RDG for DPF Host Trusted with OVN-Kubernetes and HBN Services

Created on December 24, 2025

Scope

This Reference Deployment Guide (RDG) provides detailed instructions for deploying a Kubernetes (K8s) cluster using NVIDIA® BlueField®-3 DPUs and DOCA Platform Framework (DPF) in Host Trusted mode. The guide covers setting up accelerated OVN-Kubernetes, Host-Based Networking (HBN) services, and additional services on NVIDIA® BlueField®-3 DPUs.

As a reference implementation, this guide focuses on using open-source components and outlines the entire deployment process, including bare metal and virtual machine provisioning with KVM virtualization and MaaS. It also addresses performance tuning to achieve optimal results.

Leveraging NVIDIA's DPF, administrators can provision and manage DPU resources within a Kubernetes cluster while deploying and orchestrating HBN and accelerated OVN-Kubernetes services. This approach enables full utilization of NVIDIA DPU hardware acceleration and offloading capabilities, maximizing data center workload efficiency and performance.

This guide is designed for experienced system administrators, system engineers, and solution architects who seek to deploy high-performance Kubernetes clusters and enable NVIDIA BlueField DPUs.

This reference implementation, as the name implies, is a specific, opinionated deployment example designed to address the use case described above.
While other approaches may exist to implement similar solutions, this document provides a detailed guide for this particular method.

Abbreviations and Acronyms

Term	Definition	Term	Definition
BFB	BlueField Bootstream	K8S	Kubernetes
BGP	Border Gateway Protocol	MAAS	Metal as a Service
CNI	Container Network Interface	OVN	Open Virtual Network
CSI	Container Storage Interface	RDG	Reference Deployment Guide
DOCA	Data Center Infrastructure-on-a-Chip Architecture	RDMA	Remote Direct Memory Access
DPF	DOCA Platform Framework	SFC	Service Function Chaining
DPU	Data Processing Unit	SR-IOV	Single Root Input/Output Virtualization
DTS	DOCA Telemetry Service	TOR	Top of Rack
GENEVE	Generic Network Virtualization Encapsulation	VLAN	Virtual LAN (Local Area Network)
HBN	Host Based Networking	VRR	Virtual Router Redundancy
IPAM	IP Address Management	VTEP	Virtual Tunnel End Point

Introduction

The NVIDIA BlueField-3 data processing unit (DPU) is a 400 Gb/s infrastructure compute platform designed for line-rate processing of software-defined networking, storage, and cybersecurity. BlueField-3 combines powerful computing, high-speed networking, and extensive programmability to deliver hardware-accelerated, software-defined solutions for demanding workloads.

NVIDIA DOCA unlocks the full potential of the NVIDIA BlueField platform, enabling rapid development of applications and services that offload, accelerate, and isolate data center workloads.

Host-based Networking (HBN) is a DOCA service that allows network architects to design networks based on layer-3 (L3) protocols. HBN enables routing to run on the server side by using BlueField as a BGP router. The HBN solution encapsulates a set of network functions inside a container, which is deployed as a service pod on BlueField's Arm cores.

OVN-Kubernetes is a Kubernetes CNI network plugin that provides robust networking for Kubernetes clusters. Built on Open Virtual Network (OVN) and Open vSwitch (OVS), it supports hardware acceleration to offload OVS packet processing to NIC/DPU hardware. With OVS-DOCA, an extension of traditional OVS-DPDK and OVS-Kernel, accelerated OVN-Kubernetes delivers industry-leading performance, functionality, and efficiency. Running OVN-Kubernetes on the DPU reserves host CPUs exclusively for workloads, maximizing system resources.

Deploying and managing DPUs and their associated DOCA services—especially at scale—can be challenging. Without a provisioning and orchestration system, the complexity of managing the DPU lifecycle, deploying DOCA services, and providing the necessary network configuration on the DPU to redirect the network traffic via those services (service function chaining, or SFC) becomes a significant burden for cluster and system administrators; which is where the DOCA Platform Framework (DPF) comes into play.

DPF simplifies DPU management by providing orchestration through a Kubernetes API. It handles the provisioning and lifecycle management of DPUs, orchestrates specialized DPU services, and automates tasks such as service function chaining (SFC). This ensures seamless deployment of DOCA services like OVN-Kubernetes and HBN, allowing traffic to be efficiently offloaded and routed through HBN's data plane.

With DPF, users can efficiently manage and scale DPUs within their clusters while automating critical processes. DPF orchestrates the deployment of OVN-Kubernetes and HBN, optimizing performance with features such as offloaded OVN-Kubernetes CNI and accelerated traffic routing through HBN.

This RDG provides a comprehensive, practical example of installing the DPF system on a Kubernetes cluster. It also demonstrates performance optimizations, including Jumbo frame implementation, with results validated through standard RDMA and TCP workload tests.

References

Solution Architecture

Key Components and Technologies

NVIDIA BlueField® Data Processing Unit (DPU)
The NVIDIA® BlueField® data processing unit (DPU) ignites unprecedented innovation for modern data centers and supercomputing clusters. With its robust compute power and integrated software-defined hardware accelerators for networking, storage, and security, BlueField creates a secure and accelerated infrastructure for any workload in any environment, ushering in a new era of accelerated computing and AI.

