Understanding Clustering

A cluster is a metasystem composed of two appliance controllers and shared storage. The clustering subsystem consists of three main building blocks:

  • Cluster I/O subsystem - The cluster I/O subsystem and the hardware device (cluster I/O ports) provide a transport for inter-controller communication within the cluster and are responsible for monitoring each peer's state. See Cluster Interconnect I/O.

  • Resource manager - The transport is used by the resource manager, which allows data service providers and other management subsystems to interface with the clustering system. See Cluster Resource Management.

  • Cluster management user interfaces - The cluster management user interfaces provide setup, resource allocation and assignment, monitoring, and takeover and failback operations.

A cluster improves availability and reduces business disruption in the following ways:

  • If one of the controllers experiences certain hardware or software failures, the second controller provides service while repair or replacement of the first controller is performed.

  • Rolling software upgrade can reduce the business disruption associated with migrating to newer software.

Each controller can be assigned storage, networking, and other resources from the set of resources that is available to the cluster. A cluster has one of the following topologies, depending on how resources are assigned to the controllers:

  • active-active - Each controller in the cluster is assigned at least one storage pool and the network resources needed by clients to reach the data stored in that pool.

  • active-passive - A single storage pool and the network resources needed by clients to reach the data stored in that pool is assigned to the controller that is designated as the active controller.

If a controller fails, the other controller (the peer) takes control of all known resources and provides the services associated with those resources.