Following are sample report pages. For an example report header, see Sample Report Headers.
CONTROL CARD(S) .........1.........2.........3.........4.........5.........6.........7.........8 REPORT TYPE=DATANE,TABLE=REO,DBID=1022 FUNCTION=REPORT DBID=01022 TABLE=REO TYPE=DATANE
This page of the report shows the following:
TBL AREA KEY CMP USER COMPRESSION RECLN RECORDS BLKSIZE TRACKS TOT BLKS INUSE BLKS PART BLKS REO REO REOK1 N 112 31,950 4,096 150 1,800 890 1 CALC MAX ROWS/BLOCK 36
This page of the report shows the following information about the table that is being processed:
Displays the heading for the columns containing information describing the selected data table.
Displays the maximum row per block (MRB) which was calculated by dividing the expanded row size plus RCE (record control element) into the data block size.
*--- BLOCK --* PBC 887 MRB *---- MAX ROW/BLK EFFICIENCY RATES WITH BUFFERS ----* GROUP ROWS BLOCKS MAX MIN AVG BLOCK EFF 2 4 8 12 16 20 24 28 32 48 64 96 128 256 READ W/DATA ROWS ROWS ROWS CHANGES % % % % % % % % % % % % % % % ________________________________________________________________________________________________________________________ FULL TABLE MRB 31,950 888 36 18 36 6,931 13 14 16 AMRB 31,950 888 36 18 36 6,931 13 14 16
This page of the report shows the following efficiency ratings for the FULL TABLE:
Displays the PERFECT BLOCK CHANGE (PBC) count of 887. The PBC is the number of block changes that would occur when a table was loaded with all rows in sequential order and all blocks had the maximum number of rows per block. The next lines are the headings for the detail line area of the report. Because RECORDS was not specified, no detail lines were generated.
Displays the summary section of the report. Two lines are created for the table.
The MRB line provides the sequential efficiency rating of this table compared to an ideal or perfect situation where each block is loaded to the maximum (rows) and all the rows are in sequential order.
The AMRB (Adjusted MRB) line provides the sequential efficiency rating of this table using only the sequential order of the rows as the ideal. This rating does not penalize the efficiency rating for available free space that can be in the data blocks.
Note: Having large free space amounts could substantially reduce sequential performance by limiting the rows returned in a data block.
Indicates the number of rows in the table.
Indicates the number of data blocks that contained at least one data row for this table. The value varies slightly from the BLOCKS INUSE figure on page 2. The difference is data blocks that are used as data area control blocks.
Indicates the high-level mark (MAX. ROWS) and the low-level mark (MIN. ROWS) of each block having a row for this table. The report also divides the total rows of the table by the blocks in-use to calculate the average rows (AVG. ROWS).
Indicates the BLOCK CHANGES that occur when the data rows are processed using the selected key sequence. A BLOCK CHANGE is counted each time the next row (in the sequence given by the index chosen with the KEYNAME=) is not within the same block as the current row.
The BLOCK CHANGES is the key factor in measuring sequential efficiency. When a table is in sequential order, all data rows are processed in a given data block before moving on to the next block. When rows are not in order, the sequential process bounces back and forth between data blocks to retrieve all data rows in the selected key sequence. The bouncing back and forth adds overhead and delays to any process that is reading the data in sequential order.
Indicates the MRB efficiency which is calculated by taking the ideal situation where all data blocks are only processed once as data rows are read sequentially (PBC 887) and dividing that by the actual number of block changes (6,931) that would have occurred. In other words, it is the efficiency of reading the data with a single buffer.
Displays the calculated MRB efficiency with buffers. Because buffering can significantly reduce the physical I/O used when a previously read data block is already in storage (a buffer), the MRB efficiency percentage is calculated using two buffers, four buffers, and so on.
A table that shows a high MRB efficiency percent (more than 80 percent) with buffers of 8 or less still provides a reasonable good sequential processing. There is processing overhead to find data blocks in the buffers, but it is significantly less than having to reread the data blocks from the DASD.
Some sites can have higher buffer availability for batch processing. For these sites, additional efficiency ratings are generated for 12 buffers, 16 buffers, and so on.
By monitoring the amount of sequential activity and the sequential efficiency of a table, you are able to make more informed decisions on when to do a data reorganization.
Note: The FULL TABLE or summary lines look at the table as a whole and their statistical calculations are based on the whole table.
|
Copyright © 2015 CA Technologies.
All rights reserved.
|
|