Here is a list of the actions (tests) and the conditions to set up the tests.  Each test action should be run with each test condition. 
The logfile should record appropriate entries for each test.

Test Actions:
1) Scheduler start  (/etc/init.d/fossology start) 
Result: Normally, scheduler and fo_watchdog should start.

2) Stop (/etc/init.d/fossology stop)
Result: Normally, scheduler and fo_watchdog should terminate and not restart.

3) Restart (/etc/init.d/fossology restart)
Result: Normally, scheduler should die and then start.  fo_watchdog process
should not change and not should not try to start a scheduler during this
process.

4) kill -9 fossology-scheduler
Result: scheduler should die, fo_watchdog remains and restarts the
scheduler in a few minutes.

5) Hang scheduler
(simulate by using instrumented code that sleeps on SIGUSR1)
This will continue to update the scheduler_status table, but not perform
any useful work with the agents.

6) Dead agent process.  Kill an agent while it is active.

Test Conditions:
1) No pid in scheduler lockfile (see /dev/shm for lock files)
2) No pid in fo_watchdog lockfile
3) Non-existant pid in scheduler lockfile
4) Non-existant pid in fo_watchdog lockfile
5) No scheduler lockfile
6) No fo_watchdog lockfile

