ION Accelerator enables a powerful and effective High Availability (HA) environment for your shared storage. HA clustering provides an important option for customers who prefer array-based HA over application-based mirroring. This can be especially useful if your application does not provide logical volume management, such as with all VMware environments and most implementations of Microsoft Clustering.
Clustering relies on point-to-point connections – HA networking across geographically distributed sites is not supported.
The diagram on the left shows a simple HA clustering setup, using a 40Gb Ethernet connection between appliances. The diagram on the right shows an ION Accelerator mirroring configuration without clustering, an alternative approach.
The diagram below shows basic LUN access (exported volumes) in an HA configuration
In this simplified example, Node 1 presents LUN 0 to the application, while Node 2 presents LUN 1 to the application. All writes to LUN 0 are synchronously replicated to Node 2. All writes to LUN 1 are synchronously replicated to Node 1. When the replication is complete, the original node acknowledges the write by the application.
Replication occurs over the 40Gb Ethernet interconnect between the two nodes, which consists of two dual-ported ConnectX®-3 adapters. In the event a node fails, all data will be available from the remaining active unit.