Beginning with Rancher 2.2.0 and continuing with future versions, we will ship a series of preview builds before the final GA release to allow our users a chance to experiment with new features and to assist with identifying potential issues.
Preview releases are tagged with the suffix
-alpha<number> and are bleeding edge. They should not be run in a production environment. Upgrading to or from an alpha release is not tested or supported, even if it appears to work successfully.
When notable features or fixes land in a particular alpha, we will call them out in the release notes. We will also call out any major known issues, but the list will not be exhaustive.
We truly value the feedback/comments/suggestions we’ve received throughout our years of building Rancher, and we feel this is another step to allow you to provide more input prior to a GA release. Thank you all for your support!
Major Features and Enhancements
- Support for multi-cluster applications [#16016] - Rancher now supports deploying, managing, and upgrading applications across multiple clusters and projects. By augmenting the functionality of Helm with Rancher’s multi-cluster management capabilities, users are now able to seamlessly manage their applications across clusters. Use this feature for high availability, disaster recovery, or ensuring your clusters are all running a common set of applications.
- Support for global DNS [#16014] - Rancher now supports configuring and managing DNS endpoints across clusters and projects. By combining Rancher’s multi-cluster management capabilities with Kubernetes’ external-dns controller, users now have unprecedented control over managing the DNS entries of their applications across availability zones, data centers, and even cloud providers.
- Support for certificate rotation for clusters provisioned by RKE [#450] - Rancher now supports rotating the certificates generated by RKE - Rancher Kubernetes Engine. Use this feature to regenerate expired or compromised certificates to ensure secure and reliable communication between Kubernetes components in RKE clusters.
- Support for etcd backup/restore for clusters provisioned by RKE with an ability to store a backup on S3 compatible storage provider [#13688] [#16163] - Rancher now supports not only configuring an etcd backup from the UI, but also restoring from it, as well as saving the backups to any S3 compatible storage provider.
- Features added in [alpha3] - These include cluster and project level monitoring, pluggable cluster drivers, and support for new Kubernetes providers along with many bug fixes and improvements
Major known issues
- Full RBAC functionality for multi-cluster app and global DNS is not yet implemented [#17565] [#17564] - Once implemented you will be able to share these resources with other users and groups. For now, only the creator can manage these resources.
- MultiClusterApp rolling upgrade: batch size is not respected [#17856]
- MultiClusterApp upon upgrade creates some extra revisions [#17818]
- Updating Global DNS annotation on a target ingress doesn’t take effect [#17831]
- Unable to correct ectd s3 backup password once it has been entered incorrectly [#17822]
- Kubectl shell breaks after etcdbackup restore [#17521]
NOTE - Image Name Changes: Please note that as of v2.0.0, our images will be rancher/rancher and rancher/rancher-agent. If you are using v1.6, please continue to use rancher/server and rancher/agent.
NOTE - We’ve introduced a new helm repository to facilitate the deployment of alpha releases. For more details, see our installation documentation.