Environmental description:
When kubeadm installs a single master cluster and installs the flannel network plug-in, the errors encountered are as follows:
The image of node1 has been downloaded and is in the Running state, but the master reports an error init:ImagePullBackOff
[root@k8smaster ~]# kubectl get pods -n kube-system NAME READY STATUS RESTARTS AGE kube-flannel-ds-jfkv6 1/1 Running 0 12m kube-flannel-ds-s6wnb 0/1 Init:ImagePullBackOff 0 12m
The reason is: failed to pull the image
Troubleshooting steps:
Method 1: The master directly pulls the image, the version number can be found in the kube-flannel.yml file;
[root@k8smaster ~]# docker pull quay.io/coreos/flannel:v0.13.1-rc2
Method 2: On node1, save the image and copy it to the master
[root@k8snode1 ~]# docker save -o flannel.tar.gz quay.io/coreos/flannel:v0.13.1-rc2 [root@k8snode1 ~]# ls -lh -rw------- 1 root root 63M 2月 7 17:06 flannel.tar.gz [root@k8snode1 ~]# scp ./flannel.tar.gz 172.16.80.20:/root/
Load mirror on master
[root@k8smaster ~]# docker load -i flannel.tar.gz
After successfully loading the image, check that the pods become Running again, and the nodes are also Ready.