Logging

Meridian stores logs based on the name of the daemon that was acting on a process. For example, logs for service assurance checks that are done by pollerd are stored in the pollerd.log file.

The log files are available in ${OPENNMS_HOME}/logs, which is a symlink to /var/log/opennms.

Check the list of daemons to see which log files are associated with each daemon.

Change logging

Meridian uses the Apache Log4j framework to manage how the logs are written, structured, and rotated.

The log configuration file is available in ${OPENNMS_HOME}/etc/opennms/log4j2.xml.

Each daemon can have an individual logging level.

The corresponding configuration in this file can be found in the XML tree path: configuration/loggers/root/DynamicThresholdFilter/KeyValuePair.

The default logging level is WARN; for troubleshooting, use DEBUG. See the table below for a description of the available log levels. Changes to logging levels are applied immediately and do not require any service or daemon restarts.

Don’t forget to set the logging level back to WARN after changing it to a higher severity. Depending on the scale, Meridian can generate a lot of logs, which requires more disk space.

Logging levels

Table 1. Service daemon logging level options
Level Description

INFO

Anything likely to be useful to the average user to determine the current general behavior/state of the system.

WARN

Messages relating to code that has a failure that might be temporary or recoverable, or code that is misconfigured but can continue running in a regressed state. This is the recommended default log level.

ERROR

Only things that cause code to stop working and be unrecoverable.

DEBUG

Additional information that can help diagnose issues.

TRACE

Even more additional information that is often useful only to developers. This mode will output very detailed logs and can cause performance issues if left enabled. Use sparingly as necessary.