Special Usage Considerations › Interpreting SNAPDUMP Output › Notes on Common Problems

Notes on Common Problems

1. If you have a large number of SNAPs with a reason code of D, you should review the processing sequence that created your SMF input data set to determine at what point the blocks in the data set might have been truncated.

   Another possibility is that incomplete blocks are being read by the Report Writer. This happens if the program receives an incorrect blocksize specification for the CAIJSSMF file. Check that the blocksize specified on the CAIJSSMF DD statement, if any, is at least as large as the actual blocksize of the file. If a blocksize has not been specified, verify that the blocksize in the data set label is correct. If in doubt, increase the blocksize on the CAIJSSMF DD statement and rerun the job.

2. If you have a large number of SNAPs with a reason code of G, the Report Writer is probably not allocating a large enough work area to hold the largest SMF records in your input file. The size of the work area is determined by the LRECL value of the input data set, obtained either from the LRECL subparameter on the CAIJSSMF DD statement, if any, or from the data set label.

   In addition, if the RECFM attribute of the CAIJSSMF data set is VB, specified on the DD statement or on the data set label, then the program does not allocate any work area at all for spanned record assembly. (The work area size being used by the program can be determined by examining the GETCB in the SNAP.)

   If necessary, change the RECFM subparameter on the CAIJSSMF DD statement to RECFM=VBS, or increase the record length specified by the LRECL subparameter and rerun the job.