Remove the part and set it in a device programmer ?
On Sat, Jul 30, 2016 at 11:35 AM, Pontus Pihlgren <pontus at update.uu.se>
wrote:
On Sat, Jul 30, 2016 at 01:58:49PM -0400,
alexmcwhirter at triadic.us wrote:
I know nothing about this machine in particular,
but i know a decent
amount
about other unix machines of the era. Chances are
that the copy of RTU on
that box is licensed to the serial / id number programmed in nvram.
Because
the nvram is dead, those numbers no longer match
and the OS panics from
an
invalid license.
I think you may very well be right. I noticed that the "show" command in
the console displays the serial number. I went back and compared it with
the serial number printed on the back of the machine. Well, it doesn't
match one bit. So.. I either need to figure out to reprogram the NVRAM
(simply set serial_number doesn't work and the manual lists the
environment variable as "permanent") or I suppose I could figure out
where on disk the serial number is.. but it doesn't sound easy.
The TOD clock typically part of the nvram chip
and loses
it's value after every reset. If i had to guess, i would say replace the
battery / nvram chip (if it's a self contained chip like the old sun
boxes)
and see if you can get enough data together to
reprogram it. Whether or
not
the machine in question has a facility to do that
like the old sun's do
i am
not sure.
I've battled the NVRAM death and corresponding TOD problems in SGI, SUN
and DEC machines before but only succeded because the "set"
functionality of the console was enough... this time I'm not so sure.
/P