Transient errors, such as network failures, memory allocation failures, and force unmounts are handled automatically with the next replication job run. If the last replication job run was incomplete, the target file system is rolled back to the last known good Storage Checkpoint and all of the changes are reapplied.

In the event of machine failure, if enabled, the replication daemons will be started by the init scripts at the system boot time. The scheduler daemon periodically scans the Veritas File System (VxFS) mount points and starts the replication job if they were started when system went down. When scheduler daemon is killed, it must be started manually using the init script or the vfradmin command.

See Performing a VFR switchover.

See Performing a VFR failover after a disaster.

See Recovering a failed site if the failed source node comes up again.

See Recovering a failed site if a new node is assigned as the target.