You are viewing the RapidMiner Server documentation for version 9.6 - Check here for latest version
Automatic Job Cleanup
Automated Job Cleanup should only be enabled when absolutely necessary! It may heavily delay existing job schedules. The duration needed for successful job cleanup varies depending on your used database system, hardware and initial amount of recently executed jobs which are potential candidates to clean.
Please do not interrupt the cleanup mechanism while it’s still progressing, e.g., by restarting!
RapidMiner Server automatically saves information related to recently executed jobs. This includes the user who triggered the execution, the job's state, the queue on which the job has been executed and also the date on which the process was executed - this information can be reviewed on the Executions page within RapidMiner Server. Additionally, the Job Agent which was responsible for executing the job, created a dedicated working directory for it.
Those stored pieces of information and also the working directories can grow large. In order to avoid this, RapidMiner Server
provides a job cleanup mechanism to wipe old jobs. If you like to configure it,
the following properties need to be added to the execution.properties
file within the <rapidminer-home>/configuration
folder of your RapidMiner Server home directory.
jobservice.scheduled.jobCleanup.cronExpression
: This property defines the point in time when the automatic job cleanup will be executed with the help of a cron expression. By default, the cleanup task is configured to run hourly with the cron expression0 0 * * * *
. It follows the cron pattern<second> <minute> <hour> <day> <month> <weekday>
. So0 */30 * * * *
would run the job cleanup every 30 minute whereas0 0 0 * * *
would run it daily.jobservice.scheduled.jobCleanup.maxAge
: This property defines the maximum age of jobs in minutes. By default, this is not configured. If configured, jobs older than the value set in the property will be cleaned up. Set this to any arbitrary number greater than zero. Please note that the property value needs to reflect the execution time. If you like to clean up all jobs which are older than 10 minutes, you also need to run the job every 10 minutes.