Background jobs

Cancel a running job

Some background jobs can be stopped part-way. Cancelling is graceful: the job stops at the next safe point and keeps what it already finished. Not every job can be cancelled; those that can have a cancel or stop control on the screen where you follow them.

Note: The Background Jobs tab in Organization Settings is for watching only. It has no cancel button.

Jobs you can cancel

Job Where Control
Schema import Organization Settings › Schema Imports, or the import’s detail page The cancel icon on the row, or Cancel on the detail page. Confirm with Yes, Cancel Import.
Importing objects from a database connection The import dialog on the schema Cancel import
SQL object analysis The object’s script tab The cancel icon beside the running analysis
Star schema build The schema’s Star Schema tab Stop

Jobs not in this table run to the end once started. That includes description generation, Pipeline procedure generation, ontology imports, Ontology Generator runs, and star schema design runs. For an Ontology Generator run, you can still throw the result away: nothing is saved until you accept changes in the review.

Cancel a schema import

  1. Open Settings › Organization Settings › Schema Imports.
  2. Click the cancel icon on the running import’s row (only shown while its status is Processing).
  3. Click Yes, Cancel Import. Click No, Keep Running to change your mind.

The import’s status changes to Cancelled.

What happens to finished work

  • Objects the job already created stay. Cancelling doesn’t undo them.
  • Items that hadn’t started yet are left for later.
  • A cancelled schema import can be resumed and picks up where it stopped.

Resume a cancelled import

On the Schema Imports tab, a Cancelled or CompletedWithErrors import shows a play icon. Click it and confirm Resume to continue. The detail page has Resume Import for a cancelled import and Resume All Failed to retry failed objects. See Fix failed import items.

When to cancel, and when not to

Cancel when you started the wrong thing: the wrong file, the wrong schema, or a run that’s about to overwrite something you want to keep.

Don’t cancel only because progress looks slow. Large files spend a long time in early phases before the bar moves, and a job that’s nearly done finishes faster than cancel-and-resume.