Handle cancel txn exception for slaves and device disconnect exception
During the device initialization process, if its encountered that the role for this node is slave
we cancel the txn which is built during the process.
Now since ping-pong txn does not have a cancel, we drop the txn and recreate it. This throws an exception
Since this is an expected exception, it would be advisable to catch and log it in debug.
Also when a node gets disconected from device, the EntityOwnershipService does send a notification. Which trigers a role change.
A check is made to check if the device is connected before moving ahead with the role change.
Change-Id: I76572f27274f5c5f08117fcfd1590d1042049735
Signed-off-by: Kamal Rameshan <kramesha@cisco.com>