On Thu, Aug 31 2000 at 13:34:20 +0100, Tim Walls wrote:
> My dinky 1.4.2 Sparc Classic's hard disk recently decided to spin
> down entirely without warning, accompanied by a plethora of
> unexpected disconnect errors on the console.  This was in the
> middle of a build of naffing Mozilla, for what it's worth (causing
> me more grief than I care to think of, that damn program is...)
> 
> Uptime up 'til then was about 50 odd days, so I'm assuming hardware
> failure, but does this sound like any known software bug?
NetBSD never sends out STOP UNIT commands. (I wish it would, so I could
spin down those harddisc I rarely access with software instead of mechanical
switches...) - so there's no way NetBSD could accidently spin down a
harddisk. (AFAIK)

> If it's not a known software bug, I'm looking to change the disk
> (an IBM 6GB SCSI) - I managed to get a tape backup of everything
> last night, thank God (time to buy my own tape drive methinks.)
There's not much room in a sparc classic and only very little airflow around
the harddisk - I bet heat was the problem.

I think some IBM harddiscs have a feature that they simply switch off when they
think they got too hot. At least I noticed that with a 6GB IDE disc mounted
in a Acorn RISC PC (Not much airflow in a RISC PC either):
You hear a "clonk" and then it's dead.  This happened a few times around
two years ago - the harddisc still runs fine (with better cooling now.)

Because of that experience most harddisks I run 24/7 have additional cooling
fans now :-)

-- 
Dies ist Thilos Unix Signature! Viel Spass damit.