RKE2 v1.31

This matrix is revised as of v1.31.4+rke2r1

For additional information about RKE2 releases, check out our docs.

Out-of-Tree CPI and CSI Support

Caution: Do not upgrade to this minor version if using in-tree cloud providers.

Container Runtime Support

Windows Worker Node Support

Note: Windows clusters are Linux Clusters with Windows worker nodes only.

All clusters using windows works MUST use Calico as the CNI.

A "Yes" in a column means this combination is supported, an "N/A" means it is not. If any combination of versions is not shown in the table, it's not supported. 

Windows Worker Nodes is based off of:

Kubernetes Official Docs regarding Windows Server Support

Microsoft Official Docs on Windows Server Support 

Microsoft Official Docs on Windows Servicing Channels 

^† All “latest” tagged releases are intended for the Rancher community users to test-drive a new release and provide feedback. These “latest” tagged releases whilst covered by Rancher SLA are not generally meant for production use cases. Kindly use only the “stable” tagged releases for your own production use cases.

⁰ For open source components not listed in the matrix above, support is limited to troubleshooting for root cause up to Rancher’s drivers and interfaces to those components. Root causes that are identified to be beyond this limit will need to be pursued by Company with the maintainers and providers of commercial support for those components.

⁰⁰ for ensuring best support and clarity on supportability, Company is recommended to publish to Rancher a list of components that are critical to its deployment but not not explicitly called out in the support matrix

¹ certified on default OS kernel. Issues resulting from third party tools (typically, for reasons such as security) interfering with Docker or other necessary system calls are deemed resolved should disabling such tools restore functionality

² certified configurations are based on default settings of individual components. Where Company has deviated from certified configurations, Rancher Labs reserves the right to recommend the Company to revert to a certified configuration to resolve the reported issue.

³ whilst running kubernetes clusters in uncertified configurations may be possible, it is not recommended

⁴ upgrade of any individual component to a different version is likely to result in system downtime

⁵ changes to default settings of individual components may be necessary on exceptional basis, for reasons such as fine tuning for performance and scale. Engagement with Rancher Consulting or a partner may be recommended.

⁶ SLA on node drivers is limited to those that are built-in and active by default when Rancher is installed

⁷ SLA does not apply to issues within an application and its lifecycle management

⁸ issues may be referred to the partner as appropriate

⁹ "custom" means creating a cluster by generating a command line and pasting this command into existing VMs. "node driver" means creating VMs directly from Rancher when it connects to vSphere or other hypervisor

¹⁰ Currently Rancher Server running on k3s under SLE Micro has two support exceptions: (1) Rancher Logging has a permission issue on SLE Micro 5.1 and is not functioning (2) Rancher CIS scanning tool has issues with the way it is checking journalctl

^‡ SLA is limited to running workload clusters on hosted kubernetes provider and does not apply to running the Rancher control plane on one of the listed hosted kubernetes providers for all Rancher versions older than Rancher v2.5.x. In Rancher v2.5.x, SLA applies to running Rancher control plane on the listed kubernetes distributions and hosted providers.