Fixed too tight timeouts for deconfiguration
Timeouts of 10 seconds for deconfiguring the device is too
tight when there are other scaling suites executed before
and the memory leak bug 4514 is present. The needed timeout
is 120 seconds as I had seen 20 to 30 second garbage
collection pauses on just a 1GB heap usage and the bug
causes the ODL to leak about 5GB if two scaling suites
are present.
There are two timeouts, one is the Karaf console timeout
and the other one is the timeout waiting for the device
to disappear after it was deconfigured. The large timeout
is necessary in the second case because the "device
disappeared" check is executed immediately after the request
to remove it from the configuration was sent.
To avoid "test running for 2 weeks if something bad happens"
type of situation a global timeout logic was also added. The
default timeout is 30 minutes but the scaling suite is
giving less time to the loops.
Change-Id: Ie59a2f9d7a6ee9999f202ba209caf03efd05a801
Signed-off-by: Jozef Behran <jbehran@cisco.com>