Detalhes da rede, do namespace e do servidor virtual

A tabela a seguir resume o exemplo de layout de rede.

Diagrama da configuração final

Cada bloco de código do guia pode ser copiado para um arquivo e aplicado com o comando oc apply -f .yml``. Alguns blocos de código podem ultrapassar os limites das páginas nos formatos PDF e Word, e algumas linhas longas podem quebrar nesses formatos.

O guia utiliza “ ic ” como abreviação de “ ibmcloud ” nos comandos da CLI do IBM Cloud. Defina alias ic=ibmcloud no seu shell ou substitua ibmcloud por ic nos comandos a seguir.

Criação de endereços IP reservados em cada sub-rede

Primeiro, crie um endereço IP reservado em cada sub-rede para seus servidores virtuais:

ic is subnet-reserved-ip-create zone1-vm-prod --address 198.18.0.20 --name zone1-web00 --auto-delete false
ic is subnet-reserved-ip-create zone1-vm-db --address 198.18.1.20 --name zone1-db00 --auto-delete false
ic is subnet-reserved-ip-create zone1-vm-app --address 198.18.2.20 --name zone1-app00 --auto-delete false

Criação de VNIs usando os endereços IP reservados

Agora, crie as VNIs usando os nomes de IP reservados na etapa anterior:

ic is virtual-network-interface-create --name zone1-web00 --allow-ip-spoofing false --auto-delete false --protocol-state-filtering-mode disabled --enable-infrastructure-nat true --subnet zone1-vm-prod --vpc my-vpc --rip zone1-web00 --sgs handclap-roundworm-clique-eccentric --resource-group-name Default
ic is virtual-network-interface-create --name zone1-db00 --allow-ip-spoofing false --auto-delete false --protocol-state-filtering-mode disabled --enable-infrastructure-nat true --subnet zone1-vm-db --vpc my-vpc --rip zone1-db00 --sgs handclap-roundworm-clique-eccentric --resource-group-name Default
ic is virtual-network-interface-create --name zone1-app00 --allow-ip-spoofing false --auto-delete false --protocol-state-filtering-mode disabled --enable-infrastructure-nat true --subnet zone1-vm-app --vpc my-vpc --rip zone1-app00 --sgs handclap-roundworm-clique-eccentric --resource-group-name Default

Servidor virtual plant-web00 no namespace green em vlan20-prod

# Bare YAML to stand up a Red Hat OpenShift on IBM Cloud virtual machine
---
apiVersion: kubevirt.io/v1
kind: VirtualMachine
metadata:
  name: "plant-web00"
  namespace: "green"
  annotations:
    description: "example vm plant-web00"
  labels:
    app: "plant-web00-green-server"
    kubevirt.io/dynamic-credentials-support: 'true'
    vm.kubevirt.io/template: "centos-stream9-server-small"
    vm.kubevirt.io/template.namespace: openshift
    vm.kubevirt.io/template.revision: '1'
    vm.kubevirt.io/template.version: v0.34.0
spec:
  dataVolumeTemplates:
    - apiVersion: cdi.kubevirt.io/v1beta1
      kind: DataVolume
      metadata:
        creationTimestamp: null
        name: "plant-web00-green-server"
      spec:
        sourceRef:
          kind: DataSource
          name: "centos-stream9"
          namespace: openshift-virtualization-os-images
        storage:
          resources:
            requests:
              storage: 30Gi
  runStrategy: RerunOnFailure
  template:
    metadata:
      annotations:
        vm.kubevirt.io/flavor: "small"
        vm.kubevirt.io/os: "centos-stream9"
        vm.kubevirt.io/workload: "server"
      labels:
        kubevirt.io/domain: example
        kubevirt.io/size: "small"
    spec:
      domain:
        cpu:
          cores: 1
          sockets: 1
          threads: 1
        devices:
          disks:
            - disk:
                bus: virtio
              name: rootdisk
            - disk:
                bus: virtio
              name: cloudinitdisk
          interfaces:
            - model: virtio
              name: default
              state: up
              bridge: {}
              macAddress: 02:00:01:00:98:A0 # <- replace with the MAC of VNI zone1-web00
          networkInterfaceMultiqueue: true
          rng: {}
        memory:
          guest: "2Gi"
      hostname: "plant-web00"
      networks:
        - name: default
          multus:
            networkName: vlan20-prod
      terminationGracePeriodSeconds: 180
      volumes:
        - dataVolume:
            name: "plant-web00-green-server"
          name: rootdisk
        - cloudInitNoCloud:
            userData: |-
              #cloud-config
              user: admin
              password: "aRandomPassword9292"
              ssh_authorized_keys:
                 - "ssh-ed25519 AAAA....BBBB....CCCCC.....DDDD put-your-key-here@example.local"
              chpasswd: { expire: False }
              runcmd:
                - [ dnf, install, -y, epel-release ]
                - [ dnf, install, -y, iperf3, nmap-ncat, darkhttpd ]
          name: cloudinitdisk

