Troubleshooting and Limitations – NNMi Integration

This section describes troubleshooting and limitations for NNMi Integration.

• Problem: The NNMi Web service responds with a cannot interrogate model message.

  Solution: This message usually indicates that the Web services request made to the NNMi server is incorrect or too complex to process. Check the NNMi jbossServer.log file for details.

• Problem: If an excessive number of nodes are to be updated with the same UCMDB ID, it may take a while for the update adapter to complete.

  Solution: The volume of data retrieved from the NNMi server might be large. The recommended memory requirements for the Data Probe process is 1024 MB. Since the NNMi Web service enables updating the individual nodes one at a time, the time to update the nodes may take a while.

• Problem: The Layer 2 by NNM job finishes with the following warning: Failed to get any Layer 2 links from NNM.

  Solution: Refer to technical article KM629927 on the HPE support Web site at http://support.openview.hp.com.

• Problem: Either of the NNMi integration jobs fails with the following error in the DFM log files: com.hp.ov.nms.sdk.node.NmsNodeFault: Cannot interrogate model.

  Solution: This error typically means that the NNMi server failed to process the Web services call. Check the following two logs on the NNMi server for exceptions when the integration was activated:

  • jbossServer.log

  • sdk.0.0.log

• Problem: Either of the NNMi integration jobs fail with the following error: Could not find Discovery Probe 'DefaultProbe'. Task for TriggerCI will not be created.

  Solution:

  1. Right-click the job and select Go To Adapter.

  2. Click the Adapter Management tab.

  3. Select the Override default Probe selection check box, and enter the name of the Probe used for the NNMi integration in the Probe field.

  4. Click Save to save the adapter, then reactivate the job against the IpAddress CI of the NNMi server.

• Problem: NNMi integration fails when using integration probe.

  Solution: In case of using integration service, for the NNM ID Update integration, the JDBC driver should be specified in the classpath as well.

  If your UCMDB server uses Oracle, you need to add ../lib/mcoracle.jar to the remoteJVMClasspath field as Adapter Properties when configuring the integration point; If your UCMDB server uses MS SQL, add ../lib/mcsqlserver.jar to the remoteJVMClasspath field.

• Problem: NNMi integration job fails with the EOFException error when using the integration service or using the Data Flow Probe that is upgraded from a previous version.

  Solution:

  1. When using the integration service, locate and open the <UCMDB_Server_Home>\integrations\conf\DataFlowProbe.properties file using a text editor.

     When using an upgraded Data Flow Probe, locate and open the <DataFlowProbe_Home>\conf\DataFlowProbe.properties file using a text editor.

     For details about this file, see "DataFlowProbe.properties File" in the HPE Universal CMDB Data Flow Management Guide.

  2. Search for the basic_discovery_minimal_classpath parameter.

  3. Change the following two values of this parameter as illustrated below:

     From	To
     ../lib/commons-dbcp.jar	../lib/commons-dbcp2.jar
     ../lib/commons-pool.jar	../lib/commons-pool2.jar

  4. Save the changes.

  5. If the original Data Flow Probe version is 10.31 or later,

     • When using the integration service, open the <UCMDB_Server_Home>\integrations\conf\DataFlowProbeOverride.properties file and follow steps b through d.
     • When using the Data Flow Probe, open the <DataFlowProbe_Home>\conf\DataFlowProbeOverride.properties file and follow steps b through d.

     For details about this file, see "DataFlowProbeOverride.properties File" in the HPE Universal CMDB Data Flow Management Guide.

  6. Restart the integration service or Data Flow Probe service.

Limitation

When integrating with multiple NNMi servers, each node (host) may only be managed by one NNMi server at the same time.