- If the JobScheduler connection to the database fails then JobScheduler tries to reopen the connection after 60 seconds. During this situation the JobScheduler is not accessible via TCP.
- How to reproduce this behavior: If I send the command <terminate/> then
- I get a Connection refused message.
- the current JobScheduler process doesn't terminate
- a new scheduler.log is written, in which a new PID is logged.
- a process with this new pid is not in the process list.
- only a kill removes the original process of the JobScheduler
- JobScheduler works as designed. While waiting for a successful reconnect to the database no commands are allowed that modify the JobScheduler status.
- However, any informational commands can be sent, e.g. all commands that start with <show_.../>
- Should a JobScheduler be terminated while it is waiting for a database connection then the JobScheduler start script can be used with the "kill" command.