Issue
When a Container Service for Kubernetes (ACK) cluster uses CentOS 7.6, the following error message appears in the logs of the kubelet:
Reason:KubeletNotReady Message:PLEG is not healthy:
Cause
If the systemd-219-62.el7_6.6.x86_64 package used by CentOS 7.6 is defective, the error occurs. For more information, see Node flapping between Ready/NotReady with PLEG issues.
Solution
Back up your data by using snapshots. Run the following command on each node of the ACK cluster to update the systemd package to the latest version. Then, rerun systemd.
yum update -y systemd && systemctl daemon-reexec
Applicable scope
- Dedicated Kubernetes clusters in ACK
- Managed Kubernetes clusters in ACK