Cisco HX CVM VMs are unable to be edited - Red DOT on VM CPU field
search cancel

Cisco HX CVM VMs are unable to be edited - Red DOT on VM CPU field

book

Article ID: 392917

calendar_today

Updated On:

Products

VMware vSphere ESXi 7.0 VMware vCenter Server 8.0

Issue/Introduction

 

In Cisco hyperconverged environment.

In order to upgrade ESXi to 8.0, Cisco requires secure boot for the ESXi host hardware to be enabled (via Cisco's HyperFlex Connect). This allows for the Cisco HX CVM VMs to be upgraded, and the ESXi to be upgraded.

  • The issue is experienced after enabling secure boot in the BIOS of the ESXi host hardware.

You may experience the following:

  • From vCenter vSphere client, when accessing the VM edit window for the Cisco HX CVM VMs, the CPU section shows red exclamation mark, while all other VMs on the host do not.

  • Any changes to the VM configuration are unable to be saved.
  • The ESXi hostd.log shows below error regarding the configuration change attempt.

    Cannot edit PCI passthrough device while VM is powered on.

Environment

VMware ESXi 7.x
VMware vCenter Server 7.x
VMware ESXi 8.x
VMware vCenter Server 8.x

Cause

Virtual machine configuration is locked by a vCenter Server extension or ESX Agent Manager (EAM) agency.  Storage Controller VMs are explicitly protected via API locks (such as Config.Disable.Change or managedBy properties) to prevent structural changes to hyperconverged storage nodes via the standard vCenter Server interface.

Resolution

Manually edit the CVM via its .vmx configuration file.  To manually edit the storage controller VM's memory

  • SSH to the ESXI host where the storage controller VM is registered
  • Put the Host in Cisco HX Maintenance Mode through the Cisco UI
  • Get the VMID for unregistering later

vim-cmd vmsvc/getallvms

  • Change directories to the VM directory

cd /vmfs/volumes/<datastore-UUID>/<vm-Name>

  • List the files in the VM directory

ls -lah

  • Take a backup of the .vmx file for the VM before editing and place it in a subdirectory inside the VM directory:

mkdir bk
cp /vmfs/volumes/<datastore-UUID>/<vm-Name>.vmx /vmfs/volumes/<datastore-UUID>/bk/<vm-Name>.vmx.bak

  • Edit the file to place the new memory size in the following parameters:
    • memSize
    • sched.mem.min
    • sched.mem.minSize

vi /vmfs/volumes/<datastore-UUID>/<vm-Name>.vmx

  • Press i for Insert Mode
  • Multiply the new GB value by 1024 to get the value for replacement
    • Old value in memSize = 72*1024 = 73728
    • New value for parameters 96*1024 98304
  • Edit each parameter and change 73728 and 98304
  • Save the edit using ESX key then :wq!
  • Unregister the VM using the VMID found in the getallvms step above

vim-cmd vmsvc/unregister <VMID>

  • Confirm the VM is unregistered

vim-cmd vmsvc/getallvms

  • Return to the ESXi host client UI and register the VM back to the Host
    • Storage > Datastore Browser > <datastore-Name> > <vm-Name>
    • Right click the .vmx file and select Register VM
  • Take the Host out of Cisco HX Maintenance Mode through the Cisco UI
    • The CVM should power back on
  • SSH to the Cisco cluster and run the following the check the health/healing status

sysmtool --ns cluster --cmd healthdetail

Powered by Wolken Software