Map ports from node1 to node3
You must make sure that the physical ports on node1 map correctly to the physical ports on node3, which will let node3 communicate with other nodes in the cluster and with the network after the upgrade.
You must already have information about the ports on the new nodes from the Hardware Universe. (Go to References to link to the Hardware Universe). You use the information later in this section and in Map ports from node2 to node4.
The software configuration of node3 must match the physical connectivity of node3, and IP connectivity must be restored before you continue with the upgrade.
Port settings might vary, depending on the model of the nodes.
You must make the original node’s port and LIF configuration compatible with what you plan the new node’s configuration to be. This is because the new node replays the same configuration when it boots, which means that when you boot node3, ONTAP will try to host LIFs on the same ports that were used on node1.
Therefore, if the physical ports on node1 do not map directly to the physical ports on node3, then software configuration changes will be required to restore cluster, management, and network connectivity after the boot. In addition, if the cluster ports on node1 do not directly map to the cluster ports on node3, node3 might not automatically rejoin quorum when it is rebooted until a software configuration change is made to host the cluster LIFs on the correct physical ports.
LIF Node1 ports Node1 IPspaces Node1 broadcast domain Node3 ports Node3 ports Node3 broadcast domains
Refer to Record node1 information for the steps to obtains this information.
Record all the cabling information for node3, the ports, broadcast domains, and IPspaces in the previous table using the same procedure in Record node1 information.
Set the privilege level to advanced:
cluster::> set -privilege advanced
Verify if the setup is a two-node switchless cluster:
network options switchless-cluster show
cluster::*> network options switchless-cluster show Enable Switchless Cluster: false/true
The value of this command must match the physical state of the system.
Return to the administration privilege level:
cluster::*> set -privilege admin cluster::>
Boot node3. See Install and boot node3 to boot the node if you have not already done so.
Verify that the new cluster ports are in the Cluster broadcast domain:
network port show -node node-name -port port-name -fields broadcast-domain
The following example shows that port "e0a" is in the "Cluster" domain on node3:
cluster::> network port show -node node3 -port e0a -fields broadcast-domain node port broadcast-domain ---------- ---- ---------------- node3 e1a Cluster
Add the correct ports to the Cluster broadcast domain:
network port modify -node node-name -port port-name -ipspace Cluster -mtu 9000
This example adds Cluster port "e1b" on node3:
network port modify -node node3 -port e1b -ipspace Cluster -mtu 9000
For a MetroCluster configuration, you might not be able to change the broadcast domain of a port because it is associated with a port hosting the LIF of a sync-destination SVM and see errors similar to, but not restricted to the following message`:
command failed: This operation is not permitted on a Vserver that is configured as the destination of a MetroCluster Vserver relationship.
Enter the following command from the corresponding sync-source SVM on the remote site to reallocate the sync-destination LIF to an appropriate port:
metrocluster vserver resync -vserver Vserver-name
Migrate the cluster LIFs to the new ports, once for each LIF:
network interface migrate -vserver Cluster -lif LIF-name -source-node node3 -destination-node node3 -destination-port port-name
Modify the home port of the cluster LIFs:
network interface modify -vserver Cluster -lif LIF-name –home-port port-name
If the cluster ports are not in the Cluster broadcast-domain, add them:
network port broadcast-domain add-ports -ipspace Cluster -broadcast-domain Cluster -ports node:port
Remove the old ports from the Cluster broadcast domain:
network port broadcast-domain remove-ports
The following example removes port "e0d" on node3:
network port broadcast-domain remove-ports -ipspace Cluster -broadcast-domain Cluster ‑ports <node3:e0d>
Verify that node3 has rejoined quorum:
cluster show -node node3 -fields health
Adjust the broadcast domains hosting your cluster LIFs and node-management and/or cluster-management LIFs. Confirm that each broadcast domain contains the correct ports. A port cannot be moved between broadcast domains if it is hosting or is home to a LIF, so you might need to migrate and modify the LIFs as follows:
Display the home port of a LIF:
network interface show -fields home-node,home-port
Display the broadcast domain containing this port:
network port broadcast-domain show -ports node_name:port_name
Add or remove ports from broadcast domains:
network port broadcast-domain add-ports
network port broadcast-domain remove-ports
Modify a LIF’s home port:
network interface modify -vserver Vserver-name -lif LIF-name –home-port port-name
Adjust the intercluster broadcast domains and migrate the intercluster LIFs, if necessary, using the same commands shown in Step 5.
Adjust any other broadcast domains and migrate the data LIFs, if necessary, using the same commands shown in Step 5.
Access the advanced privilege level on either node:
set -privilege advanced
Delete the ports:
network port delete -node node-name -port port-name
Return to the admin level:
set -privilege admin
network interface modify -failover-group failover-group -failover-policy failover-policy
The following example sets the failover policy to "broadcast-domain-wide" and uses the ports in failover group "fg1" as failover targets for LIF "data1" on "node3":
network interface modify -vserver node3 -lif data1 failover-policy broadcast-domainwide -failover-group fg1
Go to References to link to Network Management or the ONTAP 9 Commands: Manual Page Reference for more information.
Verify the changes on node3:
network port show -node node3
Each cluster LIF must be listening on port 7700. Verify that the cluster LIFs are listening on port 7700:
::> network connections listening show -vserver Cluster
Port 7700 listening on cluster ports is the expected outcome as shown in the following example for a two-node cluster:
Cluster::> network connections listening show -vserver Cluster Vserver Name Interface Name:Local Port Protocol/Service ---------------- ---------------------------- ------------------- Node: NodeA Cluster NodeA_clus1:7700 TCP/ctlopcp Cluster NodeA_clus2:7700 TCP/ctlopcp Node: NodeB Cluster NodeB_clus1:7700 TCP/ctlopcp Cluster NodeB_clus2:7700 TCP/ctlopcp 4 entries were displayed.
If necessary, for each cluster LIF that is not listening on port 7700, set the administrative status of the LIF to
::> net int modify -vserver Cluster -lif cluster-lif -status-admin down; net int modify -vserver Cluster -lif cluster-lif -status-admin up
Repeat Step 11 to verify that the cluster LIF is now listening on port 7700.