NVIDIA DOCA Software Framework
NVIDIA DOCA™ unlocks the potential of the NVIDIA® BlueField® networking platform. By harnessing the power of BlueField DPUs and SuperNICs, DOCA enables the rapid creation of applications and services that offload, accelerate, and isolate data center workloads. It lets developers create software-defined, cloud-native, DPU- and SuperNIC-accelerated services with zero-trust protection, addressing the performance and security demands of modern data centers.

NVIDIA ConnectX SmartNICs
10/25/40/50/100/200 and 400G Ethernet Network Adapters
The industry-leading NVIDIA® ConnectX® family of smart network interface cards (SmartNICs) offer advanced hardware offloads and accelerations.
NVIDIA Ethernet adapters enable the highest ROI and lowest Total Cost of Ownership for hyperscale, public and private clouds, storage, machine learning, AI, big data, and telco platforms.

NVIDIA LinkX Cables
The NVIDIA® LinkX® product family of cables and transceivers provides the industry’s most complete line of 10, 25, 40, 50, 100, 200, and 400GbE in Ethernet and 100, 200 and 400Gb/s InfiniBand products for Cloud, HPC, hyperscale, Enterprise, telco, storage and artificial intelligence, data center applications.

NVIDIA Spectrum Ethernet Switches
Flexible form-factors with 16 to 128 physical ports, supporting 1GbE through 400GbE speeds.
Based on a ground-breaking silicon technology optimized for performance and scalability, NVIDIA Spectrum switches are ideal for building high-performance, cost-effective, and efficient Cloud Data Center Networks, Ethernet Storage Fabric, and Deep Learning Interconnects.
NVIDIA combines the benefits of NVIDIA Spectrum^™ switches, based on an industry-leading application-specific integrated circuit (ASIC) technology, with a wide variety of modern network operating system choices, including NVIDIA Cumulus^® Linux, SONiC and NVIDIA Onyx^®.

NVIDIA Cumulus Linux
NVIDIA® Cumulus® Linux is the industry's most innovative open network operating system that allows you to automate, customize, and scale your data center network like no other.

NVIDIA Network Operator
The NVIDIA Network Operator simplifies the provisioning and management of NVIDIA networking resources in a Kubernetes cluster. The operator automatically installs the required host networking software - bringing together all the needed components to provide high-speed network connectivity. These components include the NVIDIA networking driver, Kubernetes device plugin, CNI plugins, IP address management (IPAM) plugin and others. The NVIDIA Network Operator works in conjunction with the NVIDIA GPU Operator to deliver high-throughput, low-latency networking for scale-out, GPU computing clusters.

Kubernetes
Kubernetes is an open-source container orchestration platform for deployment automation, scaling, and management of containerized applications.

Kubespray
Kubespray is a composition of Ansible playbooks, inventory, provisioning tools, and domain knowledge for generic OS/Kubernetes clusters configuration management tasks and provides:A highly available clusterComposable attributesSupport for most popular Linux distributions

OVN-Kubernetes
OVN-Kubernetes (Open Virtual Networking - Kubernetes) is an open-source project that provides a robust networking solution for Kubernetes clusters with OVN (Open Virtual Networking) and Open vSwitch (Open Virtual Switch) at its core. It is a Kubernetes networking conformant plugin written according to the CNI (Container Network Interface) specifications.

RDMA
RDMA is a technology that allows computers in a network to exchange data without involving the processor, cache or operating system of either computer.
Like locally based DMA, RDMA improves throughput and performance and frees up compute resources.

Solution Design

Solution Logical Design

The logical design includes the following components:

1 x Hypervisor node (KVM based) with ConnectX-7
- 1 x Firewall VM
- 1 x Jump VM
- 1 x MaaS VM
- 3 x K8s Master VMs running all K8s management components
2 x Worker nodes (PCI Gen5), each with a 1 x BlueField-3 NIC
Single High-Speed (HS) switch, 1 x L3 HS underlay network
1 Gb Host Management network

K8s Cluster Logical Design

The following K8s logical design illustration demonstrates the main components of the DPF system, among them:

3 x K8s Master VMs running all K8s management components
2 x K8s Worker nodes (x86)
2 x K8s DPU Workers running DOCA services (OVN-K8s, HBN, DTS, BlueMan)
1 x Kamaji (K8s Control-Plane Manager)
1 x DPU Control Plane (Tenant Cluster)
Connectivity to High-Speed/1Gb networks

Firewall Design

The pfSense firewall in this solution serves a dual purpose:

Firewall – provides an isolated environment for the DPF system, ensuring secure operations
Router – enables internet access and connectivity between the host management network and the high-speed network

Port-forwarding rules for SSH and RDP are configured on the firewall to route traffic to the jump node’s IP address in the host management network. From the jump node, administrators can manage and access various devices in the setup, as well as handle the deployment of the Kubernetes (K8s) cluster and DPF components.

The following diagram illustrates the firewall design used in this solution:

Software Stack Components

Make sure to use the exact same versions for the software stack as described above.

