RE: Telnet server multiple instances?

I know I'm late to the party, but we've had good luck in picking the subsystem that provides the signon screen based on the order the subsystems start. It may not be official, but it seems to work reliably.

Normally, we have a variety of QINTER subsystems that pick up all interactive sessions. We also have a number of 'laser tag' hand-held scanners that we want assigned to their own subsystem. At IPL time we start the QINTER subsystems first (which have device entries that will issue signon screens to the laser tag devices), then after a few minutes, start the laser tag subsystem. The laser scanner devices get a new signon screen from the laser tag subsystem.

The laser tag subsystem has its own signon screen. It appears that the last subsystem that starts that has a matching device entry can 'steal' the allocated device from another subsystem. This seems to work even if the device is not visible to the system at that time. I just tried it with a test QINTER, and as I fired up the test subsystem, the signon screen changed to match the newly-started subsystem. A 2nd trial, with the session shut down while the subsystem started did the same thing.

We've been doing this for years. One of my sessions has a matching entry in the laser tag subsystem. After an IPL, it's in that subsystem.

That said, we also have a backstop in the initial program for those devices, which will reroute the job if it's in the wrong subsystem. I don't have any statistics on how many times the reroute has to be used.

Paul E Musselman
<mailto:PaulMmn@xxxxxxxxxxxxx>PaulMmn@xxxxxxxxxxxxx