This document shows how to backup and restore the etcd store for a user cluster. This page also provides a script that you can use to automatically back up a cluster's etcd store.
You can create a backup file for recovery from foreseen disasters that might damage your cluster's etcd data. Store the backup file in a location that is outside of the cluster and is not dependent on the cluster's operation.
Limitations
This procedure does not back up application-specific data.
This procedure does not back up your PersistentVolumes.
Workloads scheduled after you create a backup aren't restored with that backup.
You cannot restore a cluster after a failed upgrade.
This procedure is not intended to restore a deleted cluster.
Backing up a user cluster
A user cluster backup is a snapshot of the user cluster's etcd store. The etcd store contains all of the Kubernetes objects and custom objects required to manage cluster state. The snapshot contains the data required to recreate the cluster's components and workloads.
To create a snapshot of the etcd store, perform the following steps:
Get a shell into the
kube-etcd
container:kubectl --kubeconfig ADMIN_CLUSTER_KUBECONFIG exec -it \ kube-etcd-0 --container kube-etcd --namespace USER_CLUSTER_NAME \ -- bin/sh
where:
- ADMIN_CLUSTER_KUBECONFIG is the admin cluster's kubeconfig file.
- USER_CLUSTER_NAME is the name of the user cluster.
In your shell, in the root directory, create backup named
snapshot.db
:ETCDCTL_API=3 etcdctl \ --endpoints=https://127.0.0.1:2379 \ --cacert=/etcd.local.config/certificates/etcdCA.crt \ --cert=/etcd.local.config/certificates/etcd.crt \ --key=/etcd.local.config/certificates/etcd.key \ snapshot save snapshot.db
In your shell, enter
exit
to exit the shell.Copy
snapshot.db
from thekube-etcd
container to the current directory:kubectl --kubeconfig ADMIN_CLUSTER_KUBECONFIG cp \ USER_CLUSTER_NAME/kube-etcd-0:snapshot.db \ --container kube-etcd snapshot.db
Restoring a user cluster from a backup (non-HA)
Before you use a backup file to restore your user cluster's etcd store, diagnose your cluster and resolve existing issues. Using a backup to restore a problematic cluster might re-create or exacerbate issues. Contact the Google Distributed Cloud support team for further assistance with restoring your clusters.
The following instructions explain how to use a backup file to restore a user cluster in cases where the cluster's etcd data has become damaged and the user cluster's etcd Pod is crashlooping.
You can restore the etcd data by deploying a utility Pod that overwrites the damaged data with the backup. The admin cluster's API server must be running and the admin cluster's scheduler must be able to schedule new Pods.
Find the name of the Secret used by etcd in the user cluster:
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONFIG -n USER_CLUSTER_NAME \ get secrets | grep kube-etcd-certs
Copy the following Pod manifest to a file named
etcd-utility.yaml
. Replace the following:NODE_NAME
: the node where thekube-etcd-0
Pod is running.ADMIN_CLUSTER_KUBECONFIG
: the admin cluster's kubeconfig file.USER_CLUSTER_NAME
: the name of the user cluster.GKE_ON_PREM_VERSION
: the version of the cluster where you want to perform the etcd restore (for example, 1.5.0-gke.0).KUBE_ETCD_SECRET_NAME
: the name of the Secret used by etcd in the user cluster, starting withkube-etcd-certs
.
