On Mon, Sep 11, 2006 at 07:14:45PM +0200, Dominik Schips wrote:
> 
> For your information.
> The holding disk is on another harddisk as the data harddisk to back up.
> So the holding disk dumps are from a RAID5 (S-ATA) were the backup files
> are to another harddisk.
>
> I have to check them with bonnie for their speed.
> 
> 
> I tested the backup without holding disk since a few days but the speed is 
> slower than with the holding disk. I only save the time for the dumps.
> 
> I have to check about I/O problems.
> The RAID5 device and the tape device use their own SCSI controller but I 
> should check the settings again.

A "SCSI" controller for a "RAID5 (S-ATA)"?  There is your bottleneck :))

Your data without and with holding disk:

                        without       with
> Dump Time (hrs:min)     8:19        7:46
> Tape Time (hrs:min)     8:19        3:31
> Original Size (meg)   375747.5    370976.6

Similar amounts of data.
Similar dump times, slightly favoring with holding disk.
But taping already completed dumps from the holding disk
is only 1/2 the dump time.

I think it would be worth the experiment to increase maxdumps
to two.  There could be such increased disk activity to swamp
their i/o capability, but seems to me worth a shot.

If that doesn't help, and particularly if you see substantial
bonnie++ differences, perhaps you could try to swap the data
to the raid and the holding space to the other disk.

If the raid gives superior read results, perhaps move the data
there and backup without a holding disk.

-- 
Jon H. LaBadie                  jon AT jgcomp DOT com
 JG Computing
 4455 Province Line Road        (609) 252-0159
 Princeton, NJ  08540-4322      (609) 683-7220 (fax)