Storage Migration Service can inventory a Windows file server, copy its files and share configuration, then optionally transfer the server's identity to a replacement. These are separate phases with different risk. An inventory or initial copy can usually run before the outage; identity cutover changes the production access path and requires a maintenance window, application coordination and a clear recovery decision.
This procedure describes a standalone domain-member file server moving to a supported Windows Server destination. It is not a domain-controller replacement procedure, a storage-array migration or a promise that every installed application moves with the files. Storage Migration Service can transfer files from some domain-controller sources, but it cannot perform identity cutover from a domain controller.
Define what the migration includes
Inventory shares, local paths, data size, file counts, ownership, NTFS permissions, share permissions, quotas and applications that write to the server. Identify encrypted files, open databases, reparse points and specialized storage features. Check the supported source and destination combinations in Microsoft's current requirements, particularly for clustered, Samba or appliance sources.
Decide whether clients will keep the old UNC name or move to a new namespace. Optional identity cutover is useful when many applications depend on \\FS01\Accounts, but it creates name, IP and Active Directory considerations. A DFS namespace or separately planned alias migration may have different requirements. Avoid combining several identity changes in one untested outage.
Confirm that the replacement has adequate capacity and performance for live users, growth, backups and temporary transfer needs. A destination that barely accommodates the current data can fail during the final delta copy or shortly after go-live. Consider both volume free space and the underlying storage's latency and resilience.
Prepare the orchestrator and destination
Use a patched Windows Server 2019 or later orchestrator and a supported destination, with current Windows Admin Center and its Storage Migration Service extension. For one migration, the destination can also act as orchestrator; several concurrent migrations usually justify a separate management server. Verify the current supported release list before selecting a build.
Install Storage Migration Service on the orchestrator and its proxy on a suitable Windows Server destination through Windows Admin Center or the approved feature-management process. The proxy improves the transfer path for supported newer destinations. Do not infer installation success solely from a role checkbox: confirm service health and the extension can connect.
Prepare migration credentials with the documented administrative access to the source, orchestrator and destination. Limit their use and lifetime. Open only the required management, SMB and related firewall paths between the relevant systems. The transfer and inventory paths must both work; an administrator opening a share from a workstation does not validate every server-to-server operation.
Protect the source and the change record
Take a recoverable backup of the source and preserve its share and network configuration. Record the current computer names, IP addresses, DNS records and domain membership. Document local users or groups involved in access and how they will be handled. Check Microsoft's current identity-transfer requirements for the domain arrangement.
Keep recovery information outside both migration servers. If the orchestrator or destination fails during cutover, the operator still needs access to the original identity map and restoration steps. Confirm console access to each server so a temporary networking change does not prevent recovery.
Coordinate with application and backup owners. A file server hosting an active application database needs an application-consistent plan; ordinary file copying is not automatically a supported way to move its open database files. Arrange a controlled write freeze for the final transfer and cutover.
Create the migration job and inspect inventory
In Windows Admin Center, connect to the orchestrator, open Storage Migration Service and create a named job. Select the correct source type and supply the authorized source credentials. Add the intended source server, run validation and start inventory.
Read the inventory results before proceeding. Confirm that expected shares and volumes are present and investigate exclusions or access failures. A successful job status is not a substitute for checking that the important business directories were included. Export or retain the inventory evidence with the change record.
Resolve permissions, name-resolution or connectivity failures at this stage. Do not continue to transfer on the assumption that a missing share will appear automatically later. If the source has several roles, separate file-migration requirements from any other service that must be rebuilt independently.
Map volumes and perform the initial transfer
Select the destination and map source volumes or paths to the intended destination storage. Review the mappings carefully: the example old data volume may not have the same drive letter or capacity as the new one. Confirm that no existing destination data will be unintentionally replaced and that the destination is reserved for the planned migration.
Validate the transfer configuration, then start the initial copy. Users can continue working only to the extent that the application and migration plan allow; the initial transfer is not the final consistency point. Monitor throughput, free space and transfer errors. Save detailed failure lists rather than reporting only a percentage complete.
Afterward, inspect representative directory trees and security settings on the destination. Compare file counts and sizes with awareness of ongoing changes and excluded items. Select important files for content comparison where appropriate. Do not manually “fix” every apparent difference without understanding how the tool handles protected or system-managed locations.
Prepare the final write freeze and cutover
Schedule the outage and stop the applications or user access that can create new changes. Confirm the write freeze through the application's supported method, not merely a notice asking users to stop. Run the required final transfer or retransfers and resolve failures before switching the production identity.
Review the cutover validation, new source identity, destination identity and network settings presented by the tool. The source is retained under a different identity and is no longer the normal production access point after successful cutover. It should not be reintroduced casually with the old name or IP while the destination owns them.
Identity changes can affect DHCP reservations, DNS caching, monitoring, backup jobs, SPNs and application configuration. Confirm how each dependency will follow the migrated server. Lowering a DNS TTL can help some planned changes but does not control every client cache or existing SMB session.
Verify access as real users
After cutover, test the original UNC path from representative client networks. Confirm name resolution reaches the destination and create, read, modify and delete test files only in an approved test location. Use ordinary authorized users and a deliberately unauthorized account to check both permitted and denied access.
Verify share and NTFS permissions, ownership, access-based enumeration, quotas and any required SMB security settings. A migration that copies data successfully can still expose it incorrectly if local group mappings or application identity assumptions were missed. Check event logs and failed-access reports during the first business cycle.
Repoint monitoring and backups where necessary and verify a destination backup completes. Validate scheduled jobs and applications that use stored UNC paths. Keep the source protected from ordinary user writes while the destination is being accepted, so the two copies do not diverge invisibly.
Handle failure without creating two active file servers
Before cutover, a failed copy can generally be corrected and repeated while the source remains authoritative. After cutover, rollback requires coordinated identity, network and data decisions. Files written to the destination after go-live may need reconciliation; simply turning the old server back on can lose work or create conflicting names.
Use Microsoft's cutover troubleshooting procedure and the saved job evidence for interrupted identity transitions. Identify which machine currently owns each name and address before changing anything. Preserve the original data until acceptance and retention requirements are met, then decommission the source through the approved process.
Microsoft's migration walkthrough and cutover explanation describe the supported phase boundaries. Use them with the current requirements for the exact server combination.
Related procedures
How to Plan an In-Place Upgrade to Windows Server 2025; How to Configure SMB Encryption for a Windows File Share; How to Enable Access-Based Enumeration on an SMB Share.