Kubernetes cluster upgrade version docker
Principles: a complete upgrade to normal after the next rise in the next stage.
Work Node
A, pod on migration (guarantee business, but there will be jitter period)
kubectl drain $NODE --ignore-daemonsets --delete-local-data --force
During the transition can not be deleted if prompted pod, you need to manually delete.
Second, the docker upgrade to version 18.09.9
1. stopped kubelet and docker services
systemctl stop kubelet systemctl stop docker
2. Uninstall the old version
sudo yum remove docker docker-common docker-selinux docker-engine
Related software and rely only removes the data directory / var / lib / docker will be retained
3. Installation depends
sudo yum install -y yum-utils device-mapper-persistent-data lvm2
Add yum repository
sudo yum-config-manager --add-repo https://download.docker.com/linux/centos/docker-ce.repo
4. Install new version
sudo yum install -y docker-ce-18.09.9-3.el7 docker-ce-cli-18.09.9-3.el7
5. Start docker Service
systemctl start docker
6. Check the version
docker version
7. Start kubelet
systemctl start kubelet
8. prohibited scheduling removed
omelets uncordon $ NODE
K8s Master
work node joins the cluster upgrade completed properly, it is recommended to do now master server snapshot, just in case problems arise upgrade.
One,
The master run essentially static pod, mater related components, no migration can not migrate.
Second, the docker upgrade to version 18.09.9
Repeat the procedure above.