[ofa-general] Multicast traffic generates Bad P_Key trap in SM when working in partial member setup

[Olga Shern (Voltaire)]

On 7/8/08, Hal Rosenstock <hrosenstock at xsigo.com> wrote:
> Hi Olga,
>
> On Tue, 2008-07-08 at 18:27 +0300, Olga Shern (Voltaire) wrote:
> > Hi,
> >
> > I have tested the same scenario with OpenSM.
> > What I have expected to see is trap 259 because of switch enforcement
> > that is done by OpenSM. I don't see this trap, therefore I assume that
> > ANAFA
>
> Do you mean Anafa II ?
>
yes
> >  doesn't generate this trap (this is not mandatory only
> > optional).
>
> Right, this trap is optional and I don't know whether or not it is
> supported. Perhaps someone from Mellanox can comment on this to be sure.
>
> > I only see trap 257 that is generated by the node (with partial pkey)
> > that sends the multicast traffic.
>
> Only when there is no switch enforcement ?
>
no, when there is switch enforcement, I see this trap generated by the
node that sends multicast packets

> > The conclusion is that if there is switch enforcement the issue is not
> > severe and we can live with it :)
>
> Would that be the case if those optional traps were supported ?
>
I think no
> Thanks.


> -- Hal
>
> > Olga
> >
> > On 6/26/08, Hal Rosenstock <hrosenstock at xsigo.com> wrote:
> > > On Thu, 2008-06-12 at 16:44 +0300, Olga Shern (Voltaire) wrote:
> > > >         Would you try this with OpenSM (and validate your theory about
> > > >         getting
> > > >         switch bad PKey traps v. end port bad PKey traps) or does VSM
> > > >         have such
> > > >         a mode (ingress/egress partition filtering) ?
> > > >
> > > >         -- Hal
> > > >
> > > > Yes, I will test it with OpenSM
> > >
> > > Any update on this ? Thanks.
> > >
> > > -- Hal
> > >
> > >
>
>