Administering virtual disks


A cluster consists of a set of contiguous blocks of data written to a physical hard disk within a disk array.

When data is written to a disk array, it is split into clusters that are allocated to different physical drives. For example, Example striped array (RAID 0), shows an array of five drives. If a file is written to this array, it is split consecutively into stripes containing five clusters each. The first cluster is written to the first drive, the second cluster to the second drive, and so on. In this way, the stripe is written across the entire disk array (from left to right across the disks in the diagram). The cluster size is configurable from 512 bytes to 64KB in increments of 512 bytes.

If the file spans many stripes, every fifth cluster is written to the same drive. For example, on the diagram, clusters 0, 5, and 10 are written to the first drive; clusters 1, 6, and 11 are written to the second drive, and so on.

In Example striped array (RAID 0), if Disk 1 had a capacity of 100KB, a cluster size of 5KB would divide each disk into twenty 5KB clusters. For example, blocks 0, 5, 10,  ... 95 are clusters on Disk 1. Disk 2 contains blocks 1, 6, 11, and so on.

Next topic: RAID
Previous topic: Hot spares

© 2003 Caldera International, Inc. All rights reserved.
SCO OpenServer Release 5.0.7 -- 11 February 2003