Bill of Materials

Deployment and Configuration

Node and Switch Definitions

These are the definitions and parameters used for deploying the demonstrated fabric:

Switch Port Usage
Hostname	Rack ID	Ports
`hs-switch`	1	swp1,11-14
`mgmt-switch`	1	swp1-3

Hosts
Rack	Server Type	Server Name	Switch Port	IP and NICs	Default Gateway
Rack1	Hypervisor Node	`hypervisor`	mgmt-switch: `swp1` hs-switch: `swp1`	lab-br (interface eno1): Trusted LAN IP mgmt-br (interface eno2): - hs-br (interface ens2f0np0): -	Trusted LAN GW
Rack1	Worker Node	`worker1`	mgmt-switch: `swp2` hs-switch: `swp11`-`swp12`	ens15f0: 10.0.110.21/24 ens5f0np0/ens5f1np1: 10.0.120.0/22	10.0.110.254
Rack1	Worker Node	`worker2`	mgmt-switch: `swp3` hs-switch: `swp13`-`swp14`	ens15f0: 10.0.110.22/24 ens5f0np0/ens5f1np1: 10.0.120.0/22	10.0.110.254
Rack1	Firewall (Virtual)	`fw`	-	WAN (lab-br): Trusted LAN IP LAN (mgmt-br): 10.0.110.254/24 OPT1 (hs-br): 172.169.50.1/30	Trusted LAN GW
Rack1	Jump Node (Virtual)	`jump`	-	enp1s0: 10.0.110.253/24	10.0.110.254
Rack1	MaaS (Virtual)	`maas`	-	enp1s0: 10.0.110.252/24	10.0.110.254
Rack1	Master Node (Virtual)	`master1`	-	enp1s0: 10.0.110.1/24	10.0.110.254
Rack1	Master Node (Virtual)	`master2`	-	enp1s0: 10.0.110.2/24	10.0.110.254
Rack1	Master Node (Virtual)	`master3`	-	enp1s0: 10.0.110.3/24	10.0.110.254

Wiring

Hypervisor Node

K8s Worker Node

Fabric Configuration

Updating Cumulus Linux

As a best practice, make sure to use the latest released Cumulus Linux NOS version.

For information on how to upgrade Cumulus Linux, refer to the Cumulus Linux User Guide.

Configuring the Cumulus Linux Switch

For the SN3700 switch (hs-switch), is configured as follows:

The following commands configure BGP unnumbered on hs-switch.
Cumulus Linux enables the BGP equal-cost multipathing (ECMP) option by default.

SN3700 Switch Console

nv set interface lo ipv4 address 11.0.0.101/32
nv set interface lo type loopback
nv set interface swp1 ipv4 address 172.169.50.2/30
nv set interface swp1,11-14 link state up
nv set interface swp1,11-14 type swp
nv set router bgp autonomous-system 65001
nv set router bgp state enabled
nv set router bgp graceful-restart mode full
nv set router bgp router-id 11.0.0.101
nv set vrf default router bgp address-family ipv4-unicast state enabled
nv set vrf default router bgp address-family ipv4-unicast redistribute connected state enabled
nv set vrf default router bgp address-family ipv4-unicast redistribute static state enabled
nv set vrf default router bgp address-family ipv6-unicast state enabled
nv set vrf default router bgp address-family ipv6-unicast redistribute connected state enabled
nv set vrf default router bgp state enabled
nv set vrf default router bgp neighbor swp11-14 peer-group hbn
nv set vrf default router bgp neighbor swp11-14 type unnumbered
nv set vrf default router bgp path-selection multipath aspath-ignore enabled
nv set vrf default router bgp peer-group hbn remote-as external
nv set vrf default router static 0.0.0.0/0 address-family ipv4-unicast
nv set vrf default router static 0.0.0.0/0 via 172.169.50.1 type ipv4-address
nv set vrf default router static 10.0.110.0/24 address-family ipv4-unicast
nv set vrf default router static 10.0.110.0/24 via 172.169.50.1 type ipv4-address
nv config apply -y

The SN2201 switch (mgmt-switch) is configured as follows:

SN2201 Switch Console

nv set bridge domain br_default untagged 1
nv set interface swp1-3 link state up
nv set interface swp1-3 type swp
nv set interface swp1-3 bridge domain br_default
nv config apply -y

Host Configuration

Make sure that the BIOS settings on the worker node servers have SR-IOV enabled and that the servers are tuned for maximum performance.

All worker nodes must have the same PCIe placement for the BlueField-3 NIC and must show the same interface name.

Hypervisor Installation and Configuration

The hypervisor used in this Reference Deployment Guide (RDG) is based on Ubuntu 24.04 with KVM.

While this document does not detail the KVM installation process, it is important to note that the setup requires the following ISOs to deploy the Firewall, Jump, and MaaS virtual machines (VMs):

Ubuntu 24.04
pfSense-CE-2.7.2

To implement the solution, three Linux bridges must be created on the hypervisor:

Ensure a DHCP record is configured for the lab-br bridge interface in your trusted LAN to assign it an IP address.

lab-br – connects the Firewall VM to the trusted LAN.
mgmt-br – Connects the various VMs to the host management network.
hs-br – Connects the Firewall VM to the high-speed network.

Additionally, an MTU of 9000 must be configured on the management and high-speed bridges (mgmt-br and hs-br) as well as their uplink interfaces to ensure optimal performance.

