Skip to content

Latest commit

 

History

History
269 lines (192 loc) · 11.5 KB

getting-started.md

File metadata and controls

269 lines (192 loc) · 11.5 KB
Raw

Getting Started

Antrea is super easy to install. All the Antrea components are containerized and can be installed using the Kubernetes deployment manifest.

antrea-demo

Ensuring requirements are satisfied

NodeIPAM

Antrea relies on NodeIPAM for per-Node CIDR allocation. NodeIPAM can run within the Kubernetes kube-controller-manager, or within the Antrea Controller.

NodeIPAM within kube-controller-manager

When using kubeadm to create the Kubernetes cluster, passing --pod-network-cidr=<CIDR Range for Pods> to kubeadm init will enable NodeIpamController. Clusters created with kubeadm will always have CNI plugins enabled. Refer to Creating a cluster with kubeadm for more information about setting up a Kubernetes cluster with kubeadm.

When the cluster is deployed by other means then:

  • To enable NodeIpamController, kube-controller-manager should be started with the following flags:

    • --cluster-cidr=<CIDR Range for Pods>
    • --allocate-node-cidrs=true

  • To enable CNI network plugins, kubelet should be started with the --network-plugin=cni flag.

  • To enable masquerading of traffic for Service cluster IP via iptables, kube-proxy should be started with the --cluster-cidr=<CIDR Range for Pods> flag.

NodeIPAM within Antrea Controller

For further info about running NodeIPAM within Antrea Controller, see Antrea IPAM Capabilities

Open vSwitch

As for OVS, when using the built-in kernel module, kernel version >= 4.6 is required. On the other hand, when building it from OVS sources, OVS version >= 2.6.0 is required.

Red Hat Enterprise Linux and CentOS 7.x use kernel 3.10, but as changes to OVS kernel modules are regularly backported to these kernel versions, they should work with Antrea, starting with version 7.4.

In case a node does not have a supported OVS module installed, you can install it following the instructions at: Installing Open vSwitch. Please be aware that the vport-stt module is not in the Linux tree and needs to be built from source, please build and load it manually before STT tunneling is enabled.

Some experimental features disabled by default may have additional requirements, please refer to the Feature Gates documentation to determine whether it applies to you.

Antrea will work out-of-the-box on most popular Operating Systems. Known issues encountered when running Antrea on specific OSes are documented here.

There are also a few network prerequisites which need to be satisfied, and they depend on the tunnel mode you choose, please check network requirements.

Installation / Upgrade

To deploy a released version of Antrea, pick a deployment manifest from the list of releases. For any given release <TAG> (e.g. v0.1.0), you can deploy Antrea as follows:

kubectl apply -f https://github.com/antrea-io/antrea/releases/download/<TAG>/antrea.yml

To deploy the latest version of Antrea (built from the main branch), use the checked-in deployment yaml:

kubectl apply -f https://raw.githubusercontent.com/antrea-io/antrea/main/build/yamls/antrea.yml

You can use the same kubectl apply command to upgrade to a more recent version of Antrea.

Antrea supports some experimental features that can be enabled or disabled, please refer to the Feature Gates documentation for more information.

Windows support

If you want to add Windows Nodes to your cluster, please refer to these installation instructions.

ARM support

Starting with v1.0, Antrea supports arm64 and arm/v7 Nodes. The installation instructions do not change when some (or all) Linux Nodes in a cluster use an ARM architecture: the same deployment YAML can be used, as the antrea/antrea-agent-ubuntu and antrea/antrea-controller-ubuntu Docker images are actually manifest lists with support for the amd64, arm64 and arm/v7 architectures.

Note that while we do run a subset of the Kubernetes conformance tests on both the arm/v7 and arm64 Docker images (using k3s as the Kubernetes distribution), our testing is not as thorough as for the amd64 image. However, we do not anticipate any issue.

Install with Helm

Starting with v1.8, Antrea can be installed and updated with Helm. Please refer to these installation instructions.

Deploying Antrea on a Cluster with Existing CNI

The instructions above only apply when deploying Antrea in a new cluster. If you need to migrate your existing cluster from another CNI plugin to Antrea, you will need to do the following:

  • Delete previous CNI, including all resources (K8s objects, iptables rules, interfaces, ...) created by that CNI.
  • Deploy Antrea.
  • Restart all Pods in the CNI network in order for Antrea to set-up networking for them. This does not apply to Pods which use the Node's network namespace (i.e. Pods configured with hostNetwork: true). You may use kubectl drain to drain each Node or reboot all your Nodes.

While this is in-progress, networking will be disrupted in your cluster. After deleting the previous CNI, existing Pods may not be reachable anymore.

