On Thu, Feb 15, 2007 at 12:09:54PM +0100, Volkmar Seifert wrote:
> Hm...I have tried that yesterday, and parity rebuild - just like
> after accidently having typed "reboot", took about an hour for an
> 80GB two-disk RAID-1. Okay, an hour is quite some time, but it's
> definitely not forever... ;-)

Well, an 80GB RAID 1 is tiny. My server has two 250GB RAID 1 volumes.
With the performance provided by your machine and disks it would take
at least six hours to rewrite the parity.

Ongoing access to the disks of course slows down the parity rewrite
a lot. A friend's machine usuallay needs three hours for two 120GB
hard disks mainly because it is a mail server which results in constant
disk access.

> ... just like after accidently having typed "reboot"

That's not really the worst case. Your machine "only" had to rebuild
the parity. If it also performs a filesystem check while rebuilding
the RAID parity things become much worse.

> And the machine does not belong to the SotA "real fast wonderboxes",
> see for yourself below...

CPU doesn't matter much in this case. I/O bandwidth from and to the disks
in the bottleneck.

> Info about my disks:
> 
[...]
> wd0: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 5 (Ultra/100)
> wd0(viaide0:0:0): using PIO mode 4, Ultra-DMA mode 5 (Ultra/100) (using DMA)
[...]
> wd1: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 5 (Ultra/100)
> wd1(viaide0:1:0): using PIO mode 4, Ultra-DMA mode 5 (Ultra/100) (using DMA)

The IDE hostadapter and your disks support Ultra/100 DMA which means that
synchronisation will runs at a reasonable speed (unless your disks run
at only 5400RPM). Modern disks will be a faster (maybe 50%) but have a
significantly higher capacity (upto 800%).

	Kind regards

-- 
Matthias Scheler                                  http://zhadum.org.uk/