The hype is that with ZFS you don't need an expensive/extra point of failure RAID card and RAID card config, plus a logical volume manager with it's own utilities plus a filesystem plus a quota management system - you plug disks in, ZFS does the RAID and the volume management and is the filesystem and the quotas all in one set of tools (quite few / simple commands for basic usage, at least).
On top of that, ZFS stores all data along with checksums, so when you read a file it will pick up errors and correct them. How do you know the photos you haven't accessed recently are OK? That they will not be corrupt? You don't - you assume that if the disk shows up and the directory list shows up then they will be there. Wih ZFS, you set up scheduled scrubs and it verifies all the data is present, readable and correct - and because of this regular check, it picks up hardware failures more quickly.
Also, it has instant (well, small constant time) snapshots, which doubles as a time-machine/filesystem versioning system so you can see files as they were, and you can clone snapshots to take backups / send them over a network to keep a clone in sync.
Also, it was going to be Zettabyte File System (rumour) because it was designed to take filesystem size limits out of consideration.
Did I mention it writes data, then updates a pointer to the data in one operation so as long as your disks /controller do not cheap out and misrepresent what they are doing, data is always in the old or the new state - never half written.
Also, because of the lack of RAID controller, you can move the disks to another system and attach the zpool with leas requirements in an emergency.
That's pretty awesome - not just acronyms and buzzwords.
On top of that, ZFS stores all data along with checksums, so when you read a file it will pick up errors and correct them. How do you know the photos you haven't accessed recently are OK? That they will not be corrupt? You don't - you assume that if the disk shows up and the directory list shows up then they will be there. Wih ZFS, you set up scheduled scrubs and it verifies all the data is present, readable and correct - and because of this regular check, it picks up hardware failures more quickly.
Also, it has instant (well, small constant time) snapshots, which doubles as a time-machine/filesystem versioning system so you can see files as they were, and you can clone snapshots to take backups / send them over a network to keep a clone in sync.
Also, it was going to be Zettabyte File System (rumour) because it was designed to take filesystem size limits out of consideration.
Did I mention it writes data, then updates a pointer to the data in one operation so as long as your disks /controller do not cheap out and misrepresent what they are doing, data is always in the old or the new state - never half written.
Also, because of the lack of RAID controller, you can move the disks to another system and attach the zpool with leas requirements in an emergency.
That's pretty awesome - not just acronyms and buzzwords.