Sysplex configurations
IBM® recommends that you provide redundant channels, signaling paths, Sysplex Timers, and shared couple data sets in order to avoid having a single point of failure in the sysplex.
For example, if a 3088 fails and there is not a second 3088 to maintain signaling connectivity between MVS™ systems in the sysplex, all but one system must be removed from the sysplex and put in a wait state. Or, if you do not have an alternate couple data set and the primary couple data set fails, all systems are put in a non-restartable wait state.
When a sysplex system is running under VM or in a logical partition
that is sharing processor resources, you must ensure that its processor
resources are sufficient to ensure good overall system performance.
A system that is running without sufficient resources can degrade
the performance of sysplex-wide communications functions, such as
global resource serialization, signaling, and console communications.