v0.2.1

Release 0.2.1

Added

• Experimental client for interacting with backup service
• The operator panics itself when it gets stuck unexpectedly. It relies on Kubernetes to
  get restarted.
• Add resource requirements field in Spec.Pod . Users can specify resource requirements for the
  etcd container with this new field.
• Add status endpoint to backup sidecar service.
• Service account of the etcd operator pod is passed to backup pod.
• Add backup service status into cluster status.

Changed

• Example deployments pin to the released version of the operator image
• Downward API of pod's namespace and name is required to start etcd operator pod.
  See example/deployment.yaml .

Removed

• Drop etcd operator command line flags: "masterHost", "cert-file", "key-file", "ca-file".

v0.2.0

etcd operator v0.2.0 has been built and pushed. This is our first beta release. All major planned features are complete. We will focus on improving usability and fixing reported bugs in the next release cycle. Please try out etcd operator 0.2, provide us feedback, and report any issues.

New Features:

• Backup and recovery
  • Backup etcd snapshot to S3.
  • Recover an etcd cluster from Persistent Volume.
  • Recover an etcd cluster from S3.
• Status reporting
  • Report cluster status and important condition changes in TPR object under “status” field.
• Garbage collection
  • Removing garbage pods, services that could be potentially left due to network failure after TPR object is removed.

Notable changes to existing behavior:

• the API group is changed from coreos.com to etcd.coreos.com
• the TRP kind is changed from EtcdCluster to Cluster
• the pod options is grouped together to Pod field in the spec

We start to keep a change log after this release.

v0.1.0

This is the first public release of etcd-operator.
The current stage is alpha.
See ROADMAP.md.