ErrImagePull for System Kubernetes Images
This section describes how to resolve the ErrImagePull status in system Kuberenetes images.
Problem
The Kubernetes kubelet cmdlet attempts to purge unused disk images and pods from the system, when the image file system (or file system) is under pressure (disk space reaches lesser than 15%). This results in the system Kubernetes images being purged from the system and the pods end up in an "ErrImagePull" status.
A sample "ErrImagePull" status appearing for the kube-system pods are shown below:
kube-system calico-node-lnz6g 0/1 Init:ErrImagePull 0 41m 40.40.40.154 cn-svi-tb10-service2
kube-system kube-proxy-zkk46 0/1 ErrImagePull 0 51m 40.40.40.154 cn-svi-tb10-service2
smi-vips keepalived-km4wx 0/2 ErrImagePull 0 40m 40.40.40.154 cn-svi-tb10-service2
Resolution
You can restore the local images to resolve the "ErrImagePull" status in system Kubernetes images.
To restore the local images, execute the following command.
sudo docker load < /var/tmp/k8s-offline.tgz