In my last post about how to send log messages to different log files, I’d promised to write about additivity. And, here is a post for keeping my promise. I hope that after reading my previous article, now you know how to configure Log4j for logging into multiple files using file appenders. But, there is a catch in that technique. To explain this clearly, consider this scenario:
You’ve configured a total of three appenders in your application. One for the package com.demo.moduleone and one for com.demo.moduletwo and one root logger for com.demo. The log4j configuration will look something like this (showing only the appender configuration, excluding other details)
The Log4j loggers are following hierarchies. ie, A log4j logger is said to be an ancestor of another logger if its name followed by a dot is a prefix of the descendant logger name. A log4j logger is said to be a parent of a child logger if there are no ancestors between itself and the descendant logger.
So, as per the hierarchy, our rootFileAppender is the parent appender for both moduleOneFileAppender and moduleTwoAppender. So, all the log messages that are coming to the child appenders will be propagated to the parent appenders too. So, in our scenario, the log messages from the package com.demo.moduleone will be sent to the moduleOneFileAppender plus the rootFileAppender. The same applies to the com.demo.moduletwo also. This leads to write the same log message in two different location.
How to avoid this redundancy?
In order to avoid this redundancy, we can use Log4j additivity. Just set the additivity property of an Log4j logger to false and then the log messages which are coming to that logger will not be propagated to it’s parent loggers. So, our new Log4j configuration file would be:
With the above configuration, the log messages from the com.demo.moduleone will go to the moduleOneAppender only and the rest of the log messages will go to the rootFileAppender.