 <rgoulding> #info distinguishing bootstrap from rejoin may be beneficial if we do this in ODL code vs relying on akka
16:14:17 <rgoulding> #info tpantelis claims it may be better to work with AKKA to plug through this
16:16:29 <rgoulding> #info jamoluhrsen has debug logging enabled for akka and ran this test overnight
16:17:02 <rgoulding> #info jamoluhrsen will dig through and figure out where this reproduced (if it did)
16:17:11 <rgoulding> #info the log file is very larger
16:17:15 <rgoulding> #undo
16:17:15 <odl_meetbot> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x2dc9510>
16:17:24 <rgoulding> #info the log file is very large compared to when it was run without DEBUG
16:19:07 <rgoulding> #info LuisGomez says that this doesn’t happen when the controller is brought down gracefully
16:19:32 <rgoulding> #info maybe we should change the test suite to utilize graceful shutdown
16:20:11 <rgoulding> #info luisgomez used to use “kill -9” now has only one test to catch that
16:20:33 <rgoulding> #info fails perhaps due to some race condition
16:20:55 <rgoulding> #info LuisGomez uses “stopNode” keyword (this may be wrong spelling)
16:21:22 <rgoulding> #info it is isolated already for openflow to utilize keyword
16:21:35 <rgoulding> #info tpantelis says that we should modify other tests to not utilize “kill -9"
16:21:45 <vpickard> jamoluhrsen: do you know how stopNode shuts down the node?
16:22:31 <jamoluhrsen> vpickard: what do you mean? the bin/stop cmd?
16:23:01 <vpickard> jamoluhrsen: I don't know what stopNode actually does, looking at code now. Was wondering if you knew already
16:23:33 <jamoluhrsen> vpickard: I don't know "stopNode" though? what is that?
16:23:51 <vpickard> jamoluhrsen: exactly my question. What luiis is taling about now
16:23:58 <rgoulding> sounded like a robot keyword
16:24:04 <jamoluhrsen> vpickard: he was talking about the robot keyword to stop and start odl
16:24:19 <jamoluhrsen> vpickard: that is the one that will call bin/stop or bin/start
16:24:55 <vpickard> jamoluhrsen: ok, thats what I wanted to know, and had asked. Was it calling bin/stop, and it appears yes. btw, havent found the keyword yet...
16:25:08 <jamoluhrsen> vpickard: I'll dig it up for you
16:25:25 <faseelak> jamoluhrsen, vpickard by graceful shutdown do we mean a bin/stop
16:25:47 <jamoluhrsen> vpickard: one var that can be used: https://github.com/opendaylight/integration-test/blob/master/csit/libraries/ClusterManagement.robot#L61
16:26:02 <jamoluhrsen> vpickard: or this keyword: https://github.com/opendaylight/integration-test/blob/master/csit/libraries/ClusterManagement.robot#L417
16:26:06 <vpickard> jamoluhrsen: thanks!
16:26:14 <jamoluhrsen> vpickard: this is more robust: https://github.com/opendaylight/integration-test/blob/master/csit/libraries/ClusterManagement.robot#L425
16:26:42 <vpickard> faseelak: yes, thats what Luis was saying he did, changed kill -9 to graceful shutdown, which I take it to mean bin/stop on karaf
16:27:47 <faseelak> vpickard, are we saying that CONTROLLER-1849 will go off with graceful shutdown?
16:28:47 <vpickard> faseelak: Luis said he only saw the issue 1 time in openflowplugin csit when he changed it from kill -9 to bin/stop
16:34:00 <rgoulding> #topic Move odl-jolokia from controller to aaa
16:35:58 <vpickard> LuisGomez: do you have a patch you can point us to where you changed from kill -9 to bin/stop for karaf? Just wanted to make sure I understand correctly. Thnaks
16:37:36 <vpickard> faseelak: after running a large number of jobs. I think
16:37:41 <vpickard> oops
16:37:53 <vpickard> https://git.opendaylight.org/gerrit/#/c/74139/ is the patch, I think
16:42:05 <rgoulding> #topic netconf/netconf-it
16:42:13 <rgoulding> #info dead code?
16:42:21 <rgoulding> #info netconf-it is deactivated
16:42:32 <rgoulding> #info is this going to be ressurected or should it be deleted?
16:42:44 <rgoulding> #info it still references config subsystem so it will require some cleaning
16:43:02 <rgoulding> #info jmorvay claims we can remove it
16:43:17 <rgoulding> #info rovarga has submitted a patch for this
16:43:23 <rgoulding> #info causes build warnings etc
16:43:41 <rgoulding> #topic odlparent 3.1.3/yangtools-2.0.9
16:43:48 <rgoulding> #info for fluorine and oxygen-sr3
16:44:10 <rgoulding> #info all 3.1.3 patches are prepared, but strange issue in lispflowmapping
16:46:40 <rgoulding> #info need to rebase patch and rerun
16:50:32 <rgoulding> #info should be relatively transparent
16:50:46 <rgoulding> #info may be able to merge in parallel
16:52:23 <rgoulding> #info shague brings up the fact that fluroine and oxygen-sr3 code freeze are on the same day (august 7th)
16:53:38 <rgoulding> #topic invokeRpc checkedFuture elimination in mdsal
16:53:49 <rgoulding> #info tpantelis claims it will cause breakage in controlle
16:53:55 <rgoulding> #info controller is only user
16:54:23 <rgoulding> #info tpantelis is running a multipatch with the controller adaptations to the changes in mdsal
16:55:50 <rgoulding> #info tpantelis is planning to change a lot of the ListenableFuture stuff to use FluentFuture
17:03:50 <rgoulding> #endmeeting