XRC supports large configurations where many system
data movers are required to handle the whole workload. The system
data movers may exist in a coupled XRC master session. Any delays
in one session may cause delays in other sessions. To avoid a persistent
workload imbalance between data movers, use the following guidelines:
- Balance the workload (highly active volumes, readers, and LSS)
across data movers.
- Run the maximum number of data movers that the available resources
(processor and memory) on an individual LPAR permits.
- Enable clustering on LPARs and balance data movers across the
clusters.
- Balance central electronics complex (CEC) resources across the
data mover LPARs.