Help with COPY command

#1 08-10-2008, 07:19 PM

I would appreciate a bit of help with the COPY command under
CMD.EXE

Hardware: E8200 core 2 duo CPU on ASUS P5B motherboard
with 2 Seagate 320 SATA II hard drives
Software: Windows XP with 4 FAT32 partitions on each drive

When I copy any very large file (greater than 1.5 GB) from drive 1
to drive 0 and check that the copy is the same, more than 10% of
the time, the files are NOT identical. I used both:

FC D:FOOBAR.GHO E:FOOBAR.GHO

D:MD5 D:FOOBAR.GHO
1st 32 bit result
D:MD5 E:FOOBAR.GHO
2nd 32 bit result is different more than 10% of the time

Note that some of the time, the FC program shows a difference
even though the MD5 program shows that the 32 bit CRC is
identical.

Am I correct to be VERY concerned?

Can anyone suggest where to start? Obviously, it may be the hard
disk drive. Might it also be the software? Could it also be the
motherboard or the memory? Anything else?

Jerome Fine

#2 08-10-2008, 09:08 PM

Why did you not specify the /b option for fc.exe? Are both volumes
NTFS? FAT32? Why not use xcopy to do the copy? Also have you used
filesync or Araxis Merge to do comparisons & copies? I use Total
Commander to copy my Symantec Ghost files. Any chance that not using
the backslash to make the pathname root relative might cause a file
with the same name in a subdirectory to be used?

Use a MD5 check and also a CRC32 and a SHA-nnn as well to validate
signatures. Reboot between the copy and comparisons to ensure the
data is really read from the drive and not remaining in the cache.
Writing your own copy routine will permit you to avoid cache issues by
using opens with caching avoided.

On Sun, 10 Aug 2008 19:19:40 -0400, "Jerome H. Fine"
<spamtrap@crayne.org> wrote:

>I would appreciate a bit of help with the COPY command under
>CMD.EXE
>
>Hardware: E8200 core 2 duo CPU on ASUS P5B motherboard
> with 2 Seagate 320 SATA II hard drives
>Software: Windows XP with 4 FAT32 partitions on each drive
>
>When I copy any very large file (greater than 1.5 GB) from drive 1
>to drive 0 and check that the copy is the same, more than 10% of
>the time, the files are NOT identical. I used both:
>
>FC D:FOOBAR.GHO E:FOOBAR.GHO
>
>D:MD5 D:FOOBAR.GHO
>1st 32 bit result
>D:MD5 E:FOOBAR.GHO
>2nd 32 bit result is different more than 10% of the time
>
>Note that some of the time, the FC program shows a difference
>even though the MD5 program shows that the 32 bit CRC is
>identical.
>
>Am I correct to be VERY concerned?
>
>Can anyone suggest where to start? Obviously, it may be the hard
>disk drive. Might it also be the software? Could it also be the
>motherboard or the memory? Anything else?
>
>Jerome Fine

#3 08-10-2008, 09:19 PM

On Sun, 10 Aug 2008 19:19:40 -0400, "Jerome H. Fine"
<spamtrap@crayne.org> wrote in comp.lang.asm.x86:

> I would appreciate a bit of help with the COPY command under
> CMD.EXE
>
> Hardware: E8200 core 2 duo CPU on ASUS P5B motherboard
> with 2 Seagate 320 SATA II hard drives
> Software: Windows XP with 4 FAT32 partitions on each drive
>
> When I copy any very large file (greater than 1.5 GB) from drive 1
> to drive 0 and check that the copy is the same, more than 10% of
> the time, the files are NOT identical. I used both:
>
> FC D:FOOBAR.GHO E:FOOBAR.GHO
>
> D:MD5 D:FOOBAR.GHO
> 1st 32 bit result
> D:MD5 E:FOOBAR.GHO
> 2nd 32 bit result is different more than 10% of the time
>
> Note that some of the time, the FC program shows a difference
> even though the MD5 program shows that the 32 bit CRC is
> identical.
>
> Am I correct to be VERY concerned?
>
> Can anyone suggest where to start? Obviously, it may be the hard
> disk drive. Might it also be the software? Could it also be the
> motherboard or the memory? Anything else?
>
> Jerome Fine
>
> <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
> <html>
> <head>
> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
> <title></title>
> </head>

[snip]

Hello,

In the future, please set your news software to post text only to
technical usenet groups, instead of text and html both.

