I have run clients on NetBSD/sparc for a long time, but never tried
the server.

I found at one point that _codaauth2 SRV records were being
interpreted byteswapped, so I put in a second SRV record at 29185.  I
think this is fixed, but I'm not certain.  I did not have to do this
with the _codasrv record.

I'd stop venus, run tcpdump -s1500 -wFILE, start venus, try things,
and then look at the traces.

-- 
        Greg Troxel <gdt_at_ir.bbn.com>