UNIX robustness

[Greg Lehey]

On Wednesday, 27 January 1999 at  5:23:58 -0800, David C. Jenner wrote:
> "Steven M. Schultz" wrote:
>>
>>         At least it's a different thread... ;-)
>>
>>         I had a UNIX system in my house that ran (and was working hard
>>         the whole time) for over 2 years before pity was taken on it
>>         and a beefier replacement system installed.  The original system
>>         was a 386/33 serving as a rather busy secondary nameserver and was
>>         never touched/rebooted/powercycled for 2+ years.  As time went on the
>>         load increased and it became apparent the system was overloaded (and
>>         paging excessively) so a P5/90 with more memory was installed.  A
>>         couple months after that the disk died ;-)
>
> I had (still have) a PDP-11/23+ that ran (still runs) 2.9BSD (with an
> honest-to-goodness AT&T binary license) that was my UUCP link to the
> net from 1989-1993.  At one stretch, it ran almost two years before a
> power failure finally got it.
>
> It was unusual to go that long without a power failure, so the limiting
> factor was clearly the power and not the hardware or software.  The
> system has a watchdog timer, so when there was a problem the system would
> usually automatically reboot, unless there was a serious hardware problem.
> Reboots were fairly rare and usually a power problem, so it would be hard
> for me to pinpoint the cause of non-power-instigated reboots.

I think you'd have to consider a reboot to be a failure.  Steveen
could say better than I, but I'd expect 2.9BSD to have told you why it
died.  2.11BSD certainly does.

Greg
--
See complete headers for address, home page and phone numbers
finger grog at lemis.com for PGP public key

Received: (from major at localhost)
	by minnie.cs.adfa.edu.au (8.9.1/8.9.1) id LAA05599
	for pups-liszt; Thu, 28 Jan 1999 11:10:03 +1100 (EST)