OSFQUE Limit Breached in CA OPS/MVS Event Management and Automation

Document ID : KB000072830
Last Modified Date : 07/05/2018
Show Technical Document Details
Issue:

We currently have our OSFQUE limit set to the default of 1700. We recently breached this threshold causing the following error. OPS4349O MESSAGE SEND TO TSO EXECUTE QUEUE FAILED, RC=4, DETECTED AT S OPRXTS+X'0000197E' We are considering doubling our OSFQUE number from 1700 to 3,400.

  • Are there any considerations that we need to take into account before increasing the OSFQUE level?

  • Are there any other parameters that we need to consider when making this change?

  • Any other application adjustments we need to consider before making this change? 

Environment:

CA OPS/MVS
z/OS
OSFQUE

Cause:

The customer experienced a problem where uncontrolled application work caused an overflow of the OSF command Queue set at 1700 commands.

They currently have a max of 15 OSF class servers. We discussed that they could increase the max number OSF servers to 30 for each class of server (TSL and TSP) where by they could have a total of 90 available servers. We also observed that some of the offending exec's ran up to 12 minutes long, which they explained was the nature of the application which uses a common dataset that every exec needs access to at some point in the code.

Since they can not 'throttle' when the exec's are invoked, the next best item would be to increase the max number of servers possible to 30 AND start making use of the TSL strictly for this application.

We also discussed increasing the OSFQUE from 1700 to 3400, this does not really address the real problem but have the effect of 'pushing' the problem further ahead in time. This is a small hit of 1 meg additional storage consumption in the OPSMAIN asid with doubling this Queue.

Resolution:

Recommended increasing the max number OSF servers to 30 for each class of server (TSL and TSP) where by they could have a total of 90 available servers.

Additional Information: