Cancelling a runaway Natural job

, , , ,
1

We have set up a “batch” environment for our Natural applications running on Windows. Developers have the ability to run large Natural jobs on their PC or on a server. Jobs are executed on the server via a CScript shell command or via batch script (DOS).

The problem we have is when a developer then wants to delete that job (because it is looping or some other reason). In Windows all the jobs run under the Natural executable (NDERUN or NATURALR) and there is no way to associate the Windows Process Id under which Natural is running with the job that needs to be deleted. This is of course assuming that there is more than 1 job running.

Has anyone come across this situation and how did you deal with it?

2

If logging is used for batch runs it is part of the information written out. The log files get created in the Natural TMP directory defined in your parms (default is %TEMP%). As part of your script you can specify NATLOG=ALL (or any of the other valid values detailed in the documentation).

The only “fun” part of this may be identifying the correct log file…

3

Chad,

Thanks for the tip. I forgot about the log files. It should not be too difficult to build an automated solution around those files.