Batch pipeline, mission 2 of 6
When a job fails
intermediate, about 10 minutes, in the Kubernetes Simulator
Start mission
Opens once you pass Run it once.
The briefing
The export Job copies yesterday's orders into the data warehouse. It has been failing: a Job retries a failed pod, backing off each time, until backoffLimit is spent — then it is Failed for good.
The failed pods are still there, and so are their logs. The source database's address lives in the export-config ConfigMap. And one thing to know: a Job's pod template can't be edited once the Job exists.
What you'll do
- Find out why the export fails
- Run an export Job that completes
- Bonus star: Discover the template is immutable: try applying the fix before deleting the Job
The mission checks the cluster, not your keystrokes: any fix that works, passes.
More Batch pipeline missions
- 1. Run it once
- 2. When a job fails
- 3. Many hands
- 4. Every night
- 5. Pile-up
- 6. The 2am failure