1 Reply Latest reply on Aug 16, 2013 2:33 AM by mpryor

    Protocol Errors and too many running VMs

    mvaughn25 Enthusiast
    vExpert

      I think I am seeing a domino effect. I have the Remote Experience Agent installed on my parent VM and HTML Access enabled on my View Connection Server. However, the customer has a group policy that disables the Windows Firewall. This causes the "VMware Blast" service on the linked clone to fail, since it lists the "Windows Firewall" service as a dependency. Desktops will initially report as "Available", then move to a status of "Protocol failure" a few minutes later. In the Agent logs, the statuses read RDP=available / PCoIP=negotiate / Blast=Unavailable with an error of "PROTOCOL_ERR_FAILURE".

       

      This appeared simple, due to the Windows Firewall service being disabled. However, I expected this to go away when I disabled HTML Access for the pool...nope. I even recomposed, but still getting the protocol failure status for the Blast service (which the pool is not configured to use).

       

      Then I noticed a new problem. In a pool of 50, with spares set to 30, all 50 were running with no users connected. Checked power settings, all looked good. Then it hit me...is it looking for 30 VMs in the status "Available"? If so, the "Protocol failure" status is keeping the pool from satisfying the spare capacity and not allowing excess VMs to power off.

       

      Essentially, I have these two questions:

      1) Is there a way to remove the "Protocol failure" related to the disabled Windows Firewall, or do I have to remove the Remote Experience Agent from the parent VM and recompose?

      2) Does the pool only recognize "Available" VMs towards the spare count, accounting for why all of my desktop VMs are running and none will power down?

        • 1. Re: Protocol Errors and too many running VMs
          mpryor Master
          VMware Employees

          > 1) Is there a way to remove the "Protocol failure" related to the disabled Windows Firewall, or do I have to remove the Remote Experience Agent from the parent VM and recompose?

           

          Yes, don't install the HTML Access component on the agent if it's not functional. If it's not installed, it won't be in the protocol list in an error state.

           

           

          > 2) Does the pool only recognize "Available" VMs towards the spare count, accounting for why all of my desktop VMs are running and none will power down?

           

          VMs in a "protocol failure" state should still count as available, as the VM is online and responding - if there's no existing session on them.