Two Maxtor Hard Drives Failed in 8 Hours on GA-7DXR

This is a strange one:

This system has four physical drives: 2 in RAID 0, two more on each primary
master IDE.
The machine is a large server tower with SIX cooling fans. Drives are
running quite cool to the touch.
Electric power is conditioned by a TrippLite 2400 conditioner. The whole
network is powered on a large APC UPS. Individual power strips with MOV
protectors form the last line of protection where each PC is plugged in.

Chronological history of this system:

Christmas Day 2002: One of the two 80gig Maxtor drives on the RAID 0 array
fails. I buy a replacement and send the failed drive for repair/replacement.
Replacement drive arrives and is in my closet.

One week ago, I upgraded the 20 gig pri master drive that holds the OS,
graphics and business data. I installed a new Maxtor 6Y120P0 120 gig drive
(8Meg cache) last week. I moved the old 20 gig drive to the second IDE
master position and moved the ancient WD 9 gig drive that was there to
another PC.

This morning, I arrived to a BSOD: KERNAL_DATA_PAGE_ERROR
When I reset the machine, the BIOS could not find the 6Y120P0 on IDE0
master. Furthermore, it could not find one of the two 80 gig drives on the
RAID controller!

I deleted the array and rebuilt, but it rebuilt with only one drive, as the
other drive vanished from the BIOS Fastrack configurator. The machine was
unbootable.


I shut the whole thing off and had lunch.
Came back an hour later, and started it up. The BIOS saw the 6Y120P0, but
Win 2K BSOD'ed during the logo screen. I reset it again, this time the BIOS
could not find the 6Y120P0. Reset, reset, reset... finally it reappeared on
the BIOS screen, the system booted, and it was like nothing had happened.
Then I shut down, and restarted and went into the Fastrack utility. Both
drives on the RAID controller reappeared, so I deleted the 1-drive array and
rebuilt the array with both drives.

The rest of the day consisted of frantically backing up all data from the
RAID and from the primary IDE.

Oddly, the system continued to run.

But I went out and bought a new 160GB Western Digital Caviar drive with 8MB
cache.

I cloned the data from the failing drive to the new drive (had to use the
command switch for Ghost that ignores drive errors) and then I ran MaxBlast
to analyze the failing drive. It generated the error code dea49db1 and
reported that the drive is failing.

Now about the RAID, it is still working. I'm waiting for it to generate an
error again so that I can locate the failing drive in the array.

First question:
What's going on here? I find it odd that two drives could fail in the same
overnight period.
Could the RAID failure be a bogus error triggered by the C drive failure?

Second question:
The new WD drive has drastically asymmetrical read/write performance.

The Maxtor drive that failed, when it was first installed, provided a
60MB/60MB read/write performance on large (256MB) files using nBench and
confirmed using RexTest.

The Western Digital drive has a little over 60MB write speed, but only 34MB
read speed. Confirmed with nBench and Rextest multiple times. With small
files, the gap widens: nBench with a 10MB file shows 59MB write and only
22MB read on the WD Caviar.

So here's the question: Is there a compatibility problem with the WD drives
and the GA-7DXR?

I can find no explanation for low read speed, but fast write speed.

--
Take care,

Mark & Mary Ann Weiss

VIDEO PRODUCTION . FILM SCANNING . AUDIO RESTORATION
Hear my Kurzweil Creations at: http://www.dv-clips.com/theater.htm
Business sites at:
www.dv-clips.com
www.mwcomms.com
www.adventuresinanimemusic.com
-

tell us about the powersupply in the huge tower case with all the hdd's and
fans in it...

NuTs

"Mark & Mary Ann Weiss" wrote in message
news
This is a strange one:

This system has four physical drives: 2 in RAID 0, two more on each
primary
master IDE.
The machine is a large server tower with SIX cooling fans. Drives are
running quite cool to the touch.
Electric power is conditioned by a TrippLite 2400 conditioner. The whole
network is powered on a large APC UPS. Individual power strips with MOV
protectors form the last line of protection where each PC is plugged in.

Chronological history of this system:

Christmas Day 2002: One of the two 80gig Maxtor drives on the RAID 0 array
fails. I buy a replacement and send the failed drive for
repair/replacement.
Replacement drive arrives and is in my closet.