Hypervisor netplan configuration

YAML

network:
    ethernets:
        eno1:
            dhcp4: false
        eno2:
            dhcp4: false
            mtu: 9000
        ens2f0np0:
            dhcp4: false
            mtu: 9000
    bridges:
      lab-br:
         interfaces: [eno1]
         dhcp4: true
      mgmt-br:
         interfaces: [eno2]
         dhcp4: false
         mtu: 9000
      hs-br:
         interfaces: [ens2f0np0]
         dhcp4: false
         mtu: 9000
    version: 2

Apply the configuration:

Hypervisor Console

sudo netplan apply

Prepare Infrastructure Servers

Firewall VM - pfSense Installation and Interface Configuration

Download the pfSense CE (Community Edition) ISO to your hypervisor and proceed with the software installation.

Suggested spec:

vCPU: 2
RAM: 2GB
Storage: 10GB
Network interfaces
- Bridge device connected to lab-br
- Bridge device connected to mgmt-br
- Bridge device connected to hs-br

The Firewall VM must be connected to all three Linux bridges on the hypervisor. Before beginning the installation, ensure that three virtual network interfaces of type "Bridge device" are configured. Each interface should be connected to a different bridge (lab-br, mgmt-br, and hs-br) as illustrated in the diagram below.

After completing the installation, the setup wizard displays a menu with several options, such as "Assign Interfaces" and "Reboot System." During this phase, you must configure the network interfaces for the Firewall VM.

Select Option 2: "Set interface(s) IP address" and configure the interfaces as follows:
- WAN (lab-br) – Trusted LAN IP (Static/DHCP)
- LAN (mgmt-br) – Static IP 10.0.110.254/24
- OPT1 (hs-br) – Static IP 172.169.50.1/30
Once the interface configuration is complete, use a web browser within the host management network to access the Firewall web interface and finalize the configuration.

Next, proceed with the installation of the Jump VM. This VM will serve as a platform for running a browser to access the Firewall’s web interface for post-installation configuration.

Jump VM

Suggested specifications:

vCPU: 4
RAM: 8GB
Storage: 25GB
Network interface: Bridge device, connected to mgmt-br

Procedure:

Proceed with a standard Ubuntu 24.04 installation. Use the following login credentials across all hosts in this setup:

Username	Password
depuser	user

Enable internet connectivity and DNS resolution by creating the following Netplan configuration:

Use 10.0.110.254 as a temporary DNS nameserver until the MaaS VM is installed and configured. After completing the MaaS installation, update the Netplan file to replace this address with the MaaS IP: 10.0.110.252.

Jump Node netplan
YAML
```
network:
    ethernets:
        enp1s0:
            dhcp4: false
            addresses: [10.0.110.253/24]
            nameservers:
              search: [dpf.rdg.local.domain]
              addresses: [10.0.110.254]
            routes:
              - to: default
                via: 10.0.110.254
    version: 2
```
Apply the configuration:

Jump Node Console
```
depuser@jump:~$ sudo netplan apply 
```
Update and upgrade the system:

Jump Node Console
```
sudo apt update -y
sudo apt upgrade -y
```

Install and configure the Xfce desktop environment and XRDP (complementary packages for RDP):

Jump Node Console

sudo apt install -y xfce4 xfce4-goodies
sudo apt install -y lightdm-gtk-greeter
sudo apt install -y xrdp
echo "xfce4-session" | tee .xsession
sudo systemctl restart xrdp

Install Firefox for accessing the Firewall web interface:

Jump Node Console
```
sudo apt install -y firefox
```

Install and configure an NFS server with the /mnt/dpf_share directory:

Jump Node Console

sudo apt install -y nfs-server
sudo mkdir -m 777 /mnt/dpf_share
sudo vi /etc/exports

Add the following line to /etc/exports:

Jump Node Console

/mnt/dpf_share 10.0.110.0/24(rw,sync,no_subtree_check)

Restart the NFS server:

Jump Node Console
```
sudo systemctl restart nfs-server
```
Create the directory bfb under /mnt/dpf_share with the same permissions as the parent directory:

Jump Node Console
```
sudo mkdir -m 777 /mnt/dpf_share/bfb
```
Generate an SSH key pair for depuser in the jump node (later on will be imported to the admin user in MaaS to enable password-less login to the provisioned servers):

Jump Node Console
```
depuser@jump:~$ ssh-keygen -t rsa
```
Reboot the jump node to display the graphical user interface:

Jump Node Console
```
sudo reboot
```
After setting up port-forwarding rules on the firewall (next steps), remote login to the graphical interface of the Jump node will be available.

Concurrent login to the local graphical console and using RDP isn't possible, make sure to first log out from the local console when switching to RDP connection.

Firewall VM – Web Configuration

From your Jump node, open Firefox web browser and go to the pfSense web UI (http://10.0.110.254, default credentials are admin/pfsense). You should see a page similar to the following:

The IP addresses from the trusted LAN network under "DNS servers" and "Interfaces - WAN" are blurred.

Proceed with the following configurations:

The following screenshots display only a part of the configuration view. Make sure to not miss any of the steps mentioned below!

