Is there any way to limit the MTU per L2 interface on this switch if you configure the system MTU for Jumbo frames?
Next question assuming no, are there any Cisco stackable switches with this capability?
-
Did any answer help you? if so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you could provide and accept your own answer.– Ron Maupin ♦Commented Aug 8, 2017 at 14:23
Add a comment
|
2 Answers
Catalyst 3750/3560 Series switches support an MTU of 1998 bytes for all 10/100 interfaces. All Gigabit Ethernet interfaces support jumbo frames up to 9000 bytes. The default MTU and jumbo frame size is 1500 bytes. You cannot change the MTU on an individual interface. You must set the MTU globally. Reset the switch afterwards for the MTU change to take effect. Gigabit Ethernet ports are not affected by the system mtu command; 10/100 ports are not affected by the system mtu jumbo command. If you do not configure the system mtu jumbo command, the setting of the system mtu command applies to all Gigabit Ethernet interfaces. If Gigabit Ethernet interfaces are configured to accept frames greater than the 10/100 interfaces, jumbo frames that ingress on a Gigabit Ethernet interface and egress on a 10/100 interface are dropped.
I'm not aware of any Cisco switches that support a per L2 interface MTU size. The Nexus class switches can use class of service, So in order to do increased MTU for your iSCSI traffic, you should create a policy-map identifying your iSCSI traffic, place the traffic in a separate queue with increased MTU.
It is important to know that in small catalysts like 3750/3560 there is different MTU for routed traffic and switched traffic. 1998B is largest packet you can route, even when using 1GE ports, the 9000B is strictly for switched traffic in 1GE ports.
It's not very uncommon for Cisco switches to support per interface MTU, ancient 3500XL does it as does 4500, 6500, 7600 at least.
3850 supports per-interface IPv4/IPv6 MTU, which is sufficient, per-interface L2 MTU is not important.
4500-X would support stack of two switches (VSS). 3850 would support stack of four switches (StackWise-480).
It's good to standardize L2 MTU as large as possible. Artificially reducing L2 MTU does little good to you, as dropped L2 frames will give no information to the sender of what went wrong. It is however crucial that L3 MTU is same in both ends of the L2 network.
There are some exceptions when you may not want to have large L2 MTU, which are related to buffer carving, but they are platform dependent issues and you shouldn't worry about them until you hit the issues, they're not common.
-
I agree a standard L2 MTU is a good thing. In this environment I have a L2 hand off to many customers, each a separate vlan and one customer needs a larger MTU. I don't want to setup a larger MTU for all customers.– JayCommented Aug 13, 2013 at 12:49
-
@Jay why particularly you don't want large MTU for all customers? Our product says 15xx something, while in practice >9k will pass in most cases.– yttiCommented Aug 13, 2013 at 13:09
-
A valid point, most of our customers are local IT departments that do not have experience with networking and might not understand MTU size issues. One option I was looking at was setting up a larger MTU on the two interfaces that one customer needs (Two side of our network, large MTUs are already turned up on all of our core gear). The other option is to turn on a larger MTU on all edge switches but this would be close to 300 switches and it seems like a large change for only one group. Might be the way to go though.– JayCommented Aug 13, 2013 at 18:19
-
Your customers don't have to understand anything, you're not creating them any MTU size issues by offering larger L2 MTU. L3 MTU mismatch is an issue, L2 not.– yttiCommented Aug 13, 2013 at 18:20