Operations > Error Messages

Your job queue has been zapped; exit A-Shell now

Reviewed

A topic that has been updated has had material added or changed or corrected. That particular text is of course correct and accurate, as of the date updated, but the rest of the topic has not been reviewed. The normal case for a topic getting an "Updated" tag is that material from the A-Shell Development Notes has been added. However, that does not mean that the rest of the topic has been reviewed for accuracy, completeness or relevance. If you see the "Reviewed" tag, then the material is known to be accurate; otherwise, it is not known to be completely up to date.

Reviewed

January 2009

Short version

This occurs when someone deletes the qflock.sys and/or jobtbl.sys files (or at least clobbers your entry within the latter, perhaps with the QUTL command) while you are running. A-Shell does not notice the problem until it has reason to update the queue (such as when chaining between programs or logging to a new ppn.) If this problem happens every time you launch a second session, then the problem is most likely that A-Shell is not able to generate a unique device identifier due to a problem in your CONSOLE statement in miame.ini.

Long version

The message indicates that another job has overwritten the current job's job-control-record. (Each job has a record in the JOBTBL.SYS file where it stores a bunch of current status information about the job - this is the primary source of information used by SYSTAT, JSTAT, and most XCALLs that request information about a job.)

When a new instance is started, A-Shell generates a (hopefully) unique ID for the instance, then scans the JOBTBL to see if there is already a record for it (which happens when a workstation crashes or is rebooted without exiting A-Shell), if so, displays a brief warning message on the screen and in the log, and then re-uses the record.

You can see the instance identifiers using SYSTAT/C (3rd column). Generally they consist of 2 parts separated by a colon: an identifier of the workstation (like a machine name or possibly an IP address), plus an instance number (01, 02, etc.) to separate the instances of A-Shell coming from the same workstation.

For example:

.systat/n/c

Status of A-Shell/32 Ver. 5.1.1134.2 on Friday, January 09, 2009 15:22:26

TSKAAA TSKAAA  JACKT400:01        DSK0:1,4       RN  SYSTAT    2912  986K bytes

TSKAAB TSKAAB  JACKT400:03        DSK0:150,277   RN  MYDOC     5464  939K bytes

TSKAAE TSKAAE  JACKT400:05        DSK0:150,277   ^C  VUE       6892 1502K bytes

 

This shows 3 instances, all coming from the machine whose name is "JACKT400", and each with a separate instance # (01, 03, and 05).

In some particular environments, such as Terminal Server, or CGI, or when launching A-Shell sessions from another application, it may be necessary to ask A-Shell to add a numeric suffix (-###) to the machine name (and prior to the instance number) in order to prevent duplicates. There are two ways to accomplish this: add OPTIONS=NTTS to the miame.ini file, or add the -awts switch to the A-Shell launch command line.

Besides SYSTAT/C, the best tool for analyzing this problem is the ashlog.log file, provided you add TRACE=INOUT,SIGHUP,BASERR to the miame.ini file (which we recommend as standard procedure.) This will cause A-Shell to log an entry for each session started and ended, along with the ID name, job number, user login, and license utilization information.