Re: ZFS kernel panic

On Wed, Aug 29, 2007 at 12:27:57PM +0200, Dag-Erling Smørgrav wrote:
> Max Laier <max_at_love2party.net> writes:
> > This is complete nonsense!  As you pointed out earlier zfs doesn't
> > know anything about the nature of the error.  There is only one
> > sensible way to deal with a disk error - unless it is transient - and
> > that is stopping all (write) access to the drive.  As you can't easily
> > move a mounted drive with opened files into read-only mode, a panic is
> > the only way to make sure.
> 
> Actually, remounting the disk read-only upon encountering a write error
> is standard behaviour in Linux.

And it didn't even tell, but that's another story.
Anyway - I prefer the panic unless a better solution is available or
possible, like fsync/write error or even better just switch to another
block.
In case of redundancy however it may be best if the disk is taken
completely out of use, since many kind write errors produce highe
latencies, which slows down the system.

-- 
B.Walter                http://www.bwct.de      http://www.fizon.de
bernd_at_bwct.de           info_at_bwct.de            support_at_fizon.de