Showing results for 
Search instead for 
Did you mean: 

Loop Protection in MLAG Desgin with non-exos edge

Contributor II

I'm looking for a solution to realise loop protection in MLAG design with non exos edge-switches.

- a dedicated STP-domain per edge-switch is a option when only 1 switch is in a location.
- How can I realise a comprehensive loop-protection between 2 edge-switches in a location, which are both connected via MLAG?

I've 2 use-cases...
- edge-switches are HP Procurve
- edge-switches are EOS Devices

Can somebody give me a bit input?

New Contributor III
If all Edge switches are EXOS, you can enable ELRP on them (all ports but the uplinks).

In your case, with HP and EOS switches, you'll have to go with Spanning Tree. Protect your edge ports and enable it on the LACP-enable uplinks. With no STP configured at all on your Cores, they will simply forward the BPDUs as they see it as normal Multicast traffic. Even through the MLAG links. See also here:

hmm... enabling stp on lacp uplinks would cause, that bpdu's would flood on all ports on mlag-peers ...thats not what I want...

enabling with spanguard etc on edge ports I will test; that could solve my problem with interrupts during mlag-upgrade, because there is no more STP (and so topo change) in core-connections.

Only use STP on edge port with an STP edge mode. You don't want TCN all over your network as soon as an edge port goes Up or Down.

Contributor II


On MLAG devices and lag-ports on edge-switches STP must be disabled...thats fact...

How can I prevet a loop between edge 2 & 3?