Why ZFS

ZFS was designed 2005 by Sun for Solaris with the following development goals and state:

Not only a filesystem but a complete all inclusive package with
- Diskmanagement
- Raidmanagement
- Pool/ Volumemanagement
- Partitionmanagement without fixes sized partitions
- each Partition can grow in size up to poolsize but limitable with quotas and reservations
- Sharemanagement (Solaris based systems) with NFS and kernelbased highspeed SMB Server,
Windows compatible ACL, Active Directory support and access to snaps via Windows previous version
No reachable capacity limits (128 bit filesystem)
Highest datasecurity
- Multiple vdevs with Raid level n-way mirror, Raid-Z1-3 (up to three disk failures at a time without dataloss)
- All data with checksums to detect any sort of errors
- Copy on Write (data is written correctly or not at all)
- Scrubbing (background repair for silent data erros based on checksums, repairs open files)

see http://en.wikipedia.org/wiki/ZFS#Data_Integrity
High Performance striped Filesystem (Performance scales with vdevs)
- controller independent Software-Raid without Write Hole Problem (no battery needed)
Realtime blockbased deduplication
Encryption (ZFS based in Solaris 11)
- with encrypted files/ devices on other OpenSolaris based systems like OpenIndiana
Compression
Snapshots without time delay and initial space consumption due to CopyOnWrite
Even thousands of short intervall snaps example every 15 Min are possible
iSCSI targets with datacenter features
dozens of more ZFS properties

Common Filesystem for

Compare Filesytems

ZFS vs HFS+: http://arstechnica.com/apple/2011/07/mac-os-x-10-7/12/#hfs-problems
Wikipedia: http://en.wikipedia.org/wiki/Comparison of file systems

Pools are compatible if you use ZFS V. 28. On FreeBSD you must use GEOM partitions, GPT partitions are not interchangeable

napp-it 01.06.2025