Overview of Manual Failover Steps When Group Leader Is Unavailable

Performing a manual failover when the group leader is not available involves performing several procedures. Because the array OS GUI is not available when the group leader becomes unavailable, you must perform most of these steps from the backup group leader using CLI commands.
NOTE: For instructions on performing the takeover, refer to the section Perform a Manual Takeover of the Group Leader Array in the CLI Administration Guide.
When the group leader cannot be reached, the following happens:
  • There is no access to the array OS GUI.
  • Group leader services are unavailable until a group leader takeover has been performed.
  • The group leader is not able to transition out of sync. This is because the Group Data Service is unavailable until the manual takeover has been completed.
  • There are only standby paths to the backup group leader until the takeover completes. At that point, the backup group leader becomes the group leader.
  • The backup group leader can be accessed from the CLI using the secondary management IP address.

Before you begin the manual takeover, you must make sure there is a common snapshot between the two arrays in the group. If there is not a common snapshot, you will not be able to re-enable synchronous replication.

To perform a manual failover when the group leader that was using synchronous replication is unreachable, you must:
  1. Confirm that the group leader is unavailable.
    IMPORTANT: If you can reach the group leader using the CLI command group --check_takeover, do not perform a manual takeover.
  2. If you have a clustered configuration, take down the cluster.
  3. From the backup group leader, use CLI commands to transition it to the role of group leader and start the group leader services.

    When the backup group leader becomes the new group leader, the previous group leader will transition to an out-of-sync state. At that time connections to it will be available.

    For instructions on performing this task, see the section Perform a Manual Takeover of the Group Leader Array in the CLI Administration Guide.

  4. Use CLI commands to remove synchronous replication from the downstream partner to allow access to the volume. The following will happen:
    1. Existing ACLs from the upstream volume are cloned to the downstream volume
    2. The LUNs on the downstream volume get new numbers

    For instructions on performing this task, see the section Remove Synchronous Replication from the Downstream Partner in the CLI Administration Guide.

  5. If you are using a clustered configuration, bring the cluster back up.
  6. (Optional) Set up synchronous replication on the new group leader (the former backup group leader) after you have made sure that both the group leader and backup group leader are in sync.
    See Add Synchronous Replication.
    NOTE: It is a good practice to set up ASO and the Witness.
  7. (Optional) Perform a handover to restore the original direction.

    See Perform a Volume Collection Handover.

  8. (Optional) Migrate the group leader role back to the original group leader. Use CLI commands to perform this task.