One week ago, I upgraded the 20 gig pri master drive that holds the OS,
graphics and business data. I installed a new Maxtor 6Y120P0 120 gig drive
(8Meg cache) last week. I moved the old 20 gig drive to the second IDE
master position and moved the ancient WD 9 gig drive that was there to
another PC.

This morning, I arrived to a BSOD: KERNAL_DATA_PAGE_ERROR
When I reset the machine, the BIOS could not find the 6Y120P0 on IDE0
master. Furthermore, it could not find one of the two 80 gig drives on the
RAID controller!

I deleted the array and rebuilt, but it rebuilt with only one drive, as
the
other drive vanished from the BIOS Fastrack configurator. The machine was
unbootable.


I shut the whole thing off and had lunch.
Came back an hour later, and started it up. The BIOS saw the 6Y120P0, but
Win 2K BSOD'ed during the logo screen. I reset it again, this time the
BIOS
could not find the 6Y120P0. Reset, reset, reset... finally it reappeared
on
the BIOS screen, the system booted, and it was like nothing had happened.
Then I shut down, and restarted and went into the Fastrack utility. Both
drives on the RAID controller reappeared, so I deleted the 1-drive array
and
rebuilt the array with both drives.

The rest of the day consisted of frantically backing up all data from the
RAID and from the primary IDE.

Oddly, the system continued to run.

But I went out and bought a new 160GB Western Digital Caviar drive with
8MB
cache.

I cloned the data from the failing drive to the new drive (had to use the
command switch for Ghost that ignores drive errors) and then I ran
MaxBlast
to analyze the failing drive. It generated the error code dea49db1 and
reported that the drive is failing.

Now about the RAID, it is still working. I'm waiting for it to generate an
error again so that I can locate the failing drive in the array.

First question:
What's going on here? I find it odd that two drives could fail in the same
overnight period.
Could the RAID failure be a bogus error triggered by the C drive failure?

Second question:
The new WD drive has drastically asymmetrical read/write performance.

The Maxtor drive that failed, when it was first installed, provided a
60MB/60MB read/write performance on large (256MB) files using nBench and
confirmed using RexTest.

The Western Digital drive has a little over 60MB write speed, but only
34MB
read speed. Confirmed with nBench and Rextest multiple times. With small
files, the gap widens: nBench with a 10MB file shows 59MB write and only
22MB read on the WD Caviar.

So here's the question: Is there a compatibility problem with the WD
drives
and the GA-7DXR?

I can find no explanation for low read speed, but fast write speed.

--
Take care,

Mark & Mary Ann Weiss

VIDEO PRODUCTION . FILM SCANNING . AUDIO RESTORATION
Hear my Kurzweil Creations at: http://www.dv-clips.com/theater.htm
Business sites at:
www.dv-clips.com
www.mwcomms.com
www.adventuresinanimemusic.com
-

"NuT CrAcKeR" wrote in message
...
tell us about the powersupply in the huge tower case with all the hdd's
and
fans in it...

NuTs


The +12 and +5 rails are 4% low, but that's about it. Other than that,
running okay.

There's extensive line conditioning and surge protection for the ac power
that feeds this and several PCs on the network. We have a MOV surge
protector on the line feeding our enterprise grade true sinewave UPS, which
feeds a TrippLite 2400 line conditioner, which feeds several power strips,
each with their own MOV protectors in common mode and across the line
(delta) configuration, where each PC plugs into it's own surge protected
strip, isolated from the surges from other equipment, such as monitor
degaussing coils.


Just as curious is the poor performance on reads of the new Western Digital
drive, WD1600JBRTL. 60 MB write and only 34MB read on large files. The
Maxtor provided a symmetrical read/write of 60/60mB/S. This sounds like a
conflict between the motherboard drivers and the hard drive itself...
wondering if there is any data on drive brand/model compatibility with the
7DXR?



