DRF Scheduling

It controls whether to enable the Dominant Resource Fairness (DRF) scheduling algorithm, which schedules jobs based on their dominant resource share. Jobs with a smaller resource share will be scheduled with a higher priority.

The idea of DRF is that in a multi-resource environment, resource allocation should be determined by the dominant share of an entity (user or queue). volcano-scheduler observes the dominant resource requested by each job and uses it as a measure of cluster resource usage. Based on this dominant resource, volcano-scheduler calculates the share of the job. The job with a lower share has a higher scheduling priority. In this way, more jobs can be processed, and a large number of services with a small resource share can still be scheduled even though there is a service with a large resource share. DRF ensures that fairness is met when multiple resources coexist.

Value Range

true or false

Default Value

true

Modifiable

Yes

Scope

CCE standard and CCE Turbo clusters

Suggestions

Configure DRF scheduling to batch process small scale services like single AI model training and single big data computing and query, because DRF preferentially considers the throughput of services in clusters.