A VSA cluster provides automatic failover from hardware and software failures.
Each VSA datastore has two volumes. A VSA cluster member exports the main volume as the VSA datastore. Another VSA cluster member maintains the second volume as a replica. If a failure occurs to the hardware, network equipment, or the VSA cluster member of the main volume, the main volume becomes unavailable, and the replica volume takes its place without service interruption. After you fix the failure and bring the failed VSA cluster member back online, the member synchronizes the main volume with the replica to provide failover in case of further failures.
A VSA cluster provides automatic failover from the following failures:
- Failure of a single physical NIC or port, or a cable connecting the NIC port to its physical switch port
- Single physical switch failure
- Single physical host failure
- Single VSA cluster member failure
The following illustration depicts automatic failover in a VSA cluster with 2 members. The replica volume takes over the failed main volume. In this case, to make sure that more than half of the members are online, the VSA cluster service simulates a VSA cluster member.
The following illustration depicts failover in a VSA cluster with 3 members.