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

Re: Serial Connection -- handshaking

* MAROUNI Abbass <abbass.marouni@internetmemory.org> [110329 14:21]:
> I think that the discussion diverted somehow from my original question.
> My problem was  the following :
> Two identical servers Z0 & Z1 and a null modem cable. ttyS0 on Z0
> connected to ttyS1 on Z1.
> on ttyS0 on Z0 I have a getty listening for incoming connections.
> on ttyS1 on Z1 I launch a minicom toward ttyS0 on Z0 and I login normally.
> Now When I try to do the inverse :
> Kill getty on ttyS0 on Z0 and remove it from inittab. Set getty to
> listen to ttyS1 on Z1. Launch minicom on ttyS0 on Z0 toward ttyS1 on
> Z1.
> I can't login all I see is garbage. All the serial parameters are
> the same on both sides.
> I thought that logically this setup have to work since I am using
> the same setup just changing which port on which server.
> I am trying to do this for a Datacenter with hundreds of servers.
> This way I can connect two servers together with just one serial
> cable.
> Does it have something to do with the ttyS0 port? is it a software problem?
> Do you have any ideas ? (Apart from theoretical discussions about
> all the standards)

A "null modem" cable may be simple (three conductors) or complex(seven
or more conductors), depending upon whether "handshaking" is
accomplished in hardware or software.

With hardware handshaking, each end of the connection typically has a
"ready" signal and a "clear to send" signal.  If you are sending in
both directions, these signals must be provided in both directions.
The signals are set by the standard "RS-232", which was up to revision
"C" the last time I used it (that is, "RS-232C").

Of course, the software may not inspect all of the signals which the
hardware provides, and not all hardware provides a complete set of
signals.  Thus, experimentation may be required.


Reply to: