Tuning OMEGAMON® components
Once configuration is complete, you can tune some components to make them operate more efficiently. You can change data collection options, change the default value for short-term history, and tune Tivoli® Data Warehouse.
This section includes some tuning information about the components of the IBM® Z OMEGAMON® AI for Networks monitoring agent that you might want to implement once the monitoring agent is installed.
- Changing data collection options
While an IBM® Z OMEGAMON® AI for Networks monitoring agent is running, you can change the data that is being collected by using z/OS® MODIFY commands to temporarily change how often data collection is performed.
Using z/OS MODIFY commands, you can reduce the z/OS system resources used to monitor your networks while a larger than normal workload is being performed. You can also monitor more of your network resources while you are investigating a problem.
Changes made using the z/OS MODIFY command are temporary. The next time that the monitoring agent is started, the agent reverts to the definitions made using the PARMGEN configuration method. In other words, the changes made using the z/OS MODIFY command are not retained when the monitoring agent is stopped.
Changes associated with the tolerate TCP/IP recycle option reimplement the configured monitoring options whenever the state of a network monitoring interface (NMI) AF-UNIX socket changes, which may further reduce the effects of changing the monitoring options using the command line interface. Monitoring options that require permanency should be made using the PARMGEN configuration method.
The following z/OS operator command causes the monitoring agent to stop collecting applications and connections data from the TCP/IP address space named TCPIPB.
The output from this command is:MODIFY proc_name,KN3FCCMD STOP CONN TCPNAME(TCPIPB)
KLVOP191 'KN3FCCMD STOP CONN TCPNAME(TCPIP)' KN3C115I STOP FOR COMPONENT CONN ACCEPTED. TCPNAME: TCPIP KN3FC000 KN3FCCMD PROCESSING COMPLETE
Note: After you have issued this command, you might experience a delay (the default is up to 30 minutes) before this output is written to RKLVLOG. To see the output immediately, enter the following z/OS operator command:MODIFY proc_name,FLUSH
To reset to the data collection you specified during configuration of the monitoring agent, you can either stop and start the started task procedure or issue the following z/OS operator commands to stop all data collection and start the data collection as configured using the PARMGEN configuration method:
MODIFY proc_name,KN3FCCMD STOP TCPC MODIFY proc_name KN3FCCMD START TCPC
For more information, see topic KN3FCCMD command reference in the Parameter Reference Guide.
- Tuning the TMS:Engine LIMIT value
IBM® Tivoli Management Services: Engine (TMS:Engine) is a common component for both the Tivoli Enterprise Monitoring Server and for IBM® Z OMEGAMON® AI for Networks Monitoring. It has startup parameters that are defined with appropriate defaults for many customer environments, and these parameters are defined in the data set pointed to by the RKLVIN DD statement in the started task procedure.
If you see this message:
An error has occurred when the monitoring agent tried to return data for the table specified by the table variable and failed. The most common cause of this error is a query that returns a large number of rows of data, causing an out-of-memory condition.KN3I008E CALL TO ADDDATA FUNCTION FAILED, RETURN CODE=retcode, FOR TABLE=table
Options are to either modify the query so that it returns fewer rows of data or change the LIMIT and MINIMUM values in &rhilev..&midlev.RKANPARU.
The current recommendation for the IBM® Z OMEGAMON® AI for Networks Agent is MINIMUM(768000,X) and LIMIT(24,X) as specified in member (KN3SYSIN). The current recommendation for the Tivoli Enterprise Monitoring Server in (KDSSYSIN) is MINIMUM(768000,X) and a LIMIT value of 24 or greater.
The LIMIT parameter can be used to specify the largest block of primary storage or extended storage that can be allocated. The syntax for setting the largest block of extended storage is shown in the example (note that setting the limit for primary storage is not recommended): LIMIT(n,X)
This value is specified in bytes, as a power of 2. For example, if n is 22, the largest block that can be allocated is 4 MB. If the LIMIT value is too small and a process in ITMS:Engine attempts to allocate a block of storage larger allowed by LIMIT, a program interruption U0100 or U0200 results When managing a large number of connections or TN3270 sessions, use a value of LIMIT (25,X) or greater.
- Changing the default value for short-term history from 24 hours
By default, historical data from the last 24 hours is retrieved from the persistent data store for display at the Tivoli Enterprise Portal. Historical data from earlier time periods is retrieved from the Tivoli Data Warehouse.
You can alter this default by modifying the KFW_REPORT_TERM_BREAK_POINT environment variable in the kfwenv file on the Tivoli Enterprise Portal Server. Information about the environment variable is found in IBM Tivoli Monitoring: Administrator's Guide. This variable is specified in seconds. The default is 86400 (= 60 seconds * 60 minutes * 24 hours). Increasing this value will result in historical data being retrieved from the persistent data store for time periods farther in the past. Decreasing this value will result in historical data being retrieved from the Tivoli Data Warehouse for more recent time periods.
You might want to increase the value of this environment variable if you have not implemented Tivoli Data Warehouse. By increasing the value, you can view more historical data, assuming the persistent data stores are allocated large enough to contain more than 24 hours of data.
You might want to decrease the value of this environment variable if you configured the warehouse interval to be one hour. By decreasing the value, you can view all historical data with smaller persistent data stores.
- Tuning considerations for the Tivoli Data Warehouse
The default database manager settings provided by DB2® may not be sufficient to support warehousing of large amounts of monitor data from IBM® Z OMEGAMON® AI for Networks. Review the appendix
Relational database design and performance tuning for DB2 database servers
appendix in the IBM Tivoli Monitoring: Administrator's Guide to learn about DB2 tuning considerations.