Thursday, October 25, 2012

Jumbo Frames "Nexus 5548P" version 5.X code

If the need comes up for jumbo-frame, than this post will help you around the issues.

1st off the  jumboframe option allows you to pass frames bigger than  the  1514bytes MTU. The feature is globally, and effects all ports.

I'm not sure if you can even set it up port by port,  and that would not be wise if  you have ports in  various vlans and spread across various physical ports.

1st let's inspect the MTU size;

show queuing interface eth1/3

Ethernet1/3 queuing information:
  TX Queuing
    qos-group  sched-type  oper-bandwidth
        0       WRR            100

  RX Queuing
    qos-group 0
    q-size: 470080, HW MTU: 1500 (1500 configured)
    drop-type: drop, xon: 0, xoff: 2938
    Statistics:
        Pkts received over the port             : 0
        Ucast pkts sent to the cross-bar        : 0
        Mcast pkts sent to the cross-bar        : 0
        Ucast pkts received from the cross-bar  : 0
        Pkts sent to the port                   : 0
        Pkts discarded on ingress               : 0
        Per-priority-pause status               : Rx (Inactive), Tx (Inactive)


Okay, fine we see the port is set as 1500bytes MTU. Here's the NX-OS that we are running on this particular switch;

nxcr01-sw01# sh ver  | grep "System version"
  System version: 5.0(3)N2(1)

Okay now we must set a  policy-map and modify the  network-qos settings to effect the system MTU and to allow jumboframes ;

policy-map type network-qos bigfreakingmtu
  class type network-qos class-default
    mtu 9216
    multicast-optimize

As you can see, the network-qos policy-map was created named bigfreakingmtu and we applied the mtu under the  class-default class-map


Now apply this to your systems qos settings;

config t
  system qos
  service-policy type network-qos bigmtu
end

and let's recheck now;

sh queuing int eth 1/3
Ethernet1/3 queuing information:
  TX Queuing
    qos-group  sched-type  oper-bandwidth
        0       WRR            100

  RX Queuing
    qos-group 0
    q-size: 470080, HW MTU: 9216 (9216 configured)
    drop-type: drop, xon: 0, xoff: 2938
    Statistics:
        Pkts received over the port             : 0
        Ucast pkts sent to the cross-bar        : 0
        Mcast pkts sent to the cross-bar        : 0
        Ucast pkts received from the cross-bar  : 0
        Pkts sent to the port                   : 0
        Pkts discarded on ingress               : 0
        Per-priority-pause status               : Rx (Inactive), Tx (Inactive)

  Total Multicast crossbar statistics:
    Mcast pkts received from the cross-bar      : 0


Okay see what happen?

We changed from  1500 to 9216 bytes MTU. 9216 is the biggest MTU size offered and is commonly call jumbo-frames.

I hope this post was interesting. keep in mind the show interface eth X/X cli command  will never reflect the  true mtu  size. You must use the "show queuing interface eth X/X" to validate the  interface MTU size that's configured.


Ken Felix
Freelance Network Engineer
kfelix " at " hyperfeed.com

No comments:

Post a Comment