When troubleshooting traffic issues between a VNF interface and another device, it may require logging into hypervisor, and collecting statistics via the command, 'ovs-ofctl dump-ports ovs-sys.'   This article explains how to interpret this output.

• Traffic drops or throughput issues between one VNF interface to another VNF interface or external device

• The command 'ovs-ofctl dump-ports ovs-sys-br' provides statistics regarding tx and rx packets, but need clarity on how to interpret this data.

When troubleshooting traffic issues through a VNF interface, go into hypervisor, then run the command 'ovs-ofctl dump-ports ovs-sys-br <port number> '. 

Example:

Assume we are looking for traffic going through the VNF interface vsrx_eth2.  We first want to identify the OVS port number. This can be done with the following:

From hypervisor:

ofport              : 22
ofport_request      : []

From this output, it is determined that the port we are interested in is port 22.  Then run the following command to inspect the transmitted, received, and dropped packets:

OFPST_PORT reply (xid=0x2): 1 ports
  port 22: rx pkts=308961794, bytes=?, drop=?, errs=?, frame=?, over=?, crc=?  <-- from VNF_interface
           tx pkts=376031520, bytes=?, drop=64, errs=?, coll=?                 <-- to VNF_interface

The first line indicates statistics sent from the VNF interface to its destination.
The second line indicates statistics sent to the VNF interface.  

When a packet is sent to the VNF interface, the rx packet counters will increase.  When a packet is sent from the VNF interface, the tx packet count will increase.  The drop packet count will reflect the number of packet drops in that direction.