Re: [BUG?] :: "Value too large for defined data type"

[David Ford]

You need to update your various tools and libraries.  You are being 
bitten by the 32bit bug where as you need your tools to now call 64bit 
functions..i.e stat64(), open64(), ...etc.

God should know these things.

David

God wrote:

>Hello,
>
>In short, I may have done a bad thing, this I know but .....
>
>I noticed that when I tried to backup a box across the LAN via dump to a
>localy mounted drive ...
>
><mental picture>
>Box1 has:
>hda and hdb, hdb has a fat32 fs (don't ask)
>
>Box2 has 
>/mnt/backup, mounted to box2:/mnt/hdb1.  
>
></mental picture>
>
>.. dump crashed with:
>
>  DUMP: 78.80% done at 814 kB/s, finished in 0:10
>  DUMP: error reading command pipe in master: No such file or directory
>  DUMP: The ENTIRE dump is aborted.
>
>
>Trying to check where dump should have written the file on box2, I get:
>
># ls -al
>/bin/ls: test: Value too large for defined data type
>total 2097216
>drwxrwxrwx   2 root     root        32768 Aug 14 23:27 ./
>drwxrwxrwx  50 root     root        32768 Aug 14 21:02 ../
>-rwxrwxrwx   1 root     root     2147483647 Aug 14 22:11
>box1.071401.dump*
>#
>
>
>Ok so I lied, that is a paste from after my little expriment, but you can
>see the file size.  2.1G, out of a 3G drive.
>
>
>What caused the error and my emailing the list, is I created a file using
>dd (dd if=/dev/zero of=test bs=4098k count=1000), but I cannot access that
>file at all.  The only thing that seems to not care about it is "echo" :
>
># echo *
>box1.071401.dump test
>#
>
>I tried to remove the file using rm, but I get the same error as ls:
>
># rm test
>rm: cannot remove `test': Value too large for defined data type
># 
>
>The file is taking up ~4G of space (which is reflected correctly with df,
>du will not read the file).  
>
>It was a stupid thing to do, but could there be a better way for the OS to
>handle this and is there any way I can remove the file?  It's on  a 60
>Gig drive, so formatting would not be an option right about now ...
>
>The box in question (box2), is running Slack with a 2.4.3 kernel:
>
># uname -a
>Linux box2 2.4.3 #7 SMP Sun Apr 29 13:42:05 EDT 2001 i686 unknown
>#
>
>
>Any thoughts?  
>
>
>
>
>-
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to [e-mail]
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>Please read the FAQ at  http://www.tux.org/lkml/
>