BackupPC-users

[BackupPC-users] BackupPC full backup speed VS iperf measurement

2010-09-16 20:53:01
Subject: [BackupPC-users] BackupPC full backup speed VS iperf measurement
From: Daniel Sperka <djsperka AT ucdavis DOT edu>
To: backuppc-users AT lists.sourceforge DOT net
Date: Thu, 16 Sep 2010 17:27:52 -0700
Hi -

I have been looking at network transfer speeds and comparing them to the rates reported for full backups.

The full backups use rsync, server is a Mac Xserve, client is an Ubuntu Hardy machine.  This is NOT the first backup of the machine. This machine is also using checksum caching.

The web interface reports that my most recent full backup was done at a rate of about 8MB/sec (pretty typical, give or take).

During the backup (which took about an hour) I was running network speed tests with iperf. The tests where one-way, from the backup client to the backup server. During this time the rates reported by iperf were always less than 1MB/sec (e.g. 0.4Mbits/sec or 0.05 Mbytes/sec)!  (Yes, I know that's very slow - that's a part of the larger problem I'm trying to solve.)

Why the discrepancy?

If I naiively assume that a full backup (rsync+ssh XFER method) will transfer ALL FILES to the server, then there is no way to understand the discrepancy.

If, on the other hand,  I assume that rsync is doing file comparisons on both ends and transferring files as needed (i.e. those which are different than those in the pool, or which are new), then the actual amount of data transferred is less than the full amount (the web interface says its just 92MB new files). If so, then the rates I measure using iperf are closer to what backuppc/rsync actually transferred (92MB/70 min = 0.02 MB/sec).


Is either of these cases correct, or is there another way to understand this?

Will the same be true for a machine that does NOT use checksum caching?


Dan






--
Daniel J. Sperka, Ph. D.
UC Davis Center for Neuroscience

------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_______________________________________________
BackupPC-users mailing list
BackupPC-users AT lists.sourceforge DOT net
List:    https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki:    http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/
<Prev in Thread] Current Thread [Next in Thread>