> I'm wondering about my perception of sluggish performace with
> "current".

I'd agree with this. Again, just a gut feeling, and not a gripe, but
I think it's to do with vm paging. I had a lot of this on a 4MB
machine, and so I put another 4MB in, and it's lessened the occurences
of this, but (I think) it still happens.
 
> no doubt I'll get flamed for this, but I had to ask ;-)

The only other thing that I've got to add is that recently I've been
having a lot of problems when I reboot after shutting down with as
clean file systems as I can manage.  This happens about 60% of the
time with any reboots.  The / partition is usually OK, but fsck
decides that there are a lot of duplicate inodes on /dev/wd0e (my
/usr) partition.  It'll then unlink all the (supposed) duplicates and
place them in the lost+found directory.

Does anyone know of anything that's been changed that might cause
dirty inodes not to be written out, or clean inodes to be written,
or anything that might be causing a block to be written to the wrong
offset on the partition?
I'm talking about update, sync, fsync, wd driver (or does this
happen on SCSI too?)

Has anyone else seen this, or is it time to call for the men in white
coats?

Alistair

------------------------------------------------------------------------------