Try OpenEdge Now
skip to main content
Database Essentials
Administrative Planning : Data layout : Sizing your database areas
 

Sizing your database areas

When trying to determine the size of an area, you must look at the makeup of the information being stored in that area; remember, an area can contain one or more tables or indexes. The default area should generally be reserved for schema and sequence definitions, as this will make conversions easier in the future. The first step in this process, if you already have an OpenEdge database, is to do a table analysis using the PROUTIL utility. See OpenEdge Data Management: Database Administration for details.
The following sample output shows a portion of a table analysis:
              ---Record---             ---Fragment---          ---Scatter---
Table        Records  Bytes      Min  Max     Mean   Count    Factor  Factor
Work Orders  12,383   6,109,746  60   10,518  493    21,131   1.6     4.0
RECORD BLOCK SUMMARY FOR AREA "Inventory" : 8
-------------------------------------------------------
                                   -Record Size (B)- ---Fragments--- Scatter
Table              Records    Size  Min   Max  Mean   Count Factor    Factor
PUB.Bin                770   26.2K    33   35    34     770    1.0       1.4
PUB.InventTrans         75    3.6K    43   51    48      75    1.0       0.9
PUB.Item                55    7.6K   103  233   141      55    1.0       1.0
PUB.POLine            5337  217.5K    40   44    41    5337    1.0       1.0
PUB.PurchaseOrder     2129   68.6K    33   36    33    2129    1.0       1.0
PUB.Supplier            10    1.1K    92  164   117      10    1.0       1.0
PUB.SupplierItmXref     56    1.1K    18   20    19      56    1.0       1.0
PUB.Warehouse           14    1.3K    82  104    92      14    1.0       1.7
                ------------------------------------------------------------
Subtotals:            8446  327.0K    18  233    39    8446    1.0       1.1
After doing a table analysis, focus on record count (Records) and mean record size (Mean). Look at every table and split them according to mean record size. Each record contains approximately 20 bytes of record overhead that is included in the calculation of the mean.
* Block sizes
* Specifying records per block
* Example of calculating records per block
* Unused slots
* Determining space to allocate per area
* Distributing tables across storage areas
* Using extents
* Index storage
* Primary recovery area