This topic provides information about configuring a process to suspend its alert for a designated number of cycles when PATROL detects that the process is not running. This feature gives the system time to detect that a process has died and restart the process before PATROL issues an alarm. Delaying an alert also helps by giving the host group time to restart a process that died on an alternate host.
In the Monitored Process String field, type the command or a unique command string for the process that is being monitored.
Note
PATROL monitors all processes that match the string you type in this field. When you enter text in this field, you may end up monitoring multiple processes. For example, if you type vi in this field, PATROL will monitor processes named vi, view, and previous.
In the Minimum Count field, type a value to set the minimum number of process instances that may be running on the local computer or in the host group. To monitor multiple instances of the same process, this value must be set to 2 or greater. If the number of running process instances falls below this value, PATROL will issue an alert.
In the Maximum Count field, type a value to set the maximum number of process instances that may be running on the local computer or in the host group. If the number of running process instances exceeds this value, PATROL will issue an alert.
Note
The value in the Maximum Count field must be equal to or greater than the value in the Minimum Count field.
Steps 9 through 11 are optional when you are configuring PATROL to delay an alert for a process that has died. If you do not want to set values for the fields as indicated in these steps, you do not have to.
In the Acceptable Process Owners field, type the user IDs for the accounts that may own the process. Separate multiple user IDs with spaces.
Select the Use Process Owners for Filtering check box and click OK.
Note
Because the process filtering for the processes is based on owners, the owner of the processes is always a subset of the provided owner set. Thereby, the PROCPPOwnerCheck parameter is deactivated when the Use Process Owners for Filtering check box is selected.
Select the Parent Process ID Must Be 1 option button if the parent process ID (PPID) must be 1. A process with a PPID of 1 is owned by init, or the UNIX scheduler.
Select the Restart Automatically option button if you want the KM to automatically restart a process when it detects that the process count has dropped below the set minimum. If you select this option button, the KM will attempt to restart the process when it detects that the process instance count has fallen below the specified minimum threshold. The KM uses the value in the Command Execution Attempts field to determine how many times it will try to restart a process.
Note
You must specify a start command and a command execution user name in the appropriate fields on this dialog box if you want PATROL to automatically restart a process.
In the Command Execution Attempts field, type a value set the number of times the host will attempt to run a Start Process or Stop Process command before it stops trying to run the command.
In the Start Command field, type the command string that will start the process instance.
Note
You must specify a command execution user account and password if you want to use the Start command.
In the Stop Command field, type the command string that will stop the process instance. This step is optional when you are configuring PATROL to delay an alert for a process that has died.
Note
You must specify a command user account and password if you want to use the Start command.
In the Command Execution User Name field, type the user ID under which the command will be executed.
Note
You must specify a command execution user account and password if you want to use the Start and Stop commands.
In the Command Execution Password field, type the password for the user ID under which the command will be executed.
You must specify a command execution user account and password if you want to use the Start and Stop commandsIn the Alert Delay Count field, type a value to set the number of collection intervals that this host will defer an alert while it waits for the process count to be reestablished across the host or group.
If you delay the alert, the system has time to detect that a process has died and restart it automatically before PATROL issues an alarm.
Click the drop-down list button in the Alert State field and select the state change that will occur when the minimum or maximum process count is exceeded and the alert delay count reaches 0. The state change will apply to the following parameters:
Select one of the following option buttons to indicate whether PATROL should monitor the process on one host computer, or across a group of hosts:
Click OK.
The dialog box closes, and the Manage List of Monitored Processes dialog box is displayed. The process that you just added to the monitored queue is listed in the Process List group box.
Using the PATROL KM for UNIX for Process Monitoring