Trunking problem on 2900XL

oxfur

I'm having some problem with trunking 2900XL switch to router 2610. For switch, SW Version 12.0(5)WC17. I created Vlans 2, 3 and set port fa0/12 to trunk,native 1, connecting to e0/0 on 2610 router. Router e0/0 subinterfaced, e0/0.1-.3, dot1q encapsulation. Below is the output for the router e0/0.

interface Ethernet0/0
no ip address
half-duplex
!
interface Ethernet0/0.1
encapsulation dot1Q 1 native
ip address 192.168.1.1 255.255.255.0
no snmp trap link-status
!
interface Ethernet0/0.2
encapsulation dot1Q 2
ip address 192.168.2.1 255.255.255.0
no snmp trap link-status
!
interface Ethernet0/0.3
encapsulation dot1Q 3
ip address 192.168.3.1 255.255.255.0
no snmp trap link-status


2900XL ports 2,4 for Vlan 2, 3, 6 for Vlan 3. Connecting f0/12 to e0/0 on router I keep getting this error below. Not sure what's going on here. Funny thing is I do this on simulation software, and it works as expected; hosts ping their gateways and each other. Not sure what's going on here. Any suggestions?

Output for f0/12

---

!
interface FastEthernet0/11
!
interface FastEthernet0/12
switchport mode trunk
!
interface FastEthernet0/13


Error Message

---

1d11h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors1d20h: %LINK-4-ERRO
R: FastEthernet0/12 is experiencing errors
1d20h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d22h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors

tech-airman

oxfur,

oxfur wrote:

I'm having some problem with trunking 2900XL switch to router 2610. For switch, SW Version 12.0(5)WC17. I created Vlans 2, 3 and set port fa0/12 to trunk,native 1, connecting to e0/0 on 2610 router. Router e0/0 subinterfaced, e0/0.1-.3, dot1q encapsulation.

Can you please separate the exact steps you took on which networking device? I just can't follow what you did or didn't do.

oxfur wrote:

Below is the output for the router e0/0.

interface Ethernet0/0
no ip address
half-duplex
!
interface Ethernet0/0.1
encapsulation dot1Q 1 native
ip address 192.168.1.1 255.255.255.0
no snmp trap link-status
!
interface Ethernet0/0.2
encapsulation dot1Q 2
ip address 192.168.2.1 255.255.255.0
no snmp trap link-status
!
interface Ethernet0/0.3
encapsulation dot1Q 3
ip address 192.168.3.1 255.255.255.0
no snmp trap link-status

For an 802.1q implementation, what is a unique requirement for the native VLAN?

oxfur wrote:

2900XL ports 2,4 for Vlan 2, 3, 6 for Vlan 3. Connecting f0/12 to e0/0 on router I keep getting this error below. Not sure what's going on here. Funny thing is I do this on simulation software, and it works as expected; hosts ping their gateways and each other. Not sure what's going on here. Any suggestions?

Output for f0/12

---

!
interface FastEthernet0/11
!
interface FastEthernet0/12
switchport mode trunk
!
interface FastEthernet0/13


Error Message

---

1d11h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors1d20h: %LINK-4-ERRO
R: FastEthernet0/12 is experiencing errors
1d20h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d21h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors
1d22h: %LINK-4-ERROR: FastEthernet0/12 is experiencing errors

Questions:

1. What kind of cable did you connect the 2610 router's Ethernet port with the 2912XL switch?
2. Which version of IOS do you have on the 2610 router?
3. How come none of the switch ports other than FastEthernet0/12 are not configured in VLAN access mode?
4. How come you're connecting the FastEthernet0/12 on the switch to the Ethernet 0/0 port on the router? There's a hint above.

bighornsheep

tech-airman wrote:

For an 802.1q implementation, what is a unique requirement for the native VLAN?

untagged traffic.

tech-airman wrote:

[How come you're connecting the FastEthernet0/12 on the switch to the Ethernet 0/0 port on the router? There's a hint above.

Because the 2610 only have Ethernet ports unless if you add FE modules.

oxfur -> on port Fe0/12 on the 2900, issue "speed 10" and "duplex full" and see what happens.
Also, issue "duplex full" on port Eth0/0 on the 2610.

Netstudent

Also check that the 2900 is using dot1q and not ISL. Try Switchport trunk encapsulation dot1q on the switch's trunk port. As per bighorn, you can check for duplex issues by checking for collisions on the interface with the show interface command.

I think in some earlier versions of code, cisco wanted thier proprietary protocols to be the default config.

Then the 2950 came out and didn't even support ISL.

techairman, I have never witnessed a person answer so many questions with more questions ever. EVER!

oxfur

yep.. that did it. Netstudent thanks for the suggestion.

Also check that the 2900 is using dot1q and not ISL. Try Switchport trunk encapsulation dot1q on the switch's trunk port.

Your a life saver. Only, the native Vlan never routes to other vlans. Tried using Vlan 2 as native it failed, tried Vlan 3 as native, failed, by the way, Vlan 1 was failing too. All vlans but the native route to any other vlan, even to the native Vlan but not from the native out. Dont know if thats normal or another issue.

tech-airman

Netstudent wrote:

Also check that the 2900 is using dot1q and not ISL. Try Switchport trunk encapsulation dot1q on the switch's trunk port. As per bighorn, you can check for duplex issues by checking for collisions on the interface with the show interface command.

