Replication VLAN trunking issues

[cratejockey]

We are working to create a replication VLAN between our primary server room and our DR site in another building.

Our basic config is this:

CORE Switch 6509 (CatOS)This is the primary router and interconnect point for the rest of our facility.

DR Building 4507R (IOS) This switch acts as the uplink for this building and its staff to the Core as well as the switch in our DR DataCenter. HSRP is also configured for 2 edge locations that have dual links to both the 6509 and 4507 for seemless failover to the DR resources.

We have ran 2 pairs of fiber from the 6509 to the 4507. One pair acts as a trunk for normal uplink to the 6509. However the other pair is to act as a replication link only for VLAN 172. Currently we have configured both sides as follows:

6509:

clear trunk 4/5 1-171,173-199,201-1005,1025-4094

set trunk 4/5 auto dot1q 172,200

4507R:

interface GigabitEthernet3/5

description ***Replication***

switchport trunk encapsulation dot1q

switchport trunk allowed vlan 172,200

switchport mode trunk

The 6509 is showing that the link is live and that it is in trunk mode. However the 4507 is showing down and down with no indicator light and no CDP info from that link to 6509.

If I change configuration on the 4507 I actually see the trunk go down on the 6509 so there is some sort of communication happening there. I'm just not sure where to go from here.

Thanks for any input you may have.

[cratejockey]

Glen,

Thanks for your post. I want to try it right now but I ended up opening a TAC case. Last I knew the Engineer was cool with the config that we had tried to implement. He was reviewing the show Tech for both units and trying to re-create the problem in the lab.

If I am unable to get a resolution from TAC by noon I'm going to try your recommendation. Etiher way I will post the resolution or lack of one later today.

Thanks for everyones help so far.

[cratejockey]

Well I'm all fixed. Big thanks to everyone who jumped in on this. And a bigger thanks to Louis at TAC.

I'm going to add more detail to this so others who might run accross this can identify this issue.

The overall solution was using the:

speed nonegotiate

command. With this in place the 4507 immediately established link in both access mode (which we had fallen back to as a test mechanism) then as a trunk when we applied the interface trunking commands. Below are the configs as the were in the end for this trunk.

6509:

clear trunk 4/5 1-171,173-199,201-1005,1025-4094

set trunk 4/5 nonegotiate dot1q 172,200

4507R:

!

interface GigabitEthernet3/5

description ***Replication***

switchport trunk encapsulation dot1q

switchport trunk native vlan 172

switchport trunk allowed vlan 172

switchport mode trunk

switchport nonegotiate

speed nonegotiate

Here is the interesting thing about all of this. None of my other trunk or access ports require this command. Throughout this process I have replicated cofigs from exiting WORKING trunks and they do not work. In my troubleshooting with TAC we tried 3 ports on 2 differnt line cards in both the 6509 and the 4507. Each time with identical results. I was beggining to think it was a fiber issue until we used the fiber path (includeing patch cables) that was working as the path for this trunk. Again no joy.

In the end the solution above reguarding speed non-negotiation worked. Whats odd is that 1000 Base GBICS are not supposed to have speed settings. I tried the normal ones like speed 1000 and duplex full only to be told that my GBICs do not support those commands. TAC was even confused as to why speed commands would come into play with this solution.

With that said it works. TAC is keeping my case open to come to a reason as to why this happened. I'm hoping it will unlock the key to some random trunk and port-channel issues we are also having here.

Again thanks.