Overview of full-function database analysis and tuning

If you receive exceptions from Autonomics Director, you should analyze the state of the database and, if necessary, change the database configuration or tune the database.

The term full-function database includes the following database organizations: HDAM, HIDAM, HISAM, SHISAM, PHDAM, and PHIDAM. This term also pertains to the following indexes: HIDAM primary index, secondary index, PHIDAM primary index, and partitioned secondary indexes.

Start of change Each exception that is reported by Autonomics Director corresponds to one or more data elements. Therefore, when you receive exceptions from Autonomics Director, see the IMS Tools reports and charts in IMS Administration Foundation or Management Console by using its graphical user interface. These resources can help you understand the cause of the exceptions. End of change

The topics in Analyzing and tuning IMS full-function databases describe how to resolve exceptions by pointing to the data elements that you should check.

Important characteristics to monitor

The typical key objectives of administering full-function databases are preventing out-of-space conditions in database data sets, preventing performance degradation that might be caused by excessive database I/Os, and, when such conditions occur, tuning the databases to recover from such conditions. To help achieve these objectives, exceptional states of database data sets are reported in the following three categories:

Space use: Monitor space use exceptions to prevent out-of-space conditions.; A full-function database consists of one or more data sets. The size of these data sets increase as the size of data increases. However, when the size of a data set reaches a limit, and if no space is available to store data, the condition might lead to a critical situation where IMS applications cannot run. This situation is referred to as an out-of-space condition.; To prevent out-of-space conditions, you must monitor how the space in the data sets is used, estimate when a data set runs out of space, and take appropriate actions to mitigate a problem before it becomes a significant problem.
Fragmentation: Monitor fragmentation exceptions to prevent and resolve redundant space and performance degradation that might be caused by excessive database I/Os. Fragmentation exceptions warn you about scattered segments, IMS free space fragmentation, VSAM control interval (CI) and control area (CA) splits, and other factors that might increase database I/Os.
Optimization: Monitor optimization exceptions also to prevent and resolve performance degradation that might be caused by excessive database I/Os. Database I/Os increase if the database definition does not match the characteristics of the records that are stored in the database.; For HDAM and PHDAM databases, the randomizing efficiency is also a key monitoring indicator.

By monitoring these three characteristics, you can understand and analyze the state of your databases. However, because exceptions tend to be interrelated, you might receive exceptions from multiple categories. When you are determining the actions to resolve exceptions, consider the exceptions not as separate exceptions, but rather as a group of interrelated exceptions.