Best practice backup and primary link over different media STP defaults

[barkerr01]

Good afternoon,

We have a fibre 1Gbps link between two office buildings. This runs down a single 4-core fibre. The remote office building has grown over the past few years and there is a concern that the single-fibre could get damaged and render the remote offices unreachable.

Since the units are only 30 metres apart (but separated by a small river) we are planning on installing some wireless bridges. This should be a quick and reliable solution in the event the primary fibre is ever damaged.

The fibre interface is on a 1Gbps SFP. Although the bridges are 1Gbps, we will just use one of the spare copper 100Mbps interfaces on the local and remote switches.

Is is safe enough to rely solely on default STP configurations to prioritise the traffic via the 1Gbps interfaces unless there is a failure?

[Reza Sharifi]

Hi,

Yes, it is a good idea to use the fiber as primary and use the copper with the wireless bridge as backup. You can manipulate STP cost to make the fiber preferable and the copper blocked until the fiber fails.

HTH

[barkerr01]

Thank you Reza - from what I have seen, 1Gbps interfaces get a default STP port and VLAN cost of 4 and 100 Mbps is 19 so I imagined this would work exactly as required without changing any defaults. I just wasn't sure if it was be quite that easy!

Once we have it all in place this weekend, I can so some real-world testing to make sure the 100Mbps port is definitely blocked whilst the 1Gbps is active...

[Reza Sharifi · Accepted solution]

You are exactly right. The 100Mb circuit should be blocking by default and the gig forwarding. If not you can always change the STP cost. And yes, I recommend doing some testing during off hours to make sure it actually works as expected before you put it in production.

Good Luck with your project!

[barkerr01]

The link is in and operational - dropping the fibre connections fails the connection over as required and fails back when the link is restored.

Spanning-tree on the local switch is blocking all VLANs as expected for the wireless bridge interface. On the remote side, spanning-tree is not blocking the port - I assumed this is normal behaviour as only one side needs to be blocked

However, we are getting MAC-FLAP alerts on the remote switch between the fibre and copper interfaces. We are also seeing some VLAN outages on the management VLAN. In order to access the management address, we have set the trunk port to have a native VLAN in the management VLAN.

I tried removing this but then we can't access the bridge monitor/management page. 

I found this community post:

https://community.cisco.com/t5/switching/spanning-tree-mac-flapping-with-native-vlan-changed/td-p/3699810

It suggests changing the native VLAN for all trunk ports - I'm a bit wary of doing this. Would this change only be needed on the local switch and remote switch trunk ports or all trunk ports on the network in that location?

For now I've shut down the interface on the local switch.