Upgrading/downgrading kubeadm clusters between v1.8 to v1.9
This guide is for upgrading kubeadm clusters from version 1.8.x to 1.9.x, as well as 1.8.x to 1.8.y and 1.9.x to 1.9.y where y > x.
See also upgrading kubeadm clusters from 1.7 to 1.8 if you’re on a 1.7 cluster currently.
- Before you begin
- Upgrading your control plane
- Upgrading your master and node packages
- Recovering from a failure state
Before you begin
Before proceeding:
- You need to have a functional
kubeadmKubernetes cluster running version 1.8.0 or higher in order to use the process described here. Swap also needs to be disabled. - Make sure you read the release notes carefully.
kubeadm upgradenow allows you to upgrade etcd.kubeadm upgradewill also upgrade of etcd to 3.1.10 as part of upgrading from v1.8 to v1.9 by default. This is due to the fact that etcd 3.1.10 is the officially validated etcd version for Kubernetes v1.9. The upgrade is handled automatically by kubeadm for you.- Note that
kubeadm upgradewill not touch any of your workloads, only Kubernetes-internal components. As a best-practice you should back up what’s important to you. For example, any app-level state, such as a database an app might depend on (like MySQL or MongoDB) must be backed up beforehand.
Also, note that only one minor version upgrade is supported. For example, you can only upgrade from 1.8 to 1.9, not from 1.7 to 1.9.
Upgrading your control plane
Execute these commands on your master node:
- Install the most recent version of
kubeadmusingcurllike so:
$ export VERSION=$(curl -sSL https://dl.k8s.io/release/stable.txt) # or manually specify a released Kubernetes version
$ export ARCH=amd64 # or: arm, arm64, ppc64le, s390x
$ curl -sSL https://dl.k8s.io/release/${VERSION}/bin/linux/${ARCH}/kubeadm > /usr/bin/kubeadm
$ chmod a+rx /usr/bin/kubeadm
Caution: Upgrading the kubeadm package on your system prior to upgrading the control plane causes a failed upgrade.
Even though kubeadm ships in the Kubernetes repositories, it’s important to install kubeadm manually. The kubeadm
team is working on fixing this limitation.
Verify that this download of kubeadm works and has the expected version:
$ kubeadm version
- On the master node, run the following:
$ kubeadm upgrade plan
[preflight] Running pre-flight checks
[upgrade] Making sure the cluster is healthy:
[upgrade/health] Checking API Server health: Healthy
[upgrade/health] Checking Node health: All Nodes are healthy
[upgrade/health] Checking Static Pod manifests exists on disk: All manifests exist on disk
[upgrade/config] Making sure the configuration is correct:
[upgrade/config] Reading configuration from the cluster...
[upgrade/config] FYI: You can look at this config file with 'kubectl -n kube-system get cm kubeadm-config -o yaml'
[upgrade] Fetching available versions to upgrade to:
[upgrade/versions] Cluster version: v1.8.1
[upgrade/versions] kubeadm version: v1.9.0
[upgrade/versions] Latest stable version: v1.9.0
[upgrade/versions] Latest version in the v1.8 series: v1.8.6
Components that must be upgraded manually after you've upgraded the control plane with 'kubeadm upgrade apply':
COMPONENT CURRENT AVAILABLE
Kubelet 1 x v1.8.1 v1.8.6
Upgrade to the latest version in the v1.8 series:
COMPONENT CURRENT AVAILABLE
API Server v1.8.1 v1.8.6
Controller Manager v1.8.1 v1.8.6
Scheduler v1.8.1 v1.8.6
Kube Proxy v1.8.1 v1.8.6
Kube DNS 1.14.4 1.14.5
You can now apply the upgrade by executing the following command:
kubeadm upgrade apply v1.8.6
_____________________________________________________________________
Components that must be upgraded manually after you've upgraded the control plane with 'kubeadm upgrade apply':
COMPONENT CURRENT AVAILABLE
Kubelet 1 x v1.8.1 v1.9.0
Upgrade to the latest experimental version:
COMPONENT CURRENT AVAILABLE
API Server v1.8.1 v1.9.0
Controller Manager v1.8.1 v1.9.0
Scheduler v1.8.1 v1.9.0
Kube Proxy v1.8.1 v1.9.0
Kube DNS 1.14.5 1.14.7
You can now apply the upgrade by executing the following command:
kubeadm upgrade apply v1.9.0
Note: Before you do can perform this upgrade, you have to update kubeadm to v1.9.0
_____________________________________________________________________
The kubeadm upgrade plan checks that your cluster is upgradeable and fetches the versions available to upgrade to in an user-friendly way.
