Since JobScheduler release 1.6.4043 the new parameter scheduler.cluster.restart_after_emergency_abort is added.
This parameter controls the behaviour of the JobScheduler in a cluster when it doesn't get a database connection.
Look here for general information about cluster.
Normally a JobScheduler in a cluster shuts down after ca. 2 minutes if the database connection get lost.
Then a new JobScheduler process is started which waits for the database and is back in the cluster when it gets the database connection.
If you don't want that a new JobScheduler process is started to wait for the database then you can use the parameter scheduler.cluster.restart_after_emergency_abort in the ./config/scheduler.xml
<spooler> <config mail_xslt_stylesheet = "config/scheduler_mail.xsl" port = "4444"> <params> <param name="scheduler.cluster.restart_after_emergency_abort" value="false"/> </params> ... </config> </spooler>