Collecting data for IBM Content Manager Enterprise Edition library server

Answer

Collecting library server log

Library server logging setup is same for Content Manager 8.3 and 8.4. There are two ways:

• With Content Manager system administration client:
1. Go to Tools-> Log Configuration-> "Library Server".
2. Specify directory for the log file in "Log file path" field and log file name in the "Log file name" field; use "Logging level" drop down list to select logging level; use "User ID" to specify a specific ID to log.

Notes:
• Please use absolute path for the log directory. Use "/" in the path on UNIX systems and "\" on Windows. "\\" is not necessary on Windows.
• In the "General" panel of Log Configuration, you can also configure library server logging by enabling the checkbox for "Library Server." Note that the log level drop down list doesn't reflect the current logging level when you open the panel.


• If System Admin Client is not available, the logging level can be set directly in library server database. This is actually the manual way to implement what the system administration client does. From a DB2 CLP console:
1. db2 connect to icmnlsdb user icmadmin using password
2. db2 select tracefilename, tracelevel, traceuser from icmstsyscontrol
3. db2 update icmstsyscontrol set tracelevel=15
Notes:
• The select query in step 2 is to find out the current trace location, tracelevel, and traceuser.
• For tracelevel value, 0 is error level and 15 is the debug level being used normally. Some not so common values are 8 for performance logging only and 63 or 31 for more detailed debug logging.
• If traceuser is empty then all users are traced.
• The multipletracefiles column in icmstsyscontrol table should be kept as 0. It should only be used when being suggested by debugging support engineers. Enable multiple trace file may result in deadlock situations.
• The keeptraceopen column in icmstsyscontrol table should be kept as 0. It's a reserved parameter.

• Library server logging level can be dynamically changed so library server database instance doesn't have to be restarted for it to take effect. The change takes effect in one minute.
• Set logging level back to error level/0 right after the test so that the log will stop growing fast and we will get a clean log.
• Before the test, delete the existing log if it's large. The normal way is to change the trace file name so that library server will start a new log and the old log will become deletable.
• Library server monitor process locks the library server log. On Windows, library server monitor process can be found in the Services control panel. On UNIX, the process is called icmplsap if DB2 is the database or icmorlsap if Oracle is the database.
• Note down the steps when reproducing the problem. Using a special ID that's not shared by other users to logon library server will be very helpful for log reading.
• Note down the library server machine clock when the first logon to library server happens during the test. Let support engineer know the time difference between the machine time zone and GMT. Library server log uses GMT timestamp.

• IBM Support Assistant (ISA)
• Electronic Service Request (ESR)
• FTP to the Enhanced Customer Data Repository (ECuRep)