About log compression

The server supports the ability to compress log files as they are written.

This feature can significantly increase the amount of data that can be stored in a given amount of space so that log information can be kept for a longer period of time.

Because of the inherent problems with mixing compressed and uncompressed data, compression can only be enabled at the time the logger is created. Compression cannot be turned on or off when the logger is configured. Because of problems in trying to append to an existing compressed file, if the server encounters an existing log file at startup, it rotates that file and begin a new one rather than attempting to append to the previous file.

Compression is performed using the standard gzip algorithm, so compressed log files can be accessed using readily available tools. The summarize-access-log tool can also work directly on compressed log files rather than requiring them to be decompressed first.

However, because it can be useful to have a small amount of uncompressed log data available for troubleshooting purposes, administrators using compressed logging might want to have a second logger defined that does not use compression and has rotation and retention policies that minimizes the amount of space consumed by those logs while still making them useful for diagnostic purposes without the need to decompress the files before examining them.

Configure compression by setting the compression-mechanism property to have the value of gzip when creating a new logger.