UPDATE: 10/21: Our software vendor has reproduced the issue and developed a patch. We will apply this patch once we complete internal testing.

UPDATE: 10/7: Some other types of job requests were found to still trigger this preemption issue. We have implemented a workaround to ensure scavenger jobs are preempted properly while we work with our software vendor to address the underlying issue.

RESOLVED: 9/22: The scavenger queue is back online. We identified the job requests that trigger the preemption issue and implemented scheduler logic to account for them.

UPDATE: 9/18: We are working with our software vendor to resolve this issue.

We have temporarily paused running jobs through the scavenger queue while we investigate a potential issue with job preemption. If you have any questions, please contact us.