When one system involved in an active/active link fails or when the primary system for an active/passive link fails, you can fail over the link to the other system involved in the link. Failing over the link stops replication on the link and, if DNS failover is enabled, allows client requests that target the failed system to be redirected to the healthy system.
If a system failure results in a replication link being broken (for example, due to the system being rebuilt after a catastrophic failure), you need to restore the link before replication can restart or data recovery can occur on that link. This applies regardless of the link type and, for an active/passive link, regardless of whether the failed system is the primary system or the replica.
To restart replication after failing over a link, you need to fail back the link. Failing back the link restarts replication on the link and returns the HCP systems involved to normal operation. For an active/passive link, failing back includes recovering data from the replica to the primary system.
Failover can be automated for both active/active and active/passive links. Failback can be partially automated for active/passive links.
This section of the Help provides instructions and considerations for managing failover and failback manually.
To fail over a replication link, restore a link, or recover replicated data, you need the administrator role.