Virtual Machines using Distributed Virtual Switches are showing as disconnected in vCenter since ESXi host is Not Responding due to hardware failure.

Products

• VMware vSphere ESXi

Issue/Introduction

Scenario:

• Environment with vDS (Distributed Virtual Switch).
• ESXi is down since there are underlying hardware issues or other non vSphere maintenance.
• Virtual machines didn't get migrated automatically to another ESXi host of the cluster since the host was not failed unexpectedly (soft power off of the VMs).
• Virtual Machines are not responsive and showing as disconnected on vCenter.

 

Environment

vSphere all versions

Cause

Since the vSphere HA will only take action when there is an unexpected power off of the virtual machines, a graceful shutdown of the VMs for a hardware maintenance will not failover the VMs to other hosts of the cluster.

Same can occur depending on non-default adjustments in the vSphere HA or lack of resources in the cluster to power on the VMs.

 

Resolution

Manually register the virtual machines to other hosts of the cluster following these steps:

1. Take note of the disconnected VMs location (find the vmx file of each affected VM).
2. Open a new tab in the browser and access the web client of any other host from the same cluster that has access to the vmx files (there are other alternatives to register the VMs in this article).
   1. Select Storage > Datastore > Datastore Browse.
   2. Find the directory where the VM vmx file is stored.
   3. Right click to the vmx file.
   4. Select "Register VM".
   5. Close the Datastore Browse panel.
   6. Select "Virtual Machines".
   7. Find the registered VM.
   8. Select to power it on (the task may take some time to complete).
3. Back to vCenter (vSphere Web Client).
   1. Find the registered VM and make sure it is now running on a different ESXi host.
   2. Edit settings of the VM.
   3. Take note of the Port group(s) configured according to each Network Adapter.
   4. Change the port group to any other port group (this is a temporary action, preferable choose a port group not being used to avoid IP conflicts - if necessary, create a new dummy port group only for this resolution then delete it later following the guidelines of this article).
   5. Save the VM configuration.
   6. Edit settings of the VM again.
   7. Change the port group back to the original one.
   8. Save the VM configuration.
4. Validate the VM is now on the network, that may take few seconds for the guest OS to respond, if necessary reboot the guest OS.
5. Repeat the process to all affected VMs (possibly they will be on different ESXi hosts as DRS will balance the cluster)