Last year, I had the problem to bring down all tasks of one RPC server job with one terminate command (USR2073), when I used an RPC Server connecting via SVC to the Broker. The solution was, to send at least the same amount of terminate commands as number of threads and the tasks as well as the job ended.
Now, I am on a TCP/IP connection with many tasks. The terminate commands let end all tasks but the job doesn
Another test with the terminate command was successful. After the last task ended, also the job ended.
Nevertheless, I experienced this week a situation, where all tasks of one RPC Server job ended but the job was still active. Only a cancel job command worked. It was connected to a network and NET-WORK problem.
Any thoughts on this?
Thanks
Dieter Storr
To terminate all servers you may also use the TERMINATE-SERVICE function of USR2075N which terminates all servers with the same name.
If the job hangs after all servers have deregistered (which should of course not happen) we need an SVC dump of the address space to check for the reason.
Thank you
Wolfgang