Clustering for High Availability

Clustering of up to 8 active/active nodes increases system availability by shortening planned and un-planned recovery times. In the rare event that a node within the cluster ceases to function properly, the remaining nodes seamlessly assume all functions of the failed node. Cluster failover times are kept short by ensuring that replay and file system recovery functions are performed concurrently. Virtual Servers are automatically migrated to a functional node with corresponding policy information while mapping to virtual machines is maintained. This means that all data is accessible from any node in the cluster – up to 100k distant for extended campuses.

Rolling cluster upgrades for updating without downtime

Rolling cluster upgrades allow for each node in a cluster to perform sequential, minor firmware updates without disruption or downtime. Each cluster is simply updated in round robin fashion until all nodes are running the same version

Storage Server	Mercury 50	Mercury 100	Titan 3100	Titan 3200
Nodes per Cluster	2	4	8	8
Total Capacity	4PB	8PB	8PB	16PB