|Skip Navigation Links|
|Exit Print View|
|Oracle Solaris ZFS Administration Guide Oracle Solaris 10 1/13 Information Library|
The following sections describe recommended file system practices.
The following sections describe ZFS file system creation practices.
Create one file system per user for home directories
Consider using file system quotas and reservations to manage and reserve disk space for important file systems
Consider using user and group quotas to manage disk space in an environment with many users
Use ZFS property inheritance to apply properties to many descendent file systems
Consider the following file system practices when creating an Oracle database.
Match the ZFS recordsize property to the Oracle db_block_size.
Create database table and index file systems in main database pool, using an 8 KB recordsize and the default primarycache value.
Create temp data and undo table space file systems in the main database pool, using default recordsize and primarycache values.
Create archive log file system in the archive pool, enabling compression and default recordsize value and primarycache set to metadata.
For more information, see the following white paper:
You should monitor your ZFS file systems to ensure they are available and to identify space consumption issues.
Weekly, monitor file system space availability with the zpool list and zfs list commands rather than the du and df commands because legacy commands do not account for space that is consumed by descendent file systems or snapshots.
For more information, see Resolving ZFS Space Issues.
Display file system space consumption by using the zfs list -o space command.
File system space can be unknowingly consumed by snapshots. You can display all dataset information by using the following syntax:
# zfs list -t all
A separate /var file system is created automatically when a system is installed, but you should set a quota and reservation on this file system to ensure that it does not unknowingly consume root pool space.
In addition, you can use the fsstat command to display file operation activity of ZFS file systems. Activity can be reported by mount point or by file system type. The following example shows general ZFS file system activity:
# fsstat / new name name attr attr lookup rddir read read write write file remov chng get set ops ops ops bytes ops bytes 832 589 286 837K 3.23K 2.62M 20.8K 1.15M 1.75G 62.5K 348M /
Keep file system snapshots
Consider enterprise-level software for weekly and monthly backups
Store root pool snapshots on a remote system for bare metal recovery