Hi Gurus,
OS: HP-UX 10.20
Progress Version: 8.3B
Blocksize: 8192 bytes
We run a loan origination application using Progress 8.3B on a HP D-class box running HP-UX 10.20. Our database contains about 200 tables, and we purge year-old data on a monthly basis to keep the database as clean as possible.
Lately, the variable extent file has continually grown at around 4MB/day-- even though we had purged two months worth of application data a week ago (when we purge data we purge all related records in all tables).
All things being equal, and living in a perfect world, new records that are added daily should fill the space left behind by the purged records, right?
One theory is this: some of our tables have had fields added to them in the last year-- thus increasing the average record size. Perhaps these new larger records are too big to fit into the holes left by smaller, older records from the same table?
To test this theory, we wanted to see what tables's records were being appended to the variable extent file. We looked at the current .st file to get the size of all of the database fixed extents in kb (in our case, 7,472,480 kb), then divided this by 8 to get the total number of blocks in the fixed extents (934,060).
Hence, the block numbers in the variable extent should begin with 934061-- and the corresponding RECIDs should start with 934061 * 64 (since there can be a max of 64 records/block) = 59779904.
However, I ran a quick utility to find which table has a record with a RECID from 59779904 to 59779904+(100*64) [to search the first 100 blocks of the variable extent file], and turned up no hits.
Did my math go wrong somewhere here... or is there some better way to find out what records are populating the variable extent?
Thanks,
James
OS: HP-UX 10.20
Progress Version: 8.3B
Blocksize: 8192 bytes
We run a loan origination application using Progress 8.3B on a HP D-class box running HP-UX 10.20. Our database contains about 200 tables, and we purge year-old data on a monthly basis to keep the database as clean as possible.
Lately, the variable extent file has continually grown at around 4MB/day-- even though we had purged two months worth of application data a week ago (when we purge data we purge all related records in all tables).
All things being equal, and living in a perfect world, new records that are added daily should fill the space left behind by the purged records, right?
One theory is this: some of our tables have had fields added to them in the last year-- thus increasing the average record size. Perhaps these new larger records are too big to fit into the holes left by smaller, older records from the same table?
To test this theory, we wanted to see what tables's records were being appended to the variable extent file. We looked at the current .st file to get the size of all of the database fixed extents in kb (in our case, 7,472,480 kb), then divided this by 8 to get the total number of blocks in the fixed extents (934,060).
Hence, the block numbers in the variable extent should begin with 934061-- and the corresponding RECIDs should start with 934061 * 64 (since there can be a max of 64 records/block) = 59779904.
However, I ran a quick utility to find which table has a record with a RECID from 59779904 to 59779904+(100*64) [to search the first 100 blocks of the variable extent file], and turned up no hits.
Did my math go wrong somewhere here... or is there some better way to find out what records are populating the variable extent?
Thanks,
James
