mjl@emsi.priv.at (Martin J. Laubach) writes:
> |  for a long time (since about end of july) my sparc IPX dies under heavy I/O
> |  load (such as a sup -o). I/O seems blocked and the disk keep trashing.
> |  
> |  Anyone else seen this ?
> 
>   Yes, I've seen this under /i386 too. No idea what triggers it,
> I did a cvs update remotely, and suddenly the machine was dead.
> Coming home, I found it massaging the disk intensily (and it
> must have done that for hours).

I now see intense thrashing on one UFS filesystem when it fills up
past what df reports as 90% (really 80%).  I don't know if this is a
new problem or if there is something special that I just never noticed
about big filesystems.

Filesystem  1K-blocks     Used     Avail Capacity  Mounted on
/dev/wd1h    59135937 46183628   9995512    82%    /w

At a df of 90% a process writing to the FS would hang for 10 seconds
or longer and the average I0 rate as reported by "iostat 1"
approximated an old paper-tty.  Making 10% more free space sped things
up immensely.

-wolfgang
-- 
       Wolfgang Rupprecht <wolfgang+gnus@dailyplanet.wsrcc.com>
		    http://www.wsrcc.com/wolfgang/
Coming soon: GPS mapping tools for Open Systems. http://www.gnomad-mapping.com/