ABSTRACT
Storage arrays are more standard, more limited, and simpler than servers; therefore, they have higher uptime. The expense of the extra hardware/software/care-and-feeding of servers capable of failover and arrays capable of failover is simply too large as compared to the cost and likelihood of downtime. Failover with replication, also known as shared nothing, uses server-based replication software. Failover with storage-array based replication employs storage array replication configured to process all writes and replicate them to another storage array. Active-active failover relied on one active controller and a passive controller that would become active in the case of a failover. The remedy to cluster problems is to use standardized hardware configurations and virtual server configuration cloning to create the target failover server. The storage array firmware is designed to handle replication, backup, and snapshots alongside regular tasks such as servicing read and write requests, handling host connections, and handling errors.
