[Info-vax] queue errors

[Stephen Hoffman]

On 2012-10-16 16:41:36 +0000, Tom Adams said:

> We get some queue errors at a low rate.  Any idea how to troubleshoot
> these? These are small files printing to a logger. It says the files
> were retained, but I cannot locate them when I search the entire
> disk.  I wonder why that is?
> 
>  235  FOR010          NETNONPRIV        1  Retained on error
>        %PSM-E-WRITEERR, error writing !AS
>        -LAT-F-NOMSG, Message number 01769FCC
>          Submitted 28-JAN-2012 15:30:22.40 /FORM=ERROR /PRIORITY=100
>          File: _DSA0:[]FOR010.DAT; /DELETE
>          Completed 28-JAN-2012 15:30:34.48 on queue EESLOG1
> 
>     619  FOR010          NETNONPRIV        1  Retained on error
>        %SYSTEM-F-ABORT, abort
>          Submitted  1-JUN-2012 09:23:46.28 /FORM=ERROR /PRIORITY=100
>          File: _DSA0:[]FOR010.DAT; /DELETE
>          Completed  1-JUN-2012 17:52:52.91 on queue EESLOG1

It looks like some Fortran program might possibly be printing directly 
to a spooled LAT device, based solely on that filename.

Most folks just delete these entries.  There are procedures around that 
can periodically delete that stuff, too.

The files are hidden in the file structure.  ANALYZE /DISK [/REPAIR] 
will pick up any scrap files that are left around.

If you want to troubleshoot this, start looking for network or 
DECserver errors and glitches.  And look at the printing-related logic 
in the Fortran code; even if you don't switch that to the $sndjbc call, 
you can look at the diagnostics and logging from that code.

As will be the usual recommendations here, also upgrade the unmentioned 
(and probably fossil-vintage) version of VMS and apply patches, too; 
make sure your patches are current, and you're running on the LAT print 
symbiont, and get your VMS version to (somewhat) newer.  There's little 
sense in chasing old bugs, after all.  And there have been some bugs 
here over the years, including one that left these hidden files "lost" 
in the file system, undeleted, and requiring an ANALYZE /DISK /REPAIR 
to clean them out.  That bug was patched some years ago.


-- 
Pure Personal Opinion | HoffmanLabs LLC