Something happening that I don't understand, involving a child-server job launch.

In our client-server products, we use persistent connections to child-server jobs, and we use an intermediate "verification server" job to validate the user, then launch the child-server job under the user's own user profile. So it uses two socket handoffs, instead of just one.

And it works surprisingly well for something so Rube Goldberg.

But at a new installation, we're experiencing something I've never seen before: the verification server job crashes on takeoff, apparently before it even begins executing our user-validation and child-server-spawning program, throwing exceptions in something called SIG0000C00 (for which a Google search returns nothing at all), apparently crashing when it tries to do a SETATNPGM in a job that isn't a terminal session.

And yet the initial server job launched just fine.

I've never heard of a case where something else was "wedged in" on a submitted job, such that it executes ahead of what was specified on the CMD parameter of the SBMJOB. What would cause this?

--
JHHL