Set the “workload manager cache size” configuration parameter to specify the maximum amount of memory that the workload manager can use. All memory used by the workload manager comes from the memory pool sized by workload manager cache size.
Connection migration consumes memory from this pool; each configured logical cluster, route, and load profile consumes memory from this pool. Actions resulting from commands such as failover and failback consume memory from this pool and continue to do so until the action is released.
Estimate the memory usage based on these guidelines:
Four memory pages for each concurrent migrating connection
One page for three logical clusters
One page for two load profiles
One page for 30 routes
One page for 12 actions
Use sp_configure to set the maximum value of the memory pool in increments of 2K pages. For example, to set the value of workload manager cache size to 100 2K pages, enter:
sp_configure "workload manager cache size", 100
workload_manager_size is dynamic; you need not restart the server. The default value is 80, or 160KB.
The default value is sufficient for most installations. If you anticipate migrating a logical cluster with many concurrent connections, you may need to increase the size of the memory pool.
