Java logging
Java logging is the logging toolkit provided by the java.util.logging package. Java logging provides a standard logging API for the applications.
Message logging (messages) and diagnostic trace (trace) are conceptually similar, but do have important differences. These differences are important for application developers to understand to use these tools properly. The following operational definitions of messages and trace are provided.
- Message
- A message entry is an informational record intended for end users, systems administrators, and support personnel to view. The text of the message must be clear, concise, and interpretable by an end user. Messages are typically localized and displayed in the national language of the end user. Although the destination and lifetime of messages might be configurable, enable some level of message logging in normal system operation. Use message logging judiciously because of performance considerations and the size of the message repository.
- Trace
- A trace entry is an information record intended for service engineers or developers to use. As such, a trace record might be considerably more complex, verbose, and detailed than a message entry. Localization support is typically not used for trace entries. Trace entries can be fairly inscrutable, understandable only by the appropriate developer or service personnel. It is assumed that trace entries are not written during normal runtime operation, but can be enabled as needed to gather diagnostic information.
The application server redirects the system streams at the server startup. There is no way to allow the application to output logging to the console because the system streams can not be obtained by the application. If we would like to use console to monitor the application without using the console handler, we can either monitor the SystemOut.log file, or monitor a file created by another file handler.
The application server uses Java logging internally and therefore certain restrictions apply for using system streams with this logging API by applications. During server startup, the standard output and error streams are replaced with special streams that write to the logging infrastructure, in order to include the output of the system streams in the log files. Because of this, applications can not use java.util.logging.ConsoleHandler, or any handler writing to SystemErr.log or System.out streams, attached to the root logger. If the user does attach the handler to the root logger, an infinite loop is created within the logging infrastructure, leading to stack overflow and server crash.
If the use of a handler that writes to system streams is necessary, attach it to a non-root logger so that it does not publish log records to parent handlers. The data written to the system streams is then formatted and written to the corresponding system stream log file. To monitor what is being written system streams, the configured log files (SystemOut.log and SystemErr.log by default) can be monitored.
IBM recommends using the HPEL log and trace infrastructure. With HPEL, one views logs using the LogViewer command-line tool in PROFILE/bin.
Related concepts:
Loggers
Log handlers
Log filters
Log formatters
Related
Troubleshoot applications with HPEL
Reference:
Log levels