[c-nsp] L3 vs. L2 trunk connections to a 6509 core. Easyrouter-head question.

Grant Moerschel gm at wavegard.com
Mon Apr 14 16:22:46 EDT 2008


Sorry if I was not clear. My examples mixed some V8 and V10 info. To be
clear, the subject L3 devices here (3560 and 6509) have a trunk that
allows V1, 10, and 100. 

Currently On 6509:
clear trunk 9/28 2-9,11-99,101-1005,1025-4094
set trunk 9/28 on dot1q 1,10,100

Currently On 3560:
interface GigabitEthernet0/1
 switchport trunk encapsulation dot1q
 switchport trunk allowed vlan 1,10,100
 switchport mode trunk

The 3560 only has one SVI for V1. The 3560 participates successfully as
a VTP client. The 6509 has multiple SVI's for various VLANs including
V10 such as:

interface Vlan10
 description GA Floor 10
 ip address 10.16.10.1 255.255.255.0
 ip helper-address 172.16.3.41
 ip helper-address 172.16.3.48
!

I assume I only want one SVI for each VLAN and I want them at the core.
I also recognize that I am extending VLANs to the access switches.  I
guess I am getting stuck on why if ip routing and EIGRP is turned on on
the 3560 things work when in my mind since I have a trunk between the
devices I should need to be running an IP protocol when using a trunk.
I am choosing to extend the VLAN from the core to the access switch and
thus forwarding alone should work (??).

Therefore is it accurate to say that if I want the 3560 to be a Layer 2
device only I have to:
- disable ip routing and eigrp
- define a management interface (the current VLAN1 SVI)
- default-gateway for the management vlan on the 3560
- trunk as described above

Or if I want the 3560 to be a Layer 3 device I have to:
- enable IP routing
- enable EIGRP consistent with the core router
- choose an completely different VLAN for the access switches and give
them an SVI at the access switch and let EIGRP handle the routing (??)

Thanks again


-----Original Message-----
From: Mike Louis [mailto:MLouis at nwnit.com] 
Sent: Monday, April 14, 2008 3:27 PM
To: Grant Moerschel; cisco-nsp at puck.nether.net
Subject: RE: [c-nsp] L3 vs. L2 trunk connections to a 6509 core.
Easyrouter-head question.

>From your configurations that you posted there is no way that local VLAN
8 traffic would ever be able to cross the trunk. You are not allowing it
on the trunk all. Any traffic assigned to vlan 8 on the 3560 will be
switched on the 3560 only due to the trunking commands. The routing
table has nothing to do with reachability for VLAN 8 in the case of the
3560 since the default gateway for VLAN 8 resides on the 6500 not the
3560 and therefore VLAN 8 traffic on the 3560 would never have a gateway
to use that would subject it to being routed via the routing table on
the 3560. Does that make sense?

6500

interface Vlan8
 description GA Floor 8
 ip address 10.16.8.1 255.255.255.0
 ip helper-address 172.16.3.41
 ip helper-address 172.16.3.48

There is no VLAN 8 SVI configured on the 3560


3560

interface GigabitEthernet0/1
 switchport trunk encapsulation dot1q
 switchport trunk allowed vlan 1,10,100 <--VLAN 8 is not permitted to
cross this trunk
 switchport mode trunk
!
interface GigabitEthernet0/2
 switchport trunk encapsulation dot1q
 switchport trunk allowed vlan 1,10,100 <--VLAN 8 is not permitted to
cross this trunk
 switchport mode trunk


>From your quote:




-----Original Message-----
From: Grant Moerschel [mailto:gm at wavegard.com]
Sent: Monday, April 14, 2008 1:15 PM
To: Mike Louis; cisco-nsp at puck.nether.net
Subject: RE: [c-nsp] L3 vs. L2 trunk connections to a 6509 core.
Easyrouter-head question.

The only SVI configured on the 3560 is the VLAN1 interface. Is this
incorrect?  My assumption was that the V1 SVI, a default route for V1,
and "no ip routing" was all that was needed but when I did that I
couldn't talk with the other VLANs on the switch.  I may have overlooked
something.

