PVS target devices on VMware with two network adaptors hang at a black screen when booting.

book

Article ID: CTX695829

calendar_today

Updated On:

Description

Environment configuration:

  • Target device VMs are UEFI and configured with VMXNET3 network adaptors.
  • Each network adaptor is configured on a separate subnet.
    • With PVS streaming traffic isolated to the first adaptor.
    • User traffic is isolated to the second adaptor.
  • ESXi 8 and ESXi 7 hosts both showed the same behavior.
  • The DHCP solution used in the environment is Cisco Prime Network Registrar 
  • In windows on the target device, the streaming network adaptor shows two ip addresses on the network adaptor.
    • This can occur in PVS environments when the DHCP solution does not provide a consistent IP address to the same mac address for each DHCP request.
    • PVS targets when booting always do two DHCP requests on the PVS network adaptor.
      • First during VM firmware stage, the firmware DHCP client sends a DHCP request.
      • Second when windows is up, the windows DHCP client sends a DHCP request.
    • PVS expects PVS target devices to get the same IP address to all DHCP requests during its booted lifespan.

 

Network traces were then captured at both VMXNET3 adaptors, capturing all packets to and from each VMXNET3 adaptor.

Outline:

  1. Boot the VM to boot selection or BIOS/Firmware screen
  2. Following VMware documentation, start all network captures using commands formatted like below:
    • pktcap-uw --switchport <XXXXXX> --capture VnicRx --outfile vmxnet3_pvs_rcv.pcap
    • pktcap-uw --switchport <XXXXXX> --capture VnicTx--outfile vmxnet3_pvs_tx.pcap
    • pktcap-uw --switchport <XXXXXX> --capture VnicRx --outfile vmxnet3_user_rcv.pcap
    • pktcap-uw --switchport <XXXXXX> --capture VnicTx --outfile vmxnet3_user_tx.pcap
  3. Initiate booting (exit boot selection or BIOS/Firmware screen)

 

The network traces showed in a failing scenario that the DHCP solution was sending NAK response to the second DHCP request causing the PVS target device to hang.

Environment

"Citrix is not responsible for and does not endorse or accept any responsibility for the contents or your use of these third party Web sites. Citrix is providing these links to you only as a convenience, and the inclusion of any link does not imply endorsement by Citrix of the linked Web site. It is your responsibility to take precautions to ensure that whatever Web site you use is free of viruses or other harmful items." 

 

Cause

The network traces showed in a failing scenario that the DHCP solution was sending NAK response to the second DHCP request causing the PVS target device to hang.

image.png

 

Resolution

Enabling the “MAC Address only” parameter in the DHCP configuration.

This configuration of Cisco Prime Network Registrar

  • Allowed target devices to successfully always boot
  • Ensured that target devices receive only a single IP address for the PVS streaming adaptor when booting.

image.png

 

 

Issue/Introduction

PVS target devices running on VMWare hang on a black screen after booting.

  • This hang occurs before the windows logon screen would be shown.
  • When a vdisk is created, it can boot once successfully on the target device, and subsequent boots result in a hang at black screen.
  • If the second network adaptor is configured to a disconnected state, the target device can successfully boot.

 

 

Additional Information

Two IP addresses are assigned to individual PVS target devices streaming network adapters.

https://support.citrix.com/external/article/CTX200036/error-a-duplicate-ip-address-was-assigne.html

How to Capture Packets for a VMXNET3 Virtual Machine Adapter

https://techdocs.broadcom.com/us/en/vmware-cis/vsphere/vsphere/6-7/vsphere-networking-6-7/monitoring-network-packets/using-the-pktcap-uw-tool/capturing-packets/capture-packets-that-flow-through-a-virtual-machine-adapter.html

Powered by Wolken Software