570667 - MMR: simultaneous total updates on the masters cause
deadlock and data loss
https://bugzilla.redhat.com/show_bug.cgi?id=570667
Description: In the MMR topology, if a master receives a total
update request to initialize the other master and being initialized
by the other master at the same time, the 2 replication threads hang
and the replicated backend instance could be wiped out.
To prevent the server running the total update supplier and the
consumer at the same time, REPLICA_TOTAL_EXCL_SEND and _RECV bits
have been introduced. If the server is sending the total update
to other replicas, the server rejects the total update request
on the backend. But the server can send multiple total updates
to other replicas at the same time. If the total update from
other master is in progress on the server, the server rejects
another total update from yet another master as well as a request
to initialize other replicas.