HP OpenVMS Systems

ask the wizard

KRNLSTAKNV kernel stack not valid bugcheck?

» close window

The Question is:

 
Oh great Oracle, I will donate my pointed hat with stars and crescent moons if
 you can figure this one out for me.
 
I have a client with a 4100, V7.2-1, lotsa memory, IDX. The backplane has four
 SCSI controllers, attached variously to HSZ and 16-bit shelves.
 
They just added a TZ89 to two TZ87s. When testing this configuration (and
 streaming all drives), the system crashed. Errorlog (Diagnose) and crash file
 are not revealing. Crash was Kernel stack not valid.
 
Could this possibly be a BAP problem? I do not see any parameters which allow
 me to change kernel stack or INTSTKPAGES on Alpha. My only comment is that the
 poor system was doing a LOT of I/O at the moment.
 
Please, oh wise one, illuminate me.
 
Werner. L. Stunkel
Senior Consulting Paleontologist
Groundwork, LTD.

The Answer is :

 
  Without the CLUE CRASH output (or access to the system dump) and without
  rather more detailed information on the system configuratation -- and the
  OpenVMS Wizard hastens to add that the Wizard is not in a particular
  position to accept and diagnose system crashes -- it would appear that
  something scribbled on the kernel stack.  (Let it never be said that the
  Wizard lacks a stranglehold on the obvious.)   Without these and other
  details, the Wizard is left only to recommend the application of the
  currently-available ECO kits for OpenVMS Alpha V7.2-1, and a call to the
  Compaq Customer Support Center to research the cause of the crash itself.
 

  
     
     answer written or last revised on ( 29-NOV-2000 )
     » close window