To defend MXG Software, yes, it can and does process DB2 records written to GTF.
There are two probles when DB2 data is written to GTF: Post processing: The problem is that processing GTF data requires two passes of the data; the long DB2 "SMF" records are broken into many 256 byte chunks and written to GTF as individual records, and the physical order of the records guaranteed not to be in order by IBM, so the raw chunks of the real data records must be read, sorted, and then the long records can be rebuilt from the little chunks, and then the "SMF" code can read the DB2 trace records. Impact on DB2: Writing an SMF record is an SVC call to the SMF address space, with no physical I/O that has to be managed. Writing to GTF requires the DB2 Address Space to have to manage each of those tiny chunk writes. I can see no way in which the overhead to write to GTF would not be significantly more than sending the records to SMF. But this is an option based on the way I think it works, and is not based on actual measurements. Barry Merrill ---------------------------------------------------------------------- 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