Interfaces
- WAN – mark “Enable interface”, unmark “Block private networks and loopback addresses”
- LAN – mark “Enable interface”, “IPv4 configuration type”: Static IPv4 ("IPv4 Address": 10.0.110.254/24, "IPv4 Upstream Gateway": None), “MTU”: 9000
- OPT1 – mark “Enable interface”, “IPv4 configuration type”: Static IPv4 ("IPv4 Address": 172.169.50.1/30, "IPv4 Upstream Gateway": None), “MTU”: 9000
Firewall:
- NAT -> Port Forward -> Add rule -> “Interface”: WAN, “Address Family”: IPv4, “Protocol”: TCP, “Destination”: WAN address, “Destination port range”: (“From port”: SSH, “To port”: SSH), “Redirect target IP”: (“Type”: Address or Alias, “Address”: 10.0.110.253), “Redirect target port”: SSH, “Description”: NAT SSH
- NAT -> Port Forward -> Add rule -> “Interface”: WAN, “Address Family”: IPv4, “Protocol”: TCP, “Destination”: WAN address, “Destination port range”: (“From port”: MS RDP, “To port”: MS RDP), “Redirect target IP”: (“Type”: Address or Alias, “Address”: 10.0.110.253), “Redirect target port”: MS RDP, “Description”: NAT RDP

- Rules -> OPT1 -> Add rule -> “Action”: Pass, “Interface”: OPT1, “Address Family”: IPv4+IPv6, “Protocol”: Any, “Source”: Any, “Destination”: Any

System:
- Routing → Gateways → Add → “Interface”: OPT1, “Address Family”: IPv4, “Name”: switch, “Gateway”: 172.169.50.2 → Click "Save"→ Under "Default Gateway" - "Default gateway IPv4" choose WAN_DHCP → Click "Save"
  
  Note that the IP addresses from the Trusted LAN network under "Gateway" and "Monitor IP" are blurred.

- Routing → Static Routes → Add → “Destination network”: 10.0.120.0/22, “Gateway”: switch – 172.169.50.2, “Description”: To HS network → Click "Save"

MaaS VM

Suggested specifications:

vCPU: 4
RAM: 4GB
Storage: 50GB
Network interface: Bridge device, connected to mgmt-br

Procedure:

Perform a regular Ubuntu installation on the MaaS VM.

Create the following Netplan configuration to enable internet connectivity and DNS resolution:

Use 10.0.110.254 as a temporary DNS nameserver. After the MaaS installation, replace this with the MaaS IP address (10.0.110.252) in both the Jump and MaaS VM Netplan files.

MaaS netplan

YAML

network:
    ethernets:
        enp1s0:
            dhcp4: false
            addresses: [10.0.110.252/24]
            nameservers:
              search: [dpf.rdg.local.domain]
              addresses: [10.0.110.254]
            routes:
              - to: default
                via: 10.0.110.254
    version: 2

Apply the netplan configuration:

MaaS Console
```
depuser@maas:~$ sudo netplan apply 
```
Update and upgrade the system:

MaaS Console
```
sudo apt update -y
sudo apt upgrade -y
```

Install PostgreSQL and configure the database for MaaS:

MaaS Console

$ sudo -i
# apt install -y postgresql
# systemctl enable --now postgresql
# systemctl disable --now systemd-timesyncd
# export MAAS_DBUSER=maasuser
# export MAAS_DBPASS=maaspass
# export MAAS_DBNAME=maas
# sudo -i -u postgres psql -c "CREATE USER \"$MAAS_DBUSER\" WITH ENCRYPTED PASSWORD '$MAAS_DBPASS'"
# sudo -i -u postgres createdb -O "$MAAS_DBUSER" "$MAAS_DBNAME"

Install MaaS:

MaaS Console
```
snap install maas
```

Initialize MaaS:

MaaS Console

maas init region+rack --maas-url http://10.0.110.252:5240/MAAS --database-uri "postgres://$MAAS_DBUSER:$MAAS_DBPASS@localhost/$MAAS_DBNAME"

Create an admin account:

MaaS Console

maas createadmin --username admin --password admin --email admin@example.com

Save the admin API key:

MaaS Console

maas apikey --username admin > admin-apikey

MaaS Console

maas login admin http://localhost:5240/MAAS "$(cat admin-apikey)"

Configure MaaS (Substitute <Trusted_LAN_NTP_IP> and <Trusted_LAN_DNS_IP> with the IP addresses in your environment):

MaaS Console

maas admin domain update maas name="dpf.rdg.local.domain"
maas admin maas set-config name=ntp_servers value="<Trusted_LAN_NTP_IP>"
maas admin maas set-config name=network_discovery value="disabled"
maas admin maas set-config name=upstream_dns value="<Trusted_LAN_DNS_IP>"
maas admin maas set-config name=dnssec_validation value="no"
maas admin maas set-config name=default_osystem value="ubuntu"

Define and configure IP ranges and subnets:

MaaS Console

maas admin ipranges create type=dynamic start_ip="10.0.110.51" end_ip="10.0.110.120"
maas admin ipranges create type=reserved start_ip="10.0.110.10" end_ip="10.0.110.10" comment="c-plane VIP"
maas admin ipranges create type=reserved start_ip="10.0.110.200" end_ip="10.0.110.200" comment="kamaji VIP"
maas admin ipranges create type=reserved start_ip="10.0.110.251" end_ip="10.0.110.254" comment="dpfmgmt"
maas admin vlan update 0 untagged dhcp_on=True primary_rack=maas mtu=9000
maas admin dnsresources create fqdn=kube-vip.dpf.rdg.local.domain ip_addresses=10.0.110.10
maas admin dnsresources create fqdn=jump.dpf.rdg.local.domain ip_addresses=10.0.110.253
maas admin dnsresources create fqdn=fw.dpf.rdg.local.domain ip_addresses=10.0.110.254

