[TriLUG] diagnostic/stress test software

[Robert Dale]

On 6/7/07, Jason Watts <jsnonzzr at gmail.com> wrote:
> im not thinking the system is that dead. its just a fault its running across
> once in a while. the system is rebooted 3 times a week, almost always
> successfully, in the past month however, it has hung twice. which is twice
> too often, we are just trying to figure out why.

Step 1.  Immediately replace machine.  Even a cheap COTS box will have
better uptime for the next couple of years than what you have now.
When/if you fix the ailing machine, hey, what organization couldn't
use an extra box?

Step 2.  Have someone stand in front of the machine while it's
booting.  Look at the boot message it's hanging on.  Replace part.
Reboot.  Repeat.

Stress testing software almost never finds the problem even when the
item being stress tested is the problem.  More often than not, the
problem lies in something that can not be stress tested.

YMMV

-- 
Robert Dale