Expanding a Nutanix Cluster with a New Node from the Factory
This article will show you how to expand a Nutanix cluster with a new node received from the factory. This article will also highlight a few common warning alerts that you might receive, and how to address them.
Starting the Cluster Expansion
First, you’ll notice in my lab cluster below, that we have 3 nodes in the cluster (1) and we’ll add one more node to make the total of four in the cluster.
To begin the cluster expansion (which can be done non-disruptively), click on the "Gear icon" (2) then choose "Expand Cluster" (3).
Assuming that the new node has been powered on and cabled onto the same network as the cluster, you should see a new node appear in the autodiscovery window that is shown.
Entering IP Addresses
Select the node by selecting the checkmark box. The window will expand showing a place for IP addresses to be entered. For a single node expansion, use the bottom boxes for each section as the range option is disabled when only selecting one node.
Enter an IP address for the Controller VM (CVM), the Hypervisor node and the IPMI (out of band management) interface.
Click on the "Next" button to continue.
On this next screen, the software will detect which hypervisor is installed on the newly discovered node and see if that matches the cluster. If it does, you will see a message like the below screenshot, which says no imaging is required.
If the hypervisor does not match, you will be prompted with a screen to browse for the hypervisor file on your local computer (ISO for VMware, downloaded from their website, or for AHV, the "host bundle" tar.gz file from our portal here https://portal.nutanix.com/#/page/static/hypervisorDetails)
Once the file is provided, the node will be imaged with the hypervisor required, and then the expand cluster process will proceed.
Once the cluster expansion starts, a progress bar will be shown showing the current status of the expansion.
If you click on the "Open" blue link, a more detailed progress will be shown.
Once completed, all of the progress bars should be green, and the cluster expansion progress should say 100%.
New Node Shown in Prism
Now the new node should be shown in Prism. The screenshot below shows the cluster overview on the "Home" screen, showing we now have 4 nodes in the cluster, versus 3 when we started.
On AHV, this node is now part of the compute and storage cluster and no further actions are needed.
On ESXi, this node is now part of the storage cluster, and now you’ll just need to perform the following actions:
1) Mount the container as a datastore on the new node. This can be done by going to the "Storage" menu, select the container you wish to mount, choose "Update" on that container, and then choose to "Mount on all hosts".
2) Add the node into vCenter, and add it to the HA/DRS cluster as your other nodes.
Possible Network Interface Warning
You might get the following warning after a cluster expansion:
"Controller VM <ip address> is running on the network interface(s) eth0, which is/are slower than 10 Gbps. This will degrade the system performance."
This might be due to not having the virtual switch (ESX) or Bond (AHV) configuration done where the 1Gb adapters are removed from the default configuration.
To do this on AHV, see my super networking post here: https://www.virtualdennis.com/nutanix-ahv-networking-commands-how-to-super-post/
On ESX, remove the 1Gb uplink adapters from the vSwitch.