breaking a three node cluster and reducing to single node

[AndrewRUK]

Hi All.

I have a test 3 node cluster running on NSX 4.1.2.5. I would like to reduce it down to a single node to dela with vsan capacity issues. 

What would be the best way to go about doing this? 

Thanks in advance for the info 

Andrew 

 

[Sriram ChunchankatteMelukote]

Hello Andrew,

The usual way to go about this is with the help of the NSXT CLI commands and also to verify the status through the GUI. If you intend to retain the node to which the Virtual IP (VIP) is pointing to in the GUI, then please follow the steps as below:-

1. Using the command 'get cluster status' you can get the uuids of all the 3 nodes of the cluster. And decide on which 2 nodes of the cluster to remove.
2. From the CLI issue the command 'detach node <uuid of the node to be removed>.
3. From the GUI  Home > Dashboard > System, wait until now the 2-node NSX-T manager cluster has green stable status.
4. Repeat steps 2 and 3 for the other node to be removed and check that your single node cluster is up and in stable state.
5. You can again repeat step 1 command to check cluster status and also use the command 'get cluster config' to verify that all the configurations of the cluster are good.

Hope this would help your query and if so you can please mark this as solved.

Thanks

Sriram

[sdebarros]

I cover this on my video, how to convert back to a single node cluster.  You can go to the 2:45 min mark

MultiNode Failure NSX-T Management Cluster. VMware NSX

YouTube		remove preview
MultiNode Failure NSX-T Management Cluster. VMware NSX You are having a bad day and your VMware NSX-T / NSX Management cluster is down, you have experienced a Mulitnode failure in your cluster.. Do you start by polishing up your resume.. No Watch this to recover the cluster and rebuild it. It is not that bad... View this on YouTube >

[Steffen Richter]

Or just shutdown 2 of the NSX Managers, connect to the last remaining and run the command:

1. deactivate cluster
   
   

Then delete the two powered off Managers properly.

Cant remember if the two powered off NSX Managers were still in the database then, but you check and solve this then like this on the remaining node (called "singleton" cluster):

get cluster status

detach node <node-uuid-of-powered-off-manager>

get cluster status

BR
Steffen

https://docs.vmware.com/en/VMware-Cloud-Foundation/5.2/vcf-admin/GUID-80E32450-8AD6-4334-AE43-22D591736333.html