lagg Interfaces - don't do Gratuitous ARP?

[Steven Hartland]

On 22/09/2016 03:58, Gleb Smirnoff wrote:
> On Wed, Sep 21, 2016 at 09:12:11PM -0400, Ryan Stone wrote:
> R> > IMHO, the original patch was absolutely evil hack touching multiple
> R> > layers, for the sake of a very special problem.
> R> >
> R> > I think, that in order to kick forwarding table on switches, lagg
> R> > should:
> R> >
> R> > - allocate an mbuf itself
> R> > - set its source hardware address to its own
> R> > - set destination hardware to broadcast
> R> > - put some payload in there, to make packet of valid size. Why should it be
> R> >   gratuitous ARP? A machine can be running IPv6 only, or may even use
> R> > whatever
> R> >   higher level protocol, e.g. PPPoE. We shouldn't involve IP into this
> R> > Layer 2
> R> >   problem at all.
> R> > - Finally, send the prepared mbuf down the lagg member(s).
> R> >
> R> > And please don't hack half of the network stack to achieve that :)
> R>
> R> The original report in this thread is about a system where it takes almost
> R> 15 minutes for the network to start working again after a failover.  That
> R> does not sound to me like a switch problem.  That sounds to me like the ARP
> R> cache on the remote system.  To fix such a case we have to touch L3.
>
> Does lagg(4) hardware address change when it failovers?
>
It moves the address between interfaces which typically moves it between 
switches too.