Master-Master-Slave replication recovery

Hi,

We have set up a master-master-slave replication setup.But the facing the problem in recovery process.
We have Master A,Master B and Slave C.
Master A and B are dual master with log-slave-updates and skip-slave-start option enable.
C is a slave of A with log-slave-updates and skip-slave-start option enable.
So the design is
A <—> B
|
C
Replication is working fine but what happen when A fails or B fails or both A and B fails.
What are the recovery process in that case?

Regards,
fedora

There is no single answer to that question. Books can be written about that topic.

Whatever you do, only write to one server at a time! DO NOT write to both masters at once.