I think in some earlier versions of code, cisco wanted thier proprietary protocols to be the default config.

Then the 2950 came out and didn't even support ISL.

techairman, I have never witnessed a person answer so many questions with more questions ever. EVER!

Netstudent,

That's because I've learned that the answer to some problems is "it depends." Let's take a classic problem definition of "It doesn't work!" Do you mean to tell me that you're going to ONLY troubleshoot the problem based on the ONLY information provided of "It doesn't work?" Wouldn't you ask questions to find out as much as you can to help figure out what the actual problem is and so you can solve the problem?

APA

oxfur wrote:

yep.. that did it. Netstudent thanks for the suggestion.

Also check that the 2900 is using dot1q and not ISL. Try Switchport trunk encapsulation dot1q on the switch's trunk port.

Your a life saver. Only, the native Vlan never routes to other vlans. Tried using Vlan 2 as native it failed, tried Vlan 3 as native, failed, by the way, Vlan 1 was failing too. All vlans but the native route to any other vlan, even to the native Vlan but not from the native out. Dont know if thats normal or another issue.

What ever your set as the native VLAN must match on both sides of the trunk link....

e.g If you specified encapsulation dot1q 2 native on the router so VLAN 2 would be the native vlan on this trunk then on the switchport you must specify

switchport trunk native vlan 2 -> otherwise by default the switch will try and use vlan 1 as the native vlan and the trunk link will never form due to mismatching native VLAN'son either side of the trunk......

Also remember native vlan traffic is untagged only for 802.1q VLAN tagging...

ISL encapsulation still encapsulates native vlan traffic.... (26Byte ISL header and 4 Byte FCS) and as you found out on the 2900XL series switches they support ISL by default but also dot1q..

jezg76

If I am not mistaken, a trunk link can, and will, form and be operational with mismatched native VLANs on either side. I threw this in GNS3 real quick before I posted what I thought was true.

Cheating, I know.

APA

No it won't...... for 802.1q encapsulation Mismatching Native VLAN's will force a trunk link to fail.... Both sides have to agree on the native vlan which marks the vlan as untagged traffic.....

If you have mismatching native vlan's that means you have two vlans which will be untagged..... trunk link will get confused as to what traffic to tag and not tag....

Netstudent

Yeah APA you are right, but there are sometimes exceptions to the rule. IN this case, the trunk would error out and fail.

But in some instances with Q-in-Q tunneling like a Metro-E link where you tunnel through a switched service provider, a mismatched native vlan isn't that important. You could have a mismatched native vlan between your edge switch and the service provider's switch and as long as you aren't trying to communicate with a device within the service provider's network that you are tunneling through, you should be ok.

But that is different than a normal dot1q trunk.

oxfur

A.P.A

What ever your set as the native VLAN must match on both sides of the trunk link....

e.g If you specified encapsulation dot1q 2 native on the router so VLAN 2 would be the native vlan on this trunk then on the switchport you must specify

Checked that, didnt work. Pings to any vlan from non native vlan works, pings from native vlan to other vlans, no show. Changed the native to different vlans on both router and switch. All other vlans work but the native, doesnt matter what the native is.

APA

Netstudent wrote:

Yeah APA you are right, but there are sometimes exceptions to the rule. IN this case, the trunk would error out and fail.

But in some instances with Q-in-Q tunneling like a Metro-E link where you tunnel through a switched service provider, a mismatched native vlan isn't that important. You could have a mismatched native vlan between your edge switch and the service provider's switch and as long as you aren't trying to communicate with a device within the service provider's network that you are tunneling through, you should be ok.

But that is different than a normal dot1q trunk.

Yeah that's spot on.... q-in-q is a different ball game, was just helping this dude understand the basics

APA

oxfur wrote:

A.P.A

What ever your set as the native VLAN must match on both sides of the trunk link....

e.g If you specified encapsulation dot1q 2 native on the router so VLAN 2 would be the native vlan on this trunk then on the switchport you must specify

Checked that, didnt work. Pings to any vlan from non native vlan works, pings from native vlan to other vlans, no show. Changed the native to different vlans on both router and switch. All other vlans work but the native, doesnt matter what the native is.

Can you show your configs please........

Netstudent

oxfur wrote:

A.P.A

What ever your set as the native VLAN must match on both sides of the trunk link....

e.g If you specified encapsulation dot1q 2 native on the router so VLAN 2 would be the native vlan on this trunk then on the switchport you must specify

Checked that, didnt work. Pings to any vlan from non native vlan works, pings from native vlan to other vlans, no show. Changed the native to different vlans on both router and switch. All other vlans work but the native, doesnt matter what the native is.

Hey,

Try putting your native vlan subnet on the router's physical interface instead of putting a tagless vlan on a subinterface. The other day I was helping some guys out at school that couldn;t get a simple vlan lab up to save their lives. It ended up being an older version of code that didn't support the native keyword in the encap dot1q 1 native command.

I helped them put the subnet on the physical fa0/0 and it worked like a champ.