Kube-Proxy IPVS Backend: Transitioning to Nftables
The kube-proxy IPVS backend was designed to tackle performance issues in service proxy implementations, aiming to outperform the traditional iptables backend. It addressed both data plane and control plane scaling problems effectively. The data plane benefited from the kernel IPVS subsystem's use of hash tables, which allowed for faster lookups compared to linear methods. On the control plane side, it avoided the need to rewrite the entire ruleset with every update, streamlining operations significantly.
As of Kubernetes 1.35, the IPVS backend is officially deprecated, with plans for its removal in Kubernetes 1.40. If you’re running Kubernetes 1.36 or later, ensure your nodes are on a Linux distribution with kernel 5.13 or later to support the nftables backend. Transitioning from IPVS to nftables is straightforward; simply change the kube-proxy mode from IPVS to nftables in your configuration. This shift is not just a matter of compliance; it’s about leveraging a more efficient and modern networking stack.
Be aware that while IPVS remains available for now, it comes with a warning about its impending removal. This means you should prioritize migrating to nftables to avoid future disruptions. The transition is essential for maintaining optimal performance and ensuring your cluster remains up-to-date with Kubernetes best practices.
Key takeaways
- →Understand the performance benefits of the kube-proxy IPVS backend over iptables.
- →Transition to nftables by changing the kube-proxy mode from IPVS to nftables.
- →Prepare for the removal of the IPVS backend in Kubernetes 1.40.
- →Ensure your nodes run a compatible Linux distribution with kernel 5.13 or later.
Why it matters
In production, staying ahead of deprecations like the kube-proxy IPVS backend is critical. Transitioning to nftables not only ensures compliance but also enhances performance and scalability in your Kubernetes clusters.
Code examples
mode: ipvs
to
mode: nftablesWhen NOT to use this
The official docs don't call out specific anti-patterns here. Use your judgment based on your scale and requirements.
Want the complete reference?
Read official docsIndustry-standard certifications built by the people behind Linux and Kubernetes. Earn the CKA — the gold standard Kubernetes administrator cert. OpsCanary readers get 30% off year-round with code OPSCANARY3.
Get CKA certified →Streamline Your ECR Workflow with Global Authentication Tokens
Managing container images across multiple AWS Regions can be a hassle. A global authentication token simplifies this by allowing you to push and pull images in any Region with a single token. Learn how to leverage this feature effectively in your Kubernetes deployments.
Mastering Kubernetes Networking: The CKNE Certification Unpacked
The Certified Kubernetes Network Engineer (CKNE) certification is here to elevate your networking skills in Kubernetes. This performance-based exam tests your ability to handle real-world networking challenges across production environments. Dive in to understand the key domains covered and how this certification can impact your career.
CiliumCon Returns: Scaling eBPF Networking at KubeCon 2026
CiliumCon is back, and it's crucial for anyone involved in Kubernetes networking. With Cilium as the de facto CNI, understanding its eBPF foundation is key to managing workloads at scale. Don't miss out on the insights that could elevate your cloud-native architecture.
Get the daily digest
One email. 5 articles. Every morning.
No spam. Unsubscribe anytime.