vSAN Datastore Inaccessible: Disk groups unhealthy due to cluster membership mismatch

Products

• VMware vSAN

Issue/Introduction

When operating a vSAN Original Storage Architecture (OSA) cluster, you may experience the following symptoms:

• The vSAN datastore suddenly becomes inaccessible or completely disappears from the inventory of one or more ESXi hosts.

• Local vSAN disk groups on the affected hosts report as "unhealthy" or inaccessible.

• Navigating to Cluster > Monitor > vSAN > Skyline Health reveals a failure in the "vSphere cluster members match vSAN cluster members" health check.

• Expanding the health check details shows that the affected ESXi hosts have a green checkmark for the "vSphere Cluster" but a red exclamation mark for the "vSAN Cluster."

Environment

VMware vSAN 8.x(Original Storage Architecture - OSA)

Cause

This issue is caused by a synchronization failure between vCenter Server and ESXi hosts, often triggered by a temporary management network interruption or management agent (vpxa/hostd) issue. Because vCenter is the "source of truth" for the unicast agent list, any failure to push updates results in hosts losing the configuration required to communicate over the vSAN network.

Resolution

Resolution

To resolve this issue, you must force vCenter Server to refresh the cluster configuration and push a new vSAN unicast agent list down to the isolated ESXi hosts. This can be accomplished without data destruction by temporarily moving the host out of the cluster configuration.

Perform the following steps for each affected ESXi host:

1. Log in to the vSphere Client.

2. Right-click the affected ESXi host and select Maintenance Mode > Enter Maintenance Mode. (Choose Ensure Accessibility or No Data Migration depending on the current cluster state and object availability). Wait for the task to complete.

3. Once in Maintenance Mode, left-click and drag the ESXi host outside of the vSAN cluster, dropping it at the Datacenter or Folder level.

4. Wait for the "Move Host" task to complete successfully.

5. Left-click and drag the ESXi host back into the vSAN cluster.

6. Right-click the ESXi host and select Maintenance Mode > Exit Maintenance Mode.

7. Repeat Steps 1-6 for any additional hosts showing as isolated in the Skyline Health view.

By removing and re-adding the host, vCenter Server completely rebuilds the cluster membership list for that node. Once the unicast list is successfully pushed, the host will re-join the vSAN network, and the osfsd daemon will automatically remount the vSAN datastore.

Verification

Once all affected hosts have been re-added to the cluster:

1. Navigate to Cluster > Monitor > vSAN > Skyline Health.

2. Click Retest.

3. Verify that the "vSphere cluster members match vSAN cluster members" check now displays as Healthy (green) and the datastores are accessible on all hosts.