Single Point of Failure (SPOF)

To remove potential SPOFs from a cluster environment, all hardware must be redundant. Creating disk mirrors within a single disk array is a SPOF, and is not recommended (unless a fully redundant Sun StorEdge™ A3500 array is used). If the array loses power or a controller becomes inoperative, data could be lost or corrupted. The cluster software may not be able to recover from data corruption or arrays going offline.

Clustering between two domains within a single Sun Enterprise 10000 server is not recommended because the server has a single backplane which has the (unlikely) potential of being a SPOF. Another disadvantage when using a single machine for high availability is the potential of that single machine becoming ...

Get Sun™ Cluster Environment: Sun Cluster 2.2 now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.