Apologies for the late summary. Things never let up around here...
My original problem was that a 4/490 with 224-MB of RAM would
run okay for a while, then lock up tight, requiring an L1-A. I only
got one response before my last attempt at installing the board,
but some in-house people gave me some info...
To get the thing to work, I reduced MAXUSERS to 64 and installed the
kernelmap jumbo patch (for want of a better name -- #100330-06). I
don't know which one actually made it work, but my guess, and that
of several others, is that it was the patch. I may try pushing
maxusers back up sometime when I'm feeling lucky...
Thanks to the following people (apologies to anyone I forgot):
Amir Ilbeig (email@example.com)
Glenn Satchell (glenn%ups.uucp@fourx.Aus.Sun.COM)
Hal Stern (stern@ki.Japan.Sun.COM)
Nelson Fernandez (firstname.lastname@example.org)
There was one more person who replied early, whose e-mail I can't
find. I wrote back to him as soon as I got his e-mail, with thanks...
Leslie Dreyer Kalra
AT&T Bell Laboratories
----- Original Message -----
>From att!ra.mcs.anl.gov!sun-managers-relay Tue Apr 13 22:56:54 1993
Original-From: mhcnet!lbd (Leslie B. Dreyer Kalra )
Subject: memory board causes hangs
Hi, all --
The System: Sun 4/490, SunOS 4.1.1, 2 32-MB memory boards, 128 MAXUSERS,
8 IPI 1-GB disks, installed patch list available if necessary.
The Attempt: To install a Dataram 128-MB memory board, to work with
the 32-MB boards, not replace them (224 MB total). We've
tried twice, with two different boards.
The Problem: The memory board passes diagnostics with no problem (several
passes). However, several hours later, the machine locks up
so tight only an L1-A can get through. The first time this
happened, it was almost immediate, but the second time it
didn't happen until 5 hours later (in the middle of the night).
Dataram has been very helpful (sending us a replacement board overnight),
but it's looking like a software problem. I'm about to install the kernelmap
jumbo patch (#100330). I have rev 5 and rev 6 is on its way.
Also, another group in my department had some memory problems earlier this
year and found out that the more memory you have, the lower MAXUSERS has to
be. I'm thinking about changing it, but I'm not sure how low to go. Rumor
has it that Hal Stern has a chart listing how low MAXUSERS should be for
xxx amount of memory. I don't have Hal's e-mail address, but I know he reads
The Question: Will the kernelmap patch help? We've used it with some success
on hanging SPARC2's, but never on anything else. Also, how low
should MAXUSERS be on a 4/490 with 224 MB of RAM? I don't want
to go too low. Is there anything else we should tweak to get
this memory to work?
As always, thanks...
Leslie Dreyer Kalra
AT&T Bell Labs
----- End Included Message -----
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:07:47 CDT