You might be familiar with the change in licensing conditions with CM V8 whereby a concurrent user is defined as someone who has initiated a Library Server request within the past 30 minutes. Customers are required to have a license for the maximum number of such users, but it is very difficult for them to monitor complicance with this condition. CM V8.2 goes some way towards providing this compliance monitoring capability with the new Library Servier monitor task.
Every 30 minutes this task will write a record to the Library Server SysAdmin Event Log table (ICMSTSYSADMEVENTS) recording the number of concurrent users at that point in time. An appropriate SQL query would be able to work out the maximum figure recorded for any day, week month or other period. Even though this is only a snapshot every 30 minutes, it is considerably better than what we have today.
To activate this function, you need to have tracing switched on. In the System Administration Client, go to ICMNLSDB, Library Server Parameters, Contigurations, Library Server Configuration, Properties.
In the Definition tab, set Max users to a very high number (it is usually 0 to allow any number of users logged on). On the Log and trace tab, set the Trace Level to Basic trace.
You should then see an entry made in the ICMADMIN.ICMSYSADMEVENTS table every 30 minutes for record type 209 which will look as follows:
db2> select eventcode,created,userid,eventdata1 from icmstsysadmevents where eventcode = 209This record does not list the users - it merely provides a count of those users who have performed some interaction with the Library Server within the previous 30 minutes. If you don't want to wait 30 minutes, set an environment variable ICMCOUNTINTERVAL to some number of seconds, the start ICMPLSAP from the command line.
EVENTCODE CREATED USERID EVENTDATA1
209 2005-01-05-09.13.36.143234 ICMADMIN 2
Be careful with this trace: When turning on this trace on a CM 8.2 system, it might impact the federated searches on CM 8.2 and CMOD 7.1.1 in such a manner, that eventually CMOD hangs. This situation was seen with CM 8.2 system on fix level 3: Federated searches hanged and caused the ARS processes on the AIX server to hang also; eventually using up all available ARS database processes, and thus disabling federated searches. After turning off tracing and rebooting the AIX system afterwards, the problem will be gone.
If you found this useful, here's more on the same topic(s) in our blog:
- DB2 Tips
- Icmrmvolval and finding the document
- Running icmrm as a Windows service
- Tests to see if ICM is working correctly
- DB2 catalog for databases in different instances
Interested in learning more?