[Users] opensm params

[Hal Rosenstock]

No partition file is same as partition file with:
Default=0x7fff : ALL=full ;
so just full default partition.

Are smpquery commands using LID or DR path ? If it's former, LID routing is
broken somehow.

For some port stuck at FDR10, is FDR enabled on both local port and peer
port ?

What is OpenSM version string ?

On Mon, Jun 5, 2017 at 10:40 AM, Michael Di Domenico <mdidomenico4 at gmail.com
> wrote:

> On Mon, Jun 5, 2017 at 10:32 AM, Hal Rosenstock
> <hal.rosenstock at gmail.com> wrote:
> >
> > If you want to disable fdr10, there is fdr10 option in opensm.conf which
> > should be set to 0:
> >
> > # FDR10 on ports on devices that support FDR10
> > # Values are:
> > #    0: don't use fdr10 (no MLNX ExtendedPortInfo MADs)
> > #    Default 1: enable fdr10 when supported
> > #    2: disable fdr10 when supported
> > fdr10 0
> >
> > You may need to bounce the links which are at fdr10 but the link would've
> > come up at FDR over FDR10 if that negotiation was successful so I don't
> > think changing fdr10 option will have an effect. Are your cables good for
> > FDR ?
>
> Yes, i tried switching that flag, it doesn't seem to make a difference.
>
> as for the cables, they were supplied by Cray.  i haven't specifically
> checked them, but i trust they are FDR capable.
>
> i did bounce the links and the whole machine, but it seems to keep
> coming up in that state.
>
> > I have no idea what is causing bad Qkey traps to the SM in your
> > configuration.
>
> me either, i reset back to a completely default opensm config and it
> still trips Bad Q_Key and doesn't let me actually talk on the fabric.
> the port shows LinkUp, but smpquery/ipoib data fail
>
> is there a possibility that there's a programming error on the fabric
> with the partitions, because i have QDR and FDR adapters on the same
> fabric.  perhaps something it can't figure out on its own and i have
> to hard code into the config?
> _______________________________________________
> Users mailing list
> Users at lists.openfabrics.org
> http://lists.openfabrics.org/mailman/listinfo/users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/users/attachments/20170605/538eb43f/attachment.html>