[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Netatalk inconsistencies wired vs. wireless.



On Jun 9, 8:25 am, A2Aviator <a2avia...@gmail.com> wrote:
> The test 'server' is a Dell XPS 3.4GHz 2004 era brick laptop heater
> device.
> It's got inbuilt G wireless, and wired gigabit.
>
> The real server is an Intel server board dual Xeon 2.8 GHz that has
> two Gigabit NICs inbuilt.
> One is Marvel, the other is Intel.
>
> The first wireless tests were on the 'test' server. The most recent
> attempt was trying to find a PCI NIC that would work in that dual
> Xeon, with Ubuntu.
>
> That part was not successful.
>
> So I figured for laughs and craps I'd try the wireless adapter on the
> IIgs side just to see if a wired to wireless transition ANYWHERE is
> the same as having the server itself run via wireless, since I had
> success on the test server with wireless, I figured I'd do it on the
> real one, and was not able.
>
> So putting the reverse wireless adapter at the IIgs end, (one of those
> things you put on an Xbox, PS/2, etc)  made the IIgs suddenly see all
> devices on the network, and when the test server saw the IIgs when it
> was switched over to the wire, I figured I'd see what the real server
> was reporting and it too was reporting seeing the IIgs. For the first
> time ever.
>
> So I figured thats gotta be the test server aiding in the situation.
> So I removed it from the network, killed the power to the reverse
> wireless adapter and rebooted the IIgs. It came right back up on the
> real server, and the IIgs saw the server. Except, the login was
> pathetically slow, as well as all access.
>
> Which is NOT the case when the "wireless" is at the server end, via
> the test server.
>
> The next thing I should try is putting the reverse wireless adapter on
> the real server and see what happens.
>
> I also need to dump some syslog dumps for comparison.
>
> The synopsis so far is - the IIgs does not see the server unless there
> is a wireless link in the path.

This may a dumb question, but is any part of the AppleTalk stack
trying to communicate over IPV4?  From the looks of your IFCONFIG, the
wired adapter only has an IPV6 address.  Could you assign it an IPV4
address and see if it gets any farther along?  If your wireless
adapter (the external one) was smart enough  to convert IPV4 to IPV6
it could explain why you could then communicate using that interface
on the server.  Just a thought,

Dave...