Some years ago, when we migrated our first series 70 to MPE/XL, I got such
an error while reading a variable record file.  The system crashed but I do
not remember what the abort number was.
 
The error occurred while testing an in-house developed routine that would
download soft fonts.  The routine opens up the font file in "MR;NOBUF" mode
but it needed to read a few bytes from the first record of the file in
order to determine the orientation and validate that the file was indeed a
soft font.  The code would read about 40 bytes but due to the way the file
was opened the file system error that corresponds to "BAD VARIABLE BLOCK
STRUCTURE" would be returned.  On the series 70 this could be ignored
because it would still bring the requested bytes of the record, but it
would crash the MPE/XL machine.
 
I worked around the problem by determining the type of CPU and when I was
dealing with an MPE/XL cpu I would open up the file a second time with out
the "MR;NOBUF" option and read the first record.
 
So is it possible that something other than a system process is opening and
closing those spoolfiles?.  Could you have some in-house or third party
utility that opens a spoolfile without allocating a large enough buffer to
hold the spoolfile block?
 
Regards
 
Paul Christidis
 
 
 
______________________________ Reply Separator _________________________________
Subject: Background Jobs and SA615
Author:  Charles Leader <[log in to unmask]> at CCGATE
Date:    1/26/1996 8:56 AM
 
 
Has anyone gotten a large number of BAD VARIABLE BLOCK STRUCTURE on
spoolfiles followed by System Abort 615's on their systems? We have
been getting this error on a quite a few spoolfiles while they are
open and a few SA615s when the system attempts to close these
spoolfiles. HPRC said this is fixed in a patch (MPEHXT7) on 5.0, but
we just started getting these errors about 2 1/2 months ago. Anyone
else traveled this road already?
Thanks in advance,
Charles Leader
ST. Lukes Regional Medical Center
Boise, Id.