There might be a reason to be very concerned, then again there might
not be.

I am assuming that a 1.5 GB file with an extension of .gho is not a
simple ASCII text file, and that might be causing a problem.

Are you using "copy /b" for copying and "fc /b" for comparing? If
these are binary files, you really should be using the "/b" for binary
option for both commands.

If you are not using "/b", does it make a difference when you do?

--
Jack Klein
Home: http://JK-Technology.Com
FAQs for
comp.lang.c http://c-faq.com/
comp.lang.c++ http://www.parashift.com/c++-faq-lite/
alt.comp.lang.learn.c-c++
http://www.club.cc.cmu.edu/~ajo/docs/FAQ-acllc.html

#4 08-10-2008, 10:53 PM

On Sun, 10 Aug 2008 19:19:40 -0400, "Jerome H. Fine"
> I would appreciate a bit of help with the COPY command
> under CMD.EXE

I would use something else, like XCOPY.EXE . COPY may
not have very high performace if it uses MS-DOS-heritege
small buffers.

> Hardware: E8200 core 2 duo CPU on ASUS P5B motherboard
> with 2 Seagate 320 SATA II hard drives
> Software: Windows XP with 4 FAT32 partitions on each drive
>
> When I copy any very large file (greater than 1.5 GB) from drive 1
> to drive 0 and check that the copy is the same, more than 10% of
> the time, the files are NOT identical. I used both:
>
> FC D:FOOBAR.GHO E:FOOBAR.GHO
>
> D:MD5 D:FOOBAR.GHO
> 1st 32 bit result
> D:MD5 E:FOOBAR.GHO
> 2nd 32 bit result is different more than 10% of the time
>
> Note that some of the time, the FC program shows a difference
> even though the MD5 program shows that the 32 bit CRC is
> identical.

Sounds like a disk cache effect.

> Am I correct to be VERY concerned?

Yes.

> Can anyone suggest where to start? Obviously, it may be the
> hard disk drive. Might it also be the software? Could it
> also be the motherboard or the memory? Anything else?

Irregular errors (only 10%) usually is a hardware fault.
I've seen these sorts of infrequent errors on other hard disks.
I check mine with repeated md5's on an unmounted partitition
(with data on it). I've seen these rare errors prom a poor PSU.

-- Robert

#5 08-11-2008, 09:39 PM

>Jack Klein wrote:

>>On Sun, 10 Aug 2008 19:19:40 -0400, "Jerome H. Fine" wrote:
>
>>I would appreciate a bit of help with the COPY command under
>>CMD.EXE
>>When I copy any very large file (greater than 1.5 GB) from drive 1
>>to drive 0 and check that the copy is the same, more than 10% of
>>the time, the files are NOT identical. I used both:
>>FC D:FOOBAR.GHO E:FOOBAR.GHO
>>D:MD5 D:FOOBAR.GHO
>>1st 32 bit result
>>D:MD5 E:FOOBAR.GHO
>>2nd 32 bit result is different more than 10% of the time
>>
>>Note that some of the time, the FC program shows a difference
>>even though the MD5 program shows that the 32 bit CRC is
>>identical.
>>
>Are you using "copy /b" for copying and "fc /b" for comparing? If
>these are binary files, you really should be using the "/b" for binary
>option for both commands.
>
>If you are not using "/b", does it make a difference when you do?
>
>
YES!!!!!!!!!!!!!!!!!!!!!!!

I have been using COPY and FC under Windows 98SE (without the /b) for over
5 years without a problem, so it seemed reasonable to assume (BAD!!!)
that since
the COPY and FC (without the /b) worked some of the time under Windows XP,
it was not the lack of the /b (which I had not remembered to use in any
case) which
was causing the problem

I started testing under Windows XP last month and noted the problems did
occur some of the time, but not always. I use GHOST under DOS to perform
backups and restores, thus the FOOBAR.GHO files which are just under
2 GB in size. After I switch back to Windows XP, I make a second copy
of the image file on drive E: (just in case I lose 2 out of 3 hard
drives - which
did happen once over the 5 years with Windows 98SE - and with hard drives
at the current price, it proves cost effective to take the extra time to
make a
second copy of the image file).

When I included the /b with both commands, all of the problems disappeared.

THANK YOU!!!!!!!!!!!!

Thanks also to dave and Robert for suggesting essentially the same thing.

Sincerely yours,

Jerome Fine

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode