This directory should contain executable programs ending in ".script"
to handle CTDB event notifications.  The first and only argument
passed to each program is the event, which is one of:

  init, setup, startup, unhealthy, healthy

An example script that sends SNMP traps for unhealthy/healthy might
look like this:

  #!/bin/sh

  case "$1" in
      unhealthy)
          # Send an SNMP trap saying that the node is unhealthy:
          snmptrap -m ALL -v 1 -c public 10.1.1.105 ctdb \
              $(hostname) 0 0 $(date +"%s") ctdb.nodeHealth.0 i 1
          ;;
      healthy)
          # Send an SNMP trap saying that the node is healthy again:
          snmptrap -m ALL -v 1 -c public 10.1.1.105 ctdb \
	      $(hostname) 0 0 $(date +"%s") ctdb.nodeHealth.0 i 0
	  ;;
  esac

Alternatively, email could be sent:

  #!/bin/sh

  case "$1" in
      unhealthy)
          mail -s "$(hostname) is UNHEALTHY" foo@example.com </dev/null >/dev/null 2>&1
          ;;
      healthy)
          mail -s "$(hostname) is HEALTHY" foo@example.com </dev/null >/dev/null 2>&1
          ;;
  esac