D875PBZ w/ SATA RAID 1

I just encountered a curious coincidence.

Two mostly identical machines. Both just a bit over a year old. Both using
the D875PBZ board. Both using the Northwood 3.0 processor. Both in the same
environment. Both with the latest firmware. Blah, blah, blah, blah ...

One is using a pair of identical Hitachi SATA drives and the other is using
a pair of identical Samsung SATA drives. Both (fortunately) configured as
RAID 1.

About a month ago one of the Hitachi drives failed ... SATA Port 0.

Today, one of the Samsung drives failed on the other machine ... SATA Port
0.

Total coincidence? Or is there something about Port 0?

"MyndPhlyp" wrote in message
nk.net...
I just encountered a curious coincidence.

Two mostly identical machines. Both just a bit over a year old. Both using
the D875PBZ board. Both using the Northwood 3.0 processor. Both in the same
environment. Both with the latest firmware. Blah, blah, blah, blah ...

One is using a pair of identical Hitachi SATA drives and the other is using
a pair of identical Samsung SATA drives. Both (fortunately) configured as
RAID 1.

About a month ago one of the Hitachi drives failed ... SATA Port 0.

Today, one of the Samsung drives failed on the other machine ... SATA Port
0.

Total coincidence? Or is there something about Port 0?


1. If both systems are mostly identical, both systems may have inadequate
cooling for the disk drives. If the SATA 0 drive in both machines is located
directly above the SATA 1 drive, the heat from the lower drive may be killing
the upper drive. Consider investing in some drive coolers and extra case fans.

2. The other possibility is that the RAID 1 firmware does not divide read
requests evenly between the two drives, putting more stress on the SATA 0
drive in each machine.

Bob

"Bob Bailin" wrote in message
m...

1. If both systems are mostly identical, both systems may have inadequate
cooling for the disk drives. If the SATA 0 drive in both machines is
located
directly above the SATA 1 drive, the heat from the lower drive may be
killing
the upper drive. Consider investing in some drive coolers and extra case
fans.

2. The other possibility is that the RAID 1 firmware does not divide read
requests evenly between the two drives, putting more stress on the SATA 0
drive in each machine.

Thanx for the feedback.

I think it is safe to rule out #1. Inside case temperatures are quite low on
both systems. Typical sensor readings place the processor at 49C and the
other zones at 37C - comfortably in the green. Drive proximity can be ruled
out because Port 0 is mounted below Port 1.

I can't rule out #2 though. One thing I do not know for sure is if the
failure was on the drive's controller or if it was something more physical.
Head crash is probably out - these machines don't get moved or bumped. That
doesn't rule out a slow spin rate, jammed head mechanism or similar
problems. With all our technological advances the HD industry (and
Microsoft) can't seem to find a way to fully report the failure. I suspect
though that at least one of the failures is due to a failed HD controller.

"MyndPhlyp" wrote in message
k.net...

"Bob Bailin" wrote in message
m...

1. If both systems are mostly identical, both systems may have inadequate
cooling for the disk drives. If the SATA 0 drive in both machines is
located
directly above the SATA 1 drive, the heat from the lower drive may be
killing
the upper drive. Consider investing in some drive coolers and extra case
fans.

2. The other possibility is that the RAID 1 firmware does not divide read
requests evenly between the two drives, putting more stress on the SATA 0
drive in each machine.

Thanx for the feedback.

I think it is safe to rule out #1. Inside case temperatures are quite low on
both systems. Typical sensor readings place the processor at 49C and the
other zones at 37C - comfortably in the green. Drive proximity can be ruled
out because Port 0 is mounted below Port 1.

I can't rule out #2 though. One thing I do not know for sure is if the
failure was on the drive's controller or if it was something more physical.
Head crash is probably out - these machines don't get moved or bumped. That
doesn't rule out a slow spin rate, jammed head mechanism or similar
problems. With all our technological advances the HD industry (and
Microsoft) can't seem to find a way to fully report the failure. I suspect
though that at least one of the failures is due to a failed HD controller.

Install the failed drive into another (non-RAID) system as a 2nd hard
disk. Then run one of the many SMART analysis programs available on
the web, or download the manufacturer's drive analysis program and
run it against the bad drive. This assumes that the drive is well-enough
to allow the BIOS to recognize it at system boot.

Bob

"Bob Bailin" wrote in message
t...

Install the failed drive into another (non-RAID) system as a 2nd hard
disk. Then run one of the many SMART analysis programs available on
the web, or download the manufacturer's drive analysis program and
run it against the bad drive. This assumes that the drive is well-enough
to allow the BIOS to recognize it at system boot.

Thanx again. I'll have to do some searching for an available SATA system.
For the time being the drives get to sit on the spare parts pile (along with
many ISA boards and 2400 baud modems).