--

Take care,

Mark & Mary Ann Weiss

Business sites at:
www.dv-clips.com NOW ONLINE!
www.mwcomms.com
www.adventuresinanimemusic.com NEW Streaming Archives!
-

Have you checked /replaced the H/D s cables and checked that there is good
contact of the
plugs into HD sockets....I've had some very big & odd come-&-go problems
with poor connections.
Luv Trimble

Mark & Mary,

Inline:

"Mark & Mary Ann Weiss" wrote in message
news
This is a strange one:

This system has four physical drives: 2 in RAID 0, two more on each
primary
master IDE.

RAID 0 on a server is inadvisable - I see in your signature that you work on
Video, so this may be logical in your case. You will have a more resilient
system if you use plain discs. Given the obvious investment you have made in
the configuration, if you really do need performance then consider SCSI RAID
1, RAID 10, or SATA IDE RAID 1 using WD Raptors. Consider hotswap drives -
not elcheapo disc drive boxes.

The machine is a large server tower with SIX cooling fans. Drives are
running quite cool to the touch.
Electric power is conditioned by a TrippLite 2400 conditioner. The whole
network is powered on a large APC UPS. Individual power strips with MOV
protectors form the last line of protection where each PC is plugged in.


Do you have a severe power problem? If an APC UPS is adequate to protect
your power then I take it you have had a registered electrician in to
recommend this configuration? The TrippLite conditioner appears from web
searching to be a consumer appliance. Why do you have two? Unless I purchase
such things myself, I always look at them suspiciously. All it would take
would be a wobbly joint in one of the units to undo all the benefits you
intend. If the TrippLite's are closer to the Server power than the UPS (IE
in between), then I would either remove them or move them upstream. All your
problems may be due to power - the PSU within the Server case could be the
problem also.

Chronological history of this system:

Christmas Day 2002: One of the two 80gig Maxtor drives on the RAID 0 array
fails. I buy a replacement and send the failed drive for
repair/replacement.
Replacement drive arrives and is in my closet.

One week ago, I upgraded the 20 gig pri master drive that holds the OS,
graphics and business data. I installed a new Maxtor 6Y120P0 120 gig drive
(8Meg cache) last week. I moved the old 20 gig drive to the second IDE
master position and moved the ancient WD 9 gig drive that was there to
another PC.

Again, consider RAID 1, 10, 5 or 50 for - particularly OS and Data. If you
can afford to buy a replacement drive and leave it sitting on a shelf, you
are better off having a RAID 1 controller with the drive installed and
providing data / OS resilience. With SCSI and some up market IDE Raid you
can configure Hot Spares.


This morning, I arrived to a BSOD: KERNAL_DATA_PAGE_ERROR
When I reset the machine, the BIOS could not find the 6Y120P0 on IDE0
master. Furthermore, it could not find one of the two 80 gig drives on the
RAID controller!


From MSDN:
See:
http://www.microsoft.com/resources/d...l/reskit/en-us
/Default.asp?url=/resources/documentation/Windows/XP/all/reskit/en-us/prmd_s
tp_fvlq.asp

1. Stop 0x7A can be caused by bad sectors in the virtual memory paging file,
disk controller error, virus infection, or memory hardware problems. ....
2. Another cause of Stop 0x7A messages is defective, malfunctioning, or
failed memory hardware,....
3. Check the hardware manufacturer's Web site for updates to disk adapter
firmware or drivers that improve compatibility...
4. The problem might also be due to cracks, scratched traces, or defective
components on the motherboard.

Given the other problems, I would suspect your motherboard has a fault. It
could be coincidental that you have this problem now, your BIOS could have
become corrupt. This is much more likely if you have a flakey power
conditioner.

I deleted the array and rebuilt, but it rebuilt with only one drive, as
the
other drive vanished from the BIOS Fastrack configurator.

I have seen this happen: Drive power connectors failing. Cables shot, or
again it could be the mobo...

The machine was
unbootable.


I shut the whole thing off and had lunch.
Came back an hour later, and started it up. The BIOS saw the 6Y120P0, but
Win 2K BSOD'ed during the logo screen. I reset it again, this time the
BIOS
could not find the 6Y120P0. Reset, reset, reset... finally it reappeared
on
the BIOS screen, the system booted, and it was like nothing had happened.
Then I shut down, and restarted and went into the Fastrack utility. Both
drives on the RAID controller reappeared, so I deleted the 1-drive array
and
rebuilt the array with both drives.

