Configuring recovery actions
Understanding recovery actions
For the sake of discussion, the recovery actions that you define in the KM using the PATROL console are referred to as PATROL native recovery actions. PATROL native recovery actions are different from PATROL KM for Event Management recovery actions; this difference is described as follows:
- PATROL native recovery actions: When you define PATROL native recovery actions in the PATROL console, you associate the recovery actions with alarm and border ranges. These recovery actions run when the PATROL parameter value enters the specified range. The parameter might be in an OK, WARN, or ALARM state when the recovery action runs, depending on how you configure the parameter.
- PATROL KM for Event Management recovery actions: Unlike PATROL native recovery actions, the PATROL KM for Event Management Recovery actions run only when a parameter changes status. For example, when a parameter goes from an OK state to a WARN or ALARM state, or even when a parameter goes from an ALARM to an OK state.
If you do not want the parameter to alarm until recovery actions have been attempted, you must use PATROL native recovery actions, rather than PATROL KM for Event Management recovery actions. However, you can use both types of recovery actions.
For example, you can define PATROL native recovery actions and specify that the parameter enters a WARN or ALARM state only after all recovery actions fail. Then you can create a PATROL KM for Event Management recovery action that runs only if the PATROL native recovery actions fail.
For more information about using PATROL KM for Event Management recovery actions, see the PATROL KM for Event Management online documentation portal.
Built-in native recovery actions
By default, PATROL for Microsoft Windows Servers provides the following built-in recovery actions that are associated with the specified parameter:
Built-in recovery actions
Recovery action | Parameter | Description | Runs automatically? |
|---|---|---|---|
Backup and Clear Event Log (PATROL KM for Microsoft Windows OS) | NT_EVLOGFILES\ ELMEvFileFreeSpacePercent | Backs up the event log file and clears all events. | Yes |
Start Windows Management Instrumentation Service Check (PATROL KM for Microsoft Windows OS) | NT_HEALTH\ WMIAvailability | Restarts the WINMGMT service when PATROL determines that it is unavailable. | Yes |
Clean Temporary Directories (PATROL KM for Microsoft Windows OS) | NT_LOGICAL_DISKS\ LDldFreeSpacePercent | Clears the temp directory. | No |
Terminate Process (PATROL KM for Microsoft Windows OS) | NT_PROCESS\ PROCProcessorTimePercent | Attempts to stop a runaway process. | No |
Restart Process (PATROL KM for Microsoft Windows OS) | NT_PROCESS\ PROCStatus | Attempts to restart the process. Note: The process is restarted under the PATROL Agent default account, even if the process was previously started under a different account. | Yes |
Restart Service (PATROL KM for Microsoft Windows OS) | NT_SERVICES\ ServiceStatus | Attempts to restart the service. | Yes |
Restarting a PATROL Agent on a remote server recovery action (PATROL KM for Windows Domain) | NT_REMOTE_SERVERS\ MsPatrolAgentStatus | Attempts to restart the PATROL Agent on the remote computer after alarming for 2 collection cycles. | No |
Increase connections to DFS root recovery action (PATROL KM for Windows Domain) | NT_DFS_ROOT\ DfsConnectionPercent | Increases the connection share limit to DFS Root after alarming for 2 collection cycles. | No |
Replication Failure: Initiate WINS Scavenging (PATROL KM for Windows Domain) | NT_WINS_PARTNER\ WpReplicationFailures | Cleans up the WINS database after alarming for 2 collection cycles. | No |
Increase connections allowed to share (PATROL KM for Windows Domain) | NT_Shares\ ShConnPercent | Increases the share connection limit after the ShConnPercent parameter alarms for 2 consecutive collection cycles. | No |
PAWorkRateExecsMin Recovery Action (PATROL KM for Microsoft Windows OS) | PATROL_NT\ PAWorkRateExecsMin | Sets the scheduling policy value to 9 (Schedule Force Delta and Schedule From End). When the parameter goes out of the alarm state, the scheduling policy value returns to the default value of 1. | Yes |
For more information about the configuration tasks, see: