[clug] fastest way to copy a fs

Peter Barker pbarker at barker.dropbear.id.au
Thu Jan 18 22:24:21 GMT 2007

On Fri, 19 Jan 2007, Tony and Robyn Lewis wrote:

> I'm not sure this is going to save you anything.  Each run of rsync is going 
> to completely read both source and dest, in order to checksum all the chunks. 
> So unless your reads are significantly faster than your writes, you're not 
> going to save anything in your downtime.
> I could be wrong though...

Slightly.  By default rsync doesn't checksum all files on a re-sync.  It 
will compare dates-and-times and size first, and if they haven't changed, 
assume the content hasn't changed either.  See the --checksum option.

You still have to traverse the filing system, though, and that's kind of 

One option which hasn't been discussed is snapshotting.  I've never played 
with it (just one of a million interesting things...), but would it be 
possible to:
  a) snapshot, sync the snapshot followed by
  b) generate diffs between the snapshot and the current state of the fs
  c) copy the diffs to the snapshot to the machine holding the copy and 
  d) apply those to the copy-of-the-snapshot.

As I said, I've never played with snapshotting, and I wouldn't have 
/thought/ the fs would bear the changes-to-the-snapshot information to the 
user.  Just a thought.

> Tony

Peter Barker                          |   N    _--_|\ /---- Barham, Vic
Programmer,Sysadmin,Geek              | W + E /     /\
pbarker at barker.dropbear.id.au         |   S   \_,--?_*<-- Canberra
You need a bigger hammer.             |             v    [35S, 149E]
"Peter is apathetic, and I'm vaguely apathetic" -- Rachel, thinking of organising a movie trip

More information about the linux mailing list