Servidor virtual plant-db00 no namespace green em vlan21-db

---
apiVersion: kubevirt.io/v1
kind: VirtualMachine
metadata:
  name: "plant-db00"
  namespace: "green"
  annotations:
    description: "example vm plant-db00"
  labels:
    app: "plant-db00-green-server"
    kubevirt.io/dynamic-credentials-support: 'true'
    vm.kubevirt.io/template: "centos-stream9-server-small"
    vm.kubevirt.io/template.namespace: openshift
    vm.kubevirt.io/template.revision: '1'
    vm.kubevirt.io/template.version: v0.34.0
spec:
  dataVolumeTemplates:
    - apiVersion: cdi.kubevirt.io/v1beta1
      kind: DataVolume
      metadata:
        creationTimestamp: null
        name: "plant-db00-green-server"
      spec:
        sourceRef:
          kind: DataSource
          name: "centos-stream9"
          namespace: openshift-virtualization-os-images
        storage:
          resources:
            requests:
              storage: 30Gi
  runStrategy: RerunOnFailure
  template:
    metadata:
      annotations:
        vm.kubevirt.io/flavor: "small"
        vm.kubevirt.io/os: "centos-stream9"
        vm.kubevirt.io/workload: "server"
      labels:
        kubevirt.io/domain: example
        kubevirt.io/size: "small"
    spec:
      domain:
        cpu:
          cores: 1
          sockets: 1
          threads: 1
        devices:
          disks:
            - disk:
                bus: virtio
              name: rootdisk
            - disk:
                bus: virtio
              name: cloudinitdisk
          interfaces:
            - model: virtio
              name: default
              state: up
              bridge: {}
              macAddress: 02:00:01:00:98:A5 # <- replace with the MAC of VNI zone1-db00
          networkInterfaceMultiqueue: true
          rng: {}
        memory:
          guest: "2Gi"
      hostname: "plant-db00"
      networks:
        - name: default
          multus:
            networkName: vlan21-db
      terminationGracePeriodSeconds: 180
      volumes:
        - dataVolume:
            name: "plant-db00-green-server"
          name: rootdisk
        - cloudInitNoCloud:
            userData: |-
              #cloud-config
              user: admin
              password: "aRandomPassword9292"
              ssh_authorized_keys:
                 - "ssh-ed25519 AAAA....BBBB....CCCCC.....DDDD put-your-key-here@example.local"
              chpasswd: { expire: False }
              runcmd:
                - [ dnf, install, -y, epel-release ]
                - [ dnf, install, -y, iperf3, nmap-ncat, darkhttpd ]
          name: cloudinitdisk

