Requirements for geographic mirroring

Geographic mirroring increases CPU load, so there must be sufficient excess CPU capacity. Add processors as needed to increase CPU capacity.
For optimal performance for geographic mirroring, particularly during synchronization, increase your machine pool size by the amount given by the following formula:
The amount of extra machine pool storage is: 271.5 MB + .2MB * Number of disk units in independent ASPs.
The extra machine pool storage is required on the target node. However, because the target node changes when doing switchovers and failovers, you should increase the machine pool on all nodes in the Cluster Resource Group. To prevent the performance adjuster from reducing the machine pool size, you should do one of the following:
1. Set the machine pool minimum size to the calculated amount (the current size plus the extra size for geographic mirroring from the formula) using Work with Shared Storage Pools (WRKSHRPOOL) command or Change Shared Storage Pool (CHGSHRPOOL) command.
  Note: It is recommended to use this option with the Work with Shared Storage Pools (WRKSHRPOOL) option.
2. Set QPFRADJ to zero which prohibits the performance adjuster from changing size of the machine pool.
Configure a separate storage pool for the jobs using geographic mirrored independent disk pools, especially if you specify a long recovery timeout.
Geographic mirroring is performed when the disk pool is available. When geographic mirroring is being performed, the system value for the time of day (QTIME) should not be changed.
Communications requirements for independent disk pools are particularly critical as they affect throughput. See Communications requirements for more information.
All independent disk pool requirements must be met. See Plan for independent disk pools for more information.