Upgrading CNFs

Overview

This document discusses how to upgrade the various Cloud-Native Network Functions (CNFs) software components. Each section represents the independent upgradable components.

_images/spk_warn.png Important: F5 recommends that you perform software upgrades during a planned maintenance window.

Requirements

Ensure you have:

BIG-IP Controller

The BIG-IP Controller installs using the f5ingress Helm chart, released as a versioned tarball, for example f5ingress-9.0.0.tgz. When extracted from the tarball, the f5ingress Helm chart also contains multiple subcharts that will in the future enable upgrading CNFs Pods independently. The table below lists the Pods and their charts or subcharts:

Charts/Subcharts

Chart/Subchart Pod Containers
f5ingress f5ingress f5ingress and f5-license-helper
f5ingress f5-tmm f5-tmm and f5-debug-sidecar
f5-afm f5-afm f5-l4p-engine and f5-fluentbit
f5-ipsd f5-ipds f5-nsec-ips-daemon and f5-fluentbit
f5-stats_collector otel-collector opentelemetry-collector

TMM RollingUpdate

When upgrading CNFs installations with multiple f5-tmm replicas (Pods) running, CNFS uses the Kubernetes RollingUpdate feature; terminating and upgrading one f5-tmm Pod at a time, leaving the remaining f5-tmm replicas available to process application traffic.

Full upgrade

Use these steps to upgrade the BIG-IP Controller and all of the Pods listed in the Charts/Subcharts table above.

  1. Change into the directory containing the latest CNFs Software, and list the f5ingress Helm chart:

    cd cnfinstall; ls -1 tar | grep f5ingress
    
    f5ingress-9.0.0.tgz
    
  2. Obtain the Helm release name for the current CNFs installation:

    In this example, the Helm release is in the cnf-gateway namespace.

    helm list -n cnf-gateway
    

    In this example, the Helm release is using CHART version f5ingress-8.0.0

    NAME        NAMESPACE     REVISION    STATUS     CHART
    f5ingress   cnf-gateway   1           deployed   f5ingress-8.0.0
    
  3. Use the new Helm chart to upgrade the installation:

    helm upgrade f5ingress tar/f5ingress-<version>.tgz \
    -f <values>.yaml -n namespace
    

    In this example, all of the CNFs Pods will be upgraded using the f5ingress-9.0.0.tgz Helm chart.

    helm upgrade f5ingress tar/f5ingress-9.0.0.tgz \
    -f values.yaml -n cnf-gateway
    
  4. Verify the Helm CHART and REVISION have incremented:

    helm list -n cnf-gateway
    

    In this example, the Helm CHART is now 9.0.0 and the REVISION is now 2.

    NAME         NAMESPACE      REVISION     STATUS     CHART
    f5ingress    cnf-gateway    2            deployed   f5ingress-9.0.0
    
  5. Verify the Pods have a STATUS of Running:

    kubectl get pods -n cnf-gateway
    

    In this example, the upgraded Pods have a STATUS of Running.

    NAME                                   READY   STATUS
    f5-afm-54d66946cd-5dmvs                2/2     Running
    f5-ipsd-54df45c9cf-tmvqv               2/2     Running
    f5-tmm-6fbbbcfb8-8jvkh                 4/4     Running
    f5ingress-f5ingress-597894b455-4hglb   2/2     Running
    

Partial upgrade

The ability to perform parital, or single CNFs Pod upgrades is not yet supported.

Cert Manager

The CNFs Cert Manager installs using the f5-cert-manager Helm chart, released as a versioned tarball, for example f5-cert-manager-1.0.0.tgz. The f5-cert-manager Helm chart manages the f5-cert-manager, f5-cert-manager-cainjector and f5-cert-manager-webhook Pods.

Use these steps to upgrade Cert Manager and the Pods mentioned above.

  1. Change into the directory containing the latest CNFs Software, and list the f5-cert-manager Helm chart:

    cd cnfinstall; ls -1 tar | grep f5ingress
    
    f5-cert-manager-2.0.0.tgz
    
  2. Obtain the Helm release name for the current CNFs installation:

    In this example, the Helm release is in the cnf-cert-manager namespace.

    helm list -n cnf-cert-manager
    

    In this example, the Helm release NAME is cnf-cert-manager and the CHART is version f5-cert-manager-1.0.0.

    NAME               NAMESPACE         REVISION    STATUS     CHART
    cnf-cert-manager   cnf-cert-manager  1           deployed   f5-cert-manager-1.0.0
    
  3. Use the new Helm chart to upgrade the installation:

    helm upgrade cnf-cert-manager tar/f5-cert-manager-<version>.tgz \
    -f <values>.yaml -n namespace
    

    In this example, the upgrade will use the f5-cert-manager-2.0.0.tgz Helm chart.

    helm upgrade f5-cert-manager tar/f5-cert-manager-2.0.0.tgz \
    -f values.yaml -n cnf-cert-manager
    
  4. Verify the Helm CHART and REVISION have incremented:

    helm list -n cnf-gateway
    

    In this example, the Helm CHART is now 2.0.0 and the REVISION is now 2.

