Multiple device duplicates in the OC Inventory
Article ID: 444170
Updated On:
Products
Issue/Introduction
We see multiple duplicates in the OC Inventory, so we would like to combine them and avoid duplication in inventory and check the discovery results as a possible cause.
We are observing many duplicate entries in UIM with or without IP addresses and same name and we tried the steps below to troubleshoot the issue:
We took the following steps:
Under the rules section of the discovery_server.cfg, add the following and then save it.
<name_mac_ip_origin>
enabled = true
position = 100
match_on_all = "defined:AllCorrelationNames, defined:IpAddresses, defined:Origins"
unmatch_on_any = "SysName, VirtualID, VirtualManagerId"
</name_mac_ip_origin>
Cold start the discovery_server (Deactivate-Activate):
Select the discovery_server probe and open the probe utility by pressing Ctrl-P
Run the discovery_server probe callback-> 'reimport_devices_by_cs_ids' for both the duplicate cs_id's using a comma to separate the values.
(The ID's a.k.a. the cs_ids were obtained by clicking on the device/node in the OC Inventory view and checking the 'Device Details' window for each one of the two duplicate entries, that had the same name and IP address, and were displaying as type Host and Virtual Machine.)
Search the device name again in OC inventory and check the device count (now there should be only one device in inventory with the given name).
We also tried:
Use "Prevent Discovery" on Deletion
When cleaning up duplicates in the Operator Console (OC):
Select the duplicate devices and click Delete.
CRITICAL: Ensure the Prevent Discovery checkbox is enabled before confirming. This prevents the discovery_server from automatically recreating the entry during the next scan
Clear NIS Cache
In cases where stale data persists, you may need to clear the niscache on the robots involved in the discovery. This forces the robot to send fresh, clean hardware/identity information to the discovery_server
Environment
- DX UIM 23.4 CU7
Cause
VirtualID and VirtualManagerId are different and discovery_server treat thems as different devices as per its correlation rules.
If the customer wants them to treat them as the same the correlation rule cab be changed but that would be specific to the customer.
Resolution
It is the out of the box/default configuration to not treat them as the same/as one. That is the default. The configuration can be changed to treat them as the same but all the devices would need to be re-imported.
They have different VirtualID and VirtualManagerIDs because those devices are moving across vCenters.
The customer decided since they are defined to be separate by default, its better to leave it that way and not change it.
Feedback
