JobScheduler variables are stored persistently

Current Situation

• JobScheduler stores variables in memory that are configured
  • on start-up of JobScheduler via the <config><params> element in scheduler.xml and
  • dynamically by scripts that use the API object spooler.variables().
• If JobScheduler is terminated or the JobScheduler environment would crash then any variables that have been configured dynamically will be lost.

Desired Behavior

• JobScheduler should persistently store variables that have been dynamically configured. This will happen each time that a variable is being set.
• JobScheduler should restore such variables on start-up.
• This feature should be available for JobScheduler Master and Agent instances.

Maintainer Notes

• This feature is not available for cluster operations.
• As we estimate this feature to be somewhat limited without cluster support we therefore believe it would be a better idea
  • to preserve this feature as designed - representing a lightweight solution for volatile information that is not required after a restart.
  • to design a new feature for persistent JobScheduler cluster variables that starts from a different set of requirements.
• Feel free to add comments about your requirements for improved use of JobScheduler variables.