K8s Operator: DaskCluster spec updates do not propagate to pods

Deploying Dask
,
1

Hi everyone,

We’ve deployed the disk-operator and custom resources (DaskCluster and DaskAutoscaler) via helm. When updating DaskCluster specs (e.g. scheduler container resources requests - cpu and memory) via helm upgrade, DaskCluster custom resource does reflect the correct configuration, but the pods are not redeployed.

Is this expected?

Thank you!

2

UPD: it appears to be expected for now. The cluster needs to be deleted and recreated.
See k8s Operator: DaskCluster spec updates do not propagate to pods · Issue #636 · dask/dask-kubernetes · GitHub