BIG-IP Next for Kubernetes (DPU) 2.2.0 release notes

About this release

Introduction

These release notes describe the differences between BIG-IP Next for Kubernetes (DPU) 2.1.0 and 2.2.0.

Supported container-orchestration platforms

This release is officially supported on these container-orchestration platforms and versions:

Platform

Version

Recommended Helm version

Kubernetes

1.30.10

3.18

Kubernetes

1.30.14

3.18

Security updates

(No security updates)

New features

Dynamically manage log levels via REST API and ConfigMap watching

This new feature allows runtime log level control without pod restarts. CWC now exposes a REST API to adjust log levels per pod. Requests update a pod-specific ConfigMap, which propagates via minlevel.yaml; the F5 logging library monitors this file and applies the new level automatically. For more information, see Dynamic log levels and API Documentation.

Automated orphaned log folder cleanup

Added Folder Cleaner, an optional feature that automatically removes orphaned log directories to prevent storage exhaustion. Cleanup runs as a scheduled job (CronJob) inside the container. For more information, see Fluentd logging.

Enhancements

Added support for 1 GB huge pages

You can now configure the huge page size to 1Gi using the new tmm.hugepages.preferredhugepagesize parameter while defining the TMM resources. For more information, see Manual CPU Allocation.

Cancelling QKView Requests by ID

Cancel a running QKView via API; on cancellation, the system stops collection and generates a QKView that includes only data gathered before the cancel request. For more information, see QKView and API Documentation.

Configuring per-table aggregation

You can now use aggregation mode to control metrics for a table. For each table, you can choose the Aggregation mode options (Aggregated, Semi Aggregated, or Diagnostic) and set an export interval. The operator enforces these settings via Jobs from the f5-observer-operator-config ConfigMap. For more information, see Aggregation mode section in Distributed Toda for Stats Aggregation and Metric table.

Model changes

  • CR name changes:

    • BNKGatewayClass CR has been renamed to CNEInstance CR.

    • Similarly, BNKManifest is now CNEmanifest.

  • Image repository field rename:

    • In CNEInstance, the image under spec has been replaced with registry. Furthermore:

      • image.repository is now registry.uri.

      • All other sub-fields under registry remain unchanged.

  • Removed fields in CNEInstance:

    • The ContainerPlatform field has been removed from CNEInstance. It is now moved to Helm values.

    • The sharedComponentNamespace field has also been moved to Helm values.

  • Several new fields have been added to CNEInstance:

    • ProductType is now a mandatory field in CNEInstance.

    • New optional modules have been introduced to extend functionality. These components must be explicitly enabled in the CNEInstance spec by configuring the relevant enablement fields.

      • IPSD: F5-IPSD provides deep packet inspection by validating configurations, compiling SNORT-compatible signature blobs, and sending them to TMM for enforcement. Installation CR enablement field: intrusionPrevention.enabled.

      • f5-analyzer: The analyzer pod analyzes and routes LLM-related traffic, enabling dynamic routing based on complexity and load, primarily for DPU installations. Installation CR enablement field: intelligentLB.enabled.

      • f5-dwbld: It encodes IP address feedlists into binary blobs for TMM to perform real-time IP reputation checks and traffic decisions. Installation CR enablement field: ipIntelligence.enabled.

      • f5-downloader: Downloader assists other pods by downloading and storing large files from external sources for shared access. Installation CR enablement field: None (enabled indirectly).

      • CNF and SPK CRDs: Additional CRDs for CNF and SPK are now installed based on the form factor specified in the installation CR.

  • New field in F5 Lifecycle Operator values:

    • The ServiceIPFamily field has been added to the F5 Lifecycle Operator Helm values to support IP version details for the cluster (primarily used for Fluentd configuration).

  • Container name changes:

    • f5-cne-controller container name changed from f5-cne-controller-<cr_name>-<namespace> to f5-cne-controller.

    • f5-spk-csrc container name is now spk-csrc.

