The need for the DDCONS option was discovered by Diane Eppestein in 1987; from MXG Newsletter TWENTY-FIVE:
In 1987, Diane Eppestine at Southwestern Bell saw that whenever their SAR job (a SYSOUT processing subsystem) was cancelled, the CPU went to 100% busy for 30-60 minutes. Instruction traces found the "loop" was in DD Consolidation. SAR dynamically allocates a DD for each SYSOUT file it processes; by the end of the week that step had over 75,000 DD entries! DD consolidation reads the first DD segment, scans the remaining 74,999 segments for a match, reads the second and scans the remaining 74,998 for a match, etc. etc., etc., all at DPRTY=FE! In response to Diane's discovery, Bill Richardson, IBM SMF Development, subsequently provided a new SMF option, DDCONS(NO), specified in SYS1.PARMLIB(SMFPRMxx), so that you can disable this very unwise (in my opinion) algorithm, and thereby eliminate its wasted CPITCBTM and CPISRBTM CPU time. The time of DDCONS(YES) is measurable because SMFTIME timestamps when each record is moved (memory-to-memory) to the SMF ASID. For step events with more than 1635 DD segments, multiple physical type 30 records are created, each with its own time stamp. One specific case found a subtype 2 interval event created seven type 30 records at Time of Day of .19 .19 .19 .22 .32 .36 & .46 TOD seconds (i.e., it took only 270 milliseconds to create these seven records, since there is no DDCONS in the creation of the subtype 2). These seven records contained 9182 DD segments that had been allocated during that interval. The step then terminated at 18.89 TOD seconds, creating two subtype 3 records both with that timestamp, containing 1929 DD segments. DDCONS then began, and it then took until 36.50 TOD seconds to create the first subtype 4 record, and then the last of the eight subtype 4 records was not created until 40.93 TOD seconds. These subtype 4s had 11,070 DD segments, while the subtype 2s and 3s had 11,111 total DD segments. Thus this invocation of DDCONS took 22.04 elapsed seconds (and recorded CPITCBTM of 22.00 CPU seconds!) from the end of step until the step actually ended, and this DDCONS invocation could remove only 31 DD segments from the step record! Examine a day's TYPE30_4 (or PDB.STEPS) data and sum the CPITCBTM and CPISRBTM, then specify DDCONS(NO) and show management how many CPU seconds you have been wasting due to DD consolidation! NOTE: THE SMF DEFAULT IS DDCONS(YES). YOU MUST CHANGE YOUR SMFPRMxx MEMBER TO SPECIFY DDCONS(NO). The IBM fields SMF30ICU and SMF30ISB in SMF 30 subtype 4 are the MXG variables CPITCBTM and CPISRBTM respectively. Barry ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
