CMDB identification created duplicate CI: IRE deduplication failed

Identification rules are missing required attributes or have incorrect criteria that prevent matching existing CIs with discovered data. (high likelihood)

Coalesce attributes are not properly configured on the CI class, causing the system to create new records instead of updating existing ones. (high likelihood)

Data quality issues where key identification attributes like serial numbers, MAC addresses, or names contain inconsistent formatting or null values. (medium likelihood)

Multiple identification rules conflict with each other, preventing proper CI matching and causing duplicate creation. (medium likelihood)

IRE table entries are corrupted or contain stale data from previous discovery runs that interfere with proper identification. (low likelihood)

Step 1: Navigate to Configuration > Identification and Reconciliation > Identification Rules and review the rules for the affected CI class to ensure they have appropriate matching criteria.

Step 2: Check the CMDB Identification Entries table (cmdb_ci_identification) to examine what identification data was captured for the duplicate CIs and identify discrepancies.

Step 3: Verify coalesce attributes are properly configured by going to System Definition > Tables, finding your CI class table, and checking the Dictionary entries for correct coalesce settings.

Step 4: Review data quality of key identifying attributes on existing CIs in the affected class, looking for null values, inconsistent formatting, or special characters that might prevent matching.

Step 5: Use the IRE Simulator at Configuration > Identification and Reconciliation > Simulator to test your identification rules against sample data before applying changes.

Step 6: If duplicates already exist, manually merge the duplicate CIs using the CMDB Duplicate CI Management or delete the incorrect duplicates after data validation.

Step 7: Clear IRE cache by navigating to System Definition > Cache Administration and flushing the IRE-related caches, then rerun discovery to test the fix.