Hi2All

Rick, i use this command

ping 192.168.1.5 -l 1500

best wishes

Anthony

Hi2All,

ping 192.168.1.5 -l 1500  I get reply wihtout any timeout, does this means MTU is supporting 1500

best wishes

Anthony

Hi2All

Rick i get this as output

Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.

best wishes

Anthony

Anthony

That is proof that the MTU of the connection between the PCs is less than 1500.

HTH

Rick

Hi2All

How do I get the exact MTU size

best wishes

Anthony

Anthony

I described how to find the size in my first response. If ping at 1500 is too big, then ping using smaller sizes till you find one that works (remembering to use the -f to specify no fragmentation). If you want to be very exact and precise then ping using 1499, then 1498, then 1497 etc till you get to a value that succeeds. Or you could make the process quicker (but slightly less precise) by ping using 1495, then 1490, than 1485 etc till you get to a value that succeeds.

HTH

Rick

Hi2All

Thanks a lot rick for supporting.

I got another weird issue on Error Disable, if possible to help me.

You can look at details, i posted in the forum today.

best wishes

Anthony

Hi Rick,

a) In the MS Windows world, ping a.b.c.d -l 1472 -f  is the correct syntax to start as Windows are calculating 28 Bytes of the header length.

So ping a.b.c.d -l 1472 -f has the same result as Cisco CLI command ping a.b.c.d  size 1500 df-bit - an ICMP Echo Request packet of 1500 Bytes size sent with Don't Fragment bit sent.

b) To speed up the procedure of finding the exact maximum packet size which is passing,  you can start with 1472, then try 1372, 1272, etc.

When the first packet passes (let's say 1272), increase the size by 50. If passes, increase by another 25, if not passes, decrease by 25. Etc, etc.

This method (called interval halving in maths)  is quite effective and fast.

BR,

Milan

Milan

I was not aware of the Windows calculation about header. Thanks for that information (and +5 for you).

I am aware of alternatives such as interval halving (I have also heard a similar technique referred to as successive approximations) and thought about suggesting that. Then I decided in the interest of simplicity to just suggest stepping through successive values. Either approach works and your approach gets the answer more quickly when the difference between where we start and the correct answer is large.

HTH

Rick

Hi2All

I did couple of test and found that maximum i get reply with 1270 ; does this mean my Wireless AP supports max MTU size of 1270

ping 192.168.1.5 -l 1270 -f

Pinging 192.168.1.5 with 1270 bytes of data:
Reply from 192.168.1.5: bytes=1270 time=6ms TTL=254
Reply from 192.168.1.5: bytes=1270 time=6ms TTL=254
Reply from 192.168.1.5: bytes=1270 time=6ms TTL=254
Reply from 192.168.1.5: bytes=1270 time=7ms TTL=254

Ping statistics for 192.168.1.5:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 6ms, Maximum = 7ms, Average = 6ms

ping 192.168.1.5 -l 1275 -f

Pinging 192.168.1.5 with 1275 bytes of data:
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.

best wishes