Skip to content

Latest commit

 

History

History
99 lines (78 loc) · 2.53 KB

migrate_cr_070.md

File metadata and controls

99 lines (78 loc) · 2.53 KB
Raw

Custom Resource Migration Guide for 0.7.0 relase

In 0.7.0 release, BackupPolicy and RestorePolicy are removed from EtcdCluster spec. This doc provides the migration guide for the change.

Migrate EtcdCluster

This has to be done before upgrading operator.

Update any EtcdCluster Custom Resource to remove BackupPolicy and RestorePolicy:

kubectl get etcdcluster <cluster-name> -o json | \
	jq 'del(.spec.backup)' | jq 'del(.spec.restore)' | \
	kubectl replace -f -

Migrate BackupPolicy to EtcdBackup CR

Read etcd backup operator doc for how to save backup for etcd cluster. It currently only supports S3 backup for non-TLS etcd cluster.

For previous backup policy, for example:

apiVersion: "etcd.database.coreos.com/v1beta2"
kind: "EtcdCluster"
metadata:
  name: "example-etcd-cluster"
spec:
  ...
  backup:
    storageType: "S3"
    s3:
      s3Bucket: <s3-bucket-name>
      awsSecret: <aws-secret-name>

The equivalent EtcdBackup CR is:

apiVersion: "etcd.database.coreos.com/v1beta2"
kind: "EtcdBackup"
metadata:
  name: example-etcd-cluster-backup
spec:
  clusterName: example-etcd-cluster
  storageType: "S3"
  s3:
    s3Bucket: <s3-bucket-name>
    awsSecret: <aws-secret-name>

Create the above CR will make one-time backup attempt to target etcd cluster. After backup is saved, it is safe to delete the CR and doing that won’t delete the S3 backup.

Migrate RestoreSpec to EtcdRestore CR

Read etcd restore operator doc for how to restore etcd cluster from backup. It currently only supports S3 backup and non-TLS etcd cluster.

For previous restore policy for example:

apiVersion: "etcd.database.coreos.com/v1beta2"
kind: "EtcdCluster"
metadata:
  name: "restored-etcd-cluster"
spec:
  ...
  restore:
    storageType: "S3"
    backupClustername: “example-etcd-cluster”

The equivalent EtcdRestore CR is:

apiVersion: "etcd.database.coreos.com/v1beta2"
kind: "EtcdRestore"
metadata:
  # an EtcdCluster with the same name will be created
  name: "restored-etcd-cluster"
spec:
  clusterSpec:
    size: 3
    version: "3.1.8"
  s3:
    # e.g: "etcd-snapshot-bucket/v1/default/example-etcd-cluster/3.1.8_0000000000000001_etcd.backup"
    path: <full-s3-path>
    awsSecret: <aws-secret>

Create the above CR to trigger restore.

Note: restore process will create a new EtcdCluster CR.

Wait until etcd cluster reaches given size. The it is safe to delete the EtcdRestore CR and doing that won’t delete the EtcdCluster CR.