CMDB Migration: What to Move, Clean, or Rebuild When You Change ITSM Platforms
A practitioner planning a move to ServiceNow described a BMC Remedy CMDB drawing on roughly 20 data sources. In a ServiceNow Community discussion, they asked how to move those records into the target tables.
The discussion does not report the migration outcome. It illustrates the decisions a CMDB owner faces before mapping fields. Each source may contribute useful records, outdated attributes, or references that other teams still need.
During a CMDB migration when changing ITSM platforms, assess each configuration item class against discovery coverage and operational use. Some records may be ready to move, while others need cleanup or a discovery-led rebuild.
What Does CMDB Migration Involve?
CMDB data migration moves configuration item (CI) records and their relationships between platforms. The project also needs a plan for references held by incidents, changes, assets, and other records. Whether those references move, remain accessible in an archive, or need remapping depends on the systems involved.
Cleanup is a separate task that can happen before or during migration. A team might correct duplicate records, confirm owners, or remove obsolete values before a load. It may also flag uncertain records for investigation after the target system is ready.
Virima’s CMDB migration guidance explains that a migration moves what the source tool already tracked. The team needs a separate way to identify infrastructure the source never recorded. Decide the treatment for each class before investing in detailed field mappings.
Why Does Carrying an Unreliable CMDB Forward Cost More Than It Saves?
A complete export can make the first load appear straightforward. The difficulty arrives when teams use inherited records for change planning, incident response, and reporting. Review these failure modes before selecting the records to carry forward.
1. Duplicate records travel together. The same server may appear in several source records with different identifiers. A target platform can create separate CIs if its matching rules cannot recognize the overlap. Result: Administrators must investigate conflicting records before relying on counts or relationships.
2. Load timestamps can obscure older evidence. A target record may show when it was imported, depending on the configured fields and transform. That timestamp alone does not establish when the source last observed the asset. Result: Reviewers need the original observation date to assess freshness.
3. Relationships can attach to the wrong records. A source relationship may refer to a CI that was duplicated, omitted, or mapped to a different class. Successful row counts cannot establish that both endpoints are correct. Result: A dependency view may show a connection that deserves review.
4. Owner fields can arrive empty. The AWS migration blog emphasizes ownership and governance when establishing a useful CMDB. An empty owner field leaves no clear reviewer for a disputed record. Result: The CMDB team must find an accountable person before resolving the exception.
5. Quality claims lack a comparison point. A count of imported rows measures loading progress, not the quality of the records. Compare selected classes with discovery evidence and owner review before claiming an improvement. Result: The migration team can identify what still needs work.
Why Is a Full Wipe Rarely the Answer Either?
Starting fresh can be practical before a new CMDB becomes operational. A team may also rebuild a class whose records have no required references and can be rediscovered reliably. Both decisions depend on checking the use of the existing data first.
In a ServiceNow Community answer, a practitioner warns against wiping an operational CMDB. They note possible loss of CI details referenced by incidents, problems, and changes. They also describe effects on asset-to-CI synchronization in ServiceNow.
Treat that answer as platform-specific practitioner guidance rather than a guarantee about every system. Before deleting records, inspect references, retention requirements, asset links, and the behavior of your own platform. Test the consequences in a nonproduction environment.
An archive preserves a way to inspect historical records, but it does not automatically restore working links in the new platform. Plan both the archive and any required reference mapping. Assign someone to verify that the records remain accessible to the teams that use them.
Which CI Classes Should You Migrate, Clean First or Rebuild?
Use the following table as a starting point, then test the proposed path on a sample. A single class can contain exceptions, including CIs linked to active work. Review those records individually before applying a class-wide decision.
| CI class or group | Decision test | Starting path | Information to preserve |
|---|---|---|---|
| Servers, virtual machines, and network devices | Can configured discovery reach and identify them? | Consider rebuilding from discovery | Validated owner, criticality, and relevant references |
| Endpoints | Does the selected source cover the devices? | Consider rebuilding where coverage is sufficient | Assignment, warranty, and lifecycle details |
| Software installations | Can the source identify installations and versions? | Compare discovery with existing records | Relevant entitlement and contract evidence |
| Business services and relationships | Who can validate the service and its dependencies? | Migrate and validate the definitions | Service ownership and approved relationships |
| CIs linked to active work | Which tickets or changes refer to each CI? | Preserve or remap required references | Target identifiers and historical context |
| Unowned or unseen CIs | Can another source or owner confirm their status? | Investigate before lifecycle action | Evidence, review decisions, and retention history |
Discovery evidence can establish technical details for the assets it reaches. People still need to confirm business owners, criticality, and other fields outside the scan’s scope. A missed scan alone does not prove that an asset was decommissioned.
For records confirmed as retired, define an archive and deletion process with the appropriate retention owners. Check legal holds, historical references, and local policy before deleting anything. Keep the team’s decisions and their supporting evidence available for review.
[VISUAL: Conceptual diagram showing CI classes routed to migration, cleanup, discovery-led rebuild, or investigation before retirement]
What Changes When You Move From Remedy to ServiceNow?
A Remedy to ServiceNow CMDB migration requires the team to compare two data models. A source class may need mapping to more than one target class. Agree on the attributes that distinguish those destinations before building the transform.
