Migrate ESXi to new physical switch.

Products

• VMware vSphere ESXi

Issue/Introduction

This article provides guidelines and a structured procedure for moving a VMware ESXi host from an aging or existing physical network switch to a new physical switch. Improper migration can result in a total loss of management connectivity, storage disconnects (NFS/iSCSI), and virtual machine (VM) downtime. This process is applicable for both standalone hosts and those managed by vCenter.

Environment

Hypervisor: VMware vSphere ESXi (all versions)
Management: VMware vCenter Server or Standalone ESXi Host Client
Network: Standard vSwitch (vSS) or vSphere Distributed Switch (vDS)
Physical Hardware: Any supported physical network switches (e.g., Cisco, Dell, Arista)

Cause

• The need to migrate physical switches typically stems from:
  • Hardware lifecycle replacement (upgrading aging switches).
  • Data center physical relocation or rack reorganization.
  • Migrating from 1GbE to 10GbE or higher-speed infrastructure.
  • Standardizing network configurations, such as moving from unmanaged to managed VLAN-tagged environments.

Resolution

To ensure a successful migration with minimal or no downtime, follow these steps:

1. Pre-Migration Checklist
   a) Validate Switch Configuration: Ensure the new physical switch ports are configured identically to the old ones regarding VLANs (Access vs. Trunk), MTU (Jumbo Frames), and Spanning Tree settings (PortFast/Edge Port).
   b) Document Settings: Record current vmnic assignments, VMkernel IP addresses, and VLAN IDs.
   c) Out-of-Band Access: Verify console access via iDRAC, iLO, or KVM in case management connectivity is lost.
2. Migration Procedure (Zero-Downtime Approach)
   a) If your host has multiple redundant uplinks, migrate them one at a time to maintain connectivity.
   b) Prepare a Standby Uplink: In vCenter or the Host Client, go to Networking > Virtual Switches. Add a second physical NIC (vmnic) that is connected to the new switch as an uplink to the existing vSwitch.
   c) Test Connectivity: Temporarily set the new vmnic as the "Active" adapter or perform a vmkping from the ESXi shell to ensure the host can reach its gateway through the new switch.
   d) Migrate Management & Traffic: Once the new link is verified, move the remaining physical cables to the new switch.
   e) Update Uplink Policy: Remove the old, disconnected vmnics from the vSwitch configuration and finalize the new vmnics as the primary active adapters.
3. Migration Procedure (Maintenance Window Approach)
   a) Recommended for complex changes or single-uplink hosts.
   b) Maintenance Mode: Place the ESXi host in Maintenance Mode to evacuate or shut down all VMs.
   c) Physical Cutover: Power down the host (optional but safer) and move all physical network cables to the new switch.
   d) Verify via DCUI: If connectivity does not return automatically, use the Direct Console User Interface (DCUI) to re-select the active network adapters under Configure Management Network > Network Adapters.
   e) Final Validation: Exit Maintenance Mode and test VM connectivity before resuming full production.
4. Troubleshooting Post-Migration
   a) VLAN Mismatch: If the host is unreachable, verify that the physical switch port is set to "Trunk" if the ESXi port group has a VLAN ID assigned.
   b) MTU Issues: If storage (iSCSI/NFS) is disconnected, ensure the new switch supports and is configured for the same MTU (e.g., 9000) as the ESXi host.
   c) Load Balancing: If using Link Aggregation (LACP/EtherChannel), ensure the physical switch hash policy matches the ESXi vSwitch teaming policy.