17.2.2 Recovery method when the KAVE00493-E message is output and services cannot start

When a physical host name used as the monitoring host name of PFM - Manager, PFM - Agent, or PFM - RM is changed improperly, subsequent service startup is suppressed. In addition, the KAVE00493-E message is output to avoid data being corrupted. In such a situation, identify the current physical host name and then recover from the event by using the methods shown below according to the type of service to start.

Reference note: If you cannot identify the current physical host name, execute the jpcconf host hostmode -display command to check the method of obtaining the monitoring host name. You can see the command for checking the physical host name from the method of obtaining the monitoring host name. For details, see the chapter that explains commands in the manual Job Management Partner 1/Performance Management Reference.

Recovery method when a PFM - Manager service cannot start

Recover from the event according to the contents of (1) and (2).

Recovery method when a PFM - Agent or PFM - RM service cannot start

You can recover from the event by correctly performing the procedure of changing the physical host name of the corresponding host again.

For details on how to change the physical host name, see the chapters on installation and setup in the Job Management Partner 1/Performance Management Planning and Configuration Guide.

Note that the operation to change the physical host name itself is not needed because it has already been performed.

Additionally, if the KAVE05217-E message is output during service stop and the service fails to stop, edit the jpchosts file so that the old host name can be resolved, and then re-execute the command. (If the service stops, restore the contents of the jpchosts file.)

Organization of this subsection

(1) Recovery procedure when the KAVE00493-E message is output and services cannot start (PFM - Manager)
(2) Recovery procedure when the KAVE00493-E message is output and services cannot start

(1) Recovery procedure when the KAVE00493-E message is output and services cannot start (PFM - Manager)

To recover from this event, you need to perform the following operation on each host of the monitoring manager, monitoring console server, and monitoring agent, as well as on the monitoring console.

The following figure shows the procedure of the operation.

Figure 17‒1: Recovery procedure when the KAVE00493-E message is output and services cannot start

To Page Top

(2) Recovery procedure when the KAVE00493-E message is output and services cannot start

To recover from this event, you need to match the current physical host name with the monitoring host name to be used in the operation monitoring system, and delete the old information.

Prerequisite conditions

For details on commands to be used for operation, see the chapter that explains commands in the manual Job Management Partner 1/Performance Management Reference.
If a host that you operate is running as a logical host, start or stop services by using the cluster software.
There are commands that need a specific option (-lhost logical-host-name) when a host on which the commands are executed is running as a logical host. Check the details of a command and then execute the command (the example of command execution in the procedure below is for a host running as a physical host).

Operation

Stop services on the PFM - Web Console host.

Stop all the Performance Management programs and services on the PFM - Web Console host connecting to PFM - Manager whose host name was changed. To do so, use the jpcwstop command.

In a multiple-monitoring configuration, perform this step on both the primary and secondary managers.
Stop services on the PFM - Agent or PFM - RM host.

Stop all the Performance Management programs and services on the PFM - Agent or PFM - RM host connecting to PFM - Manager whose host name was changed. To do so, use the jpcspm stop command.
Stop services on the PFM - Manager host.

Stop all the Performance Management programs and services on the PFM - Manager host whose host name was changed. To do so, use the jpcspm stop command.

In a multiple-monitoring configuration, perform this step on both the primary and secondary managers.

Note that if the KAVE05217-E message is output during command execution and the service fails to stop, edit the jpchosts file so that the old host name can be resolved, and then re-execute the command. (If the service can stop, restore the contents of the jpchosts file.)
Change the monitoring host name of the PFM - Manager host.

Change the monitoring host name by executing the jpcconf host hostname command on the PFM - Manager host whose host name was changed.

The following is an example of executing the command to change the physical host name to hostB:

In Windows:
```
jpcconf host hostname -newhost hostB -d d:\backup -dbconvert convert
```
In UNIX:
```
jpcconf host hostname -newhost hostB -d /var/tmp/backup -dbconvert convert
```
Note that when the jpcconf host hostname command is executed, information such as definition information or performance information before the change is all inherited.
Change the settings of the connection-target PFM - Manager (for multiple-monitoring configuration)

Change the settings of the connection-target PFM - Manager by executing the jpcconf mgrhost define command on the PFM - Manager host whose host name is not changed.
If necessary, change the jpchosts file, hosts file, and the DNS settings so that the new host name can be resolved in the Performance Management system.

