All Products
Search
Document Center

What Can I Do if Nodes Fail to be Removed from an ACK Cluster?

Last Updated: Jun 29, 2021

Problem description

A node fails to be removed from the node pool of a Container Service for Kubernetes (ACK) cluster. The following error message is returned:

Failed to drain node [cn-beijing.10.0.18.xx], err: drain-node job execute timeout, last logs is: node/cn-beijing.10.0.18.xx already cordoned WARNING: ignoring DaemonSet-managed Pods: arms-prom/node-exporter-2mx45, kube-system/flexvolume-72bxr, kube-system/kube-flannel-ds-4j7xb, kube-system/kube-proxy-worker-k648t, kube-system/logtail-ds-nb4hw evicting pod "kube-state-metrics-bccdcfb54-x2kpl"

Cause

Pods on the node to be removed cannot be scheduled to other nodes, and therefore the node fails to be removed.

Solution

  • If pods can be scheduled to other nodes, scale out the node pool to ensure sufficient resources. For more information, see Manage a node pool. Then, remove the node. For more information, see Remove nodes from an ACK cluster.
  • If pods cannot be scheduled to other nodes, clear the Drain the Node check box when you remove the node.

Applicable scope

  • ACK