> sd0(ncr0:0:0): COMMAND FAILED (4 a8) @f86bd200
> [2 assert failures from lines 5172 and 5173 in sys/dev/pci/ncr.c]
> ncr0 targ 0?: ERROR(80:100:29) (8/13) @3870cc:e0000000
> ncr0: restart (fatal error)
> sd0(ncr0:0:0): COMMAND FAILED (9 ff) @f86bd200

I have a similar problem, though my combination used to work fine, until
I built a new kernel on Monday.  Now even my netbsd.old kernel reports
the error!  Possibly my filesystem was corrupted or something?
Here is what I get:

sd0(ncr0:0:0): COMMAND FAILED (9 0) @f867c000.
sd0(ncr0:0:0): extraneous data discarded.

This repeats endlessly.  Everything probes fine, and this error doesn't
show up until right after the "dump on ....." message, presumably when
it tries to mount the root partition.

> The problem appears to occur when a FAST SCSI-2 drive is attached to the 
> 53c810 controller. 

Mine is a FAST SCSI-2 drive also, but reports it is a "DEC", even though
I know it is a quantum.. :-)  I also know it has worked fine!

> Has anyone seen something like this? Any workarounds?

Unfortunately, "Yes, and Not that I know of"

-Andrew
-- 
============================ Real 32bit multi-tasking UN*X System
Andrew Gillham             | TCP/IP,NFS,PPP,4.4BSD-lite,multi-user
gillham@andrews.edu        | i386,sparc,alpha,mac68k,amiga,others
LAN/WAN/NW/UN*X specialist |   ---> http://www.NetBSD.org <---