SEO

[ passing on a request ]

We have a batch of Sun Javastation's, basically a series of thin clients
before they released the sunrays. Even more interesting is that these
machines are Intel based - something Sun apparently don't easily admit
too (from the Javastation FAQ, these machines 'dont exist').

Ok, to complicate matters - these machines have OpenBoot 3 installed on
them, which is a bootloading / interpreter environment that is
frequently used on their sparc boxes. To make matters even worse - it's
a restricted environment; normal openboot systems have a shell much like
grub with great functionality, these don't. You can enable it, but it
takes soldering a resistor onto the motherboard.

So, it does have a 'normal' network boot process - running ethereal you
can see that it searches for a dhcp server, grabs an IP, then tries to
grab a file from a tftp server.

That's where we hit a glitch - *what* file should it be fed to work
properly. We've tried everything we can find, from linux kernels
(compressed and not), to etherboot images, pxe files, second stage
bootloaders, etc. Nothing works.

What we have learned though is that anything compressed gives an
'unrecognised program format' error. Uncompressed elf (as in
uncompressed kernel) seems to work best, although the response is either
the machine locking up, or rebooting (but at least no error messages

Any ideas from the boot experts out there?

On Tuesday 06 July 2004 5:32 pm in comp.sys.sun.hardware Andy Rabagliati
wrote:

> We have a batch of Sun Javastation's, basically a series of thin clients
> before they released the sunrays. Even more interesting is that these
> machines are Intel based - something Sun apparently don't easily admit
> too (from the Javastation FAQ, these machines 'dont exist').

Are you absolutely sure that these are Intel based machines ?
If so they are _very_ rare, the existance of such a machine was rumoured,
but there are no reliable reports of them being seen "in the wild".


--
My real address is crn (at) netunix (dot) com
WARNING all messages containing attachments or html will be silently
deleted. Send only plain text.

According to Chris Newport <me@see-my-sig.invalid>:
> On Tuesday 06 July 2004 5:32 pm in comp.sys.sun.hardware Andy Rabagliati
> wrote:
>
> > We have a batch of Sun Javastation's, basically a series of thin clients
> > before they released the sunrays. Even more interesting is that these
> > machines are Intel based - something Sun apparently don't easily admit
> > too (from the Javastation FAQ, these machines 'dont exist').
>
> Are you absolutely sure that these are Intel based machines ?
> If so they are _very_ rare, the existance of such a machine was rumoured,
> but there are no reliable reports of them being seen "in the wild".

Yes. Shipped to South Africa.

About /One Thousand/ of them.

Destined for the junk pile - unless we can make them Thin Clients for
schools here. They look quite sweet.

http://www.slug.org.za/
http://wizzy.org.za/

Cheers, Andy!

Andy Rabagliati <andyr@wizzy.com> wrote:

> Destined for the junk pile - unless we can make them Thin Clients for
> schools here. They look quite sweet.

Do you know which model there are?

There's a list of them in JavaStation-HOWTO (which should help you with
your questions too!) at
http://www.linux.se/doc/HOWTO/JavaSt...NMODELSSECTION
(line may have been wrapped)

Pictures would be nice too And in case you ever feel like selling one
(or few) do let me know. I have a JavaStation "Krups" myself and it
would like some company.

-Sami

Andy Rabagliati wrote:

> According to Chris Newport <me@see-my-sig.invalid>:
>
>>On Tuesday 06 July 2004 5:32 pm in comp.sys.sun.hardware Andy Rabagliati
>>wrote:
>>
>>
>>>We have a batch of Sun Javastation's, basically a series of thin clients
>>>before they released the sunrays. Even more interesting is that these
>>>machines are Intel based - something Sun apparently don't easily admit
>>>too (from the Javastation FAQ, these machines 'dont exist').
>>
>>Are you absolutely sure that these are Intel based machines ?
>>If so they are _very_ rare, the existance of such a machine was rumoured,
>>but there are no reliable reports of them being seen "in the wild".
>
>
> Yes. Shipped to South Africa.
>
> About /One Thousand/ of them.
>
> Destined for the junk pile - unless we can make them Thin Clients for
> schools here. They look quite sweet.
>
> http://www.slug.org.za/
> http://wizzy.org.za/
>
> Cheers, Andy!

I'd like to have one (or a few) as a collector's item.

--
The e-mail address in our reply-to line is reversed in an attempt to
minimize spam. Our true address is of the form che...@prodigy.net.

On Tuesday 06 July 2004 6:23 pm in comp.sys.sun.hardware Andy Rabagliati
wrote:

> According to Chris Newport <me@see-my-sig.invalid>:
>> On Tuesday 06 July 2004 5:32 pm in comp.sys.sun.hardware Andy Rabagliati
>> wrote:
>>
>> > We have a batch of Sun Javastation's, basically a series of thin
>> > clients before they released the sunrays. Even more interesting is that
>> > these machines are Intel based - something Sun apparently don't easily
>> > admit too (from the Javastation FAQ, these machines 'dont exist').
>>
>> Are you absolutely sure that these are Intel based machines ?
>> If so they are _very_ rare, the existance of such a machine was rumoured,
>> but there are no reliable reports of them being seen "in the wild".
>
> Yes. Shipped to South Africa.
>
> About /One Thousand/ of them.
>
> Destined for the junk pile - unless we can make them Thin Clients for
> schools here. They look quite sweet.

Hmmmm - I had heard rumours of these boxes being in South Africa, but apart
from that I have few clues, I was back in the UK before they would have
arrived.

I think your best bet may be to track down where they came from, possibly a
government department or a university. Whenever something wierd like this
comes up the word WITS is often a good guess. You might also try to track
down some of the older helpdesk or field engineer chaps at Sun SA, someone,
somewhere must have some information.

