Open chia7712 opened 3 years ago
related to https://github.com/palantir/k8s-spark-scheduler/issues/152#issuecomment-877627500
As I did not include the overhead when setting memory annotations, the scheduler tries to assign pods to nodes which have no enough (memory) resources. Also, I can observe following message from logs.
not enough capacity to reschedule the executor
related to https://github.com/palantir/k8s-spark-scheduler/issues/152#issuecomment-877627500
As I did not include the overhead when setting memory annotations, the scheduler tries to assign pods to nodes which have no enough (memory) resources. Also, I can observe following message from logs.