[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: 2.4.18 ext3 problems



Hi,

On Fri, Nov 08, 2002 at 12:16:38PM +0000, glen wrote:
 
> Gettings lots of nasty ext3 errors, as follows:
> 
> attempt to access beyond end of device
> 03:06: rw=0, want=1989391900, limit=55167178

That's a data corruption.  It usually turns out to be hardware
problems in my experience, but you could also try a current Marcelo or
-ac kernel, as there are some fixes in there which might be relevant.

Basically, there's a metadata block on disk which tells ext3 where other
blocks are, and that metadata has been corrupted for reasons unknown.
It usually turns out to be something like IDE write caching problems
or bad memory.  But because there's important metadata which is
garbled, ext3 is being told to look in All The Wrong Places for other
parts of the file, resulting in all the other IO failures caused by
illegal blocks getting referenced.

> attempt to access beyond end of device
> 03:06: rw=0, want=274736132, limit=55167178
> EXT3-fs error (device ide0(3,6)): ext3_free_branches: Read failure,
> inode=10092558, block=-1541928704
> Assertion failure in do_get_write_access() at transaction.c:708:
> "handle->h_buffer_credits > 0"

I'll look into that --- even in the presence of disk corruption, ext3
shouldn't panic.

Cheers,
 Stephen





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]