Okay, so is there anyone even remotely concerned about SMF DIRECTOR DUMP performance?
After reviewing SMF type 15 (CLOSE event) data, it appears that a BUFNO override on a //HISTORY1 DD will only be applied "IF" there happens to also be a BLKSIZE= parameter coded.
Clearly the two are not related at all, and CA has composed a PTF (confirmed it worked and soon to be published) so that if a //HISTORYn DD or a //HISTxxn DD (where "xx" relates to a SUBSTREAM ID) is coded, then the elapsed time for an SMFD / SMFDLS DUMP (either to tape or DASD) will likely improve.
I realize that using these "model" DD statements for various types of HISTORY files is not that well-documented, nonetheless the functionality does exist and stands to help improve SMF DIRECTOR DUMP elapsed time, at the small increased cost of REGION.
Scott Barry
SBBWorks, Inc.