Group job performance tips

 

This topic provides you with some tips on maintaining good system performance when using group jobs.

• The effect on the system for a large number of suspended jobs is normally small if the dedicated main storage requirement is not a factor.

• When a TFRGRPJOB command runs and a new job must be started, the overhead involved is roughly the same as signing on to the system. When the command is run and the group job is already started, the overhead required is roughly the same as using the transfer to a secondary job option on the System Request menu when the secondary job is already active.

• If a group job is to be run with any frequency, it is desirable to prevent it from ending. That is, do not end the program, but issue a TFRGRPJOB command to prevent job starting each time the group job function is needed.

• The SETATNPGM command causes the current display to be saved when the Attention key is pressed, and to be restored when the Attention key-handling-program ends. This is roughly the same as using of the System Request menu and has a more noticeable effect on remote workstations.

• The controls on the number of jobs active in the system (the MAXJOBS parameter on the CRTSBSD command) are not affected by the number of group jobs active at any time.

• All system values that control the creation of job structures (QACTJOB and QADLACTJ, and QTOTJOB and QADLTOTJ) are affected; these values may need to be increased to allow for the addition of group jobs.

 

Parent topic:

Group jobs