VMware vSAN: How to replace a cache disk using vSphere Web Client

How to Replace a Cache Disk (vSAN 8.x and 9.x OSA)

If a problem occurs on a cache disk, it affects the entire disk group. The disk group must be removed and re-created with the replacement cache disk.

Replacement Steps

1. In the vSphere Client, navigate to the affected host and place it in maintenance mode. Select the appropriate data evacuation mode for your environment:

   • Full data migration — Migrates all data off the host. This is the safest option but requires sufficient cluster capacity and takes the longest time.
   • Ensure data accessibility — Ensures at least one copy of the data remains accessible. Faster than full migration but may temporarily reduce redundancy.
   • No data migration — Enters maintenance mode immediately without migrating data. Use only for brief maintenance windows when data redundancy is already sufficient across other hosts.

   For detailed guidance on maintenance mode options, refer to the Broadcom vSAN documentation for your version: Configure > vSAN > Disk Management on the host object in vCenter.

2. Navigate to the affected host in the vSphere Client: Configure > vSAN > Disk Management. Select the disk group containing the failed cache disk and click Remove (the remove icon in the toolbar). Removing the cache disk removes the entire disk group.

3. Replace the failed cache disk physically.

4. After the physical replacement, return to Configure > vSAN > Disk Management in the vSphere Client and click the Create a new disk group icon in the toolbar.
   

5. In the Create Disk Group dialog, select the replacement cache disk from the upper (cache tier) list, select the capacity disks from the lower (capacity tier) list, and click OK.
   

6. Once the newly created disk group is displayed in Disk Management, exit maintenance mode on the host.
   

7. Confirm that the capacity of the re-created disk group is added to the total capacity of the vSAN datastore.

8. Since health alerts may remain even after the problem is resolved, navigate to Monitor > vSAN > Health and click Acknowledge or Reset to Green to clear the alerts manually.
   

Additional Information: Unable to Remove Disk Group

The disk group icon may not appear in the vSphere Client, or removal of the disk group may fail through the GUI. In that case, use the CLI to remove the disk group or individual capacity disks before replacing the cache disk.

Method 1: Remove the Disk Group by UUID (Recommended)

This is the preferred method in vSAN 6.6 and later, including vSAN 8.x and 9.x.

1. Connect to the affected host using SSH.

2. Identify the disk group UUID by running:

   esxcli vsan storage list

   Locate the vSAN Disk Group UUID for the disk group that contains the failed cache disk.

3. Remove the entire disk group by UUID:

   esxcli vsan storage remove -u <VSAN Disk Group UUID>

   Example:

   esxcli vsan storage remove -u 52a038e0-b564-e579-a498-4ceb5aa9e372

4. Replace the cache disk physically and create a new disk group using the vSphere Client as described in the replacement steps above.

Method 2: Remove Capacity Disks Individually

If the UUID-based method is unavailable or fails, remove each capacity disk individually.

1. Connect to the affected host using SSH.

2. Identify the capacity disk device names by running one of the following commands:

   esxcli vsan storage list

   or

   vdq -q -H

   Note the capacity disk identifiers in the format mpx.vmhbax:Cx:Tx:Lx.

3. Remove each capacity disk individually:

   esxcli vsan storage remove --evacuation-mode=noAction --disk=mpx.vmhbax:Cx:Tx:Lx

   Example:

   esxcli vsan storage remove --evacuation-mode=noAction --disk=mpx.vmhba1:C0:T4:L0

4. Replace the cache disk physically and create a new disk group using the vSphere Client. The replaced cache disk and the existing capacity disks that were removed are now available for selection.

Additional Resources

• How to Replace a vSAN Capacity Disk — For replacing a vSAN capacity disk instead of a cache disk.

• VxRail: How to Manually Remove and Re-create a vSAN Disk Group Using esxcli — For VxRail-specific vSAN disk group procedures.

• Requirements When Replacing Disks in a vSAN Cluster (VMware KB 67856) — Covers Deduplication and Compression requirements for disk replacement.

• Managing and Configuring a vSAN Disk Group Using esxcli Commands (Broadcom KB 315532) — Covers UUID-based disk group management commands.

• Replace a Flash Caching Device in vSAN Cluster (Broadcom TechDocs, vSAN 8.0) — Official Broadcom documentation for cache disk replacement in vSAN 8.x.