[ofa-general] Still looking for help debugging a problem.

[Roland Dreier]

 > May  9 08:38:54 compute-0-4.local kernel: mlx4_core 0000:02:00.0:
 > Internal error detected:
 > May  9 08:38:54 compute-0-4.local kernel: mlx4_core 0000:02:00.0:
 > buf[00]: ffffffff

 > The HCA in question is a Connect-X and the problem only seems to happen
 > with this node.

Sounds like a hardware problem.  Try reseating everything etc.