Configure static DHCP leases for the worker nodes (replace MAC address as appropriate with your workers MGMT interface MAC):

MaaS Console

maas admin reserved-ips create ip="10.0.110.21" mac_address="04:32:01:60:0d:da" comment="worker1"
maas admin reserved-ips create ip="10.0.110.22" mac_address="04:32:01:5f:cb:e0" comment="worker2"

Complete MaaS setup:
1. Connect to the Jump node GUI and access the MaaS UI at http://10.0.110.252:5240/MAAS.
2. On the first page, verify the "Region Name" and "DNS Forwarder," then continue.
3. On the image selection page, verify that Ubuntu 24.04 LTS (amd64) image is synced and continue.
4. Import the previously generated SSH key (id_rsa.pub) for the depuser into the MaaS admin user profile and finalize the setup.
Update the DNS nameserver IP address in both Jump and MaaS VM Netplan files from 10.0.110.254 to 10.0.110.252 and reapply the configuration.

K8s Master VMs

Suggested specifications:

vCPU: 8
RAM: 16GB
Storage: 100GB
Network interface: Bridge device, connected to mgmt-br

Before provisioning the Kubernetes (K8s) Master VMs with MaaS, create the required virtual disks with empty storage. Use the following one-liner to create three 100 GB QCOW2 virtual disks:

Hypervisor Console
```
for i in $(seq 1 3); do qemu-img create -f qcow2 /var/lib/libvirt/images/master$i.qcow2 100G; done
```
This command generates the following disks in the /var/lib/libvirt/images/ directory:
- master1.qcow2
- master2.qcow2
- master3.qcow2
Configure VMs in virt-manager:
1. Open virt-manager and create three virtual machines:
  - Assign the corresponding virtual disk (master1.qcow2, master2.qcow2, or master3.qcow2) to each VM.
  - Configure each VM with the suggested specifications (vCPU, RAM, storage, and network interface).
2. During the VM setup, ensure the NIC is selected under the Boot Options tab. This ensures the VMs can PXE boot for MaaS provisioning.
3. Once the configuration is complete, shut down all the VMs.
After the VMs are created and configured, proceed to provision them via the MaaS interface. MaaS will handle the OS installation and further setup as part of the deployment process.

Provision Master VMs and Worker Nodes Using MaaS

Master VMs

Install `virsh` and Set Up SSH Access

SSH to the MaaS VM from the Jump node:

MaaS Console

depuser@jump:~$ ssh maas
depuser@maas:~$ sudo -i

Install the virsh client to communicate with the hypervisor:

MaaS Console
```
# apt install -y libvirt-clients
```
Generate an SSH key for the root user and copy it to the hypervisor user in the libvirtd group:

MaaS Console
```
# ssh-keygen -t rsa
# ssh-copy-id ubuntu@<hypervisor_MGMT_IP>
```

Verify SSH access and virsh communication with the hypervisor:

MaaS Console

# virsh -c qemu+ssh://ubuntu@<hypervisor_MGMT_IP>/system list --all

Expected output:

MaaS Console

 Id   Name          State
------------------------------
 1    fw     running
 2    jump   running
 3    maas   running
 -    master1       shut off
 -    master2       shut off
 -    master3       shut off

Copy the SSH key to the required MaaS directory (for snap-based installations):

MaaS Console

# mkdir -p /var/snap/maas/current/root/.ssh
# cp .ssh/id_rsa* /var/snap/maas/current/root/.ssh/

Get MAC Addresses of the Master VMs

Retrieve the MAC addresses of the Master VMs:

MaaS Console

# for i in $(seq 1 3); do virsh -c qemu+ssh://ubuntu@<hypervisor_MGMT_IP>/system dumpxml master$i | grep 'mac address'; done

Example output:

MaaS Console

<mac address='52:54:00:a9:9c:ef'/>
<mac address='52:54:00:19:6b:4d'/>
<mac address='52:54:00:68:39:7f'/>

Add Master VMs to MaaS

Add the Master VMs to MaaS:

Once added, MaaS will automatically start the newly added VMs commissioning (discovery and introspection).

MaaS Console

# maas admin machines create hostname=master1 architecture=amd64/generic mac_addresses='52:54:00:a9:9c:ef' power_type=virsh power_parameters_power_address=qemu+ssh://ubuntu@<hypervisor_MGMT_IP>/system power_parameters_power_id=master1 skip_bmc_config=1 testing_scripts=none
Success.
Machine-readable output follows:
{
    "description": "",
    "status_name": "Commissioning",
...
    "status": 1, 
...
    "system_id": "c3seyq",
...
    "fqdn": "master1.dpf.rdg.local.domain",
    "power_type": "virsh",
...
    "status_message": "Commissioning",
    "resource_uri": "/MAAS/api/2.0/machines/c3seyq/"
}

# maas admin machines create hostname=master2 architecture=amd64/generic mac_addresses='52:54:00:19:6b:4d' power_type=virsh power_parameters_power_address=qemu+ssh://ubuntu@<hypervisor_MGMT_IP>/system power_parameters_power_id=master2 skip_bmc_config=1 testing_scripts=none

# maas admin machines create hostname=master3 architecture=amd64/generic mac_addresses='52:54:00:68:39:7f' power_type=virsh power_parameters_power_address=qemu+ssh://ubuntu@<hypervisor_MGMT_IP>/system power_parameters_power_id=master3 skip_bmc_config=1 testing_scripts=none

Repeat the command for master2 and master3 with their respective MAC addresses.

