When a backup or read-only fileset occupies the same site (File Server machine and aggregate) as its read/write source fileset, DFS attempts to save disk space by having the filesets share data that is the same across the different types of filesets. This data sharing is accomplished in the following way:
· When the backup or read-only fileset is created, the new fileset is filled with an array of pointers to the data housed by the read/write source.
· The identities of the read/write source and the backup or read-only fileset are then exchanged so that the read/write source becomes the backup or read-only fileset, and the backup or read-only fileset becomes the read/write source.
This technique provides full access to the data via the read/write fileset without requiring that the read/write fileset physically house the data. As long as the read-only or backup fileset remains identical to the read/write source, the disk space occupied by the read/write fileset remains small (because pointers take up much less disk space than the data to which they point). However, as changes are made to the data in the read/write fileset, the amount of space occupied by the read/write fileset increases. This is because the read/write fileset must acquire additional disk blocks to store changed data; it can no longer simply point to disk blocks housed by the read-only or backup fileset.
Because of this data-sharing arrangement, DFS provides statistics on two types of disk usage for a fileset. The first statistic, quota, identifies the amount of disk space occupied by all of the files and directories in the read/write fileset, including those files and directories in the read/write fileset that are actually pointers to disk blocks in the backup or read-only version of the fileset. The second statistic, allocation, identifies the amount of disk space occupied by those files and directories actually housed in the fileset, excluding those files and directories that are represented by pointers to disk blocks in another version of the fileset.
Users are concerned with the quota statistics only, because the quota dictates the amount of data that they can store in a read/write fileset. To check a fileset's quota statistics, issue the fts lsquota command. (See Part 2 of this guide and reference for information on the fts lsquota command.)
Administrators are also concerned with the quota statistics, especially when dealing with users. One of the most common user requests of an administrator is for an increase in the size of the user's fileset quota. To change the size of a fileset's quota, issue the fts setquota command. (See Managing Filesets for information on changing a fileset's quota.)
Administrators are also concerned with the allocation statistics when they need information on the physical disk usage of a fileset (for example, when moving filesets for load-balancing purposes). To check both the quota and allocation statistics for an individual fileset, issue the fts lsft command. To check the quota and allocation statistics for multiple filesets on a File Server machine, issue the fts lsheader command. (See Managing Filesets for information on the fts lsft and fts lsheader commands.)
Note: There is no way to set or change the allocation for a fileset. It is automatically set during the creation of the fileset and cannot be changed.