If the machine needs to restart, change the settings so that the Performance Management programs and services do not start automatically after restart. After operation on the PFM - Manager host is complete, restore the settings so that the Performance Management programs and services start automatically.

If necessary, perform a procedure specific to PFM - Agent.

In a configuration in which PFM - Agent is installed on a PFM - Manager host whose host name was changed, you might have to perform a procedure specific to PFM - Agent.

Table 17‒2: Necessity of PFM - Agent-specific procedure
Configuration		Necessity of procedures and reference
PFM - Agent 09-00 or later is installed on the PFM - Manager host.		The necessity of procedures specific to PFM - Agent differs by PFM - Agent. For details, see the chapter that describes installation and setup in each PFM - Agent manual.
PFM - Agent earlier than 09-00 is installed on the PFM - Manager host.	For the following PFM - Agent: PFM - Agent for Enterprise Applications PFM - Agent for Microsoft SQL Server	You need to perform a procedure specific to PFM - Agent. For details, see the description about procedures specific to PFM - Agent that are necessary when a physical host name is changed in the Job Management Partner 1/Performance Management Planning and Configuration Guide.
	Others	You do not need to perform a procedure specific to PFM - Agent.

If you need to perform a procedure specific to PFM - Agent, perform the procedure described in the reference shown in the table, and then go to the next step.

Start services on the PFM - Manager host.

Start the Performance Management programs and services on the PFM - Manager host. To do so, use the jpcspm start command.

In a multiple-monitoring configuration, perform this step on both the primary and secondary managers.
Change the settings of the connection-target PFM - Manager on the PFM - Web Console host.

Change the settings of the connection-target PFM - Manager on the PFM - Web Console host connecting to PFM - Manager whose host name was changed. To do so, change the initialization file (config.xml). For details, see the part that explains the initialization file (config.xml) in the appendix of the manual Job Management Partner 1/Performance Management Reference.
Start services on the PFM - Web Console host.

Start the Performance Management programs and services on the PFM - Web Console host connecting to PFM - Manager whose host name was changed. To do so, use the jpcwstart command.
Delete the health check agent settings.

If the health check function is used, delete the agent definition of a health check agent with service ID that contains the old host name from PFM - Web Console (delete from the management folder in the Agents tree and the association with alarm tables).

In a multiple-monitoring configuration, perform this step on the primary manager.

For details on how to change the agent definition, see 3. Monitoring Agents or 6. Monitoring Operations with Alarms.
Delete the PFM - Agent or PFM - RM settings.

From PFM - Web Console, delete the agent definition of the PFM - Agent or PFM - RM host with the service ID containing the old host name installed on the same host as PFM - Manager whose host name was changed. Remove this agent definition from the management folder in the Agents tree and also remove the association with alarm tables. In a multiple-monitoring configuration, perform this step on the primary manager.

For details on how to change the agent definition, see 3. Monitoring Agents or 6. Monitoring Operations with Alarms.
Delete service information on the PFM - Manager host.

Even though the PFM - Manager host name is changed, the service information of the Performance Management programs with the old host name remains the same. Therefore, you need to delete unnecessary information. In a multiple-monitoring configuration, perform this step on both the primary and secondary managers.

The types of service information that you need to delete and the method of checking the service information are described as follows:

Service information on the host with the old host name

All the items that are displayed by executing the following command must be deleted:
```
jpctool service list -id "*" -host old-host-name
```
Service information whose service ID contains the old host name

Items whose Service ID column contains the old host name of the items that are displayed by executing the following command must be deleted:
```
jpctool service list -id "*"
```
Service information can be deleted by using the jpctool service delete command.

Delete service information on the host with the old host name by using the following command:
```
jpctool service delete -id "*" -host old-host-name
```
Additionally, delete service information whose service ID contains the old host name by using the following command:
```
jpctool service delete -id "???old-host-name" -host new-host-name
```
If the message KAVE05233-W is issued during command execution because of a service information deletion error, re-execute the command as follows:
```
jpctool service delete -id "*" -host old-host-name -force
jpctool service delete -id "???old-host-name" -host new-host-name -force
```
Note

Even though you execute the jpctool service list command, old service information containing the old host name that remains in the database might not be displayed. Because such service information also needs to be deleted from the database, you must execute the jpctool service delete command shown above.
Apply the service information to PFM - Manager.

