See the failoverd.pod file for technical information

06/08/99
CPU utilization was obsessed with, and lowered substantially.  This seems 
stable.  It has also been tested on a switched network, with good results.

I'm starting to look into real-time link state.  This seems to be dependant 
on the driver, so we'll see where we get.

###############################
06/02/1999
We can now detect a slow interface, as well as a dead one.  $verbose is an
option to turn off debugging logs, $low_water exists, and is explained in the
POD file, failoverd is much more reliable in general.

###############################
05/21/1999
The port and remote host are now variables inside failoverd.conf, so the notes
from 4/13/1999 are bogus regarding Failover.pm line 169.

The authentication steps now work properly.  Failoverd logs connection 
attempts, and behaves properly regarding this.

################################
04/13/1999
You need to modify Failover.pm at line 169 to reflect the host's status.

Master server:
  Failover::logger("Notifying $secondary to take over");
  my $sock = IO::Socket::INET->new(PeerAddr     =>      '1.0.0.2',

Slave server:
  Failover::logger("Notifying $primary to take over");
  my $sock = IO::Socket::INET->new(PeerAddr     =>      '1.0.0.1',


Note that we now use failoverd.conf, which needs to be in the same directory 
as the program itself.  /etc/failover/ should contain the following files:

host_type -->

# boot: defines what to do at boot
# run: is not used at present, but you never know.
boot:master
run:master

secret -->

<your secret word here>

Secret is used to validate a failure request over the network.  See the .pod 
file for more information.

-brad
