>> I did it, so can you.  It _ISN'T_ pretty, at least in the state I had it
in, but
>> it *did* work.  =) A diskless machine swaps a LOT in netbsd unless you pare
>> it down a lot as well...   Take all your daemons out, once you get it
>> working, and put them back as needed.  Course, you have a 3/60.  I suggest
>> more ram, if you can get it.  4 mb should be about $35-40, in 100 ns SIMMS.
>
>Wow!  I'm not the only one -- great!  I seem to be having a LOT more 
>problems than anyone else I've talked to about installing an 
>operating system on this machine!  

It's a definite learning experience.  Not easy, but something neat to claim
as a skill......  "Hey, I have a sun that boots X-windoze diskless... Hire
ME!!"  =)

In reference to my "I did it" line, I'm a big bad English major.  =)
Therefore.... =)

>Once things get going I have some "fancy" 1MB simms from a 386 which 
>should hopefully do the job! :-)

Hee.  I bought more for my 386, which was serving NFS and boot files to the
3/50 here, but when I put them in the board seems to have cracked or
something.  Of course, systems don't like being left on for 8 months without
a reboot and then turned off and on again a bunch....  =)


>> Hmm.  Haven't seen this problem.  I _think_ I had my /etc/exports file set to
>> insecure, (pretty sure on that one) with another parameter.. Maybe something
>> to do with Root_quash???
>
>Thanks to a couple of the posts I found out I have to mount it 
>(insecure,no_root_quash) so the root uid gets shared; I'm not quite 
>sure what the insecure is for.  I haven't tested it with the RD 
>kernel again since I found out I don't have to use it, but now that I 
>think of it, something to do with that may be causing my CURRENT 
>problems.  I installed the GEN kernel and added these to bootparams:
>(my little network actually IS 123.123.123, it's not connected to 
>anything):


I think insecure pretty much just tells the box to let another system mount
it without too much hassle....  If you have another machine that's configged
as a valid device, and you use insecure, you should be able to mount it with
not much trouble.  =)


Look for posts from sometime in.....  Early March?  I think that's ABOUT the
time I spammed to port-sun3 for a LOT of useful help.

I use 192.251.251.XXX.  No traffic there either, fortunately.  =)


>	xhost root=123.123.123.1:/usr/export/root/netbsd \
>		swap=123.123.123.1:/usr/export/swap/asdf
>
>(asdf is null-length, but I've also tried an 8mb file filled with 
>nulls).
>
>As soon as it tries to boot, the computer apparently hangs, after 
>saying "WARNING! No swap space found".  After that there is no packet 
>activity at all.

Hmm....  Don't know what to tell you, really. I think that maybe you don't have 
/usr/export/swap/ in /etc/exports??  Just a guess, but I did something like
that as well.  Different error message, but the same problem.  =)  


>> The docs that you see telling you to use some long, funky string of 
>params 
>>  aren't
>> neccessary.  something like XX=1024, XX=2048, or something.  It worked with
>> just  a couple of params.  (under slackware 3.0, on a system that had been
>> working as an Xterminal(the sun) with Xkernel.
>
>> It shouldn't.  From what I recall (My NFS server died, saving up for a new
>> motherboard to replace it and put my 3/50 back into action) the kernel
>> really doesn't eat that much ram until you start loading stuff port-init.
>> Dunno what that message you got from the SunOS kernel was, but it probably
>> was just griping about something in your setup.  =)  Always a joy, that...
>
>I was going to say that it might have been with no_quash or insecure 
>settings, but those didn't fix it.  I don't know what's up with it!  
>I'll send you the error message next time I see it if you can help?


When you get an actual error, and if its' a number, look in the linux source
tree for 
errors.h  That'll tell you what the problem is if the NFS server is griping,
generally.  =)


>> You don't need the miniroot at all.  Untar the distributed 1.1 files in a
>> directory, export them, and go.  From what I recall, you have to change the
>> params on 
>> the init binary to make it go....  I think the way I was doing it remapped
>> all the UIDs from the linux box and made them all owned by lp.adm, or
somesuch.
>> chown root.root * after the system booted in /sbin fixed most of it, at
>> least.  =)
>I should have thought about it a little more clearly!  How come you 
>have to map the UIDs -- if you have no_quash_root turned on and 
>everything is owned by root, shouldn't things work fine?
>


NetBSD and Linux use different numbering schemes for users and groups...
There probably needs to be SOME communication between the two groups on
this, if for no other reason than to get rid of this garbage.  From what I
remember,
the system works (if you log in as root) with only changes to the
permissions of /sbin/init.  


>Thanks for all the help.
>	Jeff

You're welcome.  Glad to walk you through it as best I can.