The rest of the day consisted of frantically backing up all data from the
RAID and from the primary IDE.

Oddly, the system continued to run.

But I went out and bought a new 160GB Western Digital Caviar drive with
8MB
cache.

I cloned the data from the failing drive to the new drive (had to use the
command switch for Ghost that ignores drive errors) and then I ran
MaxBlast
to analyze the failing drive. It generated the error code dea49db1 and
reported that the drive is failing.

Now about the RAID, it is still working. I'm waiting for it to generate an
error again so that I can locate the failing drive in the array.

First question:
What's going on here? I find it odd that two drives could fail in the same
overnight period.

That can happen.

Could the RAID failure be a bogus error triggered by the C drive failure?

Second question:
The new WD drive has drastically asymmetrical read/write performance.


When you get all else fixed, revisit this.

The Maxtor drive that failed, when it was first installed, provided a
60MB/60MB read/write performance on large (256MB) files using nBench and
confirmed using RexTest.

The Western Digital drive has a little over 60MB write speed, but only
34MB
read speed. Confirmed with nBench and Rextest multiple times. With small
files, the gap widens: nBench with a 10MB file shows 59MB write and only
22MB read on the WD Caviar.

So here's the question: Is there a compatibility problem with the WD
drives
and the GA-7DXR?


Doubt it. The only issue with WD is they often prefer Cable Select rather
then Master / Slave config. Read speeds are usually always higher than
writes...

I can find no explanation for low read speed, but fast write speed.

What else is on the same cable? If it is a substantially older / slower
drive this can have an effect.

____

What would I do? This... its a bit long, but will get you there.

Check anti static procedures first. IE Anti static wrist strap, no synthetic
clothing or carpet, no vacuum cleaners, care in handling of components and
so on...

Step 1: Objective - prove mobo, bios, memory & CPU stability.

Strip the system down, consider removing the mobo onto an insulated desktop
(wood, newspaper is usually OK - anti static mat can be a problem since they
are conductive) since you will be making a lot of h/w config changes &
having the system in the case will invite errors / take more time. If you
decide to leave the mobo in the case be patient.

If you disconnect anything, *always* remove the associated cable from the
mobo and power from the device (EG disc drive), remove Secondary drives
first and re-attach last as you rebuild during the test sequence since a
secondary by itself is often not happy. Check every Master / Slave / Cable
Select jumper before you connect the IO cable - read the label on the disc
drive and if WD drive use cable select if either they advise or it does not
work. If you use Cable select on 1 IDE drive, use it regardless on both.
Make sure Secondary drives only ever are attached at the secondary position
on a cable(middle).

Configure the system with floppy, CPU, memory, graphics only and run
memtest86 for 5 full extended runs. If there are memory errors, address this
issue 1st. If this is OK, then ...

Step 2: Prove the HDD's, cables, controllers.

install an HDD on Primary IDE. Do a disk check - full surface scan. Install
an OS, run the OS, install and run Prime95 with MBM5. Check CPU temps. If
all is OK, then slowly add 1 disc at a time, disc checking if you can afford
to do so otherwise run the manufacturers test utility. Continue rebuilding
by adding other discs / controllers 1 step at a time. Reboot often to look
for HDD reporting inconsistencies as you have mentioned. Test
everything........

Step 3: Server PSU. You haven't told us yet the Make, Model / Rating of this
unit. For the number of drives you have it should be 'oversized' and a
quality unit. If this is playing up then you *may* have problems reappear as
you reinstall things or sometime later. So, as you go through the above
process, you need to keep an eye out for conclusive evidence that changes
have 'Fixed' things otherwise you may be no better off. IE: Change, test,
undo change, test & Confirm the change had the benefit & the benefit is
needed. Of course, don't change Willy Nilly. If you have another PSU at hand
of equal or better specs then obviously try swapping PSU's at some point.

--
Take care,

Mark & Mary Ann Weiss

VIDEO PRODUCTION . FILM SCANNING . AUDIO RESTORATION
Hear my Kurzweil Creations at: http://www.dv-clips.com/theater.htm
Business sites at:
www.dv-clips.com
www.mwcomms.com
www.adventuresinanimemusic.com
-

the cleanest power on the plant is only as good as the weak ass powersupply
that uses it.

what

powersupply

are

you

using

????


