Re: Hold ALL jobq's

Steve McKay wrote:

I'm writing a UPS monitor program. If I can hold the job queues in the UPS monitor program before I shutdown the system (because the UPS is dying), I don' have to change our start up program. The current startup program is used, more often than not, after a normal, intentional shutdown, so it does start subsystems needed for normal work. If I had to shutdown the system with the UPS monitor program, I might have serveral jobs in a job queue (many of our job queues are single threaded) that shouldn't run because Job1 was running at the time of shutdown and those jobs depend on files created by Job1. Therefore, I prefer to hold job queues rather than rework the startup program.

Steve:

Rob has given one potentially useful answer. When you ENDSBS *CNTRLD, the subsystem (or subsystems if *ALL) stops pulling jobs off of its jobqs. Existing jobs are allowed to continue until the *CNTRLD time limit is reached. At that time, the currently running jobs are signaled to end.

One problem has been that the subsystem must eventually actually be ended in order to restart it. (Anybody know if that's changed?) A UPS monitor would want to be sure that it was really going to end stuff rather than simply start it all back up in a minute or two because power suddenly was restored.

As for your actual question, no, I haven't seen any generic kind of HLDOUTQ command. Seems pretty easy to create one though.

When your system restarts, were you intending for jobqs still to be held?

Tom Liotta