Re: How could a job escape an ENDJOB?

On 6/28/18, 10:34 AM, Mark Murphy wrote:

Are you sure it's the same job? Some server jobs will end and start
another. Maybe it's implemented as a pre-start job in which case you can
end the job, but if that takes the number of jobs below the threshold,
another one will start. Maybe there is a monitor job that will start
another server if it detects that one went away.

No; it's not a prestart job, and neither is it self-perpetuating, nor being watched by a monitor. (I should know; I wrote the CL program that started it), and while I was analyzing the situation, I was able to determine that was the one that had been told to shut down, and then should have been hit with not one, but two separate ENDJOBs.

We've got the same webapp running on the same Tomcat server, started in the same way, from the same STRTOMCAT and ENDTOMCAT CL programs, on dozens of customer boxes, as well as our own production box, mostly being shut down and restarted daily, and until now, we've never seen a server that failed to complete a shutdown survive even ONE ENDJOB, let alone two.

--
JHHL