apiVersion: v1 kind: Pod metadata: name: etcd-utility-0 namespace: USER_CLUSTER_NAME spec: containers: - command: ["/bin/sh"] args: ["-ec", "while :; do echo '.'; sleep 5 ; done"] image: gcr.io/gke-on-prem-release/etcd-util:GKE_ON_PREM_VERSION name: etcd-utility volumeMounts: - mountPath: /var/lib/etcd name: data - mountPath: /etcd.local.config/certificates name: etcd-certs nodeSelector: kubernetes.googleapis.com/cluster-name: USER_CLUSTER_NAME kubernetes.io/hostname: NODE_NAME tolerations: - effect: NoExecute key: node.kubernetes.io/not-ready operator: Exists tolerationSeconds: 300 - effect: NoExecute key: node.kubernetes.io/unreachable operator: Exists tolerationSeconds: 300 - effect: NoSchedule key: node.kubernetes.io/unschedulable operator: Exists volumes: - name: data persistentVolumeClaim: claimName: data-kube-etcd-0 - name: etcd-certs secret: defaultMode: 420 secretName: KUBE_ETCD_SECRET_NAME
Deploy the utility Pod:
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONFIG \ create -f etcd-utility.yaml --namespace USER_CLUSTER_NAME
Copy
snapshot.db
from the current directory to the root directory of the utility Pod:kubectl --kubeconfig ADMIN_CLUSTER_KUBECONFIG cp snapshot.db \ USER_CLUSTER_NAME/etcd-utility-0:snapshot.db --container etcd-utility
Get a shell into the
etcd-utility
container:kubectl --kubeconfig ADMIN_CLUSTER_KUBECONFIG exec it \ etcd-utility-0 --container etcd-utility --namespace USER_CLUSTER_NAME \ -- bin/sh
In your shell, in the root directory, run the following command to create a new folder that contains the backup:
ETCDCTL_API=3 etcdctl \ --endpoints=https://127.0.0.1:2379 \ --cacert=/etcd.local.config/certificates/etcdCA.crt \ --cert=/etcd.local.config/certificates/etcd.crt \ --key=/etcd.local.config/certificates/etcd.key \ snapshot restore snapshot.db
In your shell, delete the old etcd data:
rm -r var/lib/etcd/*
In your shell, copy the restored etcd data to its permanent location:
cp -r default.etcd/* var/lib/etcd/
In your shell, enter
exit
to exit the shell.Delete the crashing etcd Pod:
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONFIG \ delete pod kube-etcd-0 --namespace USER_CLUSTER_NAME
Verify that the etcd Pod is no longer crashing.
Delete the utility Pod:
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONFIG \ delete pod etcd-utility-0 --namespace USER_CLUSTER_NAME
Remove
etcd-utility.yaml
from the current directory:rm etcd-utility.yaml
Restoring a user cluster from a backup (HA)
This section shows how to restore the etcd data for a high-availability (HA) user cluster.
For an HA user cluster, there are three nodes in the admin cluster that serve as control planes for the user cluster. Each of those nodes runs an etcd Pod that maintains etcd data on a storage volume.
If two of the etcd Pods are healthy, and the data on the associated storage volumes is intact, then there is no need to use a backup file. That is because you still have an etcd quorum.
In the rare case that two of the etcd storage volumes have corrupt data, you need to use a backup file to restore the etcd data.
To do the steps in this section, you must have already created a snapshot.db
file as described in Backing up a user cluster.
Listing your etcd Pods and nodes
List the etcd Pods that manage the etcd store for your user cluster. These Pods run in the admin cluster:
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONFIG get pods --namespace USER_CLUSTER_NAME \ --output wide | grep kube-etcd
The output shows the etcd Pods and the nodes where the Pods run. The nodes shown in the output are nodes in the admin cluster that serve as control planes for your user cluster:
NAME ... NODE kube-etcd-0 ... node-xxx kube-etcd-1 ... node-yyy kube-etcd-2 ... node-zzz
Make a note of the Pod names and the control plane node names for later.
Notice that each etcd Pod is named
kube-etcd
appended with a number. This number is called the member number for the Pod. It identifies the Pod as being a particular member of the etcd cluster that holds the object data for your user cluster. This guide uses the placeholder MEMBER_NUMBER to refer to the etcd Pod member number.Also notice that each Pod in your etcd cluster runs on its own node.
Preparing to deploy the utility Pods
Save a manifest for the PodDisruptionBudget (PDB) for the user cluster's Kubernetes API server. Then delete the PDB.
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONIFG get pdb --namespace USER_CLUSTER_NAME \ kube-apiserver-pdb --output yaml > kube-apiserver-pdb.yaml kubectl --kubeconfig ADMIN_CLUSTER_KUBECONIFG delete pdb --namespace USER_CLUSTER_NAME \ kube-apiserver-pdb
Stop the Kubernetes API server and the etcd maintenance Deployment. This ensures that no components will use etcd during restoration:
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONIFG --namespace USER_CLUSTER_NAME \ scale --replicas 0 statefulset kube-apiserver kubectl --kubeconfig ADMIN_CLUSTER_KUBECONIFG --namespace USER_CLUSTER_NAME \ scale --replicas 0 deployment gke-master-etcd-maintenance
Recall the name of the container image for your etcd Pods.
Deploying the utility Pods
Recall the name of the etcd Pod and the name of the node where the Pod runs.
