ESX host to AHV | Nutanix Community
Skip to main content
Hi all,



I have a production ESX cluster and an AHV cluster (currently Dev but will be live soon) and I want to move a couple of hosts out of the ESX cluster and then expand the AHV cluster. I want to make sure I have the correct process and do not miss anything but couldn't find this specifically documented anywhere. Has anyone done this already?



current plan is -



- Cluster checks.

- maintenance mode esx host to move all VMs off.

- shutdown CVM and allow maintenance mode to complete.

- remove VMhost from the prod vmware cluster in vSphere.

- remove from Nutanix cluster using the option in Prism.

- wait for Nutanix to remove the node.

- Expand the AHV cluster adding the released ESX host.



Any comments or advise would be welcome.



Thanks.
Hi,



you have to have a running CVM on the node being removed to allow the remove host to execute.

So after you have removed the ESXi host from the vSphere cluster you may power on the CVM again ; wait for data resiliency to be o.k. and then start the removal process in PRISM.
ah yes good point, thanks :)



I'm hoping to get this done this week or next, thanks for the advise.
HI,



Have you complete this process? Can you share the work experience of that activity?



Regards

Jai

  • Cluster checks
  • move all VMs off (Will the cluster have enough resources to sustain an HA event after you remove this node?)
  • remove Node from Nutanix cluster using the option in Prism
  • wait for Nutanix to remove the node (this may take multiple hrs depending on size/speed of disks)


  • maintenance mode esx host
  • shutdown CVM and allow maintenance mode to complete
  • remove VMhost from the prod vmware cluster in vSphere


  • Verify that host/IPMI/CVM is on the same L2 VLAN as the AHV clusters and is reachable via IPv6 by the AHV cluster
  • Expand the AHV cluster adding the released ESX host.

See additional information:

https://portal.nutanix.com/#/page/docs/details?targetId=Web-Console-Guide-Prism-v510:wc-cluster-modify-wc-t.html
Hi



Thanks for the comments, I have not done this yet due to other projects but will let you know how it goes.



I did log a call with support to confirm the best practice here and they conform that DavidN is spot on.



Also to confirm, there is enough resources to cover a HA event in the ESX cluster after the node is removed.



Thanks.
I'd like to amend the steps earlier and point out a "small" detail...


  1. Verify that you know and TEST the root credential on the ESXi host to be removed from the cluster.

Confirm if default credentials must be used (either on that host/cluster/IPMI or on all hosts/IPMIs and cluster) for this to work right... not sure what the assumptions are of the migration steps.



Another 'assumption' that should be acknowledged and checked is that the links of the host to be moved to the AHV cluster are on the same subnet and VLAN... things will go a bit easier.



The migration probably needs IPv6 so make sure that it's "working" (not dropped by some fancy ACL). I wouldn't try to do this across subnets...
Terms & Conditions