If HP uCMDB is creating duplicate CIs, first step you are required to perform is to look at the reconciliation rule to see which attribute(s) specifically is causing the duplicates. It is strongly suggested to look at the description of the rule, if it exists, before going through the xml of the rule.
Nodes are identified using OS identifiers (e.g. IP addresses, Net Bios Name) or hardware identifiers (e.g. MAC addresses, Serial Number). Two nodes that match in at least one OS identifier and one hardware identifier, are considered to be the same node. Nodes not having such match, and have a conflicting value in either OS or hardware identifier are considered to be different.
OS identifiers include: * Name * 66% of their IP Addresses' * 66% of their IP Addresses' Authoritative DNS Name * Net Bios Name
Hardware identifiers include: * 66% match of their Interfaces' MAC Addresses * SNMP System Name * BIOS Serial Number * BIOS UUID * Serial Number * BIOS Asset Tag
Two similarly identified nodes will be always be considered different entities in case of mismatch of one of the following: * Operation System Family * Cluster Resource Group
Therefore, two ‘Node’ CIs will never reconcile if none of the OS identifiers match or none of the Hardware identifiers match or if both the OS identifiers and Hardware identifiers match but Operating System Family and Cluster Resource Family attributes don’t match. Once you have identified the specific attribute because of which the hosts are not reconciling, you can then determine why the attribute on the two hosts are different.
Most likely, in case of duplicate CIs issue, you will have data coming in from two different sources (source can be a discovery job, integration job etc). In the event that none of the OS identifiers match, you can look into why the ‘Name’ attribute on the two hosts is not the same. If the data is coming from let’s say two sources, integration job and a discovery job, you should look at what command is run to get the name attribute and what the result of that command is. It’s possible that the names are actually different on the two sources.
Here is an example:
Source 1: NNM
Source 2: HC by Shell.
If the two hosts are the same and you have identified that duplicates are getting created b/c let’s say the ‘Name’ attribute is not matching, it is suggested to look at what command is run to get the name attribute and what the result of that command is. It’s possible that the names are actually different on the two sources (HC by shell is returning a different ‘Name’ and NNM has a different name) or it could be a bug in the script (If so, please contact Support for further assistance).
Some other possibilities are: the command the script is trying to run is not running successfully and returning a null value (permissions issue possibly) or the part of the script that is setting the attribute is not getting called for some reason. If the duplicates are coming in from one source, look at the results of the command (in the communication logs) that is being run to get the specific attribute and find out why the command is returning two different results. See section “How to read communication log” for details on how to look at command results in the communication logs.