I am sure everyone will agree that
forcepurge is a very handy function when it comes to handling
misbehaving tasks. The earlier implementation of forcepurge had some
side effects. U abends and region hangs were some of the dreaded side
effects. Apart from the side effects, it would take quite some time
to terminate the task. Forcepurge is not just used by the
administrators, it is used by the TXSeries runtime as well, for eg.
If there is a network failure, all the tasks associated with that
client/server are forcepurged. The only way to know that something
has been forcepurged is by looking at a console log message that
reads “Abnormal termination of task...... exit 199”. Exit code
199 was the only way to identify that a task has been forcepurged.
TXSeries has made significant
improvements to the forcepurge logic since v6. It is quite stable
when compared to the previous versions. The U abends are very rare
(though not eliminated) and the task termination is almost immediate.
A bunch of new messages have been introduced to help identify the
cause and also the process that induced the forcepurge.