Increase deldb wait time.

It seems that out QA cluster performance can degrade to a point where
wiping ZooKeeper info cannot be finished in 1 second.
Increasing wait to 5 seconds as a workarouond for now. (ONOS-1242)

Change-Id: If019b3b146543a37ed9437c3e1b727779128e5fb
diff --git a/onos.sh b/onos.sh
index c3331de..810b5e9 100755
--- a/onos.sh
+++ b/onos.sh
@@ -633,9 +633,9 @@
   echo -n "Deleting RAMCloud cluster coordination info ... "
   ${RAMCLOUD_HOME}/obj.${RAMCLOUD_BRANCH}/coordinator ${coord_args} --reset &> $RAMCLOUD_COORD_LOG &
 
-  # TODO Assuming 1sec is enough. To be sure monitor log?
-  sleep 1
-  # Silently kill coordinator
+  # TODO Assuming 5 sec is enough. To be sure monitor log?
+  sleep 5
+  # kill coordinator
   (pkill -f ${RAMCLOUD_HOME}/obj.${RAMCLOUD_BRANCH}/coordinator &> /dev/null)
 
   echo "DONE"