So heaps has happened in my home lab over the past 18 months, where not only am I now running ESXi 6.7 U3, I also have an extra host with the same specs as the first, with both now acting in a 2-Node vSAN configuration.
Had a tricky issue with the vSAN Health check plugin recently, where a warning was reporting issues with some hosts not contributing vSAN performance stats.
When configuring a VSAN cluster, it is recommended to disable heartbeat datastores in your cluster, as this ensures that if only the VSAN network fails, vSphere HA will still restart the VMs on another host in the cluster (more info on why the heartbeat datastore should be disabled can be found in the VSAN Stretched Cluster guide here).
Now, when datastore heartbeats are disabled on your cluster, you may then see the following warning message on your hosts:
This is because vSphere HA requires a minimum of two shared datastores between all hosts in a vSphere HA enabled cluster for heartbeat detection (more info at the following VMware KB: https://kb.vmware.com/kb/2004739)
So if the only shared storage available is VSAN, then you may want to remove this warning. To do that:
- In the vSphere Web Client, right click your cluster and select Settings
- Under vSphere HA go to Edit
- Under Advanced Options add the following Configuration Parameter: das.ignoreInsufficientHbDatastore
- For its value, enter true
- Disable then re-enable HA for your cluster to apply the changes.
Of course if you don’t want VMs to fail over to another host in your cluster in the event the VSAN network is unavailable, then you will need to configure another non-VSAN datastore to use for heart beats.