SUMMARY: Second Etherboard in PI

[Dave Olson]

In <9106082201.AA10957 at nazgul.physics.mcgill.ca> loki at NAZGUL.PHYSICS.MCGILL.CA (Loki Jorgenson) writes:


| 
| 	Hey ho (again)....
| 
| 	My original problem concerned a second Etherboard (CMC ENP10)
| in the VME slot of a 4D/25.  We were observing "enp0: died and restarted"
| and, in extreme cases, kernel faults due to bus errors.  This latter
| usually accompanied attempts to ftp files over the interface in question.
| 
| 	Swapping the board with a new one has cured the problem.  All 
| works well now.  However, for those considering installing a second
| board in a PI, I have one comment to make: 
| 
| 	The physical design of the board is lacking.  The surface of the
| board, as it rises above the connector backplane, extends 1.5" closer to
| the back of the cabinet than normal.  Consequently, once the standard SGI Ether
| cable is installed, you don't have a hope in hell of closing the box
| again.  The only reasonable solution is to get ahold of an Ether cable
| with a 90 degree angled head.  DEC makes a suitable one; they are 
| fairly rare.

I have to disagree here; we have at least 4 PI's in my area with enp
boards in them, with standard thick ethernet drop cables in them, and
no problems closing up the case.  The board extends to almost
exactly the same point as the integral ethernet.  Of course, it is 
possible that some new board type is being shipped, but I hadn't
heard of this problem before.
--

	Dave Olson

Life would be so much easier if we could just look at the source code.