[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference vaxaxp::vmsnotes

Title:VAX and Alpha VMS
Notice:This is a new VMSnotes, please read note 2.1
Moderator:VAXAXP::BERNARDO
Created:Wed Jan 22 1997
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:703
Total number of notes:3722

483.0. "N6.2-1H3 ???" by LYOV00::RIVARD (We're NOT only in it for the money !) Wed Apr 16 1997 06:54

    
    Hi ...
    
    I tried to upgrade an OpenVMS V6.2-1h2 axp cluster to V6.2-1h3 using
    the kit located on BULOVA:: AXPVMSU1H3062.A .
    I also upgraded the firmware using the V3.8 CD .
    
    The cluster is an scsi cluster with 2 AlphaServer 1000a .
    
    There was no problem during vmsinstal . 
     
    When I rebooted one of the 2 nodes , the VMS banner indicated N6.2-1H3
    and the node crashed with a fatal bug check (process STACONFIG) .
    
    Then I booted conversational and set NISCS_LOAD_PEA0 to 0 and the
    machine booted correctly . 
    
    Thinking there was something corrupted in the cluster configuration , I
    then rebooted with VAXCLUSTER set to 0 and did a CLUSTER_CONFIG.COM .
     But the result was the same and both the machines crashed .
    
    I had then to restore the system disk back to V6.2-1H2 .
    
    Is there anything I did wrong , is it possible to upgrade the V6.2-1H2
    system disk to V6.2-1H3 ? I need to have the V6.2-1H3 because there
    will be a third machine to integrate wihtin this cluster , and it needs
    V6.2-1H3 .
    
    
    Thanks in advance for any help ...
    
    Philippe RIVARD .
T.RTitleUserPersonal
Name		DateLines
483.1Confirm Console Variables?XDELTA::HOFFMANSteve, OpenVMS EngineeringWed Apr 16 1997 09:5121
    When you upgraded the PALcode, did you confirm that the correct disk
    and system root were set in the post-PALcode-upgrade-environment
    variables?   (Multiple VMScluster nodes booting from the same SCSI
    system root -- SYS0 is the default -- will tend to cause problems
    and will potentially cause disk corruptions...)

    I will assume that only one node was booted from the system disk
    being upgraded -- the node that was performing the upgrade...

:    When I rebooted one of the 2 nodes , the VMS banner indicated N6.2-1H3
:    and the node crashed with a fatal bug check (process STACONFIG) .

    The upgrade had (obviously) not completed.  What was the crash?

:    Then I booted conversational and set NISCS_LOAD_PEA0 to 0 and the
:    machine booted correctly . 

    I have found that when I set VAXCLUSTER to 0, it is often best
    to also set NISCS_LOAD_PEA0 to 0.
483.2More informationLYOV00::RIVARDWe're NOT only in it for the money !Wed Apr 16 1997 16:0460
    Well Steve ,
    
    First : of course when I did the upgrade , there was only one node booted
    in the cluster . And the upgrade did not show any error message .
    
    Second : the crash allways occured with VAXCLUSTER set to 2 . Of course
    as I restored the system disk , I have no sysdump to analyze , but
    below are some hardcopy informations :
    
    (boot dkb100.1.0.2001.0 -flags 0,0)
    .....
    jumping to bootstrap code 
    
    OpenVMS (TM) Alpha Operating System , Version N6.2-1H3
    
    %CNXMAN, Using remote access method for quorum disk
    waiting to form or join a VMScluster system
    %VMScluster-I-LOADSECDB, loading the cluster security database
    System Machine Check Through Vector 00000660
    logout frame address 0x6048 code 0x208
    Machine Check Code---> 0x208 Invalid page table lookup (scatter gather)
    
    IPRs: 
    EXC_ADD:00000000001214A8  ICCSR: 001EC4F800000004 HIER:  00000001FFFFDCC0
    HIRR: some other hex codes...
    
       EDSR (Comanche):  00002000-->
       DCSR (    Epic):  802a4058--> lost iPTL
       SEAR ( Sysaddr):  00285800
       PEAR ( Pciaddr):  00000280
    
    **** OpenVMS (TM) Alpha Operating System N6.2-1H3 -BUGCHECK ****
    * Code=00000215: MACHINECHECK, Machine check while in kernel mode
    * Crash CPI: 00   Primary CPU:00 Active CPUs:00000001
    * Current Process = STACONFIG
    * Image Name =STACONFIG.EXE
    
    **** Starting Memory Dump...
    
    *******************************************************
    
    Below are the firmware information :
    
    SRM Console:	V4.7-166
    ARC Console:	4.49
    PALcode:		VMS PALcode V5.56-6, OSF PALcode X1.45-12
    Serial Rom:		V2.8
    
    Processor
    DECchip (tm) 21064A-2  266MHz
    
    Memory
         384 Meg Of System Memory
    
    
    
    Regards ,
    
    
    Philippe RIVARD.
483.3Pointers...XDELTA::HOFFMANSteve, OpenVMS EngineeringWed Apr 16 1997 16:2720
    There's another similar discussion of a machine check in 475.*.

    You might want to take a look through the following notes from
    the Mikasa notes conference, as these cover the same machine
    check code.

    83  BACHUS::VANHAVERE    18-JAN-1995    10  Machinecheck 0x208 with EISA mo
   140  CHOOKE::KLUMPES      15-MAR-1995     6  Machine Check 208
   372   CSC32::HUTMACHER    29-SEP-1995     5  208 machine checks as1000 V6.1-
   699   CSC32::BRISSETTE    15-AUG-1996     1  pal error type code 208 , repro

   You will want to submit the CLUE output from the crashdump to the
   CANASTA e-mail server -- see note 233.* for details on how to do
   this -- as well.  CANASTA can provide clues as to the cause of the
   crash, if it has been seen before...

   And, if the above notes and the CANASTA server do not find a pointer
   to a solution to the problem, start an IPMT.