    NAME               NAMESPACE         REVISION    STATUS     CHART
    cnf-cert-manager   cnf-cert-manager  2           deployed   f5-cert-manager-2.0.0
    
  5. Verify the Pods have a STATUS of Running:

    kubectl get pods -n cnf-cert-manager
    

    In this example, the upgraded Pods have a STATUS of Running.

    NAME                                          READY   STATUS  
    f5-cert-manager-5d7b7b456d-v85jc              2/2     Running
    f5-cert-manager-cainjector-848555556c-zsdt8   1/1     Running
    f5-cert-manager-webhook-75b59556b9-bvph8      1/1     Running
    

Partial upgrade

The ability to perform parital, or single CNFs Pod upgrades is not yet supported.

Cert Manager

CRDs

CRD upgrades should be performed after installing the latest BIG-IP Controller that came packaged with the CRD bundle. If the CRD bundle is newer than the BIG-IP Controller, it will not not recognize the newer CRD paramaters, and the expected configuration will not be applied to the Service Proxy TMM. Use these steps to upgrade the CRDs:

_images/spk_warn.png Important: Helm currently does not support upgrading or deleting CRDs. Refer to Helm Custom Resource Definitions.

Full upgrade

Use these steps to upgrade all of the CNFs CRDs.

  1. Change into the directory containing the latest CNFs Software, and list the CRD bundle:

    cd cnfinstall; ls -1 | grep crds
    
    f5-cnf-crds-n6lan-0.161.0-0.1.2.tgz
    
  2. Extract the Helm subcharts from the f5ingress tarball:

    tar xvf f5-cnf-crds-n6lan-0.161.0-0.1.2.tgz
    
  3. Install the CRDs:

    helm upgrade f5crds f5-cnf-crds-n6lan-0.161.0-0.1.2.tgz -n default
    

Partial upgrade

Use these steps to upgrade a single CNFs CRD.

  1. Change into the directory containing the latest CNFs Software, and list the CRD bundle:

    cd cnfinstall; ls -1 | grep crds
    
    f5-cnf-crds-n6lan-0.161.0-0.1.2.tgz
    
  2. Extract the Helm subcharts from the f5ingress tarball:

    tar xvf f5-cnf-crds-n6lan-0.161.0-0.1.2.tgz
    
  3. Install a specific CRD:

    Note: In this example, only the f5-big-nat-policy.yaml CRD is upgraded.

    kubectl apply -f f5-cnf-crds-n6lan/crds/f5-big-nat-policy.yaml
    

dSSM database

To ensure there is no loss of data or downtime, the dSSM Databases should be upgraded using the Upgrading dSSM guide.

Fluentd logging

Use these steps to upgrade the Fluentd Logging collector.

  1. Change into the directory containing the latest CNFs Software, and list the f5-fluentd Helm chart:

    cd cnfinstall; ls -1 tar | grep fluentd
    
    f5-toda-fluentd-1.23.36-0.0.1.tgz
    
  2. Obtain the Helm release name for the current CNFs installation:

    helm list -n cnf-gateway
    

    In this example, the release name is f5-fluentd.

    NAME         NAMESPACE     REVISION    STATUS     CHART
    f5-fluentd   cnf-gateway   1           deployed   f5-toda-fluentd-1.23.36-0.0.0
    
  3. Use the new Helm chart to upgrade the installation:

    helm upgrade f5-fluentd tar/f5-toda-fluentd-<version>.tgz \ 
    -f <values>.yaml -n cnf-gateway
    

    In this example, the Fluend Pod will be upgraded using the f5-toda-fluentd-1.23.36-0.0.1.tgz Helm chart.

    helm upgrade f5-fluentd tar/f5-toda-fluentd-1.23.36-0.0.1.tgz \
    -f fluentd-values.yaml -n cnf-gateway
    
  4. Verify the Helm CHART and REVISION have both incremented:

    helm list -n cnf-gateway
    

    In this example, the Helm CHART is now 1.23.36-0.0.1 and the REVISION is now 2.

    NAME         NAMESPACE      REVISION    STATUS     CHART
    f5-fluentd   cnf-gateway    2           deployed   f5-toda-fluentd-1.23.36-0.0.1
    
  5. Verify the Fluentd Pod is Running:

    kubectl get pods -n cnf-gateway | grep fluent
    

    In this example, the Fluentd Pod is Running.

    f5-toda-fluentd-5c4876d88f-99n94   1/1     Running
    

Feedback

Provide feedback to improve this document by emailing cnfdocs@f5.com.