Outsourcer has built-in job queueing and dynamic job throttling. By default, there are four concurrent jobs that can be executed at once. Jobs are left in a “queued” state until an available slot in the queue is available.
How to use the Queue
Job details (the source and target mapping or transformation SQL) are stored in your os.job table. When you want to execute one of these jobs, you simply insert the job details into the os.queue table.
If there is an error found while processing, the Queue will capture the error raised from your source SQL Server or Oracle system as well as from Greenplum.
The maximum number of jobs is dynamic and configurable by updating a table inside Greenplum. This means it is easy to optimize the workload for your environment by configuring the maximum number of jobs that can execute at once. Look at the os.variables table for the ‘max_jobs’ name. Update the value to dynamically change the maximum number of concurrent jobs.