confusing/conflicting reports re memory (non)problems

From: Mark Bartelt <mark_at_cita.utoronto.ca>
Date: Wed, 20 Nov 1996 10:54:20 -0500

We have a pair of four-CPU AlphaServer 2100 systems; one
is a 4/200, and the other is a 5/250. Yesterday the two
systems had their memory swapped (the 4/200 had 1 Gbyte
and the 5/250 had 256 Mbyte, and our users decided that
it made more sense for the faster machine to be the one
with the most memory).

After the swap, the 5/250 reported memory errors right
at power-up. It gives me an address, tells me the bank
and board number, and also reports an error type (0) and
an error syndrome (0000348). It tells me there are two
bad pages.

However, if I then do a "test" at the console prompt, it
grinds away doing a memory test, but reports no errors.
Furthermore, when I do an "init" at the console prompt,
it reports "passed" for both memory boards when it gets
to the "memory testing and configuration status" output.

I don't recall problems of this sort on either system when
the 1 Gbyte was in the 4/200, and the 5/250 had 256 Mbyte.
It seems to have started when the 1 Gbyte got installed in
the 5/250.

In case it's relevant, the 5/250 firmware is:

        SRM console V4.2-22
        VMS palcode V1.15-1
        OSF palcode V1.18-11

The system boots without problem, and seems to be running
just fine.

So what exactly is with these reports of memory errors at
power-up time? Should I be worried, or can I ignore them?
And if I can ignore them, why are they even being reported?

(And where exactly are the error types and error syndromes
documented?)

Mark Bartelt 416/978-5619
Canadian Institute for mark_at_cita.utoronto.ca
Theoretical Astrophysics http://www.cita.utoronto.ca/~mark

"Nur eine Waffel taugt!" -- Parsifal, in an Eggo commercial
Received on Wed Nov 20 1996 - 17:21:32 NZDT

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