Ticket 467 - CLEANALLRUV abort task should be able to ignore down replicas
Bug Description: The abort task previously would wait for all the replicas to be online,
and receive the extended op task. This made it impossible to abort the
cleanallruv task, if you wanted to abort because a server was down.
Fix Description: Changed the default behavior to ignore down replicas, and added a new attribute
to the abort task entry: replica-certify-all: (yes,no). Default is "no". If
set to "yes", then it will wait until all the replicas have received the abort task.
Also fixed a crash caused by running the abort task agsint a server that does not have
replication setup. The crashed was caused by freeing uninitialized pointers.
https://fedorahosted.org/389/ticket/467
Reviewed by: richm(Thanks!)