Save the following Pod manifest in the current directory in a file named
etcd-utility-MEMBER_NUMBER.yaml
:apiVersion: v1 kind: Pod metadata: name: etcd-utility-MEMBER_NUMBER namespace: USER_CLUSTER_NAME spec: containers: - command: ["/bin/sh"] args: ["-ec", "while :; do echo '.'; sleep 5 ; done"] image: gcr.io/gke-on-prem-release/etcd-util:GKE_ON_PREM_VERSION name: etcd-utility volumeMounts: - mountPath: /var/lib/etcd name: data - mountPath: /etcd.local.config/certificates name: etcd-certs nodeSelector: kubernetes.googleapis.com/cluster-name: USER_CLUSTER_NAME kubernetes.io/hostname: NODE_NAME tolerations: - effect: NoExecute key: node.kubernetes.io/not-ready operator: Exists tolerationSeconds: 300 - effect: NoExecute key: node.kubernetes.io/unreachable operator: Exists tolerationSeconds: 300 - effect: NoSchedule key: node.kubernetes.io/unschedulable operator: Exists volumes: - name: data persistentVolumeClaim: claimName: data-kube-etcd-MEMBER_NUMBER - name: etcd-certs secret: defaultMode: 420 secretName: KUBE_ETCD_SECRET_NAME
The preceding manifest describes a utility Pod that you run temporarily to restore etcd data.
Create the utility Pod in your admin cluster:
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONFIG apply -f etcd-utility-MEMBER_NUMBER.yaml
Copy your backup file,
snapshot.db
, to the root directory of your utility Pod:kubectl --kubeconfig ADMIN_CLUSTER_KUBECONIFG cp snapshot.db \ USER_CLUSTER_NAME/etcd-utility-MEMBER_NUMBER:snapshot.db
Get a shell into the
etcd-utility
container in the utility Pod:kubectl --kubeconfig ADMIN_CLUSTER_KUBECONIFG exec -it --namespace USER_CLUSTER_NAME \ etcd-utility-MEMBER_NUMBER --container etcd-utility -- bin/sh
In your shell, in the root directory, use
snapshot.db
to restore the etcd data:ETCDCTL_API=3 etcdctl \ --endpoints=https://127.0.0.1:2379 \ --cacert=/etcd.local.config/certificates/etcdCA.crt \ --cert=/etcd.local.config/certificates/etcd.crt \ --key=/etcd.local.config/certificates/etcd.key \ --name=kube-etcd-MEMBER_NUMBER \ --initial-cluster=kube-etcd-0=https://kube-etcd-0.kube-etcd:2380,kube-etcd-1=https://kube-etcd-1.kube-etcd:2380,kube-etcd-2=https://kube-etcd-2.kube-etcd:2380 \ --initial-cluster-token=etcd-cluster-1 \ --initial-advertise-peer-urls=https://kube-etcd-MEMBER_NUMBER.kube-etcd:2380 \ snapshot restore snapshot.db
The preceding command stored etcd data in the
/kube-etcd-MEMBER_NUMBER.etcd
directory.In your shell, delete the old etcd data:
rm -r var/lib/etcd/*
In your shell, copy the restored etcd data to its permanent location:
cp -r kube-etcd-MEMBER_NUMBER.etcd/* var/lib/etcd/
In your shell, remove the temporary etcd directory and the backup file:
rm -R kube-etcd-MEMBER_NUMBER.etcd/ rm snapshot.db
In your shell, enter
exit
to exit the shell.Delete the utility Pod:
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONIFG delete pod \ --namespace USER_CLUSTER_NAME etcd-utility-MEMBER_NUMBER
Restarting components
Now that you have deployed and deleted your utility Pods, you need to restart some cluster components.
Restart the Pods in the
kube-etcd
StatefulSet:kubectl --kubeconfig ADMIN_CLUSTER_KUBECONFIG rollout restart statefulset \ --namespace USER_CLUSTER_NAME kube-etcd
Start the Kubernetes API servers for your user cluster:
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONIFG scale statefulset --replicas 3 \ --namespace USER_CLUSTER_NAME kube-apiserver
Start the etcd maintenance Deployment for your user cluster:
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONIFG scale deployment --replicas 1 \ --namespace=USER_CLUSTER_NAME gke-master-etcd-maintenance
Restore the PDB for the Kubernetes API server:
kubectl --kubeconfig ADMIN_CLUSTER_KUBECONIFG apply -f kube-apiserver-pdb.yaml
Automatic cluster backup
You can use the script given here as an example on how to automatically back up your clusters. Note that the following script is not supported, and should only be used as reference to write a better, more robust and complete script. Before you run the script, fill in values for the five variables at the beginning of the script:
- Set
BACKUP_DIR
to the path where you want to store the admin and user cluster backups. This path should not exist. - Set
ADMIN_CLUSTER_KUBECONFIG
to the path of the admin cluster's kubeconfig file - Set
USER_CLUSTER_NAMESPACE
to the name of your user cluster. The name of your user cluster is a namespace in the admin cluster. - Set
EXTERNAL_IP
to the VIP that you reserved for the admin control plane service. - Set
SSH_PRIVATE_KEY
to the path of your SSH key. - If you are using a private network, set
JUMP_IP
to your network's jump server's IP address.
