OS: windows server 2003 SE 32 bit, latest service pack and updates.
I got a promise fasttrack 4310 controller with two 2TB hitachi drives.
Previously, I had two 1TB harddisk in a striping volume and everything
was working fast and smooth.
I thought that promise would be able to setup a 4TB volume using those
new drives, but now it seems that this card isn't capable of doing this.
I think the reason is that the controller supports LBA up to 32 bits.
The drives have 512 byte sectors so 2^32 x 512 = 2TB max.
These 2TB is the maximum for any logical volume, independent of
the number of drives and the raid mode used to create the volume.
Since I didn't want to buy another controller, I just set it to jbod.
On a promise card, jbod means: "just a buch of discs and nothing more".
So it really is a SEPERATE bunch of drives. No spanning.
I converted the two drives to 'dynamic' using the windows disk
manager and created a software raid volume in striping mode.
Copying files to this volume through a network connection takes a lot
more time than before. I haven't measured the exact time but my feeling
says that copy operations take at least 4 times longer than before.
The network isn't the bottleneck. The network usage is almost half
of what it was before with the faster 2x1TB array. So the network
isn't maxed out.
The cpu usage shows about 12 to 15%. Never 100.
The drives can transfer data a lot faster because they have been doing
that before. So why not now?
Nothing bad shows up in any of the windows log files nor in the promise
fasttrack manager.
So nothing seems to be maxed out and still it is terribly slow.
Why?
regards,
Marcel
__________ Information from ESET NOD32 Antivirus, version of virus signature database 4678 (20091211) __________
The message was checked by ESET NOD32 Antivirus.