Filesystem Dataset
Default dataset type. Mounts into the filesystem tree and stores files.
Solaris · ZFS Filesystem & Dataset Management
Solaris · Lesson 18
ZFS datasets are lightweight, flexible filesystems that support compression, quotas, reservations, snapshots, clones and seamless mountpoint control. This chapter teaches everything.
A ZFS dataset is a flexible filesystem OR volume created inside a ZPOOL. Datasets support advanced features:
Default dataset type. Mounts into the filesystem tree and stores files.
Block device dataset used for Oracle DB, iSCSI, or raw storage.
Instant copy-on-write snapshots and writable clones without needing extra space.
Shows all datasets, mountpoints, and usage. Most important command for day-to-day checks.
Datasets are lightweight and instant. ZFS encourages creating many small datasets.
Deletes dataset and its data instantly. Cannot destroy if children exist.
Useful when using dataset for raw backup, clones, volumes or other non-filesystem purposes.
When you want dataset mounted somewhere else than default path.
Shows 100+ tunables including compression, quota, checksum.
Quota = max size allowed. Reservation = guaranteed space.
lz4 is fast, safe and recommended for almost all datasets. Saves space.
Snapshots are read-only and almost instant. Perfect for backups and rollbacks.
Reverts dataset to exact state at snapshot. WARNING: destroys newer data.
Clones are writable datasets created from snapshots. Excellent for development/testing.
Most powerful backup tool for ZFS.
df makes it clear which dataset is mounted where and how much space it uses.
Next chapter: Snapshots & Send/Receive in advanced depth