Troubleshooting packet flow from one VNF to another VNF, or to the front panel.
This article helps to determine if packets leaving the VNF are making it across the backplane of the NFX. This to VNF interfaces that traverse the default ovs-sys-br, and does not apply to VNF interfaces that map through SR-IOV.
This solution applies to VNF interfaces that are bound to the default ovs-sys-br, and does not apply to VNF interfaces that map through SR-IOV. This also applies to NFX devices that are running nfx-3 software. For NFX devices that are running nfx-2 software, refer to KB35958 - [NFX] How to interpret 'ovs-ofctl dump-ports' output from Hypervisor. [juniper.net]
To determine if packets are being transmitted from one VNF through the NFX ovs-sys-br backplane, you will want to inspect the counters going across a VNF interface. For example, let's assume you are inspecting the VNF interface vsrx-1_eth2. You will then want to use the command:
show vmhost network nfv-back-lane ovs-sys-br vsrx-1_eth2
root@b01-41> show vmhost network nfv-back-plane ovs-sys-br vsrx-1_eth2 Network Name : ovs-sys-br Interface : vsrx-1_eth2 Type : dpdkvhostuser, Link type : Full-Duplex, MAC : 00:00:00:00:00:00 MTU : 1500, Link State :up, Admin State : up Native Vlan ID : None, Vlan mode : Trunk, Vlan Members : 10 IPV4 : None, Netmask : None IPV6 : None, IPV6 netmask : None Rx-packets : 265 <---- Packets received from VNF to back-plane Rx-drops : 0 Rx-errors : 0 Tx-packets : 27 <---- Packets sent from back-plane to VNF Tx-drops : 0 Tx-errors : 0
As you can see, this output gives you Tx-packets and Rx-packets. This is analyzing the counters from the perspective of the backplane. The Rx-packet are packets initiated from the VNF interface, vsrx-1_eth2, going towards the backplane, which will eventually forward across to the intended destination. The response packet going towards the VNF will be reflected as Tx-packets, as it is transmitting packet from the backplane towards the VNF interface.