how to test a Jumbo Frame setup?

Answered Question
Feb 24th, 2010

Hello

I am currently trying to Lab a solution using Jumbo Packets.

I have a Catalyst 3750 connected to a Catalyst 3560.  They are directly connected to each others Gigabit ports and are access ports in the same VLAN.

Both ports are up and can ping each other.  I have set up Jumbo globally on both switches:

Switch(config)#do sh system mtu

System MTU size is 1600 bytes
System Jumbo MTU size is 9000 bytes
Routing MTU size is 1600 bytes

And if I 'show' the ports they both have an mtu of 9000:

GigabitEthernet1/0/1 is up, line protocol is up (connected)
  Hardware is Gigabit Ethernet, address is 0019.0647.8681 (bia 0019.0647.8681)
  MTU 9000 bytes, BW 1000000 Kbit, DLY 10 usec,

I am trying to test using a ping with the df-bit set, but it does not work:

Sending 5, 5000-byte ICMP Echos to 10.232.196.253, timeout is 2 seconds:
Packet sent with a source address of 10.232.196.254
Packet sent with the DF bit set
.....
Success rate is 0 percent (0/5)

I cannot work out why this does not work.  Is it possible I am missing something obvious?

Many thanks for your help in advance

Ronnie Loraine

Systems Engineer

I have this problem too.
0 votes
Correct Answer by Sergei Vasilenko about 6 years 11 months ago

Hi Ronnie Loraine,

For c3750 (or other switch) to test the system jumbo MTU (applied to Gig int) setup properly

for the sizes bigger than 1998 bytes (depends on IOS version actually)

you would need a transit traffic, not generated on the system CPU itself (like in ping from the switch CLI session).

As per CCO "Catalyst 3750 Switch Cisco IOS Commands":

system mtu

http://www.cisco.com/en/US/docs/switches/lan/catalyst3750/software/release/12.2_50_se/command/reference/cli3.html#wp1948947

----

The size of frames that can be received by the switch CPU is limited to 1998 bytes, regardless of the value entered with the system mtu command. Although forwarded or routed frames are usually not received by the CPU, some packets (for example, control traffic, SNMP, Telnet, and routing protocols) are sent to the CPU.

Because the switch does not fragment packets, it drops:

switched packets larger than the packet size supported on the egress interface

routed packets larger than the routing MTU value

-----

Thanks,

Sergey

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
Sergei Vasilenko Wed, 02/24/2010 - 05:15

Hello,

If you are pinging from the windows based server/host most likely the problem is that it has default MTU of 1500,

so the ping packets might even does not get sent out net interface of the box.

In order to increase it, you migth need to enable jumbo MTU support (per int) on both end system sides,

as described for example here:

http://www.performancewiki.com/windows-tuning.html

Thanks,

Sergey

ronnie.loraine Wed, 02/24/2010 - 07:35

Hi

thank you very much for your reply.

In actual fact there are no devices plugged into the switches at present.  the ping tests I have been carrying out are extended pings from switch to switch on the console.

Regards

Ronnie

Correct Answer
Sergei Vasilenko Wed, 02/24/2010 - 08:06

Hi Ronnie Loraine,

For c3750 (or other switch) to test the system jumbo MTU (applied to Gig int) setup properly

for the sizes bigger than 1998 bytes (depends on IOS version actually)

you would need a transit traffic, not generated on the system CPU itself (like in ping from the switch CLI session).

As per CCO "Catalyst 3750 Switch Cisco IOS Commands":

system mtu

http://www.cisco.com/en/US/docs/switches/lan/catalyst3750/software/release/12.2_50_se/command/reference/cli3.html#wp1948947

----

The size of frames that can be received by the switch CPU is limited to 1998 bytes, regardless of the value entered with the system mtu command. Although forwarded or routed frames are usually not received by the CPU, some packets (for example, control traffic, SNMP, Telnet, and routing protocols) are sent to the CPU.

Because the switch does not fragment packets, it drops:

switched packets larger than the packet size supported on the egress interface

routed packets larger than the routing MTU value

-----

Thanks,

Sergey

ronnie.loraine Wed, 02/24/2010 - 09:58

That makes perfect sense, thats exactly the info i needed

thank you very much for your time

Actions

This Discussion