Copying a large file system (again)

[Tino Schwarze]

(replying to the list since it might be interesting for others as well)

On Mon, Dec 03, 2007 at 12:30:30PM -0500, Robert Terzi wrote:

> You can use find's -links option to break things up
> into multiple passes by the number of links.  Do all
> of the unique (single link files first), then do multiple
> links.  Both tar and rsync take a --files-from= option.
> 
> find . -type f -links 1 > unique-files
> find . -type f -links +2 > linked-files
> 
> tar cf - --files-from unique-files | ssh ...
> 
> If necessary you can break things up further to do
> a specific number of links, i.e. try to do the
> files with the max number of links as one pass.

This is smart. I'll probably see whether I can hack up some mini perl or
similar so I can create all the files  in one pass - I suppose, the find
alone would take about an hour or so for traversing the file system once.
And I'm not sure about the link counts - there should be hardly any
files with link count 1 (only config and logs).

BTW: Some files have tens of thousands of links.

Thanks,

Tino.

-- 
www.craniosacralzentrum.de
www.lebensraum11.de
www.spiritualdesign-chemnitz.de

Tino Schwarze * Parkstraße 17h * 09120 Chemnitz