Servidor virtual plant-app00 no namespace red em vlan22-app

---
apiVersion: kubevirt.io/v1
kind: VirtualMachine
metadata:
  name: "plant-app00"
  namespace: "red"
  annotations:
    description: "example vm plant-app00"
  labels:
    app: "plant-app00-red-server"
    kubevirt.io/dynamic-credentials-support: 'true'
    vm.kubevirt.io/template: "centos-stream9-server-small"
    vm.kubevirt.io/template.namespace: openshift
    vm.kubevirt.io/template.revision: '1'
    vm.kubevirt.io/template.version: v0.34.0
spec:
  dataVolumeTemplates:
    - apiVersion: cdi.kubevirt.io/v1beta1
      kind: DataVolume
      metadata:
        creationTimestamp: null
        name: "plant-app00-red-server"
      spec:
        sourceRef:
          kind: DataSource
          name: "centos-stream9"
          namespace: openshift-virtualization-os-images
        storage:
          resources:
            requests:
              storage: 30Gi
  runStrategy: RerunOnFailure
  template:
    metadata:
      annotations:
        vm.kubevirt.io/flavor: "small"
        vm.kubevirt.io/os: "centos-stream9"
        vm.kubevirt.io/workload: "server"
      labels:
        kubevirt.io/domain: example
        kubevirt.io/size: "small"
    spec:
      domain:
        cpu:
          cores: 1
          sockets: 1
          threads: 1
        devices:
          disks:
            - disk:
                bus: virtio
              name: rootdisk
            - disk:
                bus: virtio
              name: cloudinitdisk
          interfaces:
            - model: virtio
              name: default
              state: up
              bridge: {}
              macAddress: 02:00:01:00:98:AA # <- replace with the MAC of VNI zone1-app00
          networkInterfaceMultiqueue: true
          rng: {}
        memory:
          guest: "2Gi"
      hostname: "plant-app00"
      networks:
        - name: default
          multus:
            networkName: vlan22-app
      terminationGracePeriodSeconds: 180
      volumes:
        - dataVolume:
            name: "plant-app00-red-server"
          name: rootdisk
        - cloudInitNoCloud:
            userData: |-
              #cloud-config
              user: admin
              password: "aRandomPassword9292"
              ssh_authorized_keys:
                 - "ssh-ed25519 AAAA....BBBB....CCCCC.....DDDD put-your-key-here@example.local"
              chpasswd: { expire: False }
              runcmd:
                - [ dnf, install, -y, epel-release ]
                - [ dnf, install, -y, iperf3, nmap-ncat, darkhttpd ]
          name: cloudinitdisk

Testes

Execute os testes a seguir para verificar a configuração da rede.

1. Inicie um servidor simples do tipo “ HTTP ” em cada servidor virtual.

   • No site 198.18.0.20, acesse darkhttpd /usr/share/doc/bash --daemon.
   • No site 198.18.1.20, acesse darkhttpd /usr/share/doc/bash --daemon.
   • No site 198.18.2.20, acesse darkhttpd /usr/share/doc/bash --daemon.

2. No site 198.18.0.20, acesse curl -v http://198.18.1.20:8080.

3. No site 198.18.1.20, acesse curl -v http://198.18.0.20:8080.

4. No servidor virtual “jump”, execute curl -v http://198.18.2.20:8080.

5. No site 198.18.1.20, inicie um servidor iperf3: iperf3 -s -p 9090.

6. No 198.18.0.20, conecte-se ao servidor iperf3 no namespace green: iperf3 -c 198.18.1.20 -p 9090.

7. Faça um ping de 198.18.1.20 para 198.18.0.20 e vice-versa.

8. Execute novamente as tarefas anteriores usando 198.18.2.20 e o endereço IP do servidor virtual jump.

Aplicação de uma política de múltiplas redes e repetição dos testes

Aplique uma política de múltiplas redes que permita apenas o acesso TCP na porta 8080 para a rede vlan20-prod. Salve o manifesto a seguir em um arquivo chamado my-policy.yml. Em seguida, execute o comando oc apply -f my-policy.yml para aplicá-lo.

---
apiVersion: k8s.ovn.org/v1alpha1
kind: MultiNetworkPolicy
metadata:
  name: allow-8080-tcp
  namespace: green
  annotations:
    k8s.v1.cni.cncf.io/policy-for: vlan20-prod
spec:
  podSelector: {}
  ingress:
    - ports:
        - protocol: TCP
          port: 8080
  egress:
    - ports:
        - protocol: TCP
          port: 8080
  policyTypes:
    - Ingress
    - Egress

Execute novamente os testes no endereço plant-web00, disponível em vlan20-prod. TCP A porta 8080 ainda funciona, mas outros tipos de tráfego, como ping e o teste iperf3 na porta 9090, não funcionam mais nessa conexão de rede.

Execute o comando oc delete -f my-policy.yml para remover a política e restaurar a conectividade. Os testes foram aprovados novamente.