NuTs

"Mark & Mary Ann Weiss" wrote in message
link.net...

"NuT CrAcKeR" wrote in message
...
tell us about the powersupply in the huge tower case with all the hdd's
and
fans in it...

NuTs


The +12 and +5 rails are 4% low, but that's about it. Other than that,
running okay.

There's extensive line conditioning and surge protection for the ac power
that feeds this and several PCs on the network. We have a MOV surge
protector on the line feeding our enterprise grade true sinewave UPS,
which
feeds a TrippLite 2400 line conditioner, which feeds several power strips,
each with their own MOV protectors in common mode and across the line
(delta) configuration, where each PC plugs into it's own surge protected
strip, isolated from the surges from other equipment, such as monitor
degaussing coils.


Just as curious is the poor performance on reads of the new Western
Digital
drive, WD1600JBRTL. 60 MB write and only 34MB read on large files. The
Maxtor provided a symmetrical read/write of 60/60mB/S. This sounds like a
conflict between the motherboard drivers and the hard drive itself...
wondering if there is any data on drive brand/model compatibility with the
7DXR?



--

Take care,

Mark & Mary Ann Weiss

Business sites at:
www.dv-clips.com NOW ONLINE!
www.mwcomms.com
www.adventuresinanimemusic.com NEW Streaming Archives!
-

Hey... i own stock in elcheapo drives, inc.

dont be telling them that !!!!!

; )

NuTs

"Tim" wrote in message ...
Mark & Mary,

Inline:

"Mark & Mary Ann Weiss" wrote in message
news
This is a strange one:

This system has four physical drives: 2 in RAID 0, two more on each
primary
master IDE.

RAID 0 on a server is inadvisable - I see in your signature that you work
on
Video, so this may be logical in your case. You will have a more resilient
system if you use plain discs. Given the obvious investment you have made
in
the configuration, if you really do need performance then consider SCSI
RAID
1, RAID 10, or SATA IDE RAID 1 using WD Raptors. Consider hotswap drives -
not elcheapo disc drive boxes.

The machine is a large server tower with SIX cooling fans. Drives are
running quite cool to the touch.
Electric power is conditioned by a TrippLite 2400 conditioner. The whole
network is powered on a large APC UPS. Individual power strips with MOV
protectors form the last line of protection where each PC is plugged in.


Do you have a severe power problem? If an APC UPS is adequate to protect
your power then I take it you have had a registered electrician in to
recommend this configuration? The TrippLite conditioner appears from web
searching to be a consumer appliance. Why do you have two? Unless I
purchase
such things myself, I always look at them suspiciously. All it would take
would be a wobbly joint in one of the units to undo all the benefits you
intend. If the TrippLite's are closer to the Server power than the UPS (IE
in between), then I would either remove them or move them upstream. All
your
problems may be due to power - the PSU within the Server case could be the
problem also.

Chronological history of this system:

Christmas Day 2002: One of the two 80gig Maxtor drives on the RAID 0
array
fails. I buy a replacement and send the failed drive for
repair/replacement.
Replacement drive arrives and is in my closet.

One week ago, I upgraded the 20 gig pri master drive that holds the OS,
graphics and business data. I installed a new Maxtor 6Y120P0 120 gig
drive
(8Meg cache) last week. I moved the old 20 gig drive to the second IDE
master position and moved the ancient WD 9 gig drive that was there to
another PC.

Again, consider RAID 1, 10, 5 or 50 for - particularly OS and Data. If you
can afford to buy a replacement drive and leave it sitting on a shelf, you
are better off having a RAID 1 controller with the drive installed and
providing data / OS resilience. With SCSI and some up market IDE Raid you
can configure Hot Spares.


This morning, I arrived to a BSOD: KERNAL_DATA_PAGE_ERROR
When I reset the machine, the BIOS could not find the 6Y120P0 on IDE0
master. Furthermore, it could not find one of the two 80 gig drives on
the
RAID controller!


