Understanding volumes and qtrees

What volumes and qtrees are

A volume is a file system on a filer. It is physically defined by the disks that are assigned to it.

Root volumes

Each filer has a special volume known as a root volume. The root volume contains the filer's configuration files, and a file known as the /etc/rc file, which contains startup commands. There can be only one root volume on a filer.

Data volumes

In addition to the root volume, you can create one or more data volumes, which hold the user visible directories and files that you want to store.

qtrees

A qtree is a logically defined file system that can exist as a special subdirectory of the root directory of a volume.

When creating volumes is appropriate

You can create up to 199 additional volumes when you want to treat large amounts of data as a unit. You create one or more additional volumes under the following circumstances:

A volume has more disks or storage than it is convenient to manage as one unit.
You want to accommodate users with different security styles or oplocks requirements that current volumes do not meet and you do not want to create a qtree for that purpose.
You have a multilingual environment and want to provide one volume for each language and character set.

If your filer has a large number of disks, creating volumes and having a multivolume configuration can also aid in administration. Having multiple volumes enables you to do the following:

Perform administrative and maintenance tasks, such as backup and restore, on individual volumes rather than on a single, large file system.
Set vol option command values, such as snap sched, raidsize, minra, and no_atime_update, differently for individual volumes.
Take individual volumes offline--to perform administrative tasks on them, for example--without interrupting data availability of the other volumes.
Have more flexible quota and snapshot configurations.
Have backups and restores of individual volumes, which take less time.

However, there is a trade-off: You must use additional disks as parity disks and as hot spares rather than for data storage.

When creating qtrees is appropriate

You create qtrees for amounts of data that do not expand the size of volumes to the point that the volume requires additional disks. You can set up qtrees and remove them easily. However, there is a trade-off. Because qtrees are part of a volume, they inherit many volume-level settings.

Detailed comparison

Qtrees and volumes have other differences and similarities that can play a role in your decision, as shown in the following table.

Qtrees	Volumes
Both enable you to organize user data. Both enable you to group users with similar needs. Both have a security style and an oplocks setting. The security style determines whether files use UNIX or Windows NT permissions. The oplocks setting determines whether files and directories use CIFS opportunistic locks. Both can be used as units in a backup and a restore. Neither can be crossed by hard links.
You can add and delete qtrees on the fly; this provides flexibility.	You can add volumes by adding disks; this makes more storage available on the filer.
You can expand and shrink qtrees.	You can expand but not shrink volumes.
You cannot create individual qtree snapshots unless you use SnapMirror. You are dependent on volume snapshots.	You can create volume snapshots, which enable you to recover and mirror data.