Re: 2.4.21-pre3 - problems with ext3 (long)

["Stephen C. Tweedie" <sct@redhat.com>]

Hi,

On Tue, 2003-01-21 at 00:25, Lukasz Trabinski wrote:

> system boot  2.4.20:
> Dec 15 15:27:01 oceanic kernel: Assertion failure in journal_start_Rsmp_c2be780a
> () at transaction.c:248: "handle->h_transaction->t_journal == journal

> With earlier kernels 2.4.X (for example 2.4.20-rc2) this machine has much 
> longer uptime.

OK, which was the last one which ran stable for you?  I note that you've
got a failure marked "2.4.20" in the log.

> Dec 15 15:27:01 oceanic kernel: Kernel panic: EXT3-fs panic (device sd(8,23)): load_block_bitmap: block_group >= groups_count - block_group = 524287, groups_count = 2126

Do you have the backtrace for that?  I can't see any way that particular
error can happen unless the kernel's memory is getting corrupt, or
there's a corrupt superblock coming in from the disk.  

Also, are you sure you've been ksymoops'ing these from the right
System.map files?  The traces really don't make a lot of sense.

Finally,

> By the way, last crash was with messages:
> Jan 19 11:50:20 oceanic kernel: kernel BUG at highmem.c:159!
> Jan 19 11:50:20 oceanic kernel: invalid operand: 0000
> Jan 19 11:50:20 oceanic kernel: CPU:    1

If that happens again, serial console is the best way of getting the
full oops.  How much memory does your system have?  Have you ever seen
this error before?  

Cheers,
 Stephen