Anyone with experience in Cross-Compiling, especially to an embedded powerPC?

[Mike Bushroe]

>
> *Author: *Michael Havens <bmike1 at gmail.com>
> *Date: *2014-05-14 01:03 -700
> *To: *Main PLUG discussion list <plug-discuss at lists.phxlinux.org>
> *Subject: *Re: Anyone with experience in Cross-Compiling, especially to
> an embedded powerPC?
> wait a second..... forgive my ignorance but I thought space craft had
> nultiple computers to use as fail-safes. couldn't you shut down one, feed
> the new system into one of those, test it, and then use that to feed a
> second and then a third until all of them are updated?
>
> :-)~MIKE~(-:
>
>
Mike that is an excellent point! On the Space Station, the Command and
Guidance (GnC) MDMs are indeed made up of three identical MDMs with
identically the same programing, and all wired into the rest of the MDM
network the same way. That way there are always two spare units already to
go with just a few commands to switch one to standby or off and the other
to full operation. That is because the CnGs are part of a 'crit 1' system,
or critical level one. That means that they have an immediate effect of
life or dearth of both the astronauts and the Space Station itself.

   But the JSL is not a crit 1 system. I think I might have heard crit 3
for it, but maybe not. So that means the answer to your question is a bit
more complicated, namely "yes, partially, and no". Yes, there are fully
redundant systems on the Station, but this is not one of them, at least not
at this time. Partially because there are two LAN Edge Router boxes running
on the network at all times, with two Edge Router cards in each box. But
no, none of the are programmed or cabled to be able to replace another. And
one particular card is special in that it alone handles all the routing
functions, the other 3 cards are only ethernet switches and provide VLAN
isolation. So if any one card fails, it can not even switch to the other
card in the same chassis, let alone be backed up by the other box. So if
any card fails, the part of the network it controls will fail. If the
single routing card fails, then the whole network becomes isolated from the
ground and from each other.

  But even if we could open the boxes and pull out one card at a time, it
would still have the exact same problems I have been mentioning before. You
can't refresh files on a running system that are currently in use by the
running system. That holds true whether they are in the box and connected
to the network, or pulled out and plugged into a different box to provide
power, serial port and ethernet port access, so it does not actually make
it any easier in this case.

Mike

-- 
"Creativity is intelligence having fun." — Albert Einstein
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.phxlinux.org/pipermail/plug-discuss/attachments/20140514/daaa74d6/attachment.html>