This section describes how to upgrade NCP to 2.4.0 in a Kubernetes environment.

Procedure

  1. Upgrade NCP ReplicationController with the following command (replace <image> with the actual name of the image).
    kubectl rolling-update nsx-ncp -n nsx-system --image=<image>
  2. Upgrade NSX Node Agent daemonSet with the following commands (replace <image> with the actual name of the image).
    kubectl set image ds nsx-node-agent -n nsx-system nsx-node-agent=<image>
    kubectl set image ds nsx-node-agent -n nsx-system nsx-kube-proxy=<images>
    kubectl rollout status ds/nsx-node-agent -nnsx-system
  3. Upgrade CNI DEB/RPM package to 2.4.0 with the following command (replace <cni deb> and <cni rpm> with the actual name of the package).
    On Ubuntu:

    dkpg -i <cni deb>

    On RHEL or CentOS:

    rpm -U <cni rpm>

  4. (Optional) Upgrade NSX-T Data Center to 2.4.
    If the hypervisor is ESXi, upgrade it to at least 6.5p03 from 6.5 or 6.7ep6 from 6.7 before upgrading NSX-T Data Center.
  5. (Optional) Upgrade the hypervisor (KVM or bare metal container).
  6. (Optional) Upgrade the container host (RHEL, Ubuntu, or CentOS).
  7. (Optional) Upgrade Kubernetes.
  8. (Optional) Upgrade OVS.
    For bare metal container, upgrading NSX-T Data Center also upgrades OVS, so this step is not necessary.

    During this step, you might see transient communication failures between nsx-kube-proxy and nsx-node-agent. This is expected behavior and does not indicate a problem.