  • Helm values updates:

    • Helm values now require a license in “connected” mode to function properly. These license details must be provided under the license section of the Helm values file.

    • The following license-specific fields are required:

      operationMode: "connected"
      logLevel: "info"
      jwt: <token>
      friendlyName: "My Cluster"
      teemCertUrl: https://product-tst.apis.f5networks.net/ee/v1
      teemEntitlementUrl: https://product-s-tst.apis.f5networks.net/ee/v1
      teemInitialConfigUrl: https://product-s-tst.apis.f5networks.net/ee/v1
      kid: "v1"
      alg: "RS512"
      kty: "RSA"
      modulus: <token>
      exponent: "AQAB"
      x5c: <token>
      use: <token>
      licenseserverrootca: <token>
      

Fixed issues

(No fixed issues)

Known issues

1926869-2

Performance may be degraded when using PVA acceleration to apps on the same chassis as TMM.

Component: FSM

Symptoms: When running BIG-IP Next for Kubernetes on a Bluefield-3 DPU, and enabling PVA Acceleration of flows, traffic destined for apps running on the host in the same chassis as the DPU with TMM may see up to 50% less throughput when compared to traffic destined for apps on other nodes in the cluster.

Conditions: PVA acceleration enabled, BIG-IP Next for Kubernetes running on Bluefield-3 DPU, Apps running on host CPU are in the same chassis as TMM.

Impact: Up to 50% degradation in throughput.

1968153-3

Traffic stats missing drop counter for trunk use cases.

Component: FSM

Symptoms: Traffic stats are not present when packet are dropped when using a trunk interface.

Conditions: The trunk does not have any interfaces to forward traffic.

Impact: Missing diagnostics.

2138129-1

Fluent-bit ARM64 image hits unsupported page size issues on ARM64 AKS cluster.

Component: Toda_fluentbit

Symptoms: When configuring a Linux kernel with the page size for 64K page size: CONFIG_ARM64_64K_PAGES=y, Fluentbit fails to start with an error “Unsupported system page size”.

Conditions: When configuring a Linux kernel with the page size for 64K page size: CONFIG_ARM64_64K_PAGES=y.

Impact: Fluentbit fails to start with an error “Unsupported system page size”.

1823977-2

Logs for TMM container is unavailable through console output when fluentbit container is enabled in TMM pod.

Component: FSM

Symptoms: When the fluentbit container is enabled for the f5-tmm pod, the f5-tmm container logs will not be outputted to the console. However, the f5-tmm container logs can still be found in the f5-toda-fluentd pod (located under “/var/log/f5”).

Conditions: Fluentbit container is enabled and running in the f5-tmm pod.

Impact: F5-tmm container logs will not be outputted through console and can only be found in the f5-toda-fluentd pod.

Workaround: The f5-tmm container logs can still be found in the f5-toda-fluentd pod (located under “/var/log/f5”).

2186317-1

Hardware-accelerated connection is not counted in DDoS Profile stats.

Component: DoS

Symptoms: Hardware-accelerated TCP/UDP connection flow packets may not be counted in DDoS Profile stats.

Conditions:

  • Hardware acceleration is enabled for context.

  • Profile DDoS is assigned to context.

  • Volumetric DDoS attack happening via accelerated connections (only TCP/UDP).

Impact: Volumetric Flood TCP/UDP DDoS attack may not be mitigated.

Workaround: Use Global DDoS to mitigate such attacks or disable hardware connection flow acceleration for context.

2046925-1

TCP throughput degradation during UDP DDoS attack with DPU hardware acceleration.

Component: DoS

Symptoms: Lower TCP throughput degradation could be observed during ongoing UDP DDoS attack on DPU platforms even when Global DDoS mitigation is configured.

Conditions:

  • Global DDoS config has TCP and UDP vectors configured.

  • DDoS hardware DPU acceleration is enabled.

  • TCP and UDP traffic is going in the same time.

Impact: Up to 30% degradation in throughput.

