| From: der Mouse <mouse@Rodents.Montreal.QC.CA>

| What I would suggest is having some machine - a third machine, if
| possible - sniff the wire with etherfind or tcpdump or moral equivalent
| to see what's _really_ happening, to see whether it's the client or the
| server that's at fault.

Thanks.

After much poking around, I found that the Solaris rpcbind doesn't know
there is a rpc.bootparamd unless rpcbind is started after
rpc.bootparamd.

The tools I used were snoop, truss, and, of course, kill!


Now I've moved on to the next problem.

I wish to run diskless.  The INSTALL.sun3 file doesn't really explain
how to do this.  It assumes that you will be installing onto a local
disk.  I just unpacked everything into the root-for-redvax directory.
I did a little bit of fiddling in /etc -- fstab and hosts, for now.
I should look at /etc/passwd.  Are there any others that I should
configure for my rather simple use?

I tried to boot the netbsd-rd and netbsd-gen kernels.  Both got hung
up contemplating SCSI.  Since I wish to run diskless, I've left all
the SCSI devices powered off (two disks and a tape drive).  Here are
the messages I got:
repeated perhaps 20 times, slowly:
	si0: select found SCSI bus busy, resetting...
then, alternating quickly, until I get bored (not sure which message
came first):
	si_intr: spurious from SBC
	isr_autovec: ipl 2 no claimed

Are these messages just because the devices are turned off?  Where can
I find a kernel that will plunk into this setup and not bother with
the devices?  Is this a job for the xkernel that Curt is mirroring?
If so, would it work with the files I've set up (from the current sun3
binaries)?

Thanks,
Hugh Redelmeier
hugh@mimosa.com  voice: +1 416 482-8253