Given my description of the 3560 (routing on, one SVI for VLAN1, one
trunk to the core), "do you think that the local V8 traffic on the 3560
is being routed via VLAN1 in contrast to being tagged and forwarded as
VLAN8 via the trunk to the core"

-----Original Message-----
From: Mike Louis [mailto:MLouis at nwnit.com]
Sent: Monday, April 14, 2008 12:11 PM
To: Grant Moerschel; cisco-nsp at puck.nether.net
Subject: RE: [c-nsp] L3 vs. L2 trunk connections to a 6509 core.
Easyrouter-head question.

If you are going to use the 3560 as a L2 switch you can disable ip
routing. You will need to define a management interface and
default-gateway for the management vlan on the switch only. All VLANs
including management will be L2 only. Do not configure VLAN SVI if you
only want to use the 3560 as a layer 3 switch. Did you configure EIGRP
on the 3560 as well? If so that may be why you are getting routing
information via VLAN 3 to the 6509.

-----Original Message-----
From: cisco-nsp-bounces at puck.nether.net
[mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of Grant Moerschel
Sent: Monday, April 14, 2008 11:45 AM
To: cisco-nsp at puck.nether.net
Subject: [c-nsp] L3 vs. L2 trunk connections to a 6509 core. Easy
router-head question.

I am trying to determine what is normal here.  This is the situation. I
have a 6509 Sup1a/MSFC2 running the latest CatOS/IOS for that hardware.
I run EIGRP. I have several VLANs on the core and use VTP.  When I trunk
a 2950 L2 switch to the core I specify VTP on the 2950. No problem. All
VLANs show up on the 2950.  For our eighth floor, for example, I trunk
V1, V8, and V100 and clear all other VLANs from the trunk. V100 is for
Voice and V8 is for most PCs for Data.  My assumption is that if an
access port is "switchport access vlan 8" and the PC is plugged in to
that port, he's on V8. To get off that broadcast network to some other
destination, he hits the gateway which is the L3 interface on the core
6509 MSFC2.  I get all this.

Here's the question.  If my access switch is a 3560 which is a Layer 3
switch, it seems that I have to have it participate in EIGRP to make it
work.  But this does not make sense because for this application I still
have a single trunk to the core from the 3560. If I trunk V1, V10, and
V100 from the 3560 to the 6509, doesn't communication just go down the
trunk to the 6509 L3 gateway?  For example, if my PC is on a 3560 V10
port and needs to hit something on V8, shouldn't the 3560 forward the
frame down the trunk to the 6509 which'll route onto V8, up V8's trunk
to that access switch and forward it out to the destination?   If I look
at my routes on the 3560, they all say "to get to V10 you must go
through the 6509 V1 layer 3 interface" (I hope that makes sense).

Should I turn off "ip routing" on the 3560 in order to mimic the setup
of the 2950 <----> 6509 trunk link?  What am I missing?

Thanks

~~~~
Grant P. Moerschel
WaveGard, Inc.
gm -at- wavegard -dot- com
~~~~


_______________________________________________
cisco-nsp mailing list  cisco-nsp at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/


Note: This message and any attachments is intended solely for the use of
the individual or entity to which it is addressed and may contain
information that is non-public, proprietary, legally privileged,
confidential, and/or exempt from disclosure.  If you are not the
intended recipient, you are hereby notified that any use, dissemination,
distribution, or copying of this communication is strictly prohibited.
If you have received this communication in error, please notify the
original sender immediately by telephone or return email and destroy or
delete this message along with any attachments immediately.



Note: This message and any attachments is intended solely for the use of
the individual or entity to which it is addressed and may contain
information that is non-public, proprietary, legally privileged,
confidential, and/or exempt from disclosure.  If you are not the
intended recipient, you are hereby notified that any use, dissemination,
distribution, or copying of this communication is strictly prohibited.
If you have received this communication in error, please notify the
original sender immediately by telephone or return email and destroy or
delete this message along with any attachments immediately.



More information about the cisco-nsp mailing list