About a day ago, I wrote

> I've got an ELC.  I'm trying to move a mud I co-run to it.  The mud
> takes up nearly 30M of virtual space, so I loaded the ELC up on
> memory.

> Trouble is, it crashes randomly.

> [...]

It now appears that it's something subtle broken in the ELC support.

I've got three IPXes, and they've all been rock-solid.  I finally took
one of them out of its previous service and tried the mud there.

Across three different ELC boards, the longest I ever had the mud stay
up was about 15 minutes, from restart to fatal signal.

I've never yet seen a crash on the IPX.  The latest run has been up
over two hours - eight times the longest survival I saw on an ELC.  A
previous run lasted at least an hour and I think a good deal more.

With the very same SIMMs I was using in the ELCs.

With the very same disk, including the same bootblocks, kernel, and all
of userland, in particular all the mud-related files (executable, data
files, etc).  The first tests even used the same SCSI cabling, though
I've now moved the disk inside the IPX.

Anyone have enough idea what the differences between the ELC and IPX
are to take even a wild stab at where to look for the problem?  If it
would help I can set someone up with serial console access to an ELC as
a crash-and-burn machine.  (All the machines involved have been running
with serial console.)

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B