Log file tracking

The system writes many kinds of log files to its logs and other directories. For the most part, these logs are troubleshooting tools for software developers and not intended for end users. Experienced users, however, might gain insights into processing activity by examining certain of these files. Log files contain a complex array of data that takes practice to interpret.

Detailed information about system events that users might find useful and how to manage and route them to various log files is in System events.

The following topics describe the log files.

Event log s

This log contains selected events from the event subsystem. These are events related to message processing activity. If you want to use a log to monitor processing activity, this is the log you may want to examine. The log is server_events.log and is written to <trading engine install directory>\logs.

System logs

System logs contain formatted, time-stamped information reported by various components of the application. The logs, intended for use by software developers in troubleshooting, are not supported for end users. Activator uses the Apache Jakarta Project's log4j framework to format and manage the logs, which are generated by each Java virtual machine during runtime. The log4j.properties file in < install directory>\conf can be edited to generate debug level events in log files.

See Troubleshooting with the log4j file for information about changing and using the file.

The system names the logs based on the names of the source JVM node. They are written to < install directory>\logs. The logs are:

  • hostname.ex.log is created by the system Executive node.
  • server.log reports processing activity of Activator.

System logs can report four levels of events. These are, in order of verbosity:

  • Error messages indicate a possibly serious error affecting service.
  • Warn messages typically have operational significance, but might not affect service.
  • Info messages provide general processing information that could be useful in troubleshooting.
  • Debug messages usually have much detailed information that can be useful in troubleshooting. This level should be turned on only when necessary, as it can degrade system performance and use large amounts of disk space.

The four logging levels are cumulative. Error messages are included at the warn level, both of which are included at the info level, and everything is logged at the debug level. The debug level also produces further details about processing.

System statistics logs

All logs file appended with stats.log are Java Management Extensions (JMX) monitoring and statistics logs.

Although not for end users, these logs are an aid in troubleshooting. If you contact Technical Support for help, a technician may ask you to send copies of these files.

User interface logs

These logs are created by the user interface. Like system logs, they do not contain information useful to the typical end user. These logs are written to <install directory>\logs\ui.

The log file names use as a prefix the name of the computer on which Activator software is installed. The names have the following format: <hostname>_cn_access.log.000001. The cn stands for control node. The trailing number (000001) identifies rolling logs. Once a log reaches a certain size, events write to another log (000002). This rolling occurs up to five times before events again start writing to the first log file.

HTTP server logs

These logs are written by the embedded HTTP server when a control node or service node is started with the -Ddebug startup option.

Class path logs

Log files appended with classpath contain information about all JARs in class paths.

Console logs

Log files appended with console contain standard output ( stdout) and standard error ( stderr) output streams from embedded third-party software.

Related topics

Related Links