copy corrupting last block of file?

[Tom Horsley]

This looks suspiciously like a software problem of some kind:

I've got a file (a firmware update) I try to copy to an sdcard.

Two different copies now have shown as corrupted at the start
of the final 16K boundary of the file data:

zooty> cmp /zooty/public/S800CUpdate/firmware/firmware.bin
/tmp/tom/bad-linux-firmware/firmware.bin 
/zooty/public/S800CUpdate/firmware/firmware.bin /tmp/tom/bad-linux-firmware/firmware.bin
differ: char 127041537, line 496168

zooty> cmp /zooty/public/S800CUpdate/firmware/firmware.bin
/tmp/tom/firmware/firmware.bin 
/zooty/public/S800CUpdate/firmware/firmware.bin /tmp/tom/firmware/firmware.bin differ:
char 127041537, line 496168

That 127041537 value in hex is 0x7928001.

Worst yet, if I use rsync rather than cp to copy the file, lo and behold the
contents turn up identical!

Is the "cp" command busted? That would be insane.

[Heinz Diehl]

On 22.12.2012, Tom Horsley wrote: 

>  I switched to a different sdcard reader, and suddenly all the
>  errors went away, so (today, anyway) all the errors seem to
>  be correlated with the sdcard reader (which will probably
>  be headed to the recycling pile soon :-). 
Sorry to say, but I had such problems some time ago, and I gave the card
reader which showed a behaviour similar to what you've described in
this thread to a friend which uses Windows 7. He didn't encounter any
corruption, and the card reader works just fine..

[Tim]

Allegedly, on or about 21 December 2012, Tom Horsley sent:
>  I switched to a different sdcard reader, and suddenly all the
>  errors went away, so (today, anyway) all the errors seem to
>  be correlated with the sdcard reader (which will probably
>  be headed to the recycling pile soon :-).  
I dare say that card readers are subjected to quite a bit of physical
stress, and probably static charge problems, too.  And some of the early
memory cards, at least, warned about being careful with the cards to
avoid static electricity.

-- 
[tim@localhost ~]$ uname -rsvp
Linux 3.6.10-2.fc17.x86_64 #1 SMP Tue Dec 11 18:07:34 UTC 2012 x86_64

All mail to my mailbox is automatically deleted, there is no point
trying to privately email me, I will only read messages posted to the
public lists.

[Tom Horsley]

On Thu, 20 Dec 2012 20:47:20 -0500
Tom Horsley wrote:

>  My vacation starts when I get home tomorrow, and I'm already
>  planning a massive set of data gathering tests. Just the
>  sort of thing I love to kill time with on holidays :-). 
Well, I've been playing with it and I'm glad to find that
the correlation with the "cp" command was apparently
random.

I wrote a big test script in perl and did a slew of copies,
and the writes all worked today, but then I started getting
errors on the reads.

I switched to a different sdcard reader, and suddenly all the
errors went away, so (today, anyway) all the errors seem to
be correlated with the sdcard reader (which will probably
be headed to the recycling pile soon :-).

[Tom Horsley]

On Thu, 20 Dec 2012 16:53:43 -0800
Joe Zeff wrote:

>  Try using cp to copy one or more large files from one directory to 
>  another on your hard drive and see if it works.  If possible, try 
>  copying from one partition to another. 
My vacation starts when I get home tomorrow, and I'm already
planning a massive set of data gathering tests. Just the
sort of thing I love to kill time with on holidays :-).

[Joe Zeff]

On 12/20/2012 04:45 PM, Tom Horsley wrote:
>  When I use "cp" though, it fails every time for the same
> half dozen
>  tries. I've even shutdown the system and let it cool off and rebooted
>  and cp still fails and rsync still works :-). 
Try using cp to copy one or more large files from one directory to 
another on your hard drive and see if it works.  If possible, try 
copying from one partition to another.

[Gregory Woodbury]

On 12/20/2012 06:48 PM, Tom Horsley wrote:
>  This looks suspiciously like a software problem of some kind:
>  
>  I've got a file (a firmware update) I try to copy to an sdcard.
>  
>  Two different copies now have shown as corrupted at the start
>  of the final 16K boundary of the file data:
>  
>  zooty> cmp /zooty/public/S800CUpdate/firmware/firmware.bin
> /tmp/tom/bad-linux-firmware/firmware.bin 
>  /zooty/public/S800CUpdate/firmware/firmware.bin /tmp/tom/bad-linux-firmware/firmware.bin
> differ: char 127041537, line 496168
>  
>  zooty> cmp /zooty/public/S800CUpdate/firmware/firmware.bin
> /tmp/tom/firmware/firmware.bin 
>  /zooty/public/S800CUpdate/firmware/firmware.bin /tmp/tom/firmware/firmware.bin differ:
> char 127041537, line 496168
>  
>  That 127041537 value in hex is 0x7928001.
>  
>  Worst yet, if I use rsync rather than cp to copy the file, lo and behold the
>  contents turn up identical!
>  
>  Is the "cp" command busted? That would be insane. 
Yes, it would be insane.
It is much more likely to be a hardware problem.  Try another large file
copy to the sdcard and see if something similar happens.

Then try a different sdcard.

-- 
G.Wolfe Woodbury
FAS: redwolfe (proventesters)

[Tim]

Tom Horsley: 
> > I switched to a different sdcard reader, and suddenly all the
> > errors went away, so (today, anyway) all the errors seem to
> > be correlated with the sdcard reader (which will probably
> > be headed to the recycling pile soon :-). 
Heinz Diehl:
>  Sorry to say, but I had such problems some time ago, and I gave the
> card
>  reader which showed a behaviour similar to what you've described in
>  this thread to a friend which uses Windows 7. He didn't encounter any
>  corruption, and the card reader works just fine.. 
Though that may be due to it being used on a different computer, with a
different power supply, rather than the OS.

-- 
[tim@localhost ~]$ uname -r
2.6.27.25-78.2.56.fc9.i686

Don't send private replies to my address, the mailbox is ignored.  I
read messages from the public lists.

[Tom Horsley]

On Thu, 20 Dec 2012 19:16:59 -0500
G.Wolfe Woodbury wrote:

>  Yes, it would be insane.
>  It is much more likely to be a hardware problem.  Try another large file
>  copy to the sdcard and see if something similar happens.
>  
>  Then try a different sdcard. 
I agree it is insane, but the sdcard works fine when the copy
is done on a Windows box and when the rsync command is used
on the linux box.

I tried about a half dozen times, and those worked every time.

When I use "cp" though, it fails every time for the same half dozen
tries. I've even shutdown the system and let it cool off and rebooted
and cp still fails and rsync still works :-).

[Heinz Diehl]

On 23.12.2012, Tim wrote: 

>  Though that may be due to it being used on a different computer, with
> a
>  different power supply, rather than the OS. 
That's quite possible, of course..