Verify commissioning by waiting for the status to change to "Ready" in MaaS.
After commissioning, the next phase is the deployment (OS provisioning).

Configure OVS Bridges on Master VMs

To be able to have persistency across reboots, create an OVS-bridge from each management interface of the master nodes and assign it a static IP address.

For each Master VM:

Create an OVS bridge in the MaaS Network tab:
1. Navigate to Network → Management Interface → Create Bridge.
2. Configure as follows:
  1. Name: brenp1s0 (prefix br added to the interface name)
  2. Bridge Type: Open vSwitch (ovs)
  3. Subnet: 10.0.110.0/24
  4. IP Mode: Static (Client configured)
  5. Address: Assign 10.0.110.1 for master1, 10.0.110.2 for master2, and 10.0.110.3 for master3.
Save the interface settings for each VM.

Deploy Master VMs Using Cloud-Init

Use the following cloud-init script to configure the necessary software and ensure OVS bridge persistency:

Replace enp1s0 and brenp1s0 in the following cloud-init with your interface names as displayed in MaaS network tab.

Master nodes cloud-init

YAML

#cloud-config
system_info:
  default_user:
    name: depuser
    passwd: "$6$jOKPZPHD9XbG72lJ$evCabLvy1GEZ5OR1Rrece3NhWpZ2CnS0E3fu5P1VcZgcRO37e4es9gmriyh14b8Jx8gmGwHAJxs3ZEjB0s0kn/"
    lock_passwd: false
    groups: [adm, audio, cdrom, dialout, dip, floppy, lxd, netdev, plugdev, sudo, video]
    sudo: ["ALL=(ALL) NOPASSWD:ALL"]
    shell: /bin/bash
ssh_pwauth: True
package_upgrade: true
runcmd:
    - apt-get update
    - apt-get -y install openvswitch-switch nfs-common
    - |
      UPLINK_MAC=$(cat /sys/class/net/enp1s0/address)
      ovs-vsctl set Bridge brenp1s0 other-config:hwaddr=$UPLINK_MAC
      ovs-vsctl br-set-external-id brenp1s0 bridge-id brenp1s0 -- br-set-external-id brenp1s0 bridge-uplink enp1s0

Deploy the master VMs:
1. Select all three Master VMs → Actions → Deploy.
2. Toggle Cloud-init user-data and paste the cloud-init script.
3. Start the deployment and wait for the status to change to "Ubuntu 24.04 LTS".

Verify Deployment

SSH into the Master VMs from the Jump node:

Jump Node Console
```
depuser@jump:~$ ssh master1
depuser@master1:~$
```

Run sudo without password:

Master1 Console

depuser@master1:~$ sudo -i
root@master1:~#

Verify installed packages:

Master1 Console

root@master1:~# apt list --installed | egrep 'openvswitch-switch|nfs-common'
nfs-common/noble-updates,now 1:2.6.4-3ubuntu5.1 amd64 [installed]
openvswitch-switch/noble-updates,now 3.3.4-0ubuntu0.24.04.1 amd64 [installed]

Check OVS bridge attributes:

Master1 Console

root@master1:~# ovs-vsctl list bridge brenp1s0

Output example:

Master1 Console

...
external_ids        : {bridge-id=brenp1s0, bridge-uplink=enp1s0, netplan="true", "netplan/global/set-fail-mode"=standalone, "netplan/mcast_snooping_enable"="false", "netplan/rstp_enable"="false"}
...
other_config        : {hwaddr="52:54:00:a9:9c:ef"}
...

Verify that enp1s0 and brenp1s0 are configured with 9000 MTU (replace enp1s0 and brenp1s0 with your interface names):

Master1 Console

root@master1:~# ip a show enp1s0; ip a show brenp1s0
2: enp1s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc pfifo_fast master ovs-system state UP group default qlen 1000
    link/ether 52:54:00:a9:9c:ef brd ff:ff:ff:ff:ff:ff
4: brenp1s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UNKNOWN group default qlen 1000
    link/ether 52:54:00:a9:9c:ef brd ff:ff:ff:ff:ff:ff
    inet 10.0.110.1/24 brd 10.0.110.255 scope global brenp1s0
       valid_lft forever preferred_lft forever
    inet6 fe80::5054:ff:fea9:9cef/64 scope link
       valid_lft forever preferred_lft forever

Finalize Setup

Reboot the Master VMs to complete the provisioning.

