One way to ensure that split-brain situations do not arise
is to devise an architecture that makes an even partition of the
cluster impossible or at least extremely unlikely. A single failure
in a four-node cluster could result in two equal-sized partitions,
but a single failure in a five-node cluster could not. The fifth
node in the cluster, then, performs the job of arbitration by virtue
of the fact that it makes the number of nodes in the cluster odd.
This type of node is sometimes called an arbitrator node, although
it may be indistinguishable from other cluster nodes and is not configured
in any special way in the cluster configuration file. This kind of
arbitration is especially useful when nodes in the cluster are separated by
significant distances, as in extended distance clusters or metropolitan clusters.
Arbitrator nodes may be configured to run non-clustered applications,
or they can be set up purely as arbitrators, with no other applications
running other than Serviceguard.
The use of an arbitrator node is shown in Figure 8.
The single system in Data Center 3 is an arbitrator node.
Printable version
