[c-nsp] loop guard still useful?

Lee ler762 at gmail.com
Mon Jan 18 14:03:26 EST 2016


Thanks for the response.

On 1/18/16, Michele Bergonzoni <bergonz at labs.it> wrote:
>>  Using the dispute mechanism included in the IEEE 802.1D-2004 RSTP
>> standard... I'm wondering if there's any reason to keep loop guard
>> configured
>
> I think the dispute mechanism can detect unidirectionality where data out of
> the designated bridge is lost (which is enough to prevent loops), not the
> unidirectionality in the other direction.

Which is my point .. or question - enable RSTP on all the switches in
the network and you don't need loop guard.  Correct?


> So the dispute does half of what UDLD does, if I got it right.
>
> Loop guard is different, it protects only from self-looped ports.

My understanding is that it keeps stp blocked ports blocking if the
other side stops sending BPDUs:

  http://www.cisco.com/c/en/us/support/docs/lan-switching/spanning-tree-protocol/10596-84.html

  The loop guard feature makes additional checks. If BPDUs are not
received on a non-designated port, and loop guard is enabled, that
port is moved into the STP loop-inconsistent blocking state, instead
of the listening / learning / forwarding state. Without the loop guard
feature, the port assumes the designated port role. The port moves to
the STP forwarding state and creates a loop.

and a lot further down

  loop guard does not work on shared links or in situations where the
link has been unidirectional since the link-up.


So it seems like loop guard isn't needed if rstp is enabled.


> I don't
> know if the wording of RSTP are written in a way to protect you from that,
> but I'm sure that the original STP standard was written in such a way that
> any compliant implementation was unable to block the loop caused by a
> self-looped port.

If self-looped means the port sends a frame and then receives the same
frame, you're right, stp doesn't protect you from that.

> Most vendors quietly worked around this, and I don't know if 802.1d
> corrected this error in the previous standard. I know that it is very
> unlikely to find a switch whose STP can't protect you from such a
> situation.
>
> So I bet that if you use RSTP you can disable loopguard, and if you like
> UDLD there is still a reason to use it.

No, I don't like UDLD at all - too many bad experiences with it.  It
was a necessary evil with cat5500s and 100Mb fiber connections, but
you don't need UDLD on 1Gb fiber links.

Thanks,
Lee


More information about the cisco-nsp mailing list