This PR fixes #1182 and adds a test for the condition (verified that the test fails on current master).
This is a critical issue since currently Fluxion could assign down resources to jobs. The jobs then immediately fail and the next job is scheduled, eating through the entire queue.
This PR fixes #1182 and adds a test for the condition (verified that the test fails on current master).
This is a critical issue since currently Fluxion could assign down resources to jobs. The jobs then immediately fail and the next job is scheduled, eating through the entire queue.