Currently the test may fail on retrieving info about buckets after
stopping the master. The router.call is made before we check statuses
of these buckets.
Deleting this call always results in unreachable state for buckets,
which means that if this call is done quickly we get test flakiness, as
failover sometimes doesn't have time to update nearest available replica
object.
Let's explicitly wake up failover and wait, until replica object is
updated.
Currently the test may fail on retrieving info about buckets after stopping the master. The router.call is made before we check statuses of these buckets.
Deleting this call always results in
unreachable
state for buckets, which means that if this call is done quickly we get test flakiness, as failover sometimes doesn't have time to update nearest available replica object.Let's explicitly wake up failover and wait, until replica object is updated.
Closes https://github.com/tarantool/vshard/issues/389
NO_DOC=testfix