Deleting and recreating default bucket 2 times in a row doesn't work

This is equivalent of bugzilla bug 2908

Everything is latest as of 9 Nov 2010. Looks like bucket_engine issue.

My actions are:

• setup cluster of 4 machines (gen_cluster_script works too)
• load some data into default bucket
• failover some node and add it back & rebalance immediately

Frequently this leads to situation where node being failed over and added back
is stuck waiting on default bucket deletion. It prints message "Waiting for
"default" to finish shutting down before we start it." There is no IO or CPU
activity on the system. Everything is waiting on that default bucket removal.