SUMMARY addemdum : CPU-EXCEPTION

From: William H. Magill <magill_at_isc.upenn.edu>
Date: Fri, 25 Oct 1996 11:11:53 -0400

> PS: for the automatic analysation of the binary.errlog
> I wrote an script piping it to the analyser.. The
> result of analysing the old logs from march was:
> (as about 100.000 time the same later on:
> REPLACE THE MEMORY-BOARD). So the machine checker
> from digital was not able to find the broken
> CPU-boards...
>
With the release of DU 4.0 Digital finally (almost) admitted that
uerf didn't work anymore - - ESPECIALLY for "cup exceptions" .. ie memory
errors on 2100s.

DecEvent "analyzer mode" is the only thing which will get the appropriate
error messages and information. Analyze WILL WORK WITHOUT a keypack.
"Uerf -z" does NOT generate output which produces meaningful information.
(I have no idea why.)

DecEvent is included in tht 4.0 distribution - but you still have to get
a key pack from your field service folks to get the automatic functions.
Supposedly, it will automatically tell "dial-a-prayer" that you have a
problem and initate a service call for you. (Obviously, you need a
maintenance contract for this service, that's why the keypak is needed.)

If you can get to your maintenance branch's software support person
(not the hardware folks, and not the software folks), they know all about
the situation. However, finding that person can be fun because neither
"dial-a-prayer" support or the normal field service people pay any
attention to the "binary.errlog" problem.

T.T.F.N.
William H. Magill Senior Systems Administrator
Information Services and Computing (ISC) University of Pennsylvania
Internet: magill_at_isc.upenn.edu magill_at_acm.org
          magill_at_upenn.edu http://pobox.upenn.edu/~magill/
Received on Fri Oct 25 1996 - 17:41:49 NZDT

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:47 NZDT