| apiVersion: k8s.cni.cncf.io/v1 | |
| kind: NetworkAttachmentDefinition | |
| metadata: | |
| name: sriov-net-vlan100 | |
| namespace: default | |
| annotations: | |
| k8s.v1.cni.cncf.io/resourceName: kubevirt.io/sriov_net | |
| spec: | |
| config: | | |
| { |
| #!/bin/bash | |
| set -ex | |
| CRI=${CRI:-podman} | |
| kind="./kind" | |
| ! [ -f $kind ] && \ | |
| wget -nv "https://github.com/kubernetes-sigs/kind/releases/download/v0.11.1/kind-linux-amd64" -O $kind && \ | |
| chmod +x $kind |
| #!/bin/bash | |
| set -ex | |
| CRI=${CRI:-podman} | |
| # download kind | |
| kind="./kind" | |
| ! [ -f $kind ] && \ | |
| curl -L "https://github.com/kubernetes-sigs/kind/releases/download/v0.11.1/kind-linux-amd64" -o $kind && \ |
| #! /bin/bash | |
| set -ex | |
| function cleanup() { | |
| set +e | |
| ip netns exec blue ip link del dev veth-blue | |
| ip netns exec red ip link del dev veth-red | |
| ip netns del blue | |
| ip netns del red |
| # Author: Or Mergi <[email protected]> | |
| # | |
| # This script runs a demonstration to reproduce the following issue: | |
| # https://github.com/k8snetworkplumbingwg/sriov-cni/issues/219 | |
| # | |
| # Warnings: | |
| # * This script creates pods and re-configure you cluster node | |
| # SR-IOV Physical Function (PF) interface and its Virtual Functions (VF). | |
| # * Do not run this on production cluster. | |
| # * Make sure no VF is attached to any workload on the node you picked. |
| # This script enable faster development iterations in Kubevirt | |
| # when working on a change that might affect backward compatibility, | |
| # upgrades or legacy VMs (a VM that existed before an upgrade) disruption. | |
| # | |
| # Use case: | |
| # 1. Changing virt-launcher pods secondary networks interfaces name scheme. | |
| # Different network interfaces name-schemes breaks migration for legacy VMs | |
| # when Kubevirt is updated. | |
| # | |
| # In order to test this, its necessary to have both versions images, |
| # /bin/bash | |
| set -ex | |
| vm_name="myvmi" | |
| vm_pod_name=$(kubectl get po -o custom-columns=:.metadata.name --no-headers | grep $vm_name | grep Running) | |
| vm_node_name=$(kubectl get po $vm_pod_name -o custom-columns=:.spec.nodeName --no-headers) | |
| vm_node_virt_handler_pod=$(kubectl get pod -n kubevirt -l kubevirt.io=virt-handler -o wide --no-headers | grep "$vm_node_name" | awk '{print $1}') | |
| vm_pod_pid=$(kubectl exec -it $vm_node_virt_handler_pod -n kubevirt -c virt-handler -- bash -c "ps -ef | grep /usr/bin/virt-launcher.*--name.*$vm_name" | grep -Pv "grep|virt-launcher-monitor" | awk '{print $2}') |
| # Author: Or Mergi | |
| # | |
| # This script build basic container image to deploy Kubevirt passt binding CNI. | |
| # | |
| # | |
| #!/bin/bash | |
| curl -LO https://github.com/kubevirt/kubevirt/releases/download/v1.2.0/kubevirt-passt-binding | |
| chmod +x kubevirt-passt-binding | |
| cat <<EOF > Containerfile |
| --- | |
| apiVersion: v1 | |
| kind: ServiceAccount | |
| metadata: | |
| name: passt-binding-cni | |
| namespace: kube-system | |
| --- | |
| apiVersion: security.openshift.io/v1 | |
| kind: SecurityContextConstraints | |
| metadata: |
There are multiple solutions for connecting k8s cluster over BGP, this doc focuses on using OVN-Kubernetes FRR-k8s and the integration between them.
OVN-Kubernetes serves as the cluster network platform. FRR-k8s install FRR instance in each cluster node running BGP server.
OVN-Kubernetes support exchanging routes of the cluster networks using FRR-k8s.