#!/usr/bin/env bash
# Automates manual steps for taking backups of user and admin clusters.
# Fill in the variables below before running the script.
BACKUP_DIR="" # path to store user and admin cluster backups
ADMIN_CLUSTER_KUBECONFIG="" # path to admin cluster kubeconfig
USER_CLUSTER_NAMESPACE="" # user cluster namespace
EXTERNAL_IP="" # admin control plane node external ip - follow steps in documentation
SSH_PRIVATE_KEY="" # path to vsphere_tmp ssh private key - follow steps in documentation
JUMP_IP="" # network jump server IP - leave empty string if not using private network.
mkdir -p $BACKUP_DIR
mkdir $BACKUP_DIR/pki
# USER CLUSTER BACKUP
# Snapshot user cluster etcd
kubectl --kubeconfig=${ADMIN_CLUSTER_KUBECONFIG} exec -it -n ${USER_CLUSTER_NAMESPACE} kube-etcd-0 -c kube-etcd -- /bin/sh -ec "export ETCDCTL_API=3; etcdctl --endpoints=https://127.0.0.1:2379 --cacert=/etcd.local.config/certificates/etcdCA.crt --cert=/etcd.local.config/certificates/etcd.crt --key=/etcd.local.config/certificates/etcd.key snapshot save ${USER_CLUSTER_NAMESPACE}_snapshot.db"
kubectl --kubeconfig=${ADMIN_CLUSTER_KUBECONFIG} cp ${USER_CLUSTER_NAMESPACE}/kube-etcd-0:${USER_CLUSTER_NAMESPACE}_snapshot.db $BACKUP_DIR/user-cluster_${USER_CLUSTER_NAMESPACE}_snapshot.db
# ADMIN CLUSTER BACKUP
# Set up ssh options
SSH_OPTS=(-oStrictHostKeyChecking=no -i ${SSH_PRIVATE_KEY})
if [ "${JUMP_IP}" != "" ]; then
SSH_OPTS+=(-oProxyCommand="ssh -oStrictHostKeyChecking=no -i ${SSH_PRIVATE_KEY} -W %h:%p ubuntu@${JUMP_IP}")
fi
# Copy admin certs
ssh "${SSH_OPTS[@]}" ubuntu@${EXTERNAL_IP} 'sudo chmod -R +rw /etc/kubernetes/pki/*'
scp -r "${SSH_OPTS[@]}" ubuntu@${EXTERNAL_IP}:/etc/kubernetes/pki/* ${BACKUP_DIR}/pki/
# Snapshot admin cluster etcd
admin_etcd=$(kubectl --kubeconfig=${ADMIN_CLUSTER_KUBECONFIG} get pods -n kube-system -l component=etcd,tier=control-plane -ojsonpath='{$.items[*].metadata.name}{"\n"}')
kubectl --kubeconfig=${ADMIN_CLUSTER_KUBECONFIG} exec -it -n kube-system ${admin_etcd} -- /bin/sh -ec "export ETCDCTL_API=3; etcdctl --endpoints=https://127.0.0.1:2379 --cacert=/etc/kubernetes/pki/etcd/ca.crt --cert=/etc/kubernetes/pki/etcd/healthcheck-client.crt --key=/etc/kubernetes/pki/etcd/healthcheck-client.key snapshot save admin_snapshot.db"
kubectl --kubeconfig=${ADMIN_CLUSTER_KUBECONFIG} cp -n kube-system ${admin_etcd}:admin_snapshot.db $BACKUP_DIR/admin-cluster_snapshot.db
What's next
- Back up and restore an admin cluster
- Learn how to diagnose cluster issues
- Learn about augur, an open-source tool for restoring individual objects from etcd backups.