Performing Service Failback

Prerequisites

• The service failback takes effect only for the following replication groups in the cluster:
  1. The role of the target cluster for the service failback is active, and the replication group contains a cluster of the passive (auto takeover) type.
  2. The role of the target cluster for the service failback is active, and other active clusters exist in the replication group.
• The status of a cluster that has synchronization tasks to the local cluster is abnormal and cannot wait until data synchronization is complete (In this case, service failback will fail. Ensure that the cluster status is normal and try again).
• A cluster (for example, cluster 2) and the local cluster form cross-site DR but are not in the same replication group as the local cluster. (In the event where cluster 2 is abnormal, you are advised not to perform service failback on the local cluster)
• For a replication group consisting of an active site and passive site (automatic takeover), after service failback starts, both the active and passive sites are read-only. After service failback is complete, the active site becomes readable and writable, and the passive site becomes read-only. Object upload requests will fail during service failback. Therefore, you are advised to perform service failback after uploads are complete or upload objects after service failback is complete. For a multipart upload, combine or cancel the upload before performing service failback.

Procedure

1. Choose Data Protection > Configuration > Cross-Site DR > Topology.
2. In the local cluster area on the right of the function pane, choose Operation > Perform Service Failback.

   The Perform Service Failback page is displayed on the right.

3. Set service failback parameters, as described in Table 1.

   Table 1 Service failback parameters

   Parameter	Description
   Failback Type	Failback after data sync: Synchronize data to the local cluster before performing service failback. If data consistency is required between sites, select Failback after data sync. Immediate failback: perform service failback immediately without checking whether data synchronization is complete. Data may be inconsistent between sites temporarily. Inconsistent data will be synchronized in the background. If there is no requirement on data consistency between sites, select Immediate failback.
   Sync Acceleration	If synchronization acceleration is enabled, data synchronization from other clusters to the local cluster will be accelerated until the service failback is complete. NOTE: You can determine whether to enable synchronization acceleration only when Failback Type is set to Failback after data sync.

4. Click OK.