large file copy results in corrupted copies

[Guest]

I have a problem when copying large files (50MB+):
Both the DOS copy /B and the drag/drop or copy/paste methods in explorer
result always in a slightly different (corrupt) copy without generating any
error message. Aparently everything is fine but when checking on binary
level, there are differences. Depending on the application this result in
unreadble files.
The number of changed bytes is proportional to the length of the file.

Example: when I copy a file of about 300MB it results in a compare as follows:

D:\>fc /B aa.ape bb.ape
Comparing file aa.ape and BB.APE
0054204C: 7F 31
0054204D: D8 30
03DD478E: B8 F8
0B16184E: 7F AD
0B16184F: C4 0B
0BBC620C: CA 0D
0BBC620D: D7 F0
0BBC620E: C3 AD
0BBC620F: 7E 0B
0C6A00CC: 12 31
0C6A00CD: CA 30
0DE24C8C: E7 31
0DE24C8D: 5B 30
0DE24C8E: D2 71
0DE24C8F: 78 68
0E3D05CC: 1B 31
0E3D05CD: D9 30
0E3D05CE: 2E 71
0E3D05CF: 67 68
0FD1244C: 30 31
0FD1244D: DB 30

There are some patterns which repeat themselves. The sequence 30 31 hex
often appear in the adjacent positions xxxxxxxC xxxxxxxD, while 68 appears in
position xxxxxxxF.

I have 2 Harddisks but it does not matter which is source or destination, it
can also coincide. (I first noted it when nero in verification phase found
differences between source and masterized copy)

The errors do not occure when i use DOS copy in Windows started in save mode!

Therefore I exclude HW problems. (Anyway I already changed IDE cables and
run memtest). Also, updating the driver VIA 4in1 did not help.

please, help!

 

[Pegasus \(MVP\)]

I have a problem when copying large files (50MB+):
Both the DOS copy /B and the drag/drop or copy/paste methods in explorer
result always in a slightly different (corrupt) copy without generating any
error message. Aparently everything is fine but when checking on binary
level, there are differences. Depending on the application this result in
unreadble files.
The number of changed bytes is proportional to the length of the file.

Example: when I copy a file of about 300MB it results in a compare as follows:
There are some patterns which repeat themselves. The sequence 30 31 hex
often appear in the adjacent positions xxxxxxxC xxxxxxxD, while 68 appears in
position xxxxxxxF.

I have 2 Harddisks but it does not matter which is source or destination, it
can also coincide. (I first noted it when nero in verification phase found
differences between source and masterized copy)

The errors do not occure when i use DOS copy in Windows started in save mode!

Therefore I exclude HW problems. (Anyway I already changed IDE cables and
run memtest). Also, updating the driver VIA 4in1 did not help.

please, help!

I suggest you run the following batch file from a Command Prompt:

@echo off
xcopy /v /b "SourceFolder\SourceFile" "DestFolder\DestFile"
echo Errorlevel for xcopy was %ErrorLevel%
pause
fc /b "SourceFolder\SourceFile" "DestFolder\DestFile" | more
pause

xxcopy /v /b "SourceFolder\SourceFile" "DestFolder\DestFile"
echo Errorlevel for xxcopy was %ErrorLevel%
pause
fc /b "SourceFolder\SourceFile" "DestFolder\DestFile" | more

While this might not solve your problem, it could shed some
light on your problem. xxcopy.exe is downloadable from various
sites.

 

[Torgeir Bakken \(MVP\)]

I suggest you run the following batch file from a Command Prompt:

@echo off
xcopy /v /b "SourceFolder\SourceFile" "DestFolder\DestFile"
echo Errorlevel for xcopy was %ErrorLevel%
pause
fc /b "SourceFolder\SourceFile" "DestFolder\DestFile" | more
pause

xxcopy /v /b "SourceFolder\SourceFile" "DestFolder\DestFile"
echo Errorlevel for xxcopy was %ErrorLevel%
pause
fc /b "SourceFolder\SourceFile" "DestFolder\DestFile" | more

While this might not solve your problem, it could shed some
light on your problem. xxcopy.exe is downloadable from various
sites.

Hi,

An alternative to xxcopy is Robocopy.