From MSDN:
See:

http://www.microsoft.com/resources/d...l/reskit/en-us

/Default.asp?url=/resources/documentation/Windows/XP/all/reskit/en-us/prmd_s
tp_fvlq.asp

1. Stop 0x7A can be caused by bad sectors in the virtual memory paging
file,
disk controller error, virus infection, or memory hardware problems. ....
2. Another cause of Stop 0x7A messages is defective, malfunctioning, or
failed memory hardware,....
3. Check the hardware manufacturer's Web site for updates to disk adapter
firmware or drivers that improve compatibility...
4. The problem might also be due to cracks, scratched traces, or defective
components on the motherboard.

Given the other problems, I would suspect your motherboard has a fault. It
could be coincidental that you have this problem now, your BIOS could have
become corrupt. This is much more likely if you have a flakey power
conditioner.

I deleted the array and rebuilt, but it rebuilt with only one drive, as
the
other drive vanished from the BIOS Fastrack configurator.

I have seen this happen: Drive power connectors failing. Cables shot, or
again it could be the mobo...

The machine was
unbootable.


I shut the whole thing off and had lunch.
Came back an hour later, and started it up. The BIOS saw the 6Y120P0,
but
Win 2K BSOD'ed during the logo screen. I reset it again, this time the
BIOS
could not find the 6Y120P0. Reset, reset, reset... finally it reappeared
on
the BIOS screen, the system booted, and it was like nothing had
happened.
Then I shut down, and restarted and went into the Fastrack utility. Both
drives on the RAID controller reappeared, so I deleted the 1-drive array
and
rebuilt the array with both drives.

The rest of the day consisted of frantically backing up all data from
the
RAID and from the primary IDE.

Oddly, the system continued to run.

But I went out and bought a new 160GB Western Digital Caviar drive with
8MB
cache.

I cloned the data from the failing drive to the new drive (had to use
the
command switch for Ghost that ignores drive errors) and then I ran
MaxBlast
to analyze the failing drive. It generated the error code dea49db1 and
reported that the drive is failing.

Now about the RAID, it is still working. I'm waiting for it to generate
an
error again so that I can locate the failing drive in the array.

First question:
What's going on here? I find it odd that two drives could fail in the
same
overnight period.

That can happen.

Could the RAID failure be a bogus error triggered by the C drive
failure?

Second question:
The new WD drive has drastically asymmetrical read/write performance.


When you get all else fixed, revisit this.

The Maxtor drive that failed, when it was first installed, provided a
60MB/60MB read/write performance on large (256MB) files using nBench and
confirmed using RexTest.

The Western Digital drive has a little over 60MB write speed, but only
34MB
read speed. Confirmed with nBench and Rextest multiple times. With small
files, the gap widens: nBench with a 10MB file shows 59MB write and only
22MB read on the WD Caviar.

So here's the question: Is there a compatibility problem with the WD
drives
and the GA-7DXR?


Doubt it. The only issue with WD is they often prefer Cable Select rather
then Master / Slave config. Read speeds are usually always higher than
writes...

I can find no explanation for low read speed, but fast write speed.

What else is on the same cable? If it is a substantially older / slower
drive this can have an effect.

____

What would I do? This... its a bit long, but will get you there.

Check anti static procedures first. IE Anti static wrist strap, no
synthetic
clothing or carpet, no vacuum cleaners, care in handling of components and
so on...

Step 1: Objective - prove mobo, bios, memory & CPU stability.

Strip the system down, consider removing the mobo onto an insulated
desktop
(wood, newspaper is usually OK - anti static mat can be a problem since
they
are conductive) since you will be making a lot of h/w config changes &
having the system in the case will invite errors / take more time. If you
decide to leave the mobo in the case be patient.

If you disconnect anything, *always* remove the associated cable from the
mobo and power from the device (EG disc drive), remove Secondary drives
first and re-attach last as you rebuild during the test sequence since a
secondary by itself is often not happy. Check every Master / Slave / Cable
Select jumper before you connect the IO cable - read the label on the disc
drive and if WD drive use cable select if either they advise or it does
not
work. If you use Cable select on 1 IDE drive, use it regardless on both.
Make sure Secondary drives only ever are attached at the secondary
position
on a cable(middle).

