Permanently Decommissioning a node from a vSAN Cluster

Products

• VMware vSAN

Issue/Introduction

Procedure to decommission a vSAN node.

Environment

VMware vSAN 7.x

VMware vSAN 8.x

Resolution

1. Following the correct procedure is crucial when permanently decommissioning a vSAN node. If the disk groups are left behind or the node is not disassociated with the cluster, residual components can be left behind which can have an impact on cluster stability.

Note: Before following any troubleshooting steps please make sure you have healthy backups for all VMs in your environment.

• • Ensure that there is no resync on the vSAN cluster. Also there should be sufficient capacity and number of vSAN nodes as per the storage policy in use.
    
    
  • Right click on the host which is to be decommissioned > Maintenance mode > Enter Maintenance mode > Pre-check.
    
    This pre-check simulates what the cluster would look like if it is not contributing to the vSAN cluster. From the pre-check results:

• • • Verify that none of the vSAN objects go in reduced availability or inaccessible state.

  • Place the host into maintenance mode with full data migration and wait for resync operation to complete.
    
    
  • For vSAN OSA after the host is in maintenance mode delete the disk group(s) that reside on the host you want to decommission:

    • Select the vSAN cluster > Configure > Disk Management > Select the host to be decommissioned > View Disks > Select the disk group >Remove the Disk group

  • For vSAN ESA remove the disks that reside on the host you want to decommission:

    • Select the vSAN cluster > Configure > Disk Management > Select the host to be decommissioned > View Disks > Select the disks > Remove Disk

  • Move the ESXi host out of the cluster to disassociate it from vSAN.
    
    
    Note: The command "esxcli vsan cluster leave" can also be run from the command line of the host to leave the cluster.
    
    Incase the unicastagent entry for this host exists on the remaining hosts in the cluster, run command "esxcli vsan cluster unicastagent list" to validate on each host:

    • From the above list, validate if the vSAN IP of the removed host remains.

    • Remove it using the command: esxcli vsan cluster unicastagent remove -a <Host_VSAN_IP>

  • Shutdown the ESXi host.

     2. To decommission a vSAN node that has failed in a vSAN cluster follow below steps -

1. • SSH to all the remaining healthy vSAN nodes and run command: esxcfg-advcfg -s 0 /VSAN/IgnoreClusterMemberListUpdates


   • Try to disconnect the failed host by right clicking the host > Connection > Disconnect.

   • Now remove the failed host from the vCenter by right clicking the host > Remove from inventory.

   • Generally the above steps should be enough to remove the failed host from the vSAN cluster, but there is a chance the other hosts are still trying to establish connectivity with the failed host, if their unicast agent list still has details for the failed host. In that scenario, run the command on all the remaining healthy vSAN nodes to remove the failed host completely from the vSAN cluster :  esxcli vsan cluster unicastagent remove -a <Host_VSAN_IP>

Additional Information

To perform maintenance mode pre-check for ESXi host: Check a Host's Data Migration Capabilities

vSAN Storage Policies Using vSAN Policies

To fully disable a vSAN cluster, see Disable vSAN.