Hard Disk IDs in Linux

Michael Torrie torriem at gmail.com
Tue Mar 19 08:55:20 MDT 2013


On 03/19/2013 12:43 AM, Dan Egli wrote:
> That's interesting. I have not personally looked at ZFS except as a way to
> obtain deduplication. If there's a way to make Deduplication work on other
> file systems (and not using the script to replace files with links, but
> actual deduplication of individual data blocks) then I'll look into that
> for my system and for the work box. Do you recall where you saw the
> "opendedup" project? I'd like to look into that!

I'm curious as to what advantage deduplication gives you on a home
server?  On a large file system with thousands of users, I can see how
it could be a benefit there. But on my own computer, the large storage
is taken up by photos, music, and video and I highly doubt there are
many duplicate blocks there.

I suppose if you're PXE-booting multiple computers with an NFS root file
system, then deduplication could be of some benefit.  Or maybe virtual
machines.  Some years ago, Red Hat was working on a system to kind of of
do this, which they called "Stateless Linux[1]".  I'm not sure if it's
still an active thing or not.  For me disk was too cheap to worry about
this sort of thing.

[1] http://people.redhat.com/dmalcolm/stateless/stateless-linux-HOWTO-en/



More information about the PLUG mailing list