hesela.devZFS
ZFS is a copy-on-write filesystem and storage-management architecture that shares pooled space among datasets and integrates data-integrity checks.
The original design combines pooled allocation, transactional updates and block checksums rather than treating a filesystem as a fixed partition. [1] In OpenZFS, a pool can contain multiple filesystem datasets and block volumes (zvols); these are not separate pools. [2]
Checksums detect mismatches; repair requires a readable correct copy or sufficient reconstruction redundancy. Extra copies on one device can repair some corruption but cannot survive loss of that device. [3] A same-pool snapshot retains earlier block references, so it is not an independent backup against pool loss. This backup distinction is an inference from the shared storage dependency. [4]
Relations
- related
- ZFS virtual device (vdev), RAIDZ, ZFS Intent Log, ZFS transaction group, ZFS separate log, Copy-on-write (storage), Checksum, Storage snapshot
Citations
- Jeff Bonwick et al. The Zettabyte File System (2003), architecture and design status
- OpenZFS 2.3: zfs(8), filesystems, volumes, snapshots and bookmarks
- OpenZFS: Checksums and Their Use in ZFS
- OpenZFS: Snapshots, Clones and Bookmarks
Sources & caveats
- This record describes the shared ZFS architecture, with current semantics scoped to the cited OpenZFS documentation.
- The 2003 design paper supplies architectural context, not current support for arbitrary vdev nesting or a current performance claim.
- No pool, failure injection or recovery experiment was performed.