--
My real address is crn (at) netunix (dot) com
WARNING all messages containing attachments or html will be silently
deleted. Send only plain text.

Andy Rabagliati <andyr@wizzy.com> writes:

> [ passing on a request ]
>
> We have a batch of Sun Javastation's, basically a series of thin clients
> before they released the sunrays. Even more interesting is that these
> machines are Intel based - something Sun apparently don't easily admit
> too (from the Javastation FAQ, these machines 'dont exist').

It seems to me that what you have is a bunch of 'Dover' Javastations.
After the Mr. Coffee (disk box) and Krups (the funky tower shaped one)
there was an experimental Sparc-based unit called Expresso which was
swiftly axed in favour of the Intel-based Dover Javastation. My
involvement with the Beta program ended around this point so I never
saw a Dover javastation.

> That's where we hit a glitch - *what* file should it be fed to work
> properly. We've tried everything we can find, from linux kernels
> (compressed and not), to etherboot images, pxe files, second stage
> bootloaders, etc. Nothing works.

On the sparc versions of OBP3, you need a replacement for the PROM.
The most popular choice is Pete Zaitcev's PROLL.
Sparc versions of this are available from Pete Zaitcev's page at
http://people.redhat.com/zaitcev/linux/index.html

The source code is there too, but I don't know whether this will
solve your problem since I suspect the intel-based javastations
may not work the same way.

Pete

--
__________________________________________________ __________________
Pete Young pete@antipope.dot.org Remove .dot to reply
"Just another crouton, floating on the bouillabaisse of life"

According to Pete Young <pete@antipope.dot.org>:
> Andy Rabagliati <andyr@wizzy.com> writes:
>
> > [ passing on a request ]
> >
> > We have a batch of Sun Javastation's, basically a series of thin clients
> > before they released the sunrays.
>
> On the sparc versions of OBP3, you need a replacement for the PROM.
> The most popular choice is Pete Zaitcev's PROLL.
> Sparc versions of this are available from Pete Zaitcev's page at
> http://people.redhat.com/zaitcev/linux/index.html
>

On Wed, 07 Jul 2004, Thomas Black thomas at shuttleworthfoundation.org wrote :-

> Hi Andy
>
> We tried that - it doesn't seem to work. Proll seems to be a
> replacement rom, literally taking over control from the old rom.
> Unfortunately this means that it is machin specific, and in this case
> sparc specific. What we're trying to find is something that plays
> nicely with openboot, but loads a linux kernel.

[ Posting this as well ]

Do you have a keyboard yet - so you can do L1-A and get into the
OpenBoot prompt ?

Can you extract every little thing about the version of openboot
so folks on the newsgroup can help ?

It is most probably Forth - so try things like "version" "info" "help"

Try plugging a 9600 Baud terminal - something like minicom off
a PC with a serial crossover cable would do - into the serial port.

You might find a Openboot prompt showing up there briefly on boot.

Cheers, Andy!

Hi Andy

>Do you have a keyboard yet - so you can do L1-A and get into the
>OpenBoot prompt ?

No, they have convetional PC type keyboards, so some of the key combo's
that work with Sun keyboards don't work with these.

And...

The Forth interpreter / openboot environment has been disabled.
Apparently if you solder a 220 ohm resister on R362 on the motherboard,
you can enable it again. Try doing that to 1000 machines ;(

>Can you extract every little thing about the version of openboot
>so folks on the newsgroup can help ?
>
>
Well, it's OpenBoot 3.0, Built April 22, 1999, 07:06:21

>It is most probably Forth - so try things like "version" "info" "help"
>
>
Well, I definately think it's forth, although a non-ineractive version,
so no change of typing anything.

>Try plugging a 9600 Baud terminal - something like minicom off
>a PC with a serial crossover cable would do - into the serial port.
>
>
Yeah, we still want to try that. Now to find a laptop with a serial port

Interestingly, I got a *little* bit further with my probing - if the
tftpboot program feeds it the openblk program, which is part of openbsd
and designed for sparc64 boxes, you actually get a different error
message - "cannot find client-services". Not much further, but at least
it's different.

-Thomas

--
Thomas Black
Open Source Program Manager
The Shuttleworth Foundation

On Wednesday 07 July 2004 10:05 am in comp.sys.sun.hardware Pete Young
wrote:

> It seems to me that what you have is a bunch of 'Dover' Javastations.
> After the Mr. Coffee (disk box) and Krups (the funky tower shaped one)
> there was an experimental Sparc-based unit called Expresso which was
> swiftly axed in favour of the Intel-based Dover Javastation. My
> involvement with the Beta program ended around this point so I never
> saw a Dover javastation.
>
>> That's where we hit a glitch - what file should it be fed to work
>> properly. We've tried everything we can find, from linux kernels
>> (compressed and not), to etherboot images, pxe files, second stage
>> bootloaders, etc. Nothing works.
>
> On the sparc versions of OBP3, you need a replacement for the PROM.
> The most popular choice is Pete Zaitcev's PROLL.
> Sparc versions of this are available from Pete Zaitcev's page at
> http://people.redhat.com/zaitcev/linux/index.html
>
> The source code is there too, but I don't know whether this will
> solve your problem since I suspect the intel-based javastations
> may not work the same way.

Sounds about right to me.
PROLL is an initial bootstrap loader which allows parameters to be passed to
the Linux kernel.

If you build an Intel version of PROLL and follow the instructions in the
FAQ it should be possible to get them to run Linux.

It is certainly worth some effort to put 1000 workstations into needy
schools.

Batteries not included, Some assembly required, Consult your local kernel
hacker, YMMV, .........


--
My real address is crn (at) netunix (dot) com
WARNING all messages containing attachments or html will be silently
deleted. Send only plain text.