[j-nsp] vpls loop avoidance

[David Ball]

On 20 October 2011 16:02, Phil Mayers <p.mayers at imperial.ac.uk> wrote:
> On 10/20/2011 08:28 PM, Sebastian Wiesinger wrote:
>>
>> One problem with active/standby is mac aging on the CE switch(es).
>> When the primary PE goes down, the backup PE takes over and the
>> CE-switches learn mac addresses on the link towards the backup PE.
>> Now when the primary link comes online again, the VPLS on the primary
>> PE goes active, forcing the standby PE to deactivate VPLS. The CE
>> switches still have the mac addresses learned on the port to the
>> standby PE so you have to wait until the mac addresses age out (or a
>> packet with that source mac comes in from the primary PE).
>
> I can think of a few ways vendors could solve this. Most simply, the backup
> PE could briefly down the link, to trigger an FDB flush. Hell, you could
> probably script this using EEM in cisco-land.

  That's assuming the LER which terminates the VPLS connects directly
to the customer, and there are no L2 agg switches between the cust and
the LER.  I wondered about the idea of the new master PE sending
gratuitous ARPs towards the CE for any MACs in a given VPLS's FDB, but
depending on the network, that could be quite a load of MACs I guess.

David