[rdiff-backup-users] What are the bottlenecks in --verify? (Or how to speed up verification?)

[Thomas Harold]

What are some options for speeding up the verification of past increments?

My guess is that the CPU might be a bottleneck for the SHA1 hash 
calculation, so that's something I would check first.

But how does the verify process work?  Does it reconstruct the file in 
memory, or does it use a temporary directory?

...

It seems like if I don't have TMP, TMPDIR, or TEMP defined as 
environment variables, it is operating system dependent on where Python 
creates the temporary file.  Or unless I pass the --tempdir option to 
rdiff-backup.

http://docs.python.org/2/library/tempfile.html

And if the --verify or --verify-at-time options create lots of temporary 
files, and write/read lots of data to the temporary directory, then I 
should probably move that directory to a separate set of spindles.