ESXi holding artifacts from fibre HBA that was replaced

Products

• VMware vSphere ESXi

Issue/Introduction

After replacing a failed Fibre Channel (FC) HBA in an ESXi host, the old storage adapter and WWNs continue to appear in the vSphere Client → Host → Configure → Storage Adapters, even though the hardware has been physically replaced.

This article explains why this occurs and provides steps to remove stale (“ghost”) HBA entries.

 

Environment

ESXi 7.x and above.

Cause

• ESXi maintains persistent configuration records for storage adapters.
• When an HBA is removed, the entry doesn’t always clear automatically — especially if it was part of a multipathing setup or had active devices previously.
• This leaves a “ghost adapter” listed in the UI, even though the device is no longer present.

Resolution

Step 1: Rescan Storage Adapters

1. In vSphere Client:
• Select the host → Configure → Storage Adapters.
• Click Rescan Storage.
2. Check if the ghost HBA entry disappears.

 

Step 2: Verify Active Adapters via CLI

Run the following command on the ESXi host:

esxcli storage core adapter list

• Confirm the list of active HBAs.
• Compare against the ghost adapter shown in vSphere Client.

If the ghost adapter does not appear here, it’s a stale UI record.

 

Step 3: Remove Stale HBA Entry

If the adapter is still listed in esxcli, remove it manually:

#esxcli storage core adapter remove -A vmhbaX

Replace vmhbaX with the ghost adapter name.

Caution: Ensure you are removing only the obsolete HBA. Verify by PCI ID and WWNs.

 

Step 4: Restart Management Services

If ghost entries remain visible after removal, restart the ESXi management agents:

services.sh restart

or individually:

/etc/init.d/hostd restart
/etc/init.d/vpxa restart

 

Step 5: Validate Hardware Detection

Confirm the replacement HBA is correctly detected:

esxcfg-scsidevs -a
lspci | grep -i fibre

• Only the new adapter and its WWPNs should appear.

 

Additional Information

Post-Replacement Checklist

• Verify zoning and LUN masking are updated with the new WWPNs on the FC switches and storage array.
• Confirm datastores are accessible from the host.
• Re-register or re-map RDMs if required.
• Ghost entries are a cosmetic issue unless they interfere with storage path detection.
• This behavior is seen most often after replacing HBAs in multipath FC environments.