A performance observation.

Previous Topic
classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

A performance observation.

Erik Forsberg

Previously, I ran rdiff-backup 0.12 on the backup server, accessing my
data partitions via NFS from the actual file server. Backing my 108GB
/home partition usually took about 2h 45 minutes, eating something
between 2 and 6Mbit/s during the time.

A few days ago, i changed the scheme, instead letting the file server
contact the backup server via ssh. Now the same backup takes about 55
minutes. I'm not done with the bandwidth analysis, but it seems to
consume any measurable bandwidth only during the actual data
transfer. During the checking of which files to transfer, a much lower
bandwidth usage can be observed.

This was a pleasant surprise. Of course, a lot less data is now
transferred between the machines, but I didn't expect the backup time
to be less than half the old one. Also, this is good news for me,
since I'll move the backup server to a different location, and the
available bandwidth to that location is smaller than the one I have on
my internal network.

Additional bonus comes in form of increased security, utilizing the
--restrict-update-only parameter in my authorized_keys file. I think
I'll sharpen up the security by using S/key logins to my backup server
for administrative logins, making keyboard sniffers on administrative
clients less useful.

Rdiff-backup is truly wonderful software!

Erik Forsberg                OpenSource-based Thin Client Technology
Cendio AB                    Phone: +46-13-21 46 00    
                             Web: http://www.thinlinc.com

rdiff-backup-users mailing list at [hidden email]
Wiki URL: http://rdiff-backup.solutionsfirst.com.au/index.php/RdiffBackupWiki