News Detail

Industry Insights and Updates

SAN switch migration checklist: compatibility, cutover, and rollback

A SAN switch migration needs a verified access map, supported hardware and software, an approved cutover method, and tests that prove host and application access after each stage. Plan rollback before moving connections, and retire the old switch only after final acceptance.

A configuration backup may not load on a newer platform. A dual-fabric diagram may hide a host that depends on one working path. Find those gaps before the change window. This guide gives you route choices, port records, and clear checks for when to continue or stop.

Key takeaways

  • Choose the migration method before writing the execution procedure.
  • Verify firmware coexistence, host support, optics, and feature licenses separately.
  • Compare zoning relationships and local switch settings rather than assuming a full restore will work.
  • Test the remaining paths and application workload before removing an active path.
  • Keep each stage’s evidence, rollback conditions, and sign-off together.

SAN switch migration at a glance

san switch migration at a glance

Define the change and the service requirement

Start a SAN switch migration by stating what will change: a failed switch, an aging fabric, a vendor platform, or the wider topology. Record the business result and the allowed maintenance conditions. A switch refresh and a storage-array data migration are different projects, even when they share a window.

List what must keep working during each stage. That includes host storage access, required application service levels, and the monitoring used to detect a problem. Assign an owner to each result so that switch health alone does not decide business acceptance.

Select the migration route

Cisco’s SAN migration white paper describes remove-and-replace, cap-and-grow, and interoperability approaches. Its examples have defined platform and software contexts. Treat them as method references, then check the support conditions for your own equipment.

Route to assessWhat the plan involvesMain prerequisite
Supported fabric integrationIntroduce a compatible switch and move approved connectionsVerified coexistence and fabric-merge conditions
Independent fabric replacementPrepare the replacement fabric and transfer its connectionsValidated remaining paths or an approved outage
Parallel environmentBuild the target environment and move workloads in stagesSpace, capacity, support, and a transition plan
Cross-vendor interoperabilityUse an explicitly supported temporary connection methodApproval for the exact platforms, releases, and features

Do not choose a SAN switch migration route because its command list looks shortest. The useful comparison includes support, business impact, temporary equipment, and rollback. A direct fabric connection is only one option.

If you are comparing equipment, request a configuration-specific quote from CREA DAS TECH. Include source and target models, software releases, ports, optics, required features, and support scope. The quote should match the reviewed design.

Capture the current fabric and host baseline

Inventory the complete path

A SAN switch migration baseline should cover the host-to-array path, not just the switch being replaced. Use the actual environment records and supported discovery tools. Save a dated copy that the change team can use during validation.

RecordDetails to captureLikely reviewer
Switch and fabricModel, release, role, logical context, domain identity, linksFabric engineer
Host connectionAdapter, driver, firmware, port WWPN, path stateHost administrator
Storage connectionController, front-end ports, software, host access settingsStorage administrator
Media and featuresOptics, cables, link speeds, enabled ports, licensesFabric and procurement teams
OperationsMonitoring, management access, backups, change contactsOperations owner

For the SAN switch migration inventory, record physical and logical roles as they exist. Access Gateway, NPV, NPIV, and ordinary switching describe different functions. Check the actual mode and the upstream devices it needs before drawing the target topology.

Build the old-to-new port map

For a SAN switch migration, give every moved connection a clear source and destination. Include the cable label, host or target WWPN, current port, proposed port, fabric context, and expected speed. Keep a separate record for each independent fabric.

Trace identities through aliases, zones, and the effective configuration. Also record array-side access and the host paths expected after the move. Our Brocade zoning guide explains the configuration states and checks behind that record.

The port map should describe what a technician will find at the rack. A correct logical design can still fail if a cable is mislabeled or the destination port lacks the required feature. Review both the diagram and the physical labels before the window.

Prove the baseline is usable

Measure the workload during the periods that matter to the business. Review host paths, array state, link traffic, errors, and application response with aligned timestamps. Avoid a universal utilization threshold; use the site’s service targets and remaining-capacity requirement.

Illustrative scenario: A pre-change review finds that a host in a dual-fabric diagram has only one usable storage path. The team holds that host’s batch and revises the maintenance plan. The diagram alone was not enough to approve its SAN switch migration.