Synchronize the service information between PFM - Manager and PFM - Web Console so that the deletion of service information takes effect in PFM - Web Console. To do so, use the jpctool service sync command.

In a multiple-monitoring configuration, perform this step on both the primary and secondary managers.

The time when the service information synchronized by the jpctool service sync command takes effect depends on the version of PFM - Web Console.
Change the business group configuration if needed.

If the PFM - Manager host whose host name was changed is assigned to a business group, you need to change the configuration of the business group.

In a multiple-monitoring configuration, perform this step on the primary manager.

For details on the procedure, see 2.7 Setting and using business groups.
Change the settings for the connection-target PFM - Manager on the PFM - Agent or PFM - RM host.

Change the settings of the connection-target PFM - Manager on the PFM - Agent or PFM - RM host connecting to PFM - Manager whose host name was changed. To do so, use the jpcconf mgrhost define command. For example, if the host name of the connection-target PFM - Manager was changed to hostB, execute the following command:
```
jpcconf mgrhost define -host hostB
```
In this example, the jpcconf mgrhost define command is executed in interactive mode. However, the command can also be executed in non-interactive mode.
Start services on the PFM - Agent or PFM - RM host.

Start the Performance Management programs and services on the PFM - Agent or PFM - RM host connecting to PFM - Manager whose host name was changed. To do so, use jpcspm start command.
Reconfigure the definition of a health check agent.

If the heath check function is used, reconfigure the definition (that was deleted in step 11) of a health check agent after changing the host name.

In a multiple-monitoring configuration, perform this step on the primary manager.
Reconfigure the definition of PFM - Agent or PFM - RM.

Reconfigure the definition (that was deleted in step 12) of PFM - Agent or PFM - RM which has been installed on the same host as PFM - Manager whose host name was changed.

In a multiple-monitoring configuration, perform this step on the primary manager.
Update the alarm settings.

In the following cases, you must update the alarm settings by using the jpctool alarm command of the PFM - Manager host or the monitoring console.

In a multiple-monitoring configuration, perform this step on the primary manager.
- The action handler of the PFM - Manager host is specified for the action handler that executes actions.
  
  Edit the alarm to set PH1<new-PFM - Manager-host-name> for the action handler that executes actions.
- JP1 events are issued by actions.
  
  Reconfigure the JP1 events of actions.
For details on how to edit alarms, see 6. Monitoring Operations with Alarms.
Update the JP1 system event settings.

If either of the conditions below is met, you need to update the JP1 system event settings from PFM - Web Console.

In a multiple-monitoring configuration, perform this step on the primary manager.
- The old host name is specified as the name of the event server to which JP1/Base connects for JP1 system events.
- The old host name is specified as the name of the monitoring console host for JP1 system events.
For details on the JP1 system events, see 12. Linking with the Integrated Management Product JP1/IM for Operation Monitoring.
Match definition information between the primary and secondary managers (for multiple-monitoring configuration).

Export the definition information for multiple-monitoring from the primary manager, and then import it to the secondary manager to match the definition information between the primary and secondary managers.

For details on the procedure for matching definition information, see 11.5 Duplicating definition information.
Make a post-setting-change check.

After changing settings, perform the following checks:
- Collecting performance data
  
  Operate the system for more than twice the time specified for the collection interval of performance data, and confirm that performance data is properly collected.
- Executing the jpcrpt command
  
  Confirm that collected performance data is properly output.
- Checking the report definition and alarm definition
  
  Confirm that there is no problem with the report definition and alarm definition created in a Web browser.
- Checking action execution
  
  Confirm that a created alarm executes properly.
If operation monitoring linked to a Network Node Manager (NNM) is performed, restart the NNM.

If there is linkage to a network management product (NNM), restart the network management product (NNM).

For details on the NNM linkage facility, see 14. Linking with Network Node Manager (NNM) for Operation Monitoring.

To Page Top