Question

Which steps are being included in Hyper-V cluster expansion on Prism?

  • 5 February 2020
  • 2 replies
  • 1044 views

Userlevel 2
Badge +17

Can anyone let me know which steps are being involved when I expand Hyper-V cluster by using Prism?

I have a Hyper-V cluster that being mixed by some AHV storage nodes and looks like I cannot use the Prism cluster expansion feature to add more nodes...


2 replies

Userlevel 2
Badge +3

Hi Hienle,

The expansion process shouldn't be much different with Hyper-v, the instruction to expand a cluster is the same although you can see some notes that are made specifically for hyper-v which makes the small difference. 

For example: 

  • If the cluster is running Hyper-v, you must first configure the nodes in the VLAN before attempting to add them. Otherwise, the discovery process will not find these nodes. See the "Configuring the Cluster in a VLAN-Segmented Network" section in the Acropolis Advanced Setup Guide for VLAN configuration instructions.
  • If you want to move the VMs between hyper-v nodes, ensure that you have selected the Migrate to a physical computer with a different processor version option for each VM by navigating to Settings > Processor > Compatibility in the Action pane of the Hyper-V Manager.
  • [Hyper-V only] Specify the credentials to join the new nodes to Active Directory and to a failover cluster.

Here is the full documentation for cluster expansion via the web console: https://portal.nutanix.com/#/page/docs/details?targetId=Web-Console-Guide-Prism-v55:wc-cluster-expand-wc-t.html

Userlevel 3
Badge +4

Hello @hienle 

You normally should be able to add the Hyper-V node to the cluster via Prism, and I recommend to sort out the problem there rather than follow the manual procedures. 
 

These articles may help:
Cannot Discover New Nodes to Create Cluster or Add Nodes to existing Cluster

DiscoveryOS: Cluster Expansion - Cannot add or discover nodes : IPMI console displays phoenix prompt

Cannot Create Cluster or Add Nodes to Existing Cluster

If you follow a manual process, please ensure you only add one node at a time because part of the process will be a domain join step for the Hyper-V node and this includes a reboot. In the automated domain join this will be done before the CVM is functionally joined to the cluster, but in a manual process if you reboot multiple active nodes for domain join this could cause impact.

The discovery process is trying to find and connect with the foundation service on any unconfigured Nutanix node. 

A node should be discoverable for prism Expand Cluster process if: The node has a running CVM, that running CVM is unconfigured, the hypervisor host and CVM have IP addresses in the same subnet as eachother (configured or unconfigured can work), the new nodes are reachable on the same VLAN as the current cluster, and IPv6 link local (mDNS) discovery is enabled on the switch/vlan.

Reply