Protect VMs against storage problems using VMCP (VM Component Protection) on VMware vSphere 6.7

In the vSphere Client, you can also see which datastores are in APD status by selecting your cluster, then going to: Monitor -> vSphere HA -> Datastores under APD or PDL.
As you can see, for our host "esxi2.informatiweb.lan", an "APD detected" failure is displayed for our iSCSI datastores ("iSCSI DS 1" and "iSCSI DS 2").

On the host in error, an additional message will appear.

Plain Text

All shared datastores failed on host esxi2.

Once the 140 second delay used in the case of an APD state has expired, you will see this appear in the "/var/log/vmkernel.log" file.

Plain Text

[identifiant] has entered the All Paths Down Timeout state after being in the All Path Down state for 140 seconds. I/Os will ...

As you can see, at the moment our Windows 10 VM is still on our host 2 (in the case of an APD state).

On your cluster, additional messages will appear as host "esxi2" is no longer available for failover.
In fact, it no longer has access to the shared storage of the cluster.

Plain Text

Insufficient vSphere HA failover resources.
Insufficient resources to meet vSphere HA failover level on cluster [cluster name] in [data center name].

After a while, you will see that your virtual machine has been restarted on another host.
In our case, our "Win 10 x64 v2004" virtual machine which was previously running on our "esxi2" host was restarted by VMCP on our host 1 (which still has access to the iSCSI storage and the "iSCSI DS 2" datastore where this virtual machine is stored).

Warning : as indicated previously, VMCP restarts the impacted virtual machine using another host in the cluster.
It is therefore normal that the Windows 10 system properties window that we had opened in this virtual machine is no longer displayed.

On the other hand, this virtual machine is functional again.

Now that the test is complete, don't forget to reconnect the previously disconnected network cables.
In our case, we reconnect the network adapters of our virtualized VMware ESXi "esxi2" host.

As you can see in the images below, some messages will continue to appear for a few minutes.
But, these will eventually disappear.

As expected, the warnings and other messages have disappeared from your hosts.

For the cluster, wait a little longer.

As expected, the latest messages will disappear by themselves.

As expected, our virtual machine continues to run on the last host it was restarted on.