How OpenYurt v0.3.0 Boosts Edge Computing with NodePools and UnitedDeployment
OpenYurt v0.3.0, released in November 2021, adds non‑intrusive edge extensions to native Kubernetes—including NodePool and UnitedDeployment controllers, a cloud‑side Yurt‑App‑Manager, and performance upgrades to yurthub, yurt‑tunnel and yurtctl—making edge workloads easier to deploy, manage, and operate offline.
OpenYurt Overview
OpenYurt is an open‑source project from Alibaba Cloud that extends native Kubernetes to support edge‑computing scenarios without modifying the upstream API. It retains full Kubernetes API compatibility, works with all standard workloads, Services, CNI and CSI plugins, and provides node autonomy so that applications continue to run when edge nodes lose connectivity to the cloud.
v0.3.0 Release Highlights (8 Nov 2021)
This release introduces two new custom resources— NodePool and UnitedDeployment —and adds the cloud‑side Yurt‑App‑Manager component. Core components yurthub and yurt‑tunnel receive performance and reliability improvements. The yurtctl CLI now includes a kubeadm provider that can convert an existing kubeadm‑created cluster into an OpenYurt cluster.
Yurt‑App‑Manager
The Yurt‑App‑Manager is a standard Kubernetes extension that installs two controllers: NodePool and UnitedDeployment. These controllers enable edge‑focused node and application management from both infrastructure and workload perspectives.
NodePool
NodePool groups edge nodes that share logical attributes such as CPU architecture, ISP, or cloud provider. Instead of relying on a large number of individual labels, NodePool abstracts a set of nodes into a higher‑level unit, allowing bulk operations such as:
Updating scheduling policies for the whole pool
Applying or removing taints across all members
Querying or reporting pool‑level health metrics
UnitedDeployment
UnitedDeployment abstracts multiple regional Deployment or StatefulSet objects into a single template. For each defined region (called a *pool*), the controller creates a child workload with its own replica count and optional overrides (e.g., node selector, resource limits). This reduces operational overhead by:
Eliminating the need to edit each Deployment when upgrading container images
Removing custom naming conventions required to identify related Deployments
Keeping configuration drift low as edge complexity grows
yurt‑hub
yurt‑hubruns as a daemon on every node and acts as an outbound proxy for Kubelet, Kube‑proxy, CNI plugins, etc. It caches Kubernetes resources locally so that a node can continue to serve pods after a network outage. v0.3.0 adds:
Automatic certificate request and rotation when connecting to the cloud API server
A timeout mechanism for watches of cloud resources
Optimized responses when requested data is missing from the local cache
yurt‑tunnel
The tunnel component consists of a cloud‑side TunnelServer and an edge‑side TunnelAgent . It establishes a reverse‑proxy connection that allows the public cloud control plane to reach edge nodes behind firewalls. The v0.3.0 release improves reliability, stability, and adds extensive integration tests.
yurtctl
The yurtctl command‑line tool now provides a kubeadm provider. With a single command, users can convert a vanilla kubeadm‑created Kubernetes cluster into an OpenYurt cluster that is optimized for weak‑network edge environments.
Future Directions
Upcoming work focuses on device management, edge‑aware scheduling, governance, and improving contributor experience. The project welcomes contributions from the community.
Repository: https://github.com/alibaba/openyurt
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
Alibaba Cloud Native
We publish cloud-native tech news, curate in-depth content, host regular events and live streams, and share Alibaba product and user case studies. Join us to explore and share the cloud-native insights you need.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
