On Sat, Nov 06, 2010 at 06:52:19PM +0000, krad wrote: > > the main problem is geom and ufs isnt a like for like replacement yet. Good > as though geom is it just not as easy as zfs from an adminsistration point > of view in my opinion. It may potentinally get a block checksum class but it > will be a long time before its like for like.
I have not really spent any quality time with ZFS, so I'm a little sketchy on the details. Is there anything the checksumming capabilities of ZFS do that cannot be duplicated with an external tool -- perhaps something like a filesystem integrity auditing system? > > I've had a play around with btrfs, which is supposed to be an opensource > equivelent to zfs. It is far from ready yet though. It may mature into a > good product in the future, but its a long way off and far from polished > (dam horrible from what ive seen so far). Most of its development was backed > by oracle though from what i have read, so who knows where that will go now. > If oracle want to continue to push linux and it to have a decent fs, it may > well just be easier for them to drop the licensing issues with cddl which > was preventing zfs from making it into linux. Who knows but for anything in > the near to medium future there is nothing to rival zfs on the opensource > market. As far as I'm aware, btrfs has not been ported to any BSD Unix systems, either -- so there's a major downside to btrfs (as compared with ZFS). > > Having said all that it really depends on whether you need the extra > features of zfs. Personally I cant see how anyone with any important data > can do without checksuming. I guess that depends on what you're doing with the data and what kind of external tools you have in place to protect/duplicate it in case of a problem. -- Chad Perrin [ original content licensed OWL: http://owl.apotheon.org ]
pgpXi4bBsOBXE.pgp
Description: PGP signature
