Hi,

At 14:33 Uhr +0100 30.11.2001, Cameron Kaiser wrote:

>I think I've pushed my NetBSD 1.4.2 machine too far -- I've started
>running a rather HD-heavy process and this will bring down the system after
>extended hard drive activity.

I wouldn't bet my last dollar on the crash being caused by the system.

If it doesn't freeze up completely, you may just try to wait a minute or so
and then look if you get any messages on the console about command-timeouts
on the SCSI-bus. The kernel may eventually reset the whole bus and you may
be able to get away without rebooting.

>I've tried putting the HD into its own powered enclosure just in case
>it was a power drain, and that doesn't make a difference.

Make sure you use high quality (twisted-pair as the standard intended)
SCSI-cables for the external box, and have proper termination. For example,
Performa 450s & 475s run fine with the last drive in the external SCSI-box
terminated, but a 6100 will die right away unless you really terminate the
end of the bus.

By the way, if you just pull the internal HD, then this end of the bus is
everything else but well-terminated.

>Is 1.5 more stable in this regard?

I have 1.5.1 on a P475 and - as a stress test - pushed the system load up
to well above twenty. The thing was swapping so heavily that it would
respond to moving the mouse every other second. It stayed in this unusable
state for about half an hour and then everything calmed down again as the
processes finished. No problem.

Regards,

	Alex