Soft Partition Overview

As disks become larger, and disk arrays present ever larger logical devices to Solaris systems, users need to be able to subdivide disks into more than 8 sections, often to create manageable file systems or partition sizes.

Note:

Enhanced Storage can support up to 8192 logical volumes, but is configured for 128 (d0-d127) by default. To increase the number of logical volumes, see "Changing SVM Defaults" in the Solaris Volume Manager Administration Guide. Do not increase the number of possible logical volumes far beyond the number you will actually use, because Enhanced Storage creates a device node (/dev/dsk/md/*) and associated data structures for every metadevice permitted by the maximum value. For small Enhanced Storage configurations, these additional possible volumes can result in substantial performance impact.

You use soft partitioning to divide a disk slice into as many divisions as needed. You must provide a name for each division or soft partition, just like you do for other storage volumes, such as stripes or mirrors. A soft partition, once named, can be accessed by applications, including file systems, directly as long as it is not included in another volume. Once included in a volume, the soft partition should no longer be directly accessed.

Soft partitions can be placed directly above a disk slice, or on top of a mirror, stripe or RAID 5 volume. Nesting of a soft partition between volumes is not allowed. For example, a soft partition built on a stripe with a mirror built on the soft partition is not allowed.

Although a soft partition appears, to filesystems and other applications, to be a single contiguous logical volume, it actually comprises a series of extents that may be located at arbitrary locations on the underlying media. In addition to the soft partitions, watermarks (also called system recovery data areas) on disk record information about the soft partitions to facilitate recovery in the event of a catastropic system failure.

Requirements for Soft Partitioning

A soft partition may be used to subdivide a device that is larger than a terabyte; however it will not produce a volume larger than a terabyte.

Slices that are used for soft partitions cannot be used for other purposes without losing data.

Suggestions for Soft Partitioning

While it is technically possible to manually place extents of soft partitions at arbitrary locations on disk, allow the system to place them automatically.

When you partition a disk and build file systems on the resulting slices, you cannot later extend a slice without modifying or destroying the disk format. With soft partitions, you can extend them up to the amount of space on the underlying device without moving or destroying data on other soft partitions.

Although you can build soft partitions on any slice, creating a single slice occupying the entire disk and then creating soft partitions on that slice is the most efficient way to use soft partitions at the disk level.

To improve your ability to expand and manage storage space, build stripes on top of your disk slices, then build soft partitions on the stripes. This allows you to add new slices to the stripe later, then grow the soft partitions.

For maximum flexibility and high availability, build RAID 1 (mirror) or RAID 5 volumes on disk slices, then soft partitions on the mirror or RAID 5 volume.