Robocopy.exe is in the free Windows Server 2003 Resource Kit:

http://www.microsoft.com/downloads/...69-57ff-4ae7-96ee-b18c4790cffd&DisplayLang=en

(The kit will install on WinXP or later, the individual files can
then be copied to another computer if needed)


See rktools.chm and robocopy.doc for documentation for Robocopy.exe.

 

[billious]

I have a problem when copying large files (50MB+):
Both the DOS copy /B and the drag/drop or copy/paste methods in explorer
result always in a slightly different (corrupt) copy without generating
any
error message. Aparently everything is fine but when checking on binary
level, there are differences. Depending on the application this result in
unreadble files.
The number of changed bytes is proportional to the length of the file.

Example: when I copy a file of about 300MB it results in a compare as
follows:
There are some patterns which repeat themselves. The sequence 30 31 hex
often appear in the adjacent positions xxxxxxxC xxxxxxxD, while 68 appears
in
position xxxxxxxF.

I have 2 Harddisks but it does not matter which is source or destination,
it
can also coincide. (I first noted it when nero in verification phase found
differences between source and masterized copy)

The errors do not occure when i use DOS copy in Windows started in save
mode!

Therefore I exclude HW problems. (Anyway I already changed IDE cables and
run memtest). Also, updating the driver VIA 4in1 did not help.

please, help!

Not sure whether the D: drive you indicate is local or network.

I had problems of a similar nature about 12 years ago with Netware "lite,"
which would randomly change bytes in large copied files. The destination
file would occasionally have bytes changed - always in the same files, same
byte-offset, and the changed-to pattern was always one of a small set of
rogue bytes.

As far as I'm aware, since Novell hasn't got back to me about it as
promised, they're still working on it....

Is there anything of this nature in your setup?

HTH

....Bill

 

[David Candy]

This is a problem, while rare, that is usually caused by network card drivers. Maybe Novell drivers were like Windows drivers.

 

[Guest]

This is a problem, while rare, that is usually caused by network card drivers. Maybe Novell drivers were like Windows drivers.

No sorry, it is disk to disk, it happens even when creating a copy on the
same drive - any idea?

 

[R. McCarty]

Go to Device Manager, Disk Drives category, Double-Click your drive
From the details box click the Policies (TAB). Uncheck "Enable Write
Cache", Apply & then reboot and retry your copy operations. Also,
could you post which Vendor/Model (& Interface type) drive you're
having issues with.

 

[Kerry Brown]

No sorry, it is disk to disk, it happens even when creating a copy on the
same drive - any idea?

Check your ram www.memtest.org Make sure you have the latest chipset drivers
for your motherboard. Is one particular hard drive always involved in the
copy? Is the computer overclocked? I have even seen a bad power supply cause
this problem. It can be tough to find.

Kerry

 

[Guest]

Sorry for post this one 3 times - I am using the web interface and had some
difficulties finding my posts. Will not happen again.

I suggest you run the following batch file from a Command Prompt:

@echo off
xcopy /v /b "SourceFolder\SourceFile" "DestFolder\DestFile"
echo Errorlevel for xcopy was %ErrorLevel%
pause
fc /b "SourceFolder\SourceFile" "DestFolder\DestFile" | more
pause

I tried this(without /v /b which were not recognized switches), but the
result is exactly the same: no error and corrupt copy, same result as normal
copy.

xxcopy /v /b "SourceFolder\SourceFile" "DestFolder\DestFile"
echo Errorlevel for xxcopy was %ErrorLevel%
pause
fc /b "SourceFolder\SourceFile" "DestFolder\DestFile" | more

idem (v/ b/ not recognized, second parameter must be directory): errorcode 0
and corrupt copy.

While this might not solve your problem, it could shed some
light on your problem. xxcopy.exe is downloadable from various
sites.

I hope that the fact these tools aslo fail will lead you to some conclusion.

 

[Guest]

Go to Device Manager, Disk Drives category, Double-Click your drive
From the details box click the Policies (TAB). Uncheck "Enable Write
Cache", Apply & then reboot and retry your copy operations.

I tried it, the copy runs now much slower (110 sec against 40 sec with cache).
The output file is still corrupt, seems in less positions: I made 4 tests
and with cache I got 20, 32, 14, 15 differences while without cache I got 8,
22, 5, 6. So even if this sample is not very significative - the average is
around the half.

Also,
could you post which Vendor/Model (& Interface type) drive you're
having issues with.

