Ext3 users - Jun 2002 - Unexpected dirty buffer encountered at do_get_write_access:598

Hello,

I am not subscribed to this mailing list. My apologies if non-members 
are not allowed to post in this forum.

I have a Linux-2.4.19pre10aa4 computer and I have seen an error message 
"Unexpected dirty buffer encountered at do_get_write_access:598 (03:05 
blocknr 0)" appears in the log.

With a little bit of research I found that this message is printed from 
fs/jbd/journal.c, and understood that it comes out of EXT3 fs. I am 
afraid I don't have enough knowledge of C and Linux Kernel/EXT3 to debug 
this message.

I have seen Andrea's change log for 2.4.19pre10aa3, and found that there 
were few updates to EXT3 (he says it is from 2.4.19pre10jam2).

The one and only File system in my computer is /, and it is an EXT3 fs 
(mounted with defaults options). Is this error message is an indication 
of any serier nature? Do you want me to try an another kernel such as 
main line, -ac etc to see if I can reproduce it? (to best of my 
knowledge I didn't do anything special to cause this message)

For the record it is a,
Intel P-II 333 MHz,
128 MB RAM,
256 MB SWAP (though the swap usage generally doesn't exceed more that 
10-20 MB)
RedHat 7.3 / glibc-2.2.5 / gcc-2.96-110
3 GB EXT3 root file system, mounted with 'defaults' option
Regular KDE 3.x Work station.
 
Thanks,
Hari,
harisri@telstra.com or harisri@bigpond.com (both are same)

Hi,

On Fri, Jun 21, 2002 at 06:31:15PM +1000, harisri wrote:
 
> I have a Linux-2.4.19pre10aa4 computer and I have seen an error message 
> "Unexpected dirty buffer encountered at do_get_write_access:598 (03:05
> blocknr 0)" appears in the log.
 
> The one and only File system in my computer is /, and it is an EXT3 fs 
> (mounted with defaults options). Is this error message is an indication 
> of any serier nature? 
Usually not, but it can be (particularly in data=journal mode), which
you're not using.  It can also arise if you are running something like
"dump" or "tune2fs" on a mounted filesystem --- it should be
benign in
those cases as long as it's a recent kernel.

The current ext3 cvs contains one fix for the known instance where
this can happen unexpectedly (the one that can corrupt data=journal
systems).

Cheers, 
 Stephen