SUMMARY: Ultra5 Sunvts gives disturbing error without apparent problems

From: Fergus Wilde <fwilde_at_chethams.org.uk>
Date: Wed Apr 09 2003 - 05:50:18 EDT
Thanks to all those who wrote with helpful suggestions, Justin Stringfellow at 
Sun cracked it by notifying me that there was a known issue with VTS 
generating memory failures on Solaris 8 with certain patchlevels:

"When running SunVTS 4.1, 5.1 with Solaris 8 and kernel update patch 
108528-16, you will see error messages of: 

FATAL mem: "Unable to read pointer to memlist structure."                             

 Problem Solution                                                                                                                                 

Problem is due to the fact that the symbol "max_ce_err" was deleted from the 
kernel with the 108528-16 patch. 

Workaround: 

      Install patch 113689 from http://sunsolve.sun.com 
      Run with kernel patch 108528-15 
      Do not run pmemtest with 108528-16 or later."

I solved the problem by 1) updating to SUNWvts 5.1 and its associated packages
		2) Installing patch 1113689-10

This summary was a little delayed while I registered with sunsolve to get this 
non-public patch.  I suppose if you don't have a support contract you would 
be obliged either not to run pmemtest (still leaves another memory test 
available) or backout to a a kernel patchlevel < 108528-16

Thanks also to Joe Fletcher for reminding me that you need to do cat 
/etc/release in order to get the version number, e.g:

$ cat /etc/release
Solaris 8 10/01 s28s_u6wos_08a SPARC
           Copyright 2001 Sun Microsystems, Inc.  All Rights Reserved.
                           Assembled 12 September 2001

and to David M. Dennis for helpful input about RAM. Please forgive me if I've 
missed anyone out, I've been collecting emails on a succession of different 
machines.  My original query:

In the spirit of hope triumphing over experience, I just wanted to ask if 
anyone can see a reason not to call the Sun Doctor (the machine is on Silver 
Service) after a stop a few seconds into a run of SunVtS with this error: 

04/03/03 15:36:05 server SunVTS4.5: VTSID 8004 pmemtest.
  FATAL mem: "Unable to read pointer to memlist structure."

uname -a# SunOS server 5.8 Generic_108528-19 sun4u sparc SUNW,Ultra-5_10
SunVTS version 4.5.

Sun's own docs do seem to tell you to call them out, but I was just hoping 
someone knows a reason why the error might be in ... er ... error.

Machine has up-to-date OBP, and is giving no signs of trouble, logs are void 
of doom-laden messages.  However, since it's done very little work recently 
except fit its own recommended patch set, am I inviting trouble by doing 
nothing if I should need it soon?


Fergus



-- 
Fergus Wilde
Chetham's Library
Long Millgate
Manchester
M3  1SB

Tel: +44 161 834 7961
Fax: +44 161 839 5797

http://www.chethams.org.uk
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
Received on Wed Apr 9 06:53:30 2003

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:43:08 EST