Metacat Performance on KNB: Bring the logging level down on ecoinfo.
Dont know how much this effects the performance but we might be logging more
than required on the production server. To bring down the number of I/O
operations to the hdd involved, we should bring down the debug level on KNB
#2 Updated by Saurabh Garg almost 15 years ago
Integrating log4j in Metacat. The debug levels will be as specified before:
Debug -- > 50
Info -- 25 - 50
Warn -- 15 - 25
Error -- 05 - 15
Fatal -- < 05
The exception to this rule is any use of log4j in catch() will default to error
irrespective of the debug level set in the statement. If it is obvious that the
level should either be set to fatal or warn, then those levels can be used.
As evident from the above described levels, Info and Debug will be used more
like Debug 1 and Debug 2 levels. Hence, by default, the configuration in
log4j.properties would specify printing Fatal, Error and Warn levels. While
debugging, level can be increased to Info or Debug.
Also log4j can be configured to just print INFO and DEBUG messages in some
Currently, the format of the logs generated specified in log4j is as follows:
Metacat [<debug level>]: <message> [<class name>]
e.g. Metacat: [WARN]: The session is not new or has sessionid parameter
This can be modified to include other info like date & time, line number of the
code where message was generated.
debugMessage and debugFormattedMessage will be removed and replcaed by
printMessage. printMessage could be used for messages which are not errors,
warnings or debug messages. Example of such messages are:
Metacat: Running recycle thread or not: off
Metacat: Cycle time of recycle: 30000
Metacat: Indexing paths specified in metacat.properties....
Metacat: Path Indexing Completed
Metacat: Metacat (1.6.0) initialized.