1. Check if vCenter Server was just installed or updated
If so, update the FDM agent on the affected ESXi hosts.
To reinstall the vSphere HA agent VIB:
- Reconfigure HA on a cluster level.
- Browse to the cluster in the vSphere Web Client object navigator.
- Click the Manage tab and click Settings.
- Under Services, click Edit.
- Uncheck the Turn ON vSphere HA option.
- Click OK.
- Click Settings and select Turn ON vSphere HA.
- Click OK.
- If any issue with any host not taking the new VIB, restart management services on that host. For more information, see Restarting the Management agents in ESXi.
- If still issues after management agents restarted, disconnect and reconnect the host from the vCenter Server.
2. Check if If ESXi host was just reinstalled or updated
If so, check if issue is
After update to ESXi 7.0 Update 3 vSphere HA fails to enableOtherwise, ensure that the FDM agent on the affected host(s) matches the vCenter version
To remove the vSphere HA agent VIB:
- Put the host into maintenance mode
- Follow How to run vSphere HA agent remove script in ESXi to remove vSphere HA agent from the ESXi host.
- If the vSphere HA remove script fails, run this command to directly remove the FDM agent VIB from the ESXi host:
esxcli software vib remove -n vmware-fdm
- Exit maintenance mode.
- Reinstall the VIB per step 1, above
3. Check time settings:
- NTP settings and current time for vCenter
- NTP settings and current time for an affected ESXi host
- NTP settings and curent time for a known good ESXi host
- If any discrepancy in times, fix the settings so the devices all match and verify they are synchronized.
4. Ensure network communication
- Management network connectivity between the vCenter and the affected ESXi host
- Management network connectivity between a good ESXi host and the affected ESXi host
- If connectivity fails, troubleshoot the network issue, ensuring required ports are open and affected host settings match good host settings including distributed virtual switch (DVS) ports used.
vCenter to affected host, substituting the host management IP:
curl -v telnet://<ESXi host IP>:443
curl -v telnet://<ESXi host IP>:902
Good Host in cluster to affected host, substituting the vmk and host management IP:
vmkping -I <management vmk, usually vmk0> <affected host IP>
5. Ensure storage communication
If the cluster is a vSAN cluster:
- Ensure there are at least 3 ESXi hosts in the cluster.
- Ensure the host has vSAN vmkernel networking set up per How to configure vSAN VMkernel networking.
For any sort of cluster:
Check that the host has reliable access to the heartbeat datastores
6. Check whether the agent on the host has failed and the watchdog process is unable to restart it
7. Check whether all hosts in the cluster have a failed vSphere HA status