When a protected object is asynchronously replicated to a replica object, data replication and configuration synchronization are involved. This process generally includes the following steps:
Building a data link
A data link is built between the source cluster and the target cluster through the replication network between the source-side replication service and the target-side replication service. The data link is used to transmit data from the protected object to the replica object.
All replication services register with AOC, and AOC maintains the connection address of each service. You can associate replication services with clusters on the AOC UI. When you replicate a virtual machine in the source cluster, AOC passes the replication service associated with the cluster to the target-side replication service as the peer replication service. The target-side replication service periodically checks network connectivity with the source-side replication service. During replication, the target-side replication service establishes an available data link with the source-side replication service. If the source cluster is associated with multiple replication services, the target-side replication service randomly selects a reachable source-side replication service as the peer replication service and establishes a data link. The target-side replication service can access ABS in the source cluster through the data link.
Obtaining the current state of the protected object
The target-side replication service sends a request to AOC to create a snapshot for the protected object and obtain the metadata of the protected object snapshot. The metadata contains the configuration information of the protected object and the ABS volume snapshot, and records the state of the protected object at that point in time.
Synchronizing the protected object configuration to the replica object
Before replicating protected object data, the target-side replication service first sends a request to AOC to copy the configuration information of the protected object to the replica object, ensuring that the replica object has the same configuration as the protected object.
For initial replication, the target-side service first creates a replica object in the target cluster. The replica object has the same configuration as the protected object.
For incremental replication, the target-side service directly synchronizes the configuration information of the protected object to the replica object.
Replicating protected object data to the replica object
After the configuration information of the replication object is synchronized, the snapshot data of the protected object is synchronized to the replica object.
Initial replication uses full asynchronous replication to transmit all snapshot data of the protected object to the replica object.
Subsequent replication uses incremental asynchronous replication and compares data block fingerprints to identify data that has actually changed since the last replication, and transmits the changed data to the replica object.
Creating a restore point on the replica object
After asynchronous replication of protected object data is complete, a snapshot is created for the replica object as the restore point of asynchronous replication. Later, after the protected object fails, you can execute a failover job to start the replica object from the specified restore point state so that it can take over the protected object and continue providing services.