Master1 Console

root@master1:~# reboot

Worker Nodes

Create Worker Machines in MaaS

Add the worker nodes to MaaS using ipmi as the power type. Replace placeholders with your specific IPMI credentials and IP addresses:

MaaS Console

# maas admin machines create hostname=worker1 architecture=amd64 power_type=ipmi power_parameters_power_driver=LAN_2_0 power_parameters_power_user=<IPMI_username_worker1> power_parameters_power_pass=<IPMI_password_worker1> power_parameters_power_address=<IPMI_address_worker1> power_parameters_workaround_flags=opensesspriv power_parameters_workaround_flags=nochecksumcheck

Output example:

MaaS Console

...
Success.
Machine-readable output follows:
{
    "description": "",
    "status_name": "Commissioning",
...
    "status": 1,
...
    "system_id": "pbskd3",
...
    "fqdn": "worker1.dpf.rdg.local.domain",
...
    "power_type": "ipmi",
...
    "resource_uri": "/MAAS/api/2.0/machines/pbskd3/"
}

Repeat the command for worker2 with its respective credentials:

MaaS Console

# maas admin machines create hostname=worker2 architecture=amd64 power_type=ipmi power_parameters_power_driver=LAN_2_0 power_parameters_power_user=<IPMI_username_worker2> power_parameters_power_pass=<IPMI_password_worker2> power_parameters_power_address=<IPMI_address_worker2> power_parameters_workaround_flags=opensesspriv power_parameters_workaround_flags=nochecksumcheck

Once added, MaaS will automatically start commissioning the worker nodes (discovery and introspection).

Create a Tag for Kernel Parameters

Create an entity called "Tag" to configure kernel parameters for the worker nodes.

In the MaaS UI sidebar, go to Organization → Tags → Create New Tag and define
- "Tag name": compute_performance
- "Kernel options":
Substitute the values for isolcpus, nohz_full, and rcu_nocbs to the CPU cores in the NUMA node which the BlueField-3 is connected to:

If you are not sure in which NUMA node BlueField is connected to, select the worker node in the Machines tab, go to Network settings and check the value under TYPE NUMA NODE.

Kernel options for worker nodes
```
intel_iommu=on iommu=pt numa_balancing=disable processor.max_cstate=0 isolcpus=28-55,84-111 nohz_full=28-55,84-111 rcu_nocbs=28-55,84-111
```
Apply the tag:
1. Go to Machines → Select a worker node → Configuration → Edit Tag → Select compute_performance → Save.
2. Repeat for the other worker node.

Adjust Network Settings

For each worker node, configure the network interfaces:

Management Adapter:
- Go to Network → Select the host management adapter (e.g., ens15f0) → Create Bridge
- Name: br-dpu
- Bridge Type: Standard
- Subnet: 10.0.110.0/24
- IP Mode: Dynamic
- Save the interface

Repeat these steps for the second worker node.

Deploy Worker Nodes Using Cloud-Init

Use the following cloud-init script for deployment:

Worker node cloud-init

YAML

#cloud-config
system_info:
  default_user:
    name: depuser
    passwd: "$6$jOKPZPHD9XbG72lJ$evCabLvy1GEZ5OR1Rrece3NhWpZ2CnS0E3fu5P1VcZgcRO37e4es9gmriyh14b8Jx8gmGwHAJxs3ZEjB0s0kn/"
    lock_passwd: false
    groups: [adm, audio, cdrom, dialout, dip, floppy, lxd, netdev, plugdev, sudo, video]
    sudo: ["ALL=(ALL) NOPASSWD:ALL"]
    shell: /bin/bash
ssh_pwauth: True
package_upgrade: true
runcmd:
  - apt-get update
  - apt-get -y install nfs-common

Deploy the worker nodes by selecting the worker nodes in MaaS → Actions → Deploy → Customize options → Enable Cloud-init user-data → Paste the cloud-init script → Deploy.

Verify Deployment

After the deployment is complete verify that the worker nodes have been deployed successfully with the following commands:

SSH without password from the jump node:

Jump Node Console
```
depuser@jump:~$ ssh worker1
depuser@worker1:~$
```

Run sudo without password:

Worker1 Console

depuser@worker1:~$ sudo -i
root@worker1:~#

Validate that nfs-common package was installed:

Worker1 Console

root@worker1:~# apt list --installed | grep 'nfs-common'
nfs-common/noble-updates,now 1:2.6.4-3ubuntu5.1 amd64 [installed]

/proc/cmdline is configured with the correct parameters and that IOMMU is indeed in passthrough mode:

Worker1 Console

root@worker1:~# cat /proc/cmdline
BOOT_IMAGE=/boot/vmlinuz-6.8.0-90-generic root=UUID=d2365b16-d371-4503-a583-a1768dd27e0c ro intel_iommu=on iommu=pt numa_balancing=disable processor.max_cstate=0 isolcpus=28-55,84-111 nohz_full=28-55,84-111 rcu_nocbs=28-55,84-111

root@worker1:~# dmesg | grep 'type: Passthrough'
[    5.033173] iommu: Default domain type: Passthrough (set via kernel command line)

ens15f0 and br-dpu are with 9000 MTU (replace ens15f0 with your interface name):

Worker1 Console

root@worker1:~# ip a show ens15f0; ip a show br-dpu
2: ens15f0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc mq master br-dpu state UP group default qlen 1000
    link/ether 04:32:01:60:0d:da brd ff:ff:ff:ff:ff:ff
    altname enp53s0f0
8: br-dpu: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UP group default qlen 1000
    link/ether 04:32:01:60:0d:da brd ff:ff:ff:ff:ff:ff
    inet 10.0.110.21/24 metric 100 brd 10.0.110.255 scope global dynamic br-dpu
       valid_lft 403sec preferred_lft 403sec
    inet6 fe80::632:1ff:fe60:dda/64 scope link
       valid_lft forever preferred_lft forever

Finalize Deployment

Reboot the worker nodes:

Jump Node Console

root@worker1:~# reboot

The