How many volumes should I assign to the storage pools?

This topic provides information about assigning volumes to storage pools.

Files are distributed in partition-sized units across the available space in a given storage pool. The storage pools need volumes to distribute those file access across the volumes, thus distributing the I/O load.

Also consider the number of spindles available for volumes in the storage pool. The storage pools need enough spindles to provide adequate I/O parallelism.

Choose volumes from different storage subsystem ranks to allow more than one port to participate in I/O.

If you anticipate needing to remove volumes from a storage pool, the data on the volume to be removed will be distributed across other volumes in the storage pool. There must be enough room on the remaining volumes to contain the data.

Each storage pool manages its own volumes. File space is allocated to the volumes in a given storage pool in a round-robin algorithm (as shown in Figure 1) in logical partitions, or in blocks. Logical partitions are allocated to the system storage pool in 16-MB blocks. For user storage pools, including the default storage pool, you can allocate logical partitions in 16, 64, or 256-MB blocks. All logical partitions in the same storage pool must be the same size.


This figure illustrates how data is allocated to volumes in a storage pool.
Tip: You can set a threshold to generate an alert when a storage pool reaches or exceeds a certain percentage of its maximum capacity. By default, an alert is generated when a storage pool becomes 80% full. An alert is logged every five minutes until one or more volumes are assigned to the storage pool. You can set configuration parameters to cause an SNMP trap message to be generated as well. An SNMP trap notifies you of this condition asynchronously.

Parent topic: Planning the storage configuration

Library | Support | Terms of use | Feedback
(C) Copyright IBM Corporation 2003, 2004. All Rights Reserved.
IBM TotalStorage SAN File System v2.2