#opendaylight-integration: cluster test
Meeting started by LuisGomez at 17:08:28 UTC
(full logs).
Meeting summary
- action points (LuisGomez, 17:09:46)
- action point: clean the car-people test as well
as file bugs (LuisGomez,
17:10:26)
- need to debug the stable/lithium cluster jobs.
external problems in beryllium that are unrelated to
clustering (jamoluhrsen,
17:13:56)
- bug 4037 is one of the cluster test
failures (jamoluhrsen,
17:14:22)
- https://bugs.opendaylight.org/show_bug.cgi?id=4037
(jamoluhrsen,
17:14:44)
- ssaxena1 has fixed some test issues in the
past. still more to look at (jamoluhrsen,
17:19:46)
- -all cluster test for beryllium finally
starting, but lots of failures. (jamoluhrsen,
17:24:11)
- "wait for controller up" is timing out, but
it's not marking it as a failure. (jamoluhrsen,
17:27:01)
- checking cluster sync status is ok to know if
clustering is ready, but doesn't indicate if application is ready
for cluster configs. (jamoluhrsen,
17:37:44)
- each application needs to have their own way to
determine if they are ready for cluster testing to start
(jamoluhrsen,
17:39:19)
- cluster test plan for projects (LuisGomez, 17:39:53)
- LuisGomez has draft of cluster testplan in
robot format (jamoluhrsen,
17:41:52)
- clustering concept of leader/follower is not
coupled with SB concept (e.g. openflow) of master/slave (jamoluhrsen,
17:45:00)
- ACTION: need to
verify that first controller that a southbound device is connected
to will become the master (jamoluhrsen,
17:51:52)
- four combinations to test for CRUD
(leader/master, leader/slave, follower/master,
follower/slave) (jamoluhrsen,
17:52:49)
- will talk more cluster testing in thursday
integration call (jamoluhrsen,
18:02:17)
Meeting ended at 18:03:03 UTC
(full logs).
Action items
- need to verify that first controller that a southbound device is connected to will become the master
People present (lines said)
- jamoluhrsen (14)
- LuisGomez (6)
- odl_meetbot (4)
- dfarrell07_conf (1)
Generated by MeetBot 0.1.4.