Cloudera Docs

Switching to disaster recovery cluster using Replication Manager

Hive ACID replication policies replicate managed data or transactional data from a source cluster to a destination cluster. In a scenario where you have a production cluster and you want to perform a maintenance activity, you might want to switch over to a disaster recovery cluster for uninterrupted usage. You can use the failover functionality to make the disaster recovery cluster as your primary cluster so that the data ingestion being performed by a replication policy is uninterrupted.

  1. Go to Cloudera Manager > Replication > Replication Policies page.
  2. Identify the Hive ACID table replication policy that has the required source cluster and target cluster.
  3. Click Actions > Start Failover for the required policy.
    This initiates the failover process. After the failover process finishes successfully, the Cloudera Manager of the target cluster initiates the next steps to complete the process.
  4. During the failover process, the Actions > Revert/Complete failover option appears for the policy. Use this option for the following scenarios:
    Scenario Description
    Stop the failover process when it is in progress In this scenario, the Replication Manager performs the Revert operation.

    The revert operation ensures that the process is stopped, the changes are reverted, the target Cloudera Manager initiates the complete command to remove any changes made on the target cluster, and then Replication Manager disables the policy.
    Target cluster Cloudera Manager did not initiate the completion command after the failover process finished In this scenario, the Replication Manager performs the Complete failover operation.

    The complete failover operation initiates the complete command by the target cluster Cloudera Manager which ensures that a clean up of temporary files and necessary modifications are complete on the target cluster. After the complete command finishes, Replication Manager disables the policy.
  5. After the failover process finishes successfully, the target cluster Cloudera Manager initiates the complete command.
    This action ensures that a clean up of temporary files and necessary modifications is complete on the target cluster. After the complete command finishes, the Replication Manager disables the replication policy on the disaster recovery cluster.

To initiate data replication from the disaster recovery cluster to the original source cluster, perform the following steps on the original source cluster (for example, production cluster):

  1. Remove the original source database.
  2. Create a Hive ACID table replication policy where you choose the disaster recovery cluster as the source and the original source cluster (for example, production cluster) as the target.