In this tutorial we will look at the steps involved in removing an HA nodepair  from an existing Clustered Data Ontap system

We will simulate an HA node pair removal from a 4 node Clustered Data Ontap system with the following setup:

  • Cluster Name: VMLABNETAPP
  • Node 1: VMLABNETAPP-01
  • Node 2: VMLABNETAPP-02
  • Node 3: VMLABNETAPP-03
  • Node 4: VMLABNETAPP-04

The HA pair that we wish to remove from this setup is VMLABNETAPP-01 and VMLABNETAPP-02

Netapp Clustered Ontap Node Removal

The following is a list of pre-requisites that are required before the actual cluster unjoin command is run

Netapp – Basic Pre-Requisites

From the HA pair that you wish to decomission, do the following

  • Either re-assign all data lifs to another HA pair within the same cluster. If the LIF is no longer in use it can be disabled and then deleted
  • Modify failover-groups to remove the HA pair’s physical ports

Netapp – Reassigning Epsilon

To check which node is epsilon, you can run the following commands:

VMLABNETAPP::> set -priv advanced

VMLABNETAPP::*> cluster show -epsilon *

Netapp Cluster Epsilon
We’ll set Epsilon to false on VMLABNETAPP-01

VMLABNETAPP::*> cluster modify -node VMLABNETAPP-01 -epsilon false

and we’ll set VMLABNETAPP-03 as epsilon true

VMLABNETAPP::*> cluster modify -node VMLABNETAPP-03 -epsilon true

Let’s verify our new settings

VMLABNETAPP::*> cluster show -epsilon *

Netapp Cluster Epsilon

Netapp – Moving the Cluster Ring Master

First we’ll check which node is the Master

VMLABNETAPP::*> cluster ring show

Netapp Cluster Ring Master
As we can see from the table above, VMLABNETAPP-01 is the cluster ring master. Since we will be removing VMLABNETAPP-01 and VMLABNETAPP-02 we will want to move the master role to either VMLABNETAPP-03 or VMLABNETAPP-04. To do this we first set VMLABNETAPP-02 as inactive

VMLABNETAPP::*> system node modify -node VMLABNETAPP-02 -eligibility false

We then set VMLABNETAPP-01 eligibility to false, this will force the master role to move to either node 3 or 4

VMLABNETAPP::*> system node modify -node VMLABNETAPP-01 -eligibility false

If we have a look at the cluster ring show command again we can see that the master has changed and that VMLABNETAPP-01 and VMLABNETAPP-02 are set as inactive

VMLABNETAPP::*> cluster ring show

Netapp Cluster Ring Master

Netapp – Storage Failover

Because this is a 4-node cluster we need to disable Storage Failover (SFO) on each node. This command can be done in admin privilege mode.

VMLABNETAPP::> storage failover modify -node VMLABNETAPP-01 -enabled false

VMLABNETAPP::> storage failover modify -node VMLABNETAPP-02 -enabled false

Netapp – Cluster Unjoin

Before we remove VMLABNETAPP-01 and VMLABNETAPP-02, the only objects that should be left on these 2 nodes should be the root vol and root aggregate along with the node management lif. If you have double checked and everything is correct you can proceed to run the following command in advanced privilege node

VMLABNETAPP::*> cluster unjoin -node VMLABNETAPP-01 (Ontap 9.3)

VMLABNETAPP::*> cluster remove-node -node VMLABNETAPP-01 (Ontap 9.4 and later)

Warning: This command will unjoin node “VMLABNETAPP-01” from the cluster. You must unjoin the failover partner as well. After the node is successfully unjoined, erase its configuration
and initialize all disks by using the “Clean configuration and initialize all disks (4)” option from the boot menu.
Do you want to continue? {y|n}: y

VMLABNETAPP::*> cluster unjoin -node VMLABNETAPP-02 (Ontap 9.3)

VMLABNETAPP::*> cluster unjoin -node VMLABNETAPP-02 (Ontap 9.4 and later)

Warning: This command will unjoin node “VMLABNETAPP-02” from the cluster. You must unjoin the failover partner as well. After the node is successfully unjoined, erase its configuration
and initialize all disks by using the “Clean configuration and initialize all disks (4)” option from the boot menu.
Do you want to continue? {y|n}: y

Disclaimer:
All the tutorials included on this site are performed in a lab environment to simulate a real world production scenario. As everything is done to provide the most accurate steps to date, we take no responsibility if you implement any of these steps in a production environment.