DLPAR performance implications

There are several implications of increasing or decreasing DLPAR performance.

You can add or remove memory in multiple logical memory blocks. When removing memory from a partition, the time it takes to complete a DLPAR operation is relative to the number of memory chunks being removed. For example, a DR operation removing 4 GB of memory from an idle partition takes 1 to 2 minutes. However, dynamically partitioning large memory pages is not supported. A memory region that contains a large page cannot be removed.

The affinity logical partitioning configuration allocates CPU and memory resources in fixed patterns based on multi-chip module, MCM, boundaries. The HMC does not provide DR processor or memory support on affinity partitions. Only the I/O adapter resources can be dynamically reconfigured when you are running affinity logical partitioning.

You can also take advantage of dynamic resource allocation and deallocation by enabling applications and middleware to be DLPAR-aware. This means the application can resize itself to accommodate new hardware resources. AIX® provides DLPAR scripts and APIs to dynamically resize vendor applications. You can find instructions for using these scripts or API's in the DLPAR section of General Programming Concepts.