PingDirectory

Monitoring disk space usage

The disk space usage monitor provides information about the amount of usable disk space available for the server components.

The disk space usage monitor evaluates the free space at locations registered through the DiskSpaceConsumer interface by various components of the server. Disk space monitoring excludes disk locations that do not have server components registered. However, other disk locations might still impact server performance, such as the operating system disk, if it becomes full. When relevant to the server, these locations include the server root, the location of the /config directory, the location of every log file, all Java Runtime Environment (JRE) backend directories, the location of the changelog, the location of the replication environment database, and the location of any server extension that registers itself with the DiskSpaceConsumer interface.

The disk space usage monitor provides the ability to generate administrative alerts and take additional action if the amount of usable space drops below the defined thresholds.

You can configure three thresholds for this monitor:

Low space warning threshold

This threshold is defined as either a percentage or absolute amount of usable space. If the amount of usable space drops below this threshold, then the server generates an administrative alert but remains fully functional. It generates alerts at regular intervals that you configure (such as once a day) unless action is taken to increase the amount of usable space. The server generates additional alerts as the amount of usable space is further reduced, such as each time the amount of usable space drops below a value 10% closer to the low space error threshold. If an administrator frees up disk space or adds additional capacity, then the server should automatically recognize this and stop generating alerts.

Low space error threshold

This threshold is also defined as either a percentage or absolute size. If the amount of usable space drops below this threshold, the server generates an alert notification and begins rejecting all operations requested by non-root users with UNAVAILABLE results. The server should continue to generate alerts during this time. When the server enters this mode, an administrator has to take some kind of action, such as, running a command to invoke a task or removing a signal file, before the server resumes normal operation. This threshold must be less than or equal to the low space warning threshold. If they are equal, the server begins rejecting requests from non-root users immediately upon detecting low usable disk space.

Out of space error threshold

This threshold can also be defined as a percentage or absolute size. If the amount of usable space drops below this threshold, the server generates a final administrative alert and shuts itself down. This threshold must be less than or equal to the low space error threshold. If they are equal, the server shuts itself down rather than rejecting requests from non-root users.

The server monitors disk space consumption during processing for the export-ldif, rebuild-index, and backup tools. Space is monitored every 10 seconds if usable space for all monitored paths is greater than 15 percent of the capacity of those volumes. If usable space for any path drops below 15 percent, or below 10GB free, the space check frequency is increased to every second. Warning messages are generated if available space falls below 10 percent, or below 5GB free. If usable space for any path drops below two percent, or 1GB free, the tool processing is aborted and files can be removed to free up space.

The default configuration uses the same values for the low space error threshold and out of space error threshold. This is to prevent having the server online but rejecting requests, which causes problems with applications trying to interact with the server. The low space warning threshold generates an alert before the problem becomes serious, well in advance of available disk space dropping to a point that it is critical.

The default values might not be suitable for all disk sizes, and should be adjusted to fit the deployment. Determining the best values should factor in the size of the disk, how big the database might become, how much space log files might consume, and how many backups are stored.

The threshold values can be specified either as absolute sizes or as percentages of the total available disk space. You must specify all values as absolute values or as percentages. A mix of absolute values and percentages cannot be used. The low space warning threshold must be greater than or equal to the low space error threshold, the low space error threshold must be greater than or equal to the out of space error threshold, and the out of space error threshold must be greater than or equal to zero.

If the out of space error threshold is set to zero, then the server does not attempt to automatically shut itself down if it detects that usable disk space has become critically low. If the amount of usable space reaches zero, then the database preserves its integrity but might enter a state in which it rejects all operations with an error and requires the server, or at least the affected backends, to be restarted. If the low space error threshold is also set to zero, then the server generates periodic warnings about low available disk space but remains fully functional for as long as possible. If all three threshold values are set to zero, then the server does not attempt to warn about or otherwise react to a lack of usable disk space.