[RFC] Delayed logging

[Dave Chinner]

On Mon, Mar 15, 2010 at 02:08:41PM -0400, Christoph Hellwig wrote:
> On Mon, Mar 15, 2010 at 03:30:00PM +1100, Dave Chinner wrote:
> > 	- occasional recovery failure with recovery reading an inode
> > 	  buffer that does not contain inodes. Cause unknown, tends
> > 	  to be reproduced by xfsqa test 121 semi-reliably. Needs
> > 	  further analysis and fixing. May already be fixed with a
> > 	  recent fix to commit record synchronisation.
> 
> I just saw a corrupted inode buffer during log recovery in 087, even
> without the delaylog mount option.  It's however not easily
> reproducible.

Yeah, that's the sort of thing I'm seeing 087, 110 (IIRC) and 121
are the tests that I've been seeing occasional, random failures like
this. I hadn't noticed it without the delaylog option, so I'll see
if I can get that to occur as well...

Cheers,

Dave.
-- 
Dave Chinner
david at fromorbit.com