[rdiff-backup-users] rdiff-backup initial backup and restore are both slow

[Nick Kartsioukas]

We have a box with about 100gb of email on it that we're trying to do an
initial rdiff-backup from, and so far it's taken almost a week and
hasn't completed yet.

We also tried doing a test restoral from backup of a box that has about
50gb on it, and the restore process takes about 5-6 hours before it
actually starts any real amount of data transfer.

Right now I'm re-running the restore of the 50gb box using strace on
both the backup host and the machine that's being restored to try to see
where it's spending all of its time, but I don't know if I have the
expertise necessary to parse the (at this point 50 megs after about 45
minutes) of strace output to see what's going on and why it's taking so
long.

What else can I try to debug these issues, and are there any known
workarounds or optimizations?

We're using the Debian Etch packaged version of rdiff-backup, 1.1.5.
Don't know if architecture differences have any impact, backup server is
an AMD64 box, machines being backed up are 32bit x86 boxes.

-- 
Nick Kartsioukas                  address@hidden
Digital West Networks, Inc.         http://www.digitalwest.net
tel(805)781-9378 -- fax(805)781-9379 -- toll free(888)781-9378