Validate compatibility before connecting the new switch

validate compatibility before connecting the new switch

Separate coexistence from upgrade support

Compatibility in a SAN switch migration has several parts. A target switch may support its own software while the proposed old/new fabric combination remains unsupported. An approved firmware upgrade path also does not prove that every host and array combination is qualified.

Check these items separately:

  • Source and target hardware support for the intended releases.
  • Any required firmware upgrade path and its prerequisites.
  • Temporary switch coexistence and fabric interoperability.
  • Host adapter, driver, OS, and array support.
  • Required switching modes and licensed features.

Save the document revision or written approval for each combination. Broadcom’s SAN design guide identifies supportability, firmware planning, redundancy, and media validation as design topics. It is a planning reference, not a replacement for the applicable release notes and support matrices.

Verify optics, cables, and usable ports

Include both ends of every optical link in the SAN switch migration review. Check approved module part numbers, supported rates, fiber type, route, connectors, and reach or loss requirements. A connector that fits does not prove that the new link is supported.

Confirm which ports are enabled and which licenses or features are needed. Review inter-switch links and trunks against the proposed design. Do not assume that restoring configuration transfers a license entitlement or activates the required capacity.

If speed is also changing, use our 32G vs 64G Fibre Channel guide for the separate end-to-end speed review. Keep the migration goal tied to the supported connection and business test.

Review any fabric merge deliberately

For a SAN switch migration involving fabric integration, review domain identities, logical fabric context, zoning state, default-zone policy, security settings, and the required merge conditions. Prepare the target away from the production fabric until those checks are complete. Do not assume that an empty zone database makes every connection safe.

Cross-vendor connectivity needs its own approval. A vendor white paper’s interoperability example is not evidence that arbitrary Brocade and Cisco releases can join the same native fabric. Record the exact supported connection method and its feature limits.

Plan SAN switch migration cutover gates

Test the paths that will remain

The SAN switch migration plan must show how the workload is served while a path or fabric is unavailable. Confirm the host multipath configuration, storage behavior, and capacity of the surviving connections. Use the approved test procedure for that environment.

Do not stop at counting paths. Check whether the required paths are usable and whether the application meets its target in the planned degraded state. If it does not, arrange a supported alternative or a suitable outage before proceeding.

For the topology and capacity review, see our Brocade SAN design guide. Use the real failure domains and traffic demands in that review rather than assuming that Fabric A and Fabric B are fully independent.

Check what can be transferred

A configuration backup and a portable configuration are different things. Dell’s cross-generation configuration KB documents restrictions when loading files across Brocade generations and switch types. Check that restriction before making full restoration the basis of the plan.

For a SAN switch migration, separate local switch settings from zoning content. If full-file restoration is unsupported, review how supported settings will be recreated and how zoning will be transferred or converted. Validate the resulting access relationships, including any special zone types, rather than only comparing object names.

Illustrative scenario: A team plans to clone an older switch onto a newer generation. Its support review finds that the full configuration file is not portable to the target. It revises the plan to review local settings and zoning separately before any production connection is moved.

Define evidence for every decision

Use a SAN switch migration gate table to make the decision visible. Each gate needs a reviewer, a clear pass condition, and a stop action. The following is a planning template, not a tested customer runbook.

GateEvidence needed to continueIf the condition is not met
Before cutoverSupport approvals, target readiness, baseline, tested path planHold the affected batch
After each batchExpected logins, access, paths, and application resultStop, diagnose, or use approved rollback
Before the next fabricCurrent stage accepted and remaining-path plan confirmedKeep the next stage unchanged
Before retirementNo remaining old-switch dependencies; final acceptanceKeep required equipment available

Agree on who can stop the SAN switch migration and who accepts the business result. Define the latest point at which rollback remains practical within the window. Do not invent a fixed duration that ignores the site’s workload and recovery requirements.

Execute and verify one approved migration stage

execute and verify one approved migration stage

Confirm the connection before moving it

At the start of a SAN switch migration stage, compare the live state with the approved baseline. Check for changes made since planning, unresolved alerts, and unexpected path loss. Reconfirm the port map and the next connection to move.

Follow the equipment-specific procedure approved for that stage. Record the source port, destination port, identity, time, and operator. Where appropriate, have another team member verify labels before moving a cable.

This article supplies planning checks rather than production commands. The correct port, fabric, and configuration actions depend on the actual platform and release. Keep that execution procedure with the approved change record.

