System storage pool

The system storage pool contains the system metadata (system and file attributes, configuration information, and metadata server state) that is accessible to all metadata servers in the cluster. There is only one system storage pool that is created automatically when SAN File System is installed. The system storage pool contains the most critical data for SAN File System. The first volume that is assigned to the system storage pool, called the master volume, contains the most critical pages of metadata that SAN File System manages.

Important: Use highly-reliable and available logical unit numbers (LUNs) for the system storage pool (for example, mirroring or redundant array of independent disks (RAID), plus hot spares in the backend storage system) so that the cluster always has a robust copy of the system metadata.

Because the amount of metadata grows as the global namespace grows, you must monitor the system storage pool to ensure that there is always enough volumes assigned to it to accommodate the growth. The system storage pool typically requires approximately 2% to 5% of the total storage capacity that SAN File System manages, but this amount varies depending on your environment. Use the alert features on the system storage pool to ensure that you do not run out of space.

Tip: The minimum size of a system volume is 2 GB; therefore, the minimum size of the system storage pool is also 2 GB.

For security and reliability, the volumes that are assigned to the system storage pool should be accessible only to the cluster using a private SAN or a shared SAN with a combination of zoning, LUN masking, or special configuration. For reliability, the volumes should be virtualized RAID arrays (also known as ranks within IBM® Enterprise Storage Server®).

Parent topic: Storage pools

Terms of use | Feedback
(C) Copyright IBM Corporation 2003, 2004. All Rights Reserved.