Changing mtu speeds on switch

[spivy66]

I have a major problem. I had a nortel switch that went bad we replaced it with a cisco 3560 and snice then i'm having problems with infrequent packets being drop. I set it up the same way i had the nortel (vlans)and full duplex. When i ping the gateway device from the cisco switch i get packet loss sometimes. when i ping the gateway from a cisco 2600 router which is connect to the 3560 I get no drops.But when i ping from the same cisco router on other interface though a tunnel I get this
DR#ping ip
Target IP address: 172.29.5.129
Repeat count [5]: 100
Datagram size [100]: 1500
Timeout in seconds [2]:
Extended commands [n]: y
Source address or interface: 172.28.2.97
Type of service [0]:
Set DF bit in IP header? [no]:
Validate reply data? [no]:
Data pattern [0xABCD]:
Loose, Strict, Record, Timestamp, Verbose[none]:
Sweep range of sizes [n]:
Type escape sequence to abort.
Sending 100, 1500-byte ICMP Echos to 172.29.5.129, timeout is 2 seconds:
Packet sent with a source address of 172.28.2.97
!!!!!.!!!!.!!.!!!!!!!!!!!!.!!!!.!!!!!!!!!!!!!!!.!!!!!!!!!.!!!!!!!!!!!!
!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!
Success rate is 92 percent (92/100), round-trip min/avg/max = 12/16/44 ms
DR#


The only thing i can think of is mtu speeds,can i change the mtu speed lower then 1500 on the port that the router is connected too. I even tested to see if the cisco router was the problem,but it's not. Once i hook up the nortel i know it's going to start working again..any idea guys??

thanks

 

[GM2005]

1. Have you repeated the extended ping with a lower mtu to check your theory?
2. On a 3560, as i recall, it must be a global command: xxxx(config)# system mtu xxxx (ie 1000 (no suffix))
3. Post up the show interface output for the suspect interface. Clear the counters on the source and destination interfaces and post the output after an extended ping.
4. Is there any qos mechanism or security policy in place dropping management traffic?

 

[spivy66]

thanks for the reply..yes i did change the datagram size and still same issue. As for the mtu size i think your right it's a global command and can only change it 1500 or higher and for cisco 2600 I tried to changed it but no good(see below). i clred my counters and i got no errors but here it the port stats and inter config for the cisco 3560.AS for the qos ,nothing is set. I will also post the cisco 2600 stats cause thats where the packets are mostly droping on the secondary fasteth.

3560-Computer-Room#sh run inter f0/24
Building configuration...

Current configuration : 58 bytes
!
interface FastEthernet0/24
speed 100
duplex full
end

3560-Computer-Room#

FastEthernet0/24 is up, line protocol is up (connected)
Hardware is Fast Ethernet, address is 001a.e2a1.d61a (bia 001a.e2a1.d61a)
MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec,
reliability 255/255, txload 1/255, rxload 2/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Full-duplex, 100Mb/s, media type is 10/100BaseTX
input flow-control is off, output flow-control is unsupported
ARP type: ARPA, ARP Timeout 04:00:00
Last input never, output 00:00:01, output hang never
Last clearing of "show interface" counters 00:01:55
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 810000 bits/sec, 128 packets/sec
5 minute output rate 47000 bits/sec, 82 packets/sec
14005 packets input, 11967335 bytes, 0 no buffer
Received 0 broadcasts (0 multicast)
0 runts, 0 giants, 0 throttles
0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
0 watchdog, 0 multicast, 0 pause input
0 input packets with dribble condition detected
9381 packets output, 694480 bytes, 0 underruns
0 output errors, 0 collisions, 0 interface resets
0 babbles, 0 late collision, 0 deferred
0 lost carrier, 0 no carrier, 0 PAUSE output
0 output buffer failures, 0 output buffers swapped out


heres the cisco 2600 stats

interface FastEthernet0/0
description DR Replication Mask
ip address 172.28.2.97 255.255.255.224
speed 100
full-duplex
no cdp enable
end

DR#ping ip
Target IP address: 172.29.5.129
Repeat count [5]: 1000
Datagram size [100]: 1500
Timeout in seconds [2]:
Extended commands [n]: y
Source address or interface: 172.28.2.97
Type of service [0]:
Set DF bit in IP header? [no]:
Validate reply data? [no]:
Data pattern [0xABCD]:
Loose, Strict, Record, Timestamp, Verbose[none]:
Sweep range of sizes [n]:
Type escape sequence to abort.
Sending 1000, 1500-byte ICMP Echos to 172.29.5.129, timeout is 2 seconds:
Packet sent with a source address of 172.28.2.97
!!!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!.!!.!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!.!!!!!.!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!.!!!!
!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!.!!!
!!.!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!.!.!!!!!!!!!!!.!!!!!!!!!!.!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!.!!!!!!!!!
!!!!!.!!!!!!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!.!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!.!!!!!!.!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!..!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!.!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!
Success rate is 96 percent (967/1000), round-trip min/avg/max = 12/16/56 ms



DR(config-if)#mtu 1492
% Interface FastEthernet0/0 does not support user settable
mtu.

I cant change the mtu speed on this interface..

thanks again

 

[burtsbees]

router(config-if)#ip tcp adjust-mss 1452
on the interface you want to adjust, if you want tcp=1492, as there is 40 bytes overhead for something (can't remember).

Burt

 

[spivy66]

I changed it but it still says the mtu speed for that interface is 1500

 

[spivy66]

never mind. i did a sh run inter .What the difference between ip tcp mms and windows size?

 

[spivy66]

guys I fixed it, thanks for the input

 

[burtsbees]

I can't remember the difference, or exactly what mss is, but I had to change it on my home Cisco 837 for dsl.

Burt

 

[helpdeskdan]

What was the answer, spivy66?

 

[spivy66]

it was mss size,need to be changed to 1452

 

[plshlpme]

so you changed that setting on the router right?
you should never need to set that on a lan switch...

 

[spivy66]

right,,made the change on the router.just for my knowing ,why shoulnt you make that change on the switch?

 

[helpdeskdan]

I'm not sure exactly what kind of network topology requires you to set a mtu that low. As for a switch, you should never need to set the mtu on a switch because the standard is 1500. The only possible time I can think of would be if trunking requires transit through a switch that does not understand 802.1q. For such a purpose, I believe 1496 would account for the oversized frame.

 

[plshlpme]

if your internet connection uses PPPOE then the mtu needs to be set to 1492 and im not sure why but that mss 1452 is a must or youll see most sites will not be reachable..

by lowering the mtu your effectivly lowering the throughput so on an ethernet lan where all devices are using a 1500 byte frame lowering the mtu to 1492 or 1452 is just going to slow everything down.

the internet is where you need this adjustment (because of the overhead your ppp authentication requires) so by doing it at the edge of your network (gateway) you ensure that only the internet bound traffic affected.

 

[spivy66]

thanks guys.I'll make the change

 

[burtsbees]

ip tcp adj-mss 1452
does NOT lower mtu to 1452---it puts it at 1492

Burt

 

[helpdeskdan]

Ah, I see, thanks Burt. It's not a command I've seen before.