Configure the system with floppy, CPU, memory, graphics only and run
memtest86 for 5 full extended runs. If there are memory errors, address
this
issue 1st. If this is OK, then ...

Step 2: Prove the HDD's, cables, controllers.

install an HDD on Primary IDE. Do a disk check - full surface scan.
Install
an OS, run the OS, install and run Prime95 with MBM5. Check CPU temps. If
all is OK, then slowly add 1 disc at a time, disc checking if you can
afford
to do so otherwise run the manufacturers test utility. Continue rebuilding
by adding other discs / controllers 1 step at a time. Reboot often to look
for HDD reporting inconsistencies as you have mentioned. Test
everything........

Step 3: Server PSU. You haven't told us yet the Make, Model / Rating of
this
unit. For the number of drives you have it should be 'oversized' and a
quality unit. If this is playing up then you *may* have problems reappear
as
you reinstall things or sometime later. So, as you go through the above
process, you need to keep an eye out for conclusive evidence that changes
have 'Fixed' things otherwise you may be no better off. IE: Change, test,
undo change, test & Confirm the change had the benefit & the benefit is
needed. Of course, don't change Willy Nilly. If you have another PSU at
hand
of equal or better specs then obviously try swapping PSU's at some point.

--
Take care,

Mark & Mary Ann Weiss

VIDEO PRODUCTION . FILM SCANNING . AUDIO RESTORATION
Hear my Kurzweil Creations at: http://www.dv-clips.com/theater.htm
Business sites at:
www.dv-clips.com
www.mwcomms.com
www.adventuresinanimemusic.com
-

"Mark & Mary Ann Weiss" wrote in message hlink.net...
This is a strange one:

Oddly, the system continued to run.

First question:
What's going on here? I find it odd that two drives could fail in the same
overnight period.
Could the RAID failure be a bogus error triggered by the C drive failure?

First check IDE cables. Other is that your PSU maybe failing. Your
motherboard controller maybe failing also.


Second question:
The new WD drive has drastically asymmetrical read/write performance.

The Maxtor drive that failed, when it was first installed, provided a
60MB/60MB read/write performance on large (256MB) files using nBench and
confirmed using RexTest.

The Western Digital drive has a little over 60MB write speed, but only 34MB
read speed. Confirmed with nBench and Rextest multiple times. With small
files, the gap widens: nBench with a 10MB file shows 59MB write and only
22MB read on the WD Caviar.


I can find no explanation for low read speed, but fast write speed.

It depends on the drive.Writes are fast cause data are written first
electronically to drives cache, then magnetically to disk platter.
Small files are written instaneously in cache, while during reading
more time is needed in order to read them from the platter and move
data back to cache.
It depends on the file system also. Did you have the same (NTFS or
FAT32) on Maxtor also?
Now as to why maxtor had the same read/write perfomance, it may
cooperate better with your mobo controller. Benchmarks are also
tricked by the way some disks operate their cache and count only the
time needed to move data from disks cache to main memory, and not the
total time which include the time needed to read magnetically the data
from the platter.

This system has four physical drives: 2 in RAID 0, two more on each primary
master IDE.

[snip]

This morning, I arrived to a BSOD: KERNAL_DATA_PAGE_ERROR
When I reset the machine, the BIOS could not find the 6Y120P0 on IDE0
master. Furthermore, it could not find one of the two 80 gig drives on the
RAID controller!


Possibly relevant, maybe not: I once had a (fairly cheap) case with a
7200 RPM SCSI drive in it, which started failing intermittently (and
was occasionally undetectable to the BIOS) when I added a second 7200
RPM drive from a different manufacturer.

Thinking that the first drive was toast, I downloaded the
manufacturer's RMA testing program -- it established that the original
drive "had been dropped". I thought this was kinda weird (because I
know that it hadn't), so I disconnected the new drive, and magically,
the "broken" drive started working fine again.

I played with positioning the drives inside my case, and discovered
that the new drive was actually setting up vibrations (probably at a
fundamental frequency) that caused the original drive to fail. Moving
the two to different ends of the case fixed the problem, and years
later, that second drive is still working great.

Just a thought.
-kris