I recently installed a 1542CF on my system, and even
more recently (yesterday) installed a 700MB SCSI drive
on that.
I disklabeled the drive, newfs'd it, and copied my usr
and home partition over.  I was able to bring it up,
twice, and then the third time /usr was gone. :(

fsck'ing the partition brought hundreds of DUPS.
It then hung the system:
<numbers up to 131075> DUP I=799

CANNOT SEEK: BLK -2
CONTINUE? yes


CANNOT READ: BLK -2
CONTINUE? yes


CANNOT SEEK: BLK -2
CONTINUE? yes

Then, in pcvt's lovely white on red, a console message:
aha0: aha_scsi_cmd, more than 17 dma segs
and the system hangs.

The disk is partitioned as (pardon the formatting):
a: 1011128  	0	4.2BSD	1024	8192	16  #0 - 2605
b: 102044 1011128	 swap       		    #2606-2868
c: 1419304 	0    unused 	0	 0          #0-3657
d: 1419304 	0    unused 	0 	 0          #0-3657
e: 306132 1113172 	4.2BSD 	1024	8192	 16 #2869-3657

I have 16 MB of memory, a 486DX4 120, running NetBSD 1.1.
The machine is running win95 on 2 drives, and NetBSD on
two drives. I have an Orchid Kelvin 64.  The only obvious
answer I can see is this is a bounce buffer problem?  Is
there some other problem that this might be related to?

If you have any questions I can answer, please feel free
to ask, I'd like to be able to use this again, and all
help is appreciated.

dmesg output is available, if you'd like to see it,
ftp tremplo.gis.iastate.edu, username: anonymous.gendalia
password: your address
(the .gendalia part of the username is important.)
the file is called dmesg.

Tracy J. Di Marco                 gendalia@iastate.edu
System Administrator                          294-2279
Geographic Information System               218 Durham
Support and Research Facility    Iowa State University