Any of you guys work with switches a lot?

I use banks of switches in my fiber, set up per neighborhood.

But really they are all linked via fiber so it's pretty much like a datacenter.

I'm using a few MPLS/VPLS tunnels to the switch banks.

I'm having a real hard time tracking down a loop issue though.

It's related to MikroTik lovely change in version 6.38 for STP/RSTP bridge 
stuff.

My problem is with an IBM switch I don't know very well.

I suspect I have settings wrong for STP/RSTP etc.

The switches are a simple configuration of VLAN 1/default for customer traffic, 
then another VLAN for management from the device at their house/MDU.

I pull out the VLAN management at the mikrotik and bridge it with an area wide 
L2 MPLS network.

My problem is the IBM switches keep shutting down access to customer data or 
management or the other switches they are connected to.
The crude method of recovery is reboot the switch.

At the 'head' of the VPLS network I use an EOIP tunnel to a hosted mikrotik 
that hosts the Dude and a few other management VM's on that layer2 network.

I'm regretting I did that and didn't create a more sophisticated routed 
management network.

But the hosted mikrotik is complaining of loop packets and sometimes just goes 
wild and shuts down it's EOIP connecting interface for 60 seconds.
I can't seem to find a way to tell it to NOT do that either, seems like it's 
built in to 6.38 no matter what Loop Protection I turn off or STP I disable.
It's not like it's a lot of traffic, it's just seeing some BPDU or STP type 
packets and killing itself.

What is the best practice for switches regarding RSTP/PVRST BDPU guards and STP 
root paths etc?

The switches typically have two uplink ports going to the router or another 
switch, the rest are customer facing ports.

Reply via email to