Relationships need their own mapping and validation plan. Identify the source and target records at both ends, along with the relationship type. Confirm that each endpoint resolves to the intended target CI after the load. ServiceNow stores CMDB relationships in the cmdb_rel_ci table, but your implementation determines how those relationships are loaded.
For imported CIs, ServiceNow documentation describes applying its Identification and Reconciliation Engine to Import Sets. Identification can reduce duplicate creation when the rules and source data support a match. The documentation also lists limitations, including dependent CIs with dependent identification rules. Check the release-specific guidance before choosing an import method.
The community discussion in the introduction asks about an out-of-the-box path for one CMDB-C environment. It does not establish connector availability for every Remedy configuration. Verify supported options for your versions and sources before deciding whether custom integration is required.
Test the mapping in a sub-production instance with representative records. Include duplicate candidates, missing identifiers, relationships, and exceptions that need human review. Compare the results with the source before repeating the load at larger scale.
Isn’t Migrating Everything the Safer Choice?
A complete migration can appeal to sponsors who want to preserve access to history. It also avoids early arguments over which records deserve attention. Those benefits depend on whether the target can use the imported records and retain their references correctly.
A reversible plan can preserve the old system or an archive for reference. It can also transfer selected records needed for active service work. The retention period and archive format should follow legal, operational, and security requirements.
For a different perspective, the shared CI record discussion covers a transition where platforms overlap. Check who can access the archive and how staff will find an older ticket. A retained file offers limited value if nobody can interpret its identifiers or restore the needed context.
Keep an export and a documented rollback procedure until the migration has passed its acceptance checks. Record the point at which the old system becomes read-only. The team also needs a plan for records created in the new platform if rollback becomes necessary.
Five Steps to a Decision Before You Map Fields
1. Score the classes you use. Measure required-field completeness, owner coverage, and the age of available observations. Separate unknown values from confirmed failures. Agree on which fields matter for the team’s actual workflows.
2. Compare a discovery sample. Run a configured scan alongside the legacy CMDB. Match observed items using suitable identifiers, such as serial numbers or cloud instance IDs where available. Investigate missing matches, scan coverage gaps, and possible duplicates before concluding.
3. Record a path for each class. Choose migrate, clean first, or rebuild as the initial treatment. Note exceptions for CIs that support active incidents or changes. Assign a reviewer to approve decisions that affect those references.
4. Set lifecycle and retention rules. Identify who can confirm that a CI is retired. Document archive access, retention periods, and any hold requirements. Delete only after the organization’s applicable conditions have been met.
5. Pilot the chosen treatment. Test a representative class in a nonproduction target. Check field values, relationship endpoints, and ticket references against the source. Record unresolved exceptions and repeat the test after correcting the mapping.
Start with 100 records from one important class if that sample suits your estate. Include records with active references and uncertain ownership. Review the findings with the people responsible for the underlying services.
[VISUAL: Illustrative example of a five-step CMDB migration decision process from class scoring through pilot validation]
Using a Discovery Baseline to Decide What Moves
A discovery baseline gives the migration team another source of evidence to compare with legacy records. Scope the scan to the environments and credentials available to the team. Document systems that the selected discovery method cannot reach.
Virima supports agent-based and agentless discovery, along with cloud discovery for AWS and Azure. Its virtual machine inventory includes VMware and Hyper-V environments. Scheduled discovery cycles provide observations that can be compared with the old CMDB.
The comparison can surface a source record with no corresponding observation. It can also identify an observed asset missing from the legacy export. Each difference needs investigation, including checks for access restrictions and scan configuration.
Virima’s CMDB can store CI information and relationships from multiple sources. People should confirm ownership and business context before relying on those fields. Teams can also review the supported ITSM integrations, including ServiceNow, when planning how discovered data reaches operational workflows.
A discovery baseline does not establish that every asset has been found. Record scan coverage and exceptions alongside the results. Revisit those exceptions as access and discovery schedules change.
Start With a Single CI Class
Select a class that supports an important service and has a named reviewer. Compare a sample against discovery evidence and active references. Then document the path for the class and its exceptions.
Explore Trusted Runtime Truth to see how discovery-sourced information supports a view of what exists, how it connects, what changed, what may be affected, and who owns it. To discuss the comparison with Virima, schedule a demo.
Frequently Asked Questions
Should you clean a CMDB before migrating it?
Clean the fields and records needed for the selected migration path. For classes rebuilt from discovery, review the owner and business information that still needs to move. Keep exceptions for active references and retention requirements.
Can you migrate a CMDB without discovery?
Yes, teams can transfer existing records without running discovery. They will need other evidence to assess missing or outdated assets. A configured scan provides a comparison point for the environments and assets it can reach.
How much ticket history should move with the CMDB?
Start with active work and the references required to resolve it. Agree on historical access with service owners, legal, and records teams. A read-only archive can preserve older material when the target does not need it.
How can Virima support a CMDB migration assessment?
Virima Discovery can provide scheduled observations for configured environments. A team can compare those observations with legacy records to investigate gaps and discrepancies. Coverage depends on discovery configuration, access, and the sources included.
Can Virima connect discovery data with ServiceNow?
Virima lists a ServiceNow integration for sharing asset and dependency information. Teams should confirm the fields, configuration, and operating scope for their implementation. A separate migration plan remains necessary for historical records and ticket references.






