Edge node in disconnected state after ESXi host went into unresponsive state

Products

• VMware NSX

Issue/Introduction

• NSX Manager UI shows the edge node in an unhealthy state.
• The TEP  status for the edge node appears as degraded.
• The underlying ESXi host where the edge node was deployed is in a "not responding" state in vCenter.
• The edge node may be found in a powered-off state.
• vSphere HA does not kick in because of the affinity rules

Environment

VMware NSX 

Cause

The edge node was deployed on an ESXi host that became unresponsive in vCenter. As a result, the edge node was powered off and marked unhealthy in NSX. The degraded TEP status was due to loss of tunnel interface connectivity from the inactive edge node.

Resolution

Fix the underlying ESXi host issue and power on the edge node node.

Workaround:

• Redeploy the edge node on another ESXi host.

OR

• Re-register the Edge node using ESXi command line using DCUI console.

1. Access the ESXi DCUI console
2. From the ESXi shell check the VM Status 
   vim-cmd vmsvc/getallvms |grep <vm name> ===> Find the VM ID and path 
   vim-cmd vmsvc/power.getstate <vmid> ===> Find the VM state.
3. Check if there are any active tasks pending for the VM
   vim-cmd vmsvc/get.tasklist VMID
4. If the task is hung and needs to be cancelled, use the command:
   vim-cmd vimsvc/task_cancel task_id
5. If the VM is in powered on state , use the following command to power off the VM
   vim-cmd vmsvc/power.off VMID
6. Unregister the VM
   vim-cmd vmsvc/unregister VMID
7. Re-register the  VM from ESXI host UI - KB 335224
8. Power on the virtual machine.
9. Make sure from NSX UI that Edge node is healthy and TEPs are connected.

Note:
Before registering the edge node to a new host, make sure the affinity rules are disabled.
Also make sure edge node is in the system exclusion list of NSX.

Additional Information

NSX-EDGE-COM
KB 377076