Feature Overview

1. start a job
2. kill a hop node on the path used to control the execution host (or more generally interrupt communication through the used path)
3. the job isn't failed, it's just re-routed through the Mesh

Background, and strategic fit

Increased reliability of the cluster, less jobs failing because of the platform.

(Optional) Use Cases

n/a

Assumptions

• Re-routing can be done in the middle of a job

Out of Scope

• make a job survive the death of the execution host
• make a job survive the death of the controller (see AAP-5298)