08.07.2010 01:09, David DeSimone пишет:
I am curious if others have experienced this problem.
We are setting up new solutions that require routing (not just
switching) of jumbo frames between racks. We'll set up a simple vlan
extended across the core with 3750E's at the edges:
system mtu jumbo 9000
system mtu routing 9000
vlan 2242
mtu 9000
interface GigabitEthernet1/0/9
switchport trunk encapsulation dot1q
switchport trunk native vlan 2242
switchport trunk allowed vlan 1,2242
switchport mode trunk
Can you show output of 'sh int gi1/0/9'?
What MTU you see in it?
IMHO, you need mtu command in interface config.
(routing config not shown, simple OSPF setup)
What I've seen a couple times now is that server admins report they
cannot ping across to the other rack using jumbo packets. Standard packet
sizes are fine, just large packets fail to route.
We waste some time verifying the router configuration and determine that
it looks correct, then in desperation, we simply try re-issuing this
command:
system mtu routing 9000
Then suddenly the router "wakes up" and starts routing the jumbo frames
correctly. No other changes needed. Before and after issuing the
command, the MTU settings are the same:
# show system mtu
System MTU size is 1500 bytes
System Jumbo MTU size is 9000 bytes
Routing MTU size is 9000 bytes
My only guess is that the jumbo MTU routing attribute somehow isn't
applied to the VLAN when the router is initially configured, because
there are currently no ports showing link in the VLAN, so the VLAN is
invisibly "admin down" until a server is eventually plugged in. It
seems that I have to reissue the "mtu routing 9000" setting in order to
get that VLAN to inherit the correct routing MTU setting.
But I'm just guessing. Anyone else seen this?
--
Sincerely yours,
Artyom Viklenko.
-------------------------------------------------------
[email protected] | http://www.aws-net.org.ua/~artem
[email protected] | ================================
FreeBSD: The Power to Serve - http://www.freebsd.org
_______________________________________________
cisco-nsp mailing list [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
