Trace parameters that can be proactively activated

Activating a trace parameter in a production system causes additional load and causes log files to grow quickly, and therefore reduces the time interval covered by log rotation.

The following trace parameters can be activated in production environments even for a long time.

Parameter Description

msglog:1

Writes all on-screen messages to the log file as well

sqllimit:2 (or higher)

Writes a message about the slow execution of any SQL statement (SELECT, INSERT, UPDATE, DELETE or others)

debugdbquery:2 (or higher)

Writes a message about slow SM queries taking longer than the specified number of seconds. The adding, updating, or deleting of records is NOT logged.

Caution Do not set it to 999 because this is a special and verbose trace.

rtm:2

Writes information about how long a user transaction took. This is basically the time from a user action until the next format is displayed.

This parameter also causes information about native heap consumption to be printed to the log file so that sessions consuming a lot of memory can be identified.

log4jdebug: com.hp.ov.sm.common.oom.LowMemoryHandler

 

The log4jdebug parameter is used to enable debugging messages for different java modules. The LowMemoryHandler module may be activated during production to monitor for a low memory situation of the servlets.

The following are examples of verbose trace parameters that must NOT be activated by default in a production system:

  • rtm:3|4|5
  • debugdbquery:999
  • sqldebug:1|2
  • debughttp

Activate these trace parameters only for special case analysis. It is recommended to contact HPE Software Support if an issue cannot be analyzed otherwise. Refer to Running traces on the Service Manager server to identify the method to activate these parameters that creates the least additional system load and log file size.