Given the recent thread on this topic, has anyone arrived at a bullet-proof solution with capturing SMF data for a time-period (date specific not withstanding) when using SMF support for system logger in z/OS 1.9, so there are no duplicate SMF records generated across "dumped" output files, created by IFASMFDL? The concern here being that the logger data is a pipe/stream, the TIME parameter supports a minute-level specification and that the SMF timestamp granularity is at 1/100 of a second (somewhat incompatible with the IFASMFDP/IFASMFDL TIME parameter specification).
As a consideration, I am evaluating a CA product, CA SMF DIRECTOR, which appears to have functionality ideal with z/OS 1.9 and system logger, where a "data marker" is identified across SMF capture/unload processes (invoked IFASMFDL using exits as I understand). The object here is to reduce/eliminate the duplicate data exposure without requiring a no-duplicates SORT option (still only at a 1/100th second granularity, by the way), especially considering a large-shop where there might be 40 million or more SMF 101/110/116 records (you choose the AND | OR condition for SMF type, as appropriate), which is unreasonable. At least with IFASMFDP, there is a finite start and end with each SMF MAN dataset processing, when performing the SMF capture/unload processing, throughout the day (at a 8-minute peak interval for a DB2/CICS/MQ LPAR). Scott Barry SBBWorks, Inc. ---------------------------------------------------------------------- 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