Removing NSX edge cluster from SDDC manager UI fails with error : NSX services have been deployed on clusters

Products

• VMware SDDC Manager

Issue/Introduction

• Unable to remove the clusters from the SDDC manager UI, it failed with the error: Validates if there are any NSX services deployed on the cluster(s) blocking its deletion 

 

Environment

VMware Cloud Foundation 5.x 

Cause

• This issue was due to Out-Of-Band (OOB) configuration changes which was performed in the SDDC manager.
• OOB configuration changes are not supported and causes unexpected exceptions due to which remove cluster from domain failed.
• Example: Manually migrating edge node VMs to different host/clusters, removing edge host from NSXT-Manager directly etc.

Resolution

To delete cluster and the corresponding hosts, need to decouple the dependent Edge Cluster from the Cluster/Hosts and then proceed with the cluster removal.

Either of the following 2 options can be performed post taking a snapshot of the SDDC manager.

Option 1:  Edge Expansion and Shrinkage.

• EC- Edge Cluster
• EN- Edge Node

1. Assume EC1 has EN1 and EN2 on Cluster1
2. Expand a new node EN3 hosted by Cluster2 to EC1
3. Shrink EN1 from EC1
4. Expand a new node EN4 hosted by Cluster2 to EC1
5. Shrink EN2 from EC1
6. EC1 has two Edge Nodes hosted by Cluster2
7. Then, can delete Cluster1 and the underlying hosts

Reference KB article: Migrate NSX edge nodes in a VCF Environment

Option 2: Remove Edge Cluster

1. Invoke Edge Cluster Removal Tool via KB remove_edge_cluster to delete Edge Cluster first from SDDC manager.
2. Then, delete the cluster and the hosts