Periodically, each management server in a site checks the status of its agent computers.
There are registry keys that control how the management server determines the status of the agent computers it communicates with. These HKEY_LOCAL_MACHINE\Software registry keys are under \NetIQ\AppManager\4.0\NetIQms\Config. Because communication is handled differently for Windows agent computers and UNIX agent computers, there are separate keys for checking the status of Windows agent computers and UNIX agent computers.
By default, the machine polling thread for Windows runs on the management server every 15 minutes. At each interval, the management server receives an updated list of its current agent computers and checks the availability of the designated primary management server for those agent computers.
Before changing this interval, you should evaluate the potential impact on your environment. If you lengthen the interval, it will take longer for job property or job status changes to be passed to your agent computers if the primary or backup management server fails. If you shorten the interval and have a large number of agent computers, it will increase the processing load on the management server and may degrade throughput performance. In general, if you have a large number of agent computers, you should not change the machine polling interval.
To change the machine polling interval for Windows agent computers:
In the Windows Registry Editor, expand \HKEY_LOCAL_MACHINE, to \SOFTWARE\netiq\appmanager\4.0\netiqms\config.
Double-click Machine Poll Interval to specify the number of seconds between updates. The default is 900 seconds. If desired, you can click the Decimal option to display the current value in decimal format.
Click OK.
For UNIX and Linux computers, the management server uses the agent heartbeat to determine the status of its agent computers. The registry keys that control how the management server determines the status of the NetIQ Corporation UNIX agents are the Unix Machine Check Interval and the Unix Machine Timeout keys.
At each Unix Machine Check Interval, the management server checks the timestamp of the last heartbeat signal from each of its UNIX agents. If the timestamp indicates that the UNIX agent has not sent a heartbeat signal within the period of time specified for the Unix Machine Timeout, the management server considers the UNIX agent unavailable and passes this information back to the repository and the computer is grayed out in the Operator Console.
Before changing the interval or the timeout period, you should consider the potential impact on your environment. If you lengthen the interval or the timeout setting, it may take longer to be notified when UNIX agents stop communicating with the management server. If you shorten the interval or timeout setting and have a large number of agent computers, it will increase the processing load on the management server and may degrade throughput performance. You should also keep in mind that these registry keys work in conjunction with each other so any changes should take in account both values.
To change the Unix Machine Check Interval or the Unix Machine Timeout period:
In the Windows Registry Editor on the management server, expand \HKEY_LOCAL_MACHINE, to \SOFTWARE\netiq\appmanager\4.0\netiqms\config.
Double-click Unix Machine Check Interval to specify the number of seconds between status checks.
This interval controls how often the management server checks the timestamp of the last heartbeat signal from each of its UNIX agents. The default is 300 seconds. If desired, you can click the Decimal option to display the current value in decimal format.
Double-click Unix Machine Timeout to specify the maximum number of seconds between heartbeat signals.
If the UNIX agent does not send a heartbeat signal within this period of time, it is deemed unavailable. The default is 1200 seconds. If desired, you can click the Decimal option to display the current value in decimal format.
NOTE:If you change the UNIX heartbeat interval, you may need to adjust the Check and Timeout intervals. For example, if you set a longer heartbeat interval to conserve network bandwidth, you should lengthen the Unix Machine Check and Unix Machine Timeout intervals to prevent the UNIX agent from appearing to be unavailable between heartbeat signals.
Click OK.
After you modify the registry entries, you must restart the NetIQ Corporation AppManager Management Service (NetIQms) for the changes to take effect. To restart the NetIQ Corporation AppManager Management Service (NetIQms), use the Services Control Panel.