FILE USAGE ANALYSIS

Also analysis of file usage can be requested by placing the cursor on a job and using a

command key. Below are just two examples:

1   In only 15 seconds (from 14:58:30 to 14:58:30) this interactive job is reading

     250.615 records – the user should have used batch for running this program!

2   Files are constantly being opened and closed (134 or 84 files are open) – and 42 of

     those are never used (heading “Zero I/Os” = files with no reads or writes).

On the next screen GiAPA shows statistics based on how often each relative record number

was found during the HotSpot sampling. Results may seem surprising, but can in fact  be seen

in most installations.

1   This file showed 250+ million reads, but the difference between the highest and lowest

     relative record number accessed was only 19.588.

2   This column contains the relative numbers of all records found repeatedly in the samples

     taken.

     The later columns contain number of times found, %, accumulated %, and accumulated

     count for how many records were found used repeatedly.

3   Since data was collected at 2647 HotSpots, the statistics can be claimed to be very

     reliable, although it of course only is sampling.

4   The bottom line for this file is that 61 % of all reads (152+ millions) were used to read

     only 27 records – also here GiAPA shows where applications may be optimized!

Also the statistics for repeated reading of records can be shown in graphical form:

Above bar chart shows potential savings for the 10 files showing most re-reads of input-only

files. The blue bars show total number of reads, the red tell saving obtainable if each record

only was read once per program call.

 BACK TO TOP