Edge HA Member Status down Alarm

Products

• VMware NSX

Issue/Introduction

Title : Alarm for Edge HA Member Status Down
Event ID: edge_health.edge_ha_member_status_down
Alarm Description

• Purpose: To inform users that the edge node is set to down . The reasons could be various such as some of the following which causes the edge node to be down.
  • Datapath not running
  • Overlay connectivity issue
  • Host switch configuration issue
  • Bootup health pre-check failed
  • Edge in read-only mode, etc.

          The reason is included in the alarm.

• Impact: 

  In an Active/Active configuration, the SRs will continue to be running on the other healthy edge nodes. In an Active/Standby setup, there will be no impact if the edge cluster has a healthy edge node member that continues servicing the SRs.

Environment

• VMware NSX
• VCF 9.0 and above.

Resolution

Steps to Resolve
For VCF 9.0 and higher

Recommended Action : 

Resolution varies based on different down reasons

1. "Bootup Precheck Failed":
   
   • This can be due to one of followings:
     • "Controller Mgmt Vertical Down"/"Edge Vertical Down"/"Switching Vertical Down":
       • Check managers connectivity via CLI "get managers".
       • "restart service nsx-opsagent" to restart opsagent.
       • Ensure the edge is correctly joined to the manager.
       • Otherwise, check events from syslog further to see why MP connectivity is down. 
     • "CCP Session Down":
       • Restart controller service from managers.
       • Otherwise, troubleshoot from the manager to see if any CCP session is pushed to the edge. 
     • "NestDB Not Ready":
       • Invoke CLI "restart service nestdb" to restart NestDB service.
     • "NSD Not Ready":
       • Toggle maintenance mode to see if this down reason is resolved.
       • Otherwise, check events from syslog further to see why NSD service is unhealthy.
2. "Read-only File System":
   • Reboot the edge after resolving the storage issue.
3. "HA not in Active state":
   
   • Can be various reasons. Invoke CLI "get edge-cluster history state" to see the transition history and the last state of edge.
   • Ensure correctness of edge-cluster configuration.
   • Check events from syslog further if required to see why edge is not transitioned to active.
4. "Datapath not connected":
   
   • Invoke CLI "get service dataplane" to see running status.
   • "restart service dataplane" to restart the dataplane. 
5. "No overlay host switch configured"/"VTEP device down"/"VTEP IP not configured":
   
   • For VM edge, ensure the vnic for VTEP is connected as well as the VM network setting. For bare metal edge, verify physical connection to the NIC is secure.
   • Invoke CLI "get host-switches" to ensure correctness of host-switches configuration
6. "RTEP device down":
   
   • For VM edge, ensure the vnic for RTEP is connected as well as the VM network setting. For bare metal edge, verify physical connection to the NIC is secure.
   • Invoke CLI "get host-switches" to ensure correctness of host-switches configuration
7.  "VTEP tunnels down": 
   • Troubleshoot overlay connectivity issue further if host configuration is correctly applied

Maintenance windows required for remediation?
No. Edge is already in Down state.