SG-SCCM causing duplicate CI's

[jimmillet]

Recently we converted SCCM integration to new Service Graph SCCM integration, and now we are getting duplicate Windows Server CI's being created with Discovery Source = "SG-SCCM". I am wondering if anyone has seen this and has suggestions on how to prevent the duplicates from being created when using the SG-SCCM integration.

[SumanthDosapati]

Hi @jimmillet ,

Check Identification and reconciliation rules.

Regards,
Sumanth

[jimmillet]

"Out of the box" ID rules are in place, we have not made any customizations.  Which field is used by SG-SCCM for identifier? Is it correlation ID or one of the others (Serial Number , Name, etc)? Existing Windows Servers CI's created by Discovery. What is weird is that some servers that were moved to Retired status, are getting a new CI created by SG-SCCM. We will be talking to local SCCM team to see if they removed entries when a server is retired.

[SumanthDosapati]

So may be till then as a work around, you can write a before insert business rule and abort action if the CI already exists.

Regards,
Sumanth

[Kyle Hill]

Hello Jim,

If you check the Audit History for your original CI(s), do you have any data flipping going on between old/new on the record? Such as name, serial number, etc. I have seen this cause the creation of duplicates when the data was flipped and a match could not occur.

Also, do you have multiple SG-SCCM data sources setup, or just one?

SG-SCCM will first check the sys_object_source table looking for the correlation ID in particular (among a few other fields). On the sys_object_source table you will see in the "ID" field: <correlation_id> | <connection_id>. Correlation ID being what will come through the SG-SCCM import for the CI, and the Connection ID is the unique identifier found in the SQL statement of the Computer Identity - Data Source (for that particular Prefix) that you created during the SG-SCCM setup. When you add multiple data sources, they each get a unique Connection ID that will differentiate them in the sys_object_source table as well. So even if you have CIs with the same correlation_id, they will be differentiated by the connection_id.

If a match is not found for the CI on the sys_object_source table, it will fail over to the IRE - Hardware Rule Identifier and look for a match there. Once found it will create a sys_object_source record (since it did not exist) that will be used for the next SG-SCCM import run.