VLAN MISMATCH

[londint]

I have just recently configured a 1GB line between our local site and remote.

The line is in addition to a 100MB line. The 100MB line have been configured in the FWLON vlan 5 at the local site and configured for FWREM 10 at the remote end.

I just configured the 1GB line to the same VLAN at both the local and remote end. And this worked fine.

But now it comes up with

%CDP-4-NVLANMISMATCH:Native vlan mismatch

at the local end. The 1GB is configured switch to switch while the 100MB line is configured router to 1router.

The vtp at the remote end (remote) is different from the vtp at the local end (local). The Switch at the remote end is a server with its own vtp name remote and the switch at the local is also part of the local vtp.

I want to keep this separate so that I don't have a problem adding devices to the remote end and I did not want them to exchanged vlan information.

Please what am I doing wrong?.

Vlan 5 port 4/3 vtp local ------- vlan 10 port fa0/2 vtp remote

Thanks

[DWAM_2]

Hello,

when i pass your message on the decoder tools

1. %CDP-4-NVLANMISMATCH: Native vlan mismatch detected on port [dec]/[dec]

This message indicates that the native VLAN is different from the one set on at least one of the neighboring ports. [dec]/[dec] is the module number/port number of the mismatched port.

Recommended Action: Check the native VLAN settings. If the VLAN configuration does not match, set it appropriately. If no apparent mismatch is found, contact your technical support representative.

If i were you and if you know that you don't have error on the configuration (you have checken the vlan setting for native vlan), you can disable cdp on the both interfaces.

Regards.

[milan.kulik]

Hi,

it is not necessary to disable CDP.

You can configure the switches to use CDP ver. 1 which doesn't send the native VLAN info.

Another possibility is

set logging level cdp 3 default

on CatOS switch which stops CDP level 4 syslog messages.

Regards,

Milan

[londint]

Thanks Milan

But does this mean the way i have configured the vtp is ok. I dont need to change the vtp settings.

So why is it sending native VLAN info? As I would not like to turn off the syslog messages if there really is something wrong.

Thanks

[peter.scott]

It's important to have the native vlan the same at both ends.

Please post the configs for both ends.

Thanks

[londint]

Thanks

At the main office i have

LLLLL sh por 4/3

Port Name Status Vlan Duplex Speed Type

----- -------------------- ---------- ---------- ------ ----- -----------

4/3 connected 5

full 1000 1000BaseSX

LLLLL sh trunk 4/3

Port Mode Encapsulation Status Native vlan

-------- ----------- ------------- ------------ -----------

4/3 off dot1q not-trunking* 5

Port Vlans allowed on trunk

-------- -----------------------------------------------------------

4/3 1-1005,1025-4094

Port Vlans allowed and active in management domain

-------- -----------------------------------------------------------

4/3 5

At the branch office, i have

RRRRR# sh run

interface fa0/2

switchport access vlan 10

no ip address

RRRRR #sh interfaces Fa0/2 trunk

Port Mode Encapsulation Status Native vlan

Fa0/2 desirable negotiate not-trunking 1

Port Vlans allowed on trunk

Fa0/2 10

Port Vlans allowed and active in management domain

Fa0/2 10

Port Vlans in spanning tree forwarding state and not pruned

Fa0/2 10

LLLL is a Cat6509 and RRRR is a Cat 3550.

Thanks

[milan.kulik]

To be clear:

My understanding is that you are connecting two switches, each of them in a different VTP domain.

Both interconnected ports are access ones, one is assigned to VLAN5 in one VTP domain, the second to VLAN10 in the other VTP domain.

So the VTP configuration is correct.

But Cisco supposes using only one VTP domain in your network, so having one line end in VLAN5 and the other end in VLAN10 is incorrect under this assumption.

The missleading info in the error message is the NATIVE VLAN keyword used.

CDP ver 2 sends the native VLAN info in the CDP packet, even while the port is an access port. The value is equal to the VLAN ID the port belongs to in that case.

So the result is a syslog message about incorrect native VLAN setting.

You can just ignore the error messages. Or you can suppress them via configuration changes I suggested.

I'm using

set logging level cdp 3 default

in my network with no problem.

Regards,

Milan

[londint]

Thank you so much.

SO the fact that i have trunking off at one end (Main office) and trunking desirable (remote office) does not pose a problem?

and the * by the status on the trunk in the main office -

* - indicates vtp domain mismatch ( from cisco)

Please can you help me here.

Is it still the same issue?

Thanks

[milan.kulik]

Well,

having one line end trunking off and the other one is allowed.

But it is not a good practice. The desirable end is trying to become a trunk and keeps sending negotiation (DTP) frames. I'd configure switchport mode access on both ends.

There is also not a good Cisco aproach on Cat3550s allowing to configure switchport access VLAN10 and leaving native VLAN the default value VLAN1.

I know two native VLAN definitions by Cisco:

a) VLAN which is sent non-tagged on 802.1q trunk

b) VLAN to which the port belongs if it doesn't trunk for any reason.

In your case a) is VLAN1 and b) is VLAN10.

I'm not sure which definition is used by CDP on Cat3550.

(Try sh cdp nei det on the other line end and you should see which VLAN is advertised as the native one.)

MAYBE if a) is used and you configure switchport thrunk native vlan 5 you might stop the errror messages.

Regards,

Milan

[londint]

Thanks Milan

DOing a show cdp on the Cat 6500 shows it as

Platform: cisco WS-C3550-48

Port-ID (Port on Neighbors's Device): fa0/2

VTP Management Domain: remote

Native VLAN: 10 (Mismatch)

SO should I change the native vlan on the Cat 3550 to vlan 10. There is this option:

switchport trunk native vlan 10. WIll this sove it. I did not know there was a native vlan config before.

Is it ok if i configure the local as mode auto and remote as mode desirable.

Thanks

[milan.kulik]

"SO should I change the native vlan on the Cat 3550 to vlan 10. There is this option:

switchport trunk native vlan 10. WIll this sove it"

NO, VLAN10 is already advertised as native by Cat3550.

The only ways to stop error messages are:

a) put both line sides to the same VLAN (but you don't want to do it)

b) change the error message level as I suggested already.

"Is it ok if i configure the local as mode auto and remote as mode desirable. "

NO, you need the same VTP domain on both trunk sides if trunk negotiaition is involved. And you can't trunk anyway while only VLAN10 allowed on one side and only VLAN5 allowed on the other side.

Read http://www.cisco.com/warp/customer/473/103.html for better problem understanding.

The final recommendation:

on 3550:

conf t

int fa 0/2

switchport mode access

on 6500:

set trunk 4/3 off dot1q

set logging level cdp 3 default

Regards,

Milan

[londint]

Thank you ever so much. I have changed the logging level.

Thanks

[bins]

And Thank you all ,I've solved a prob like this!