Wow, an AlphaServer 1000A installation.

It's interesting that the kernel seemed to start up just fine. There seem
to be two possibilities. [1] (unlikely) we are referencing a bogus physical
address for some reason, or [2] (more likely) you have one or more defective
RAM addresses and we just aren't working around that as effectively as Digital
Unix is.

Did you run DU on that box recently? (Or, since the NetBSD attempt?)

If it is a real ECC error, there may be some things we can do in the kernel
to deal with it. For now, you might try removing half the RAM .. if it
still fails, put it back and remove the other half.

If that works it will give us some good information on the problem and we
can probably make a test kernel that can deal with it.

        Ross.Harvey@Computer.Org