On Fri, 17 Apr 1998 21:23:31 -0700  "Dr. John Refling" wrote:

>> What O/S are you running that prints out this message?

>NetBSD.  Ultrix logs these events through the elcsd daemon, and can be
>read by the uerf program.  NetBSD just does a kernel printf(), which
>can get captured by syslog.  For example:
>
>        Nov  2 14:00:01 vlad /netbsd: CPU memory read ECC error at 0x02b2a254   
>ECC 0x8ca98c00

The  3MAX (5000/200) and 3MAXPLUS have the same memory daughterboard
slots and similar ECC SEC/DED hardware; the only difference I know
of is the ECC registers are at different addresses.
The 5000/240 ECC-error handling code is a copy of the 5000/200 code.

In -current,  they even share the same code.

I don't see how John Refling's problems could be bad memory.  Maybe
it's due to the SCSI problem reported on 5000/200s -- causing the
kernel to fill pages with bad data on a pagein?