I have 2 Maxtor disks, one 60 and one 200 MB (new). Both behave the same!
Both attached to same IDE channel.

 

[Guest]

Check your ram www.memtest.org

Done, see initial post

Make sure you have the latest chipset drivers
for your motherboard.

Done, see initial post

Is one particular hard drive always involved in the
copy?

Source is one of my 2 HD, 60MB and 200MB Maxtor, both on same IDE channel.
Cable checked.

Is the computer overclocked?

How can I check this?

I have even seen a bad power supply cause
this problem.

How can I chewck this?

It can be tough to find.

Don't encourage me so much

Thanks all for your interest!
Berni

 

[R. McCarty]

Thanks for testing and the results. If you're still inclined to test more
I would switch your IDE Primary and Secondary connectors on the
motherboard (Test using the other Channel). You should also reset
the Write cache on both drives. I would carefully check the BIOS
settings for the IDE channels and while in Windows check settings
on the Primary, Secondary and make sure the drives are running at
UDMA Mode 5 or 6. This seems to be a motherboard issue, not a
drive related one. If they are new drives, you might consider getting
a SATA controller card and two IDE-to-SATA conversion adapters.
On machines that don't have SATA but new large/fast drives I've
starting configuring them that way, so each has it's own dedicated
channel to work from. A SATA card is fairly cheap at ~$30.00 &
the conversion modules are even less at around ~$17.00. Anyway,
this seems like a motherboard or chipset issue.

 

[Kerry Brown]

Done, see initial post


Done, see initial post


Source is one of my 2 HD, 60MB and 200MB Maxtor, both on same IDE channel.
Cable checked.


How can I check this?


How can I chewck this?

Don't encourage me so much

Thanks all for your interest!

Sorry, if I was repeating what you've already done. How long did you let
memtest run? Which memtest program did you use? I would try R. McCarty's
suggestion of trying the drives on a different channel. It does sound more
and more like a motherboard/controller problem. You can measure voltages on
a power supply but the easiest way to test is to swap in a known good one.
Unless you have a very accurate meter and test under a load the voltages
don't tell you much. The only sure way to tell if your system is overclocked
is to physically identify the CPU and make sure it is running at the right
frequency. If you didn't overclock the system in all likelihood it is not
overclocked.

Kerry

 

[Guest]

Thanks to all who try to help me out!

I will try the IDE channel test when I'm home this evening ...
Just one doubt: How can it be a HW problem, when the copy works perfectly in
case I start windows in temporary/save mode. I thought that winXP drives the
HW somewhat different?

Kerry: I used memtest86, at least one full turn around through all steps. If
I remember well it lasted about 15 minutes. The copy problem occures always!

Cheers
Berni

 

[Kerry Brown]

Thanks to all who try to help me out!

I will try the IDE channel test when I'm home this evening ...
Just one doubt: How can it be a HW problem, when the copy works perfectly
in
case I start windows in temporary/save mode. I thought that winXP drives
the
HW somewhat different?

Kerry: I used memtest86, at least one full turn around through all steps.
If
I remember well it lasted about 15 minutes. The copy problem occures
always!

You should let memtest86 run for several hours at least. It probably isn't
the problem but that possible cause should be eliminated. Windows XP demands
a lot more from the hardware than DOS or safe mode. In safe mode the hard
drives are accessed much differently than regular mode.

Kerry

 

[Guest]

You should let memtest86 run for several hours at least. It probably isn't
the problem but that possible cause should be eliminated. Windows XP demands
a lot more from the hardware than DOS or safe mode. In safe mode the hard
drives are accessed much differently than regular mode.

Kerry

Hi guys, thanks a lot for your help.
After executing the tests that R. McCarty and Kerry suggested I conclude
that it must be HW (motherboard, chipset) - I will carry my PC to a service
shop and I will see if I go for the PCI SATA Controller + adapters or for a
new motherboard ...
I will post if they find another reason -
thanks again
Berni

 

[Kerry Brown]

Hi guys, thanks a lot for your help.
After executing the tests that R. McCarty and Kerry suggested I conclude
that it must be HW (motherboard, chipset) - I will carry my PC to a
service
shop and I will see if I go for the PCI SATA Controller + adapters or for
a
new motherboard ...
I will post if they find another reason -
thanks again

Thanks for the update. Let us know what they find.

Kerry