To check CoreDNS version, include the --feature-gates=CoreDNS=true flag to verify the CoreDNS version which will be installed in place of kube-dns.
- Pick a version to upgrade to and run. For example:
$ kubeadm upgrade apply v1.9.0
[preflight] Running pre-flight checks.
[upgrade] Making sure the cluster is healthy:
[upgrade/config] Making sure the configuration is correct:
[upgrade/config] Reading configuration from the cluster...
[upgrade/config] FYI: You can look at this config file with 'kubectl -n kube-system get cm kubeadm-config -oyaml'
[upgrade/version] You have chosen to upgrade to version "v1.9.0"
[upgrade/versions] Cluster version: v1.8.1
[upgrade/versions] kubeadm version: v1.9.0
[upgrade/confirm] Are you sure you want to proceed with the upgrade? [y/N]: y
[upgrade/prepull] Will prepull images for components [kube-apiserver kube-controller-manager kube-scheduler]
[upgrade/apply] Upgrading your Static Pod-hosted control plane to version "v1.9.0"...
[etcd] Wrote Static Pod manifest for a local etcd instance to "/etc/kubernetes/tmp/kubeadm-upgraded-manifests802453804/etcd.yaml"
[upgrade/staticpods] Moved upgraded manifest to "/etc/kubernetes/manifests/etcd.yaml" and backed up old manifest to "/etc/kubernetes/tmp/kubeadm-backup-manifests502223003/etcd.yaml"
[upgrade/staticpods] Waiting for the kubelet to restart the component
[apiclient] Found 1 Pods for label selector component=etcd
[upgrade/staticpods] Component "etcd" upgraded successfully!
[upgrade/staticpods] Writing upgraded Static Pod manifests to "/etc/kubernetes/tmp/kubeadm-upgraded-manifests802453804"
[controlplane] Wrote Static Pod manifest for component kube-apiserver to "/etc/kubernetes/tmp/kubeadm-upgraded-manifests802453804/kube-apiserver.yaml"
[controlplane] Wrote Static Pod manifest for component kube-controller-manager to "/etc/kubernetes/tmp/kubeadm-upgraded-manifests802453804/kube-controller-manager.yaml"
[controlplane] Wrote Static Pod manifest for component kube-scheduler to "/etc/kubernetes/tmp/kubeadm-upgraded-manifests802453804/kube-scheduler.yaml"
[upgrade/staticpods] Moved upgraded manifest to "/etc/kubernetes/manifests/kube-apiserver.yaml" and backed up old manifest to "/etc/kubernetes/tmp/kubeadm-backup-manifests502223003/kube-apiserver.yaml"
[upgrade/staticpods] Waiting for the kubelet to restart the component
[apiclient] Found 1 Pods for label selector component=kube-apiserver
[upgrade/staticpods] Component "kube-apiserver" upgraded successfully!
[upgrade/staticpods] Moved upgraded manifest to "/etc/kubernetes/manifests/kube-controller-manager.yaml" and backed up old manifest to "/etc/kubernetes/tmp/kubeadm-backup-manifests502223003/kube-controller-manager.yaml"
[upgrade/staticpods] Waiting for the kubelet to restart the component
[apiclient] Found 1 Pods for label selector component=kube-controller-manager
[upgrade/staticpods] Component "kube-controller-manager" upgraded successfully!
