Spectrum creates duplicate models during an AutoDiscovery configuration run

Document ID : KB000049874
Last Modified Date : 14/02/2018
Show Technical Document Details

Description:

When running the Spectrum AutoDiscovery configurations duplicate models are created.

AutoDiscovery configurations include IP addresses for devices that are both new to Spectrum and already exist as models.

Duplicate models are created for those IP addresses that already exist in the Spectrum database as models.

After the duplicates are created alarms appear against those models warning the user that a duplicate model with the same IP address already is present within Spectrum.

This is a known issue introduced as of Spectrum 9.2.1 (H05).

To determine if the problem being faced the same as this is known issue, debug will need to be enabled while reproducing the issue. To do so complete the following:

  1. Identify one or more IP addresses for devices that have duplicate models created for them.

  2. Delete the existing duplicate models that were created for those devices.

  3. Enable AutoDiscovery Debug off the VNM model.

    1. Locate the VNM model for the SpectroSERVER landscape against which the problematic discoveries are run.

    2. Display the Information tab for that VNM model in the Spectrum OneClick client console.

    3. Expand the AutoDiscovery Control->Debug Options section.

    4. Set the 'Debug AutoDiscovery' option to On.

  4. Reproduce the problem by running the AutoDiscovery configuration against the SpectroSERVER involved that creates the duplicate models.

  5. After the AutoDiscovery finishes and creates the duplicates debug must be turned off. The debug log for AutoDiscovery debug isn't written to with data until it is turned off after being enabled. Go back to the VNM model where the AutoDiscovery was enabled (step 3 above) and turn it Off.

  6. Examine the debug log created on the SpectroSERVER involved in the $SPECROOT/SS directory. It will be named ADiscDebug_*_*. Within the log these are the messages we would expect to see for the IP addresses against which duplicate models were created:

    13:01:01 Identified mtype Host_Device for <IP_Address> with community name <commString> and SNMP port <Port>
    13:01:01 Device <deviceName>/<model_handle> (Host_Device) has full IP match with <IP_Address>
    13:01:01 Device <deviceName>/<model_handle> (Host_Device) does not have full MAC match with <IP_Address>
    13:01:01 Device <IP_Address> was created with model handle <duplicate_model_modelHandle>. Error Code: 0

If those messages are seen the known problem that has been reported is the same problem being faced in this situation.

If the issue is not easily reproducible by deleting existing duplicate models and running the same discovery that created them again, then a different problem may be at hand here. If that is the case a new issue with support should be opened for investigation.

Solution:

To resolve this a Programmable Test Fix (PTF) patch 09.02.01.PTF_H0604 is available for installation on Spectrum. The fix is built against 9.2.1.H06 and can only be installed on that release and patch level of Spectrum. Please be sure that is the current release running before opening a support issue to request the fix that is available.

Once the issue is reproduced on the 9.2.1.H06 release of Spectrum, please open a new issue with Spectrum support. When opening the issue please providing the following information:

  • A copy of the ADiscDebug log created

  • A copy of the SpectroSERVER .history file found in $SPECROOT/Install-Tools on the SpectroSERVER

  • (rename .history to just history.txt before attaching to issue)

  • A reference to this Knowledge Base article ID

The information provided shall be reviewed to be sure it is a match for the known issue, and the PTF patch fix will be provided for installation.