A Kubernetes Scheduler Extender to provide gang scheduling support for Spark on Kubernetes
175
stars
43
forks
source link
When scheduling dynamic executors, order nodes by usage instead of picking a node at random #230
Closed
Alexis-D closed 1 year ago
The current logic just picks a random node on which the executor will fit:
https://github.com/palantir/k8s-spark-scheduler/blob/36236cf049a5c730afa5e7107eb3ef0fb624cc35/internal/extender/resource.go#L435-L442
This PR simply tries to fit the new executor in the 'fullest' node.