|
|
|
Data component CA size is controlled by the space allocation specified for the data component. Generally, it is one cylinder, but it can be as small as one track. VSAM sets the data component CA size to the lesser of the primary allocation, the secondary allocation, and one cylinder. VSAM creates sequence sets such that they hold one compressed key per data component CI in a data component CA. This means that the data component CA size controls the number of index records VSAM retrieves on one read (one index component CI) as well as the number of index levels that are required to point to the sequence sets. Because obtaining more index records per read benefits both random and sequentially accessed files, it is generally best to use cylinder CAs. This means all but very small files should be allocated in cylinders.
An exception to the above guideline is files that have high browse update activity in CICS. In this case, a small data component CA size can reduce the number of browse enqueue locks because the number of index entries in the sequence set is smaller. Therefore, fewer entries are enqueued by browse. However, to achieve optimal performance, replace the browse operations with Get Key Greater than or Equal. This replacement eliminates the sequence set enqueue.
| Copyright © 2012 CA. All rights reserved. |
|