fp-ethX is down on boot with XXV710 NIC Down on NSX Bare Metal Edge

Products

• VMware NSX

Issue/Introduction

• Management interface (eth0.<VLAN>) pings successfully during Linux boot.
• Management interface (eth0.<VLAN>) shares eth0 with Datapath Interface fp-eth0.
• Pings to management and other interfaces timeout immediately once the dataplane (nsx-edge-datapath) starts.
• RX counters for eth0 and eth0.<VLAN> remain at zero.
• NSX Edge Form Factor: Bare Metal
• NIC Hardware: Intel XXV710 (8086:158b)
  NVM Version: 8.40
  
  Review syslog after boot to verify NIC Hardware and NVM Version:
  <TIMESTAMP> <HOSTNAME> kernel - - - [    <NUMBER>] i40e <PCI_ADDRESS>: fw <FIRMWARE_VERSION> api <API_VERSION> nvm 8.40 <ETrack ID> <PACKAGE_VERSION> [8086:158b] [<SUBSYSTEM_ID>]
  
  As an alternative, you can locate the Device ID within the edge/datapath-configuration file of the edge log bundle:
  "device_id": "0x158b",
"name": "fp-eth0",
  
  
• syslog says fp-eth0 goes down:
  <TIMESTAMP> <HOSTNAME> NSX 4738 FABRIC [nsx@6876 comp="nsx-edge" subcomp="datapathd" s2comp="phys-port" tname="dp-ipc31" level="WARN" eventId="vmwNSXPhysicalNicStatus"] {"event_state":0,"event_external_reason":"Physical port link down","event_src_comp_id":"<EVENT_SRC_COMP_ID>","event_sources":{"interface_name":"fp-eth0"}}


• syslog shows grace timer is expired for port 0:
  <TIMESTAMP1> <HOSTNAME> NSX 4738 FABRIC [nsx@6876 comp="nsx-edge" subcomp="datapathd" s2comp="phys-port" tname="dp-ipc31" level="INFO"] [mgmt lsn] starting grace timer for port 0
<TIMESTAMP2> <HOSTNAME> NSX 4738 FABRIC [nsx@6876 comp="nsx-edge" subcomp="datapathd" s2comp="phys-port" tname="dp-ipc31" level="INFO"] Timer expired for port 0
<TIMESTAMP2> <HOSTNAME> NSX 4738 FABRIC [nsx@6876 comp="nsx-edge" subcomp="datapathd" s2comp="intel-rte" tname="dp-ipc31" level="INFO"] Posted mgmt lsn message mgmt: port 0 type TIMER_EXPIRED
<TIMESTAMP2> <HOSTNAME> NSX 4738 FABRIC [nsx@6876 comp="nsx-edge" subcomp="datapathd" s2comp="phys-port" tname="dp-ipc31" level="INFO"] [mgmt lsn] Processing msg for port 0 type: TIMER_EXPIRED
<TIMESTAMP2> <HOSTNAME> NSX 4738 FABRIC [nsx@6876 comp="nsx-edge" subcomp="datapathd" s2comp="phys-port" tname="dp-ipc31" level="INFO"] [mgmt lsn] Port 0 type: TIMER_EXPIRED, relaxed: 1, link_s: 0
<TIMESTAMP2> <HOSTNAME> NSX 4738 FABRIC [nsx@6876 comp="nsx-edge" subcomp="datapathd" s2comp="phys-port" tname="dp-ipc31" level="INFO"] [mgmt lsn] grace timer expired for port 0

Environment

NSX 4.2

Cause

The issue is caused by a compatibility mismatch between the DPDK version used in NSX 4.2 and Intel NVM firmware version 8.40.

Resolution

To resolve this issue, the Intel NIC NVM firmware must be updated to version 9.0 or later.

Additional Information

NSX Edge Bare Metal Requirements