Distributed Monitoring Take 2
-----------------------------

  This system is very similar to the previous one, only it doesn't
 rely upon an installation of Apache for viewing the results of tests
 on other nodes.

  Instead we offer a simple network server running upon port 2929.

  (Please firewall this port from untrusted networks.)


About
-----

  This is a simple distributed system which is designed to perform service
 availability tests via remote testing.  The testing is carried out upon
 multiple hosts, but alerts will only be generated if >= 2 nodes see a
 service failure.

  The intention is that you'll run the deamon on a number of hosts which
 will be used to instigate network testing - remotely.  The hosts which
 have the deamon installed are known as "nodes".

  The nodes will be queried by a small client which actually schedules
 the execution of tests, and collates the results of those tests.

  In summary the software is:

    dmon-server
    -----------
    This runs non-stop on the testing hosts, and is responsible for
    performing the network tests.  It is controlled remotely.

    dmon-client
    -----------
    This runs every minute, or two, upon each node and instructs the
    server to perform the tests.  The results of the tests will be collected
    by this client and used to alert, if necessary.


Installation
------------

  1. Run "make install" upon each node, which will create a stub configuration
     in /etc/dmonitor.

  2. Update the list of nodes with the public names/IPs of the monitored nodes
     in /etc/dmonitor/nodes.

  3. Update the test hosts in /etc/dmonitor/hosts.d/

  4. Start dmon-server in a screen session.

  5. Run "dmon-client" in a cron, setting it to run every 1/2/3 minutes.



Steve
--