Operating System - OpenVMS
1828371 Members
3098 Online
109976 Solutions
New Discussion

CRASH ANALYSIS of VAX/VMS server

 
Volker Halle
Honored Contributor

Re: CRASH ANALYSIS of VAX/VMS server

This a not a real valid OpenVMS HALT restart bugcheck, the APPLICATION itself is issueing BUGW BUG$_HALT bugchecks internally !!! This explains, why the registers and the HALT code (e.g. in the RPB) do NOT provide valid values.

Here is an example instruction stream from the application:

00195D27: BRW 00195DEF
00195D2A: PUSHL #10000002
00195D30: CALLS #01,@0019C250
00195D37: BLBS R0,00195D4D <<< return address
00195D3A: PUSHL R2
00195D3C: MOVL #0EC0DE07,R2
00195D43: BUGW #0160 <<< issue HALT bugcheck !!!

The saved PC on the kernel stack should point to the application code declaring the HALT bugcheck, so please provide the following data:

SDA> EXA/INS 195FEB-10;10

The non-fatal bugcheck entries can be reproduced easily with the following piece of MACRO code:

.entry halt, ^m<>
$cmkrnl_s routin=halt_cpu
MOVZWL #1, R0
ret

.entry halt_cpu, ^m<>
bug$_halt=^x160
BUGW
.WORD ; issue a HALT bugcheck
ret

.end halt

Fix the problems in your application and those crashes will go away. If you need further assistance, please consider to contact me offline (see e-mail address in my ITRC profile).

Volker.
Sk Noorul  Hassan
Regular Advisor

Re: CRASH ANALYSIS of VAX/VMS server

Volker, Sorry for the late response. Pls find the attached SDA output as asked by u.
Volker Halle
Honored Contributor

Re: CRASH ANALYSIS of VAX/VMS server

Now what code is at 000BE3F4

SDA> EXA/INS 000BE3F4;20

I'm even more interested in:

SDA> EXA/INS 195FEB-10;20

This PC is found on the kernel stack in the crash and should reflect the last instruction executed before the crash !

Volker.
Volker Halle
Honored Contributor

Re: CRASH ANALYSIS of VAX/VMS server

continued and fully explained in:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=998774

Volker.

PS: Please consider to close this thread (9972828).