Chapter 2: Control Data Collection › Use the Database Directory

Use the Database Directory

A Performance Agent stores a node's collected data in a central database directory specified by the collection definition. Created during installation, the default database directory is assigned the logical name PSDC$DATABASE. You can specify an alternative location for a node's data file by modifying the collection definition.

If you install the Performance Agent software in a cluster environment, the PSDC$DATABASE directory should be on a permanently mounted disk, accessible to every node in the cluster because the data collection schedule file, which is common to all nodes, is in this directory. The PSDC$DATABASE directory contains the following files:

• Schedule file
• Parameters file
• Holiday file (optional)
• History files (optional)
• Performance Agent error log file (created when an error occurs)

The PSDC$DATABASE directory also contains the daily data files unless you specify an alternate directory in the collection definition.

Note: If the path is changed, the data files and history files for that node must be moved from the original location to this new location.

If a database disk has fewer than the required number of free blocks, any data collection processes stop recording data in the database. The primary Performance Agent continues to collect data but stops recording the data until there are once again enough free blocks.

If you enter the ADVISE COLLECT SHOW ALL command, the required space is displayed under “DskFreeSpace” and the Performance Agent status shows as “FreeSpaceWait.” The Performance Agent then hibernates until sufficient free space becomes available.

To adjust the free-space threshold, use the following command format:

ADVISE COLLECT MODIFY collection-name * /FREE_SPACE = number

If this threshold is already low, it is better that you free up additional space on the disk.

Normally you never want to run out of database disk space because this will cause gaps in your data and limit the use of this data in reports, graphs, or modeling. Use the collection parameters and collection schedule to control the amount of data written to the database directory. Archiving also contributes to the use of disk space; see Chapter 4 for more information.

Optionally, you can set up a secondary database on another device and modify some of the nodes' database paths to point to these directories.