mtu mismatch problem

[shahaij]

Sir,

My router's POS module met a ping lost problem.

"

R1#ping

Protocol [ip]:

Target IP address: xxx.xxx.xxx.xxx

Repeat count [5]: 1000

Datagram size [100]: 4470

Timeout in seconds [2]:

Extended commands [n]:

Sweep range of sizes [n]:

Type escape sequence to abort.

Sending 1000, 4470-byte ICMP Echos to xxx.xxx.xxx.xxx, timeout is 2 seconds:

!!!!!!!.!!!!!!!.!!!!!!!.!!!!!!!.!!!!!!!.!!!!!!!.!!!!!!!.!!!!!!!.!!!!!!

!!!!!.!!!!!!!.!!!!!!!.!!!!!!!.!!!!!!!.!!!!!!!.!!!!!!!.!!!!!!!.!!!!!!!.

= = = = = =

Success rate is 87 percent (876/1000), round-trip min/avg/max = 1/2/20 ms

"

After troubleshooting, I found the MTU was set to 1500. Then I modify it to 4470 and everyting is ok. I want to know why this happened. The packets transferred 7 successful and 1 failed. Why mtu mismatch could lead to this result.

Thanks for your advice.

Regards,

Sha

[Georg Pauwen]

Hello,

I can imagine that, since the 4470-byte packets have to be fragmented with an MTU size set to 1500, the ping timeout of 2 seconds causes some pings to be lost (which actually only means that no reply is received within those 2 seconds). Can you try to increase the timeout to 10 seconds and see if you still miss pings ?

Regards,

Georg

[shahaij]

Sir,

Why always lost the 8th packet?

4470 * 7 = 31290 - 1500 * 20 = 1290 bytes

[1290 (bytes) + 4470 (the 8th packets)], the 1290 should be transfer successful. But the 8th 4470 can't. Why only this packet can't met the 2 seconds limit?

Thanks.

Regards,

Sha

[smif101]

If you enabled SDCC for your POS interface than the default value changes to 1500. If you have the default encapsulation type of LEX than the MTU is 1500. Or someone could have just changed it to 1500. One thing I do for my POS interfaces is to make them 1500 anyways. That way they are the same as my ethernet segments. And any traffic going over it wouldn't get that high except for maybe the routing protocol.