[j-nsp] bfd on LAG

[Mark Tinka]

On Wednesday 03 December 2008 14:11:38 Richard A Steenbergen 
wrote:

> BFD on a LAG is completely pointless. The BFD packets
> will only be transmitted over a single member link (since
> BFD works at the IP layer, and will be hashed onto one
> member), and you have no guarantee that the return
> packets will come back over the same LAG member link.

I recall a horror story at a regional meeting earlier in the 
year where an operator said running BFD on a 10Gbps-based 
LAG (JunOS) was bad for them because a failure of one link 
brought down the whole LAG from the perspective of the 
routing protocol, when BFD was enabled.

I recall someone from Juniper saying they were aware about 
this issue, and didn't recommend BFD on LAG's at that time, 
but didn't stay long to finish the conversation.

We haven't had to use LAG's on Juniper yet, so I can't speak 
to this, but it'd be nice to lab this in case you do.

Cheers,

Mark.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 835 bytes
Desc: This is a digitally signed message part.
URL: <https://puck.nether.net/pipermail/juniper-nsp/attachments/20081203/d77d1855/attachment.bin>