Cluster Autoscaler: Nodes not getting "ready"
(Doc ID 2788162.1)
Last updated on JULY 05, 2021
Applies to:Oracle Cloud Infrastructure - Oracle Container Engine for Kubernetes - Version N/A to N/A
Information in this document applies to any platform.
We have several nodes which are in "Node State" "active" but stay in "unavailable" for "Kubernetes Node Condition". This behavior is switching between the nodes.
Furthermore, the instance behind the "unavailable" "active" node is sometimes in state "Terminated".
Re-creating of the node pool helps for a short time period only.
The cluster autoscaler only scales up and scales down node pools, so theoretically it shouldn’t affect whether the node goes ready or not ready. However, it’s accurate to say that the autoscaler should definitely not be scheduled on nodes that are managed by the autoscaler.
UI can be up to 5 minutes delayed behind what the true state of the cluster is, which is why you might see the “unavailable” state. It's also normal that when nodes first come up, that they start out as NotReady for approximately 1-2 minutes while the kubelet registers.
To view full details, sign in with your My Oracle Support account.
Don't have a My Oracle Support account? Click to get started!
In this Document