IC78728: FREQUENT PROGRESS MONITOR POLLING AFTER SESSION IN PREPARE STATE TAKES TOO MUCH CPU CYCLES TO PROCESS

4.2.2 Fix Pack 7 (July 2014) for Tivoli Storage Productivity Center
4.2.2 Interim Fix 2 (December 2012) for Tivoli Storage Productivity Center
4.2.2 Fix Pack 8 (March 2015) for Tivoli Storage Productivity Center
4.2.2 Fix Pack 9 (May 2015) for Tivoli Storage Productivity Center
4.2.2 Fix Pack 10 (August 2015) for Tivoli Storage Productivity Center
4.2.2 Fix Pack 3 (October 2012) for Tivoli Storage Productivity Center

APAR status

• Closed as program error.

Error description

• After a GM session reached prepared state, frequent Progress
  Monitor polling, particularly on the GM Journal FlashCopy pairs,
  causes too much CPU cycles (in the server machine) to process
  the polling, which slows down TPC-R handling of other activities
  in the machine. This problem is seen in some z/OS TPC-R servers
  

Local fix

• Manually increase the polling parameters in
  rmserver.properties file:
  
  The following values would run at 1/100 the current default
  interval (once every 45000 seconds)
  csm.server.progress.query.interval.initial=3000
  csm.server.progress.query.interval.max=90000
  csm.server.progress.query.interval.min=3000
  csm.server.progress.query.interval.step=6000
  csm.server.progress.query.size.step=500
  

Problem summary

• Frequent progress monitor polling takes too much CPU cycles to
  process
  

Problem conclusion

• Fewer polls will be taken to help improve the performance and
  use less CPU cycles
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  TPC FOR REPL SE

• Fixed component ID

  5608TRMSV

Applicable component levels

• R420 PSY

     UP