For example, when migrating from Flannel to Antrea, you will need to do the following:

  1. Delete Flannel with kubectl delete -f <path to your Flannel YAML manifest>.
  2. Delete Flannel bridge and tunnel interface with ip link delete flannel.1 && ip link delete flannel cni0 on each Node.
  3. Ensure requirements are satisfied.
  4. Deploy Antrea.
  5. Drain and uncordon Nodes one-by-one. For each Node, run kubectl drain --ignore-daemonsets <node name> && kubectl uncordon <node name>. The --ignore-daemonsets flag will ignore DaemonSet-managed Pods, including the Antrea Agent Pods. If you have any other DaemonSet-managed Pods (besides the Antrea ones and system ones such as kube-proxy), they will be ignored and will not be drained from the Node. Refer to the Kubernetes documentation for more information. Alternatively, you can also restart all the Pods yourself, or simply reboot your Nodes.

To build the image locally, you can follow the instructions in the Contributor Guide.

Deploying Antrea in Kind

To deploy Antrea in a Kind cluster, please refer to this guide.

Deploying Antrea in Minikube

To deploy Antrea in a Minikube cluster, please refer to this guide.

Deploying Antrea in Rancher Managed Cluster

To deploy Antrea in a Rancher managed cluster, please refer to this guide.

Deploying Antrea in AKS, EKS, and GKE

Antrea can work with cloud managed Kubernetes services, and can be deployed to AKS, EKS, and GKE clusters.

Deploying Antrea with Custom Certificates

By default, Antrea generates the certificates needed for itself to run. To provide your own certificates, please refer to Securing Control Plane.

Antctl: Installation and Usage

To use antctl, the Antrea command-line tool, please refer to this guide.

Features

Antrea Network Policy

Besides Kubernetes NetworkPolicy, Antrea also implements its own Network Policy CRDs, which provide advanced features including: policy priority, tiering, deny action, external entity, and policy statistics. For more information on usage of Antrea Network Policies, refer to the Antrea Network Policy document.

Egress

Antrea supports specifying which egress (SNAT) IP the traffic from the selected Pods to the external network should use and which Node the traffic should leave the cluster from. For more information, refer to the Egress document.

Network Flow Visibility

Antrea supports exporting network flow information using IPFIX, and provides a reference cookbook on how to visualize the exported network flows using Elastic Stack and Kibana dashboards. For more information, refer to the network flow visibility document.

NoEncap and Hybrid Traffic Modes

Besides the default Encap mode, in which Pod traffic across Nodes will be encapsulated and sent over tunnels, Antrea also supports NoEncap and Hybrid traffic modes. In NoEncap mode, Antrea does not encapsulate Pod traffic, but relies on the Node network to route the traffic across Nodes. In Hybrid mode, Antrea encapsulates Pod traffic when the source Node and the destination Node are in different subnets, but does not encapsulate when the source and the destination Nodes are in the same subnet. Refer to this guide to learn how to configure Antrea with NoEncap or Hybrid mode.

Antrea Web UI

Antrea comes with a web UI, which can show runtime information of Antrea components and perform Antrea Traceflow operations. Please refer to the Antrea UI repository for installation instructions and more information.

OVS Hardware Offload

Antrea can offload OVS flow processing to the NICs that support OVS kernel hardware offload using TC. The hardware offload can improve OVS performance significantly. For more information on how to configure OVS offload, refer to the OVS hardware offload guide.

Prometheus Metrics

Antrea supports exporting metrics to Prometheus. For more information, refer to the Prometheus integration document.

Support for Services of type LoadBalancer

By leveraging Antrea's Service external IP management feature or configuring MetalLB to work with Antrea, Services of type LoadBalancer can be supported without requiring an external LoadBalancer. To learn more information, please refer to the Service LoadBalancer document.

Traceflow

Traceflow is a very useful network diagnosis feature in Antrea. It can trace and report the forwarding path of a specified packet in the Antrea network. For usage of this feature, refer to the Traceflow user guide.

Traffic Encryption

Antrea supports encrypting traffic between Linux Nodes using IPsec or WireGuard. To deploy Antrea with traffic encryption enabled, please refer to this guide.

Antrea Multi-cluster

Antrea Multi-cluster implements Multi-cluster Service API, which allows users to create multi-cluster Services that can be accessed cross clusters in a ClusterSet. Antrea Multi-cluster also supports Antrea ClusterNetworkPolicy replication. Multi-cluster admins can define ClusterNetworkPolicies to be replicated across the entire ClusterSet, and enforced in all member clusters. To learn more information about Antrea Multi-cluster, please refer to the Antrea Multi-cluster user guide.