2186773-1

DNS Nxdomain DoS Detection & Mitigation Supported only on Global Context.

Component: DoS

Symptoms: Most DDoS Vectors are supported on both Global level & Per Context level. DNS Nxdomain DoS detection & mitigation feature is supported only on Global Context. However the Per Context DDoS CRD wrongly includes sub-section “dnsNxdomainQuery” defined, which is not supported.

Conditions: Per context DDoS CRD wrongly included sub-section “dnsNxdomainQuery”. However this configuration section is not applicable.

Impact: Including dnsNxdomainQuery vector in Per Context DDoS CRD, is misleading.

Workaround: Use Global DDoS dnsNxdomainQuery vector. Per Context level DoS mitigation is not supported for DNS Nxdomain Vector.

2200517-1

Hardware acceleration for IPv6 is not supported.

Component: FSM

Symptoms: On the CX7 NIC, BIG-IP Next for Kubernetes hardware acceleration is not supported for IPv6 traffic. Only IPv4 traffic offloading is available, including IPv4 flow offloading, IPv4 NAT44 offloading, and ACL offloading.

Conditions: BIG-IP Next for Kubernetes running on the CX7 NIC.

Impact: IPv6 traffic is not optimized for acceleration.

Upgrade considerations

Reinstall BIG-IP Next for Kubernetes from 2.1.0 to 2.2.0

This release introduces significant enhancements and new features to provide better performance. Direct upgrade from 2.1.0 to 2.2.0 is not supported. You must uninstall 2.1.0 and perform a clean installation of 2.2.0.

Before you install version 2.2.0, make a full backup of your current BIG-IP Next for Kubernetes (DPU) 2.1.0 setup. Back up all configurations, Helm values, and sensitive data. This will help you restore your setup if there are problems during the new installation.

Note

Perform the tasks in order.

Configuration Backup

This section covers the process of creating a backup configuration files. The backup process captures your existing system configurations and exports these configurations as YAML files to a designated network location.

  1. Access the Kubernetes cluster.

  2. Export all F5 CRs to YAML files.

    for crd in $(kubectl get crds | grep f5 | awk '{print $1}'); do
    resource=$(echo $crd | cut -d. -f1)
    kubectl get $resource -A -o yaml > backup_${resource}.yaml
    done
    
  3. Export Gateway API CRs.

    kubectl get gatewayclasses,gateways,httproutes,grpcroutes -A -o yaml > backup_gateway_api.yaml
    
  4. Export Helm values.

    helm get values flo -n f5-cne-core > backup_flo_values.yaml
    
  5. Export FAR pull secrets.

    kubectl get secret far-pull-secret -n f5-cne-core -o yaml > backup_far_pull_secret_cne_core.yaml
    kubectl get secret far-pull-secret -n <cneInstanceNamespace> -o yaml > backup_far_pull_secret_<cneInstanceNamespace>.yaml
    
  6. Export TLS certificates.

    kubectl get certificates -A -o yaml > backup_certificates_all_namespaces.yaml
    

    Now, the configuration backup files are created. Ensure these backup files are stored securely.

Uninstall BIG-IP Next for Kubernetes (DPU) 2.1.0

  1. Access the Kubernetes cluster.

  2. Delete BNKGatewayClass custom resource (bnkgatewayclasscr.yaml).

    $ kubectl delete -f bnkgatewayclasscr.yaml --namespace <cneInstanceNamespace>
    
  3. Verify that no pods are running in the CNE-instance namespace.

    $ kubectl get pods -n <cneInstanceNamespace>
    
  4. Uninstall F5 Lifecycle Operator (FLO).

    $ helm uninstall flo --namespace <floNamespace>
    
  5. Verify FLO was uninstalled.

    $ kubectl get pods -n <floNamespace>
    

    The BIG-IP Next for Kubernetes (DPU) 2.1.0 is now uninstalled.

Install BIG-IP Next for Kubernetes (DPU) 2.2.0