On 10/05/2013 09:47 AM, Charles Mills wrote: > I asked this question on the ACF2 list. The only answer I got was from the > always-helpful Lizette who indicated a low degree of certainty in her answer > and suggested that I ask here. > > I am trying to clarify some product documentation and clean up some internal > logic - I don't have an actual "problem." > > Question: Assuming an ACF2 shop wants to cut SMF 230 (or other as > configured) records, is it sufficient to specify ACF2=230 on the @SMF macro > of the ACFFDR in UM99901, or is it also necessary that the shop specify > SYS(TYPE(230)) or the equivalent in the SMFPRMxx member of the SYS1.PARMLIB > concatenation? > > We don't run ACF2 so I can't experiment (but we have customers that do). > > The documentation of SMFPRMxx TYPE() would seem to imply that SMF > "suppresses" any record type not specified in SMFPRMxx, but I know from > experiments that I can write a user SMF record successfully with SMF(E)WTM > even if the type is not in SMFPRMxx. However, that's not the whole answer > because a third party product such as ACF2 *could* use SMFRTEST to see if > the customer wanted 230's or not, and if TYPE(230) were not specified, not > write them. > > I also have a customer that I know is writing 230's that I asked to eyeball > a D SMF,O and they said 230 was not in there, but I am not sure that I trust > their answer. > > I have a question in to CA but so far they are scratching their collective > heads. > > Thanks, > > Charles >
Any application or product with proper authority can create and "write" an SMF record with SMF(E)WTM, and that creation should appear "successful" as long as the record is successfully passed to SMF. Whether that record actually makes it all the way to the SMF datasets depends on SMF Exits and SMFPRMxx PARMLIB definitions. Since there are all sorts of ways of including and excluding record types and ranges of record types in SMFPRMxx, that doesn't necessary mean you would see an explicit occurrence of "230" in SMFPRMxx. The "ACF2=230" sounds like a way to tell ACF that it should write Type 230 records. I would think it highly unlikely that this would also control whether SMF physically saves the records to its datasets, unless the ACF installation process uses the mentioned @SMF macro to generate a new SMFPRMxx that you are supposed to install as part of ACF customization. I think it more likely they would just advise one to modify or assume you would know to modify SMFPRMxx if you wanted to retain the records. A Sort of SMF data that only includes records with type field of 230 (X'E6') would give a definitive answer if you don't trust eyeball analysis. However, the free CBT REVIEW utility also allows you to eyeball specific SMF record types, and that should be pretty definitive as well. -- Joel C. Ewing, Bentonville, AR [email protected] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
