Duplicate Network adapters for servers

[Miguel Lara]

If I go to the identification rules for network adapter class, "name" and "MAC address" are used together for identification. We have a manually entered network adapter with a name and a MAC address, however, Discovery is creating a new record with the same Name and the same MAC address (only difference is capitalization in the MAC). My question is; what could cause Discovery to create a new network adapter instead of matching the one that is already in CMDB?

Thanks in advance for the help!

Miguel. 

[DuaneNMore]

The only thing that causes Discovery to create a new one is if it doesn't properly match during identification the existing network adapter.  

Question, why would you create a Related CI like that? Storage Devices, network adapters and other CIs that appear in the related lists at the bottom  should be left to discovery.

When looking at the class manager for network identifier you will see that "MAC Address, Name" but notice that it is also a dependent CI. SO part of the identification process is checking to see if there are other adapters on this particular host that has the same MAC Address and Name. 

When you created the Network Adapter Manually, did you fill in the Configuration Item reference field in the network adapter CI? Do both of these CIs show up in the network adapter related list for the device?

You could turn up identificatoin engine debugging by setting glide.cmdb.logger.source.identification_engine to debugVerbose. see https://hi.service-now.com/kb_view.do?sysparm_article=KB0750382 for more about this.

[Miguel Lara]

Hello Duane,

When we first implement ServiceNow, CMDB was populated either manually or via integrations from other systems. Right now we are implementing Discovery; that's why we already have data in CMDB.

How do "independent" or "dependent" rules affect the fact that we have already a record in CMDB with the same name and MAC address? I'm sorry for asking that, I'm pretty new to discovery and I don't really understand the difference between these two types of identification rules.

Yes, the manual entry has the CI filled and both adapters are in the server's related list. I will take a look at that identification engine debugging option. Thanks a lot for that.

Thanks for the help!

[doug_schulze · Accepted solution]

Are you using patterns? If so there are relationships that need to exist that the IRE would need to see. I talk about that here. If those relationships are not there, then it will create a duplicate.

[Miguel Lara]

Hello Doug,

Yes, we are using patterns. I'm assuming that the needed relationship is created with "Windows OS - Servers" pattern.

I went to cmdb_rel_ci and found a "Owns::Owned by" relationship between the server and the adapter for the network adapter entry that was created by Discovery. If I'm understanding correctly, the identification is failing because this relationship does not exist for the manual entry? Does this have anything to do with the fact that network adapters have a dependent id rule?

Please note: I think I just answered my question; Network adapter class has a dependent relationship with Hardware (reflected by the "Owns::Owned by" relationship), and since it has a dependent identification rule, the dependent relationship must be there in order to make a full match with the existent CI. Is this correct? I used THIS document as reference.

Thanks!

Miguel.

Thanks in advance.