VM hosts are included in one of the scopes defined for the discovery_agent and are being discovered with WMI authentication. These same devices are also discovered and monitored by the vmware probe. In the UMP USM portlet Inventory view, there are 2 entries for some of these devices. If we check the CM_COMPUTER_SYSTEM table, we find that the entry which corresponds to the device discovered with the discovery_agent probe have an IP and MAC address. The entry that corresponds to the device discovered by the vmware probe contain a NULL for both the IP and MAC addresses.
UIM server: 8.31
The problematic VM hosts either did not have VMware Tools installed on them or the installed version of VMware Tools was outdated. This prevents the vmware probe from acquiring the system information from these devices causing the probe to create probe discovery messages that do not contain an IP or MAC address. Because of this, the discovery_server is not able to correlate these devices into one device and creates a new entry for the device in the CM_COMPUTER_SYSTEM table.
1. Install a valid version of VMware Tools on all VM hosts that will be discovered/monitored by the vmware probe.
2. Clear the niscache directory on the robot where the vmware probe is installed. This can be done using the controller probe utility as follows:
From Infrastructure Manager:
a. Select the controller probe on the robot where the vmware probe is installed and enter <ctrl>p
b. Select the Options button (button with the yellow gear) and select the "Expert Mode (no confirmation required)" on the Options GUI
c. Select OK to save the change
d. From the Probe commandset drop down list, select the _nis_cache_clean command
e. Select the green arrow (play) button to execute the command
From Admin Console:
a. Left-click the drop down list for the controller probe on the robot where the vmware probe is installed and select Probe Utility
b. Select the _nis_cache_clean command from the command list
c. Select the green arrow (play) button to execute the command
3. If only a few devices have duplicate entries in the CM_COMPUTER_SYSTEM table, you can delete both entries for each device from the CM_COMPUTER_SYSTEM table using the remove_master_devices_by_cskeys callback in the discovery_server probe utility. If there are a large number of devices with duplicate entries, then do a full discovery reset.
See the following Knowledge Document for details on full discovery reset:
TEC000004861 : How to reset discovery in CA UIM 8.10 or later
SQL Query to retrieve the cs_key for devices that will be deleted from the CM_COMPUTER_SYSTEM table:
SELECT cs_key, name FROM CM_COMPUTER_SYSTEM where name like '%<vm name>%';
where <vm name> is a string that will match all entries of the device that you see in the Inventory view/list for the VM host with duplicate entries.