Skip to end of metadata
Go to start of metadata

Restart behaviour

If you restart the JobScheduler via JOC (Menu->'Terminate and restart') or with the startscript

 ./bin/jobscheduler.sh restart

then first the JobScheduler tries to stop all jobs.
The JobScheduler waits for all jobs with running tasks.

Only when all tasks are completed, then the JobScheduler stops and starts.
Until then ...

  • the JobScheduler has the status "stopping"
  • all jobs with running tasks get the status "stopping"
  • all jobs without running tasks get the status "stopped"

You can restart the JobScheduler with a timeout (in seconds) via JOC (Menu->'Terminate and restart within ~60s') or with the startscript

  ./bin/jobscheduler.sh command "<terminate restart='yes' timeout='60'/>"

In this case the JobScheduler waits 60 seconds for running tasks.
All tasks which are still running after 60 seconds are aborted with an exit code 99.
So each job of these tasks will be stopped, except the job is configured with stop_on_error="no".

Note that the timeout for terminate in JOC is configurable in the ./config/operations_gui/custom.js (see also here).

See also http://www.sos-berlin.com/doc/en/scheduler.doc/termination.xml

JobScheduler remains in the status stopping

If you have an endless task then the JobScheduler remains endless in the status "stopping" after a normal restart without a timeout.
In this case you must abort and restart the JobScheduler via JOC (Menu->'Abort immediately and restart') or via the startscript

 ./bin/jobscheduler.sh command "<modify_spooler cmd='abort_immediately_and_restart'/>"

Please avoid endless tasks!

Write a comment…