Validate the full storage path

After a SAN switch migration batch, verify device logins and the intended effective access. Confirm the array mapping and the host’s supported discovery and multipath result. A port showing Online is only one check.

Compare the result with the agreed baseline:

  • Required host and target identities are present in the right context.
  • Expected storage access and host paths are available.
  • Relevant link and fabric counters have been reviewed.
  • The agreed application workload meets its acceptance target.
  • Any required failover and recovery tests have passed.

Record exceptions rather than hiding them inside a general “migration complete” status. Keep test conditions comparable and note any unrelated workload changes. Make the go/no-go decision from the evidence.

Hold the next stage until this one passes

Illustrative scenario: New-switch ports are online after the first fabric move, but the host path result does not match the plan. The team pauses before changing the other fabric and investigates. It does not treat link status as business acceptance.

For a SAN switch migration, working on one fabric at a time is useful only within a validated design. It does not itself guarantee an interruption-free change. Confirm the current stage and the next stage’s dependencies before continuing.

Keep rollback and decommissioning practical

Preserve a usable return path

A SAN switch migration rollback plan needs more than an old configuration file. Identify the equipment, cabling, settings, licenses, and supported software state needed to restore the previous access path. Confirm that those resources remain available until the agreed decision point.

Keep the SAN switch migration rollback record with:

  • The event that triggers the decision and the person who makes it.
  • The old-to-new port map and the return connection plan.
  • Required equipment, settings, and available recovery time.
  • The tests that will prove the restored service works.

Check whether host or array changes affect the return path. Don’t assume a firmware downgrade or configuration restore is supported just because the earlier system worked.

Keep backups and restoration requirements distinct. A saved file is useful evidence, but recovery still depends on a supported target and a reviewed procedure. Test what is practical before relying on it during the window.

Retire the old switch after acceptance

Before closing the SAN switch migration, confirm that no host, target, inter-switch link, or active traffic still depends on the old equipment. Review the final topology and inventory with the teams that own them. Save the final configuration and acceptance records.

Close the SAN switch migration retirement record under site policy:

  • Update monitoring and the final topology.
  • Remove management access and credentials as required.
  • Update asset records and arrange equipment handling.

Don’t disconnect an old core just because its visible host ports have been moved. Check the remaining switch and fabric dependencies first.

Include the transition in the equipment quote

Budget for the whole SAN switch migration transition: required ports and modules, optics, cables, licenses, spares, support coverage, and any temporary parallel capacity. Separate equipment from implementation responsibilities. Our Fibre Channel switch buying guide helps define that equipment scope.

If you are gathering bids, send CREA DAS TECH the source and target equipment list through its contact page. Ask for an itemized equipment quote showing included parts, required features, support terms, and exclusions. Do not assume that an equipment quote includes execution of the production change.

Frequently asked questions

Can I migrate SAN switches without downtime?

It depends on the supported design, working multipath behavior, workload, and cutover method. A SAN switch migration needs evidence that the remaining paths can meet the service requirement. Do not promise zero interruption from a dual-fabric diagram alone.

Can I restore the old Brocade configuration on a newer switch?

Not automatically. Dell’s documented generation and switch-type restrictions show why full-file portability needs checking. Review zoning transfer and local configuration separately when a full restore is unsupported.

Should I join the old and new switches to the same fabric?

Only when coexistence and merge requirements are verified for that combination. Independent replacement or a parallel environment may be more appropriate. Choose the route before connecting the new switch.

Must firmware be upgraded first?

The answer comes from the source and target support requirements. Check upgrade paths and temporary coexistence separately. This checklist does not prescribe a universal release or upgrade sequence.

How do I know zoning survived the migration?

Verify effective access relationships and device identities, then check host paths, array access, and the application result. Configuration names and successful login alone do not prove the whole service is ready.

When should I roll back?

Use the stop and rollback criteria agreed before cutover. Consider failed access or application tests, unresolved path behavior, and the time remaining for a supported recovery. The change owner should make that decision from recorded evidence.

Complete a SAN switch migration by proving each stage, retaining a usable rollback path, and documenting final acceptance. Request a configuration-specific equipment quote with your switch, software, port, media, and support requirements so CREA DAS TECH can quote against the reviewed plan.

Previous News Brocade zoning: WWPN planning, configuration, and validation