I have two C6Ks on the same subnet and they are connected to each other.
One of the switch configure only global mtu (system jumbomtu 1500)
Another of the switch configure mtu 1521 on the interface as follows.
system jumbomtu 1500
ip address 10.1.1.1 255.255.255.0
system jumbomtu 9216 (default)
ip address 10.1.1.2 255.255.255.0
When I try to ping 10.1.1.1 size 1521 on the C6K2, I get a fauiler.
But when I change to mtu 1520 on the C6K2, I get a success.
*ping 10.1.1.1 size 1520
In addision, when I change to the system jumbomtu 1501 on the C6K1, it's success.
What is this 20-byte?
I would suggest to repeat tests using:
set DF bit on packets
you can use the sweep option in extended ping to provide a range of sizes to be explored
Extended commands [n]: y
Source address or interface:
Type of service :
Set DF bit in IP header? [no]: y
Validate reply data? [no]:
Data pattern [0xABCD]:
Loose, Strict, Record, Timestamp, Verbose[none]:
Sweep range of sizes [n]: y
Sweep min size : 1450
Sweep max size : 1522
Sweep interval :
Type escape sequence to abort.
Sending 365, [1450..1522]-byte ICMP Echos to 10.X.0.Y, timeout is 2 seconds:
Packet sent with the DF bit set
this should help in understanding what is happening.
I think you see the effect of fragmentation on one side
Hope to help
Cisco will donate $1 to the Red Cross Haiti fund for every rated post!