[upgrade/staticpods] Moved upgraded manifest to "/etc/kubernetes/manifests/kube-scheduler.yaml" and backed up old manifest to "/etc/kubernetes/tmp/kubeadm-backup-manifests502223003/kube-scheduler.yaml"
[upgrade/staticpods] Waiting for the kubelet to restart the component
[apiclient] Found 1 Pods for label selector component=kube-scheduler
[upgrade/staticpods] Component "kube-scheduler" upgraded successfully!
[uploadconfig] Storing the configuration used in ConfigMap "kubeadm-config" in the "kube-system" Namespace
[bootstraptoken] Configured RBAC rules to allow Node Bootstrap tokens to post CSRs in order for nodes to get long term certificate credentials
[bootstraptoken] Configured RBAC rules to allow the csrapprover controller automatically approve CSRs from a Node Bootstrap Token
[bootstraptoken] Configured RBAC rules to allow certificate rotation for all node client certificates in the cluster
[addons] Applied essential addon: kube-dns
[addons] Applied essential addon: kube-proxy
[upgrade/successful] SUCCESS! Your cluster was upgraded to "v1.9.0". Enjoy!
[upgrade/kubelet] Now that your control plane is upgraded, please proceed with upgrading your kubelets in turn.
To upgrade the cluster with CoreDNS as the default internal DNS, invoke kubeadm upgrade apply with the --feature-gates=CoreDNS=true flag.
kubeadm upgrade apply does the following:
- Checks that your cluster is in an upgradeable state:
- The API server is reachable,
- All nodes are in the
Readystate - The control plane is healthy
- Enforces the version skew policies.
- Makes sure the control plane images are available or available to pull to the machine.
- Upgrades the control plane components or rollbacks if any of them fails to come up.
- Applies the new
kube-dnsandkube-proxymanifests and enforces that all necessary RBAC rules are created.
-
Manually upgrade your Software Defined Network (SDN).
Your Container Network Interface (CNI) provider may have its own upgrade instructions to follow. Check the addons page to find your CNI provider and see if there are additional upgrade steps necessary.
Upgrading your master and node packages
For each host (referred to as $HOST below) in your cluster, upgrade kubelet by executing the following commands:
- Prepare the host for maintenance, marking it unschedulable and evicting the workload:
$ kubectl drain $HOST --ignore-daemonsets
When running this command against the master host, this error is expected and can be safely ignored (since there are static pods running on the master):
node "master" already cordoned
error: pods not managed by ReplicationController, ReplicaSet, Job, DaemonSet or StatefulSet (use --force to override): etcd-kubeadm, kube-apiserver-kubeadm, kube-controller-manager-kubeadm, kube-scheduler-kubeadm
- Upgrade the Kubernetes package versions on the
$HOSTnode by using a Linux distribution-specific package manager:
If the host is running a Debian-based distro such as Ubuntu, run:
$ apt-get update
$ apt-get upgrade
If the host is running CentOS or the like, run:
$ yum update
Now the new version of the kubelet should be running on the host. Verify this using the following command on $HOST:
$ systemctl status kubelet
- Bring the host back online by marking it schedulable:
$ kubectl uncordon $HOST
- After upgrading
kubeleton each host in your cluster, verify that all nodes are available again by executing the following (from anywhere, for example, from outside the cluster):
$ kubectl get nodes
If the STATUS column of the above command shows Ready for all of your hosts, you are done.
Recovering from a failure state
If kubeadm upgrade somehow fails and fails to roll back, for example due to an unexpected shutdown during execution,
you can run kubeadm upgrade again as it is idempotent and should eventually make sure the actual state is the desired state you are declaring.
You can use kubeadm upgrade to change a running cluster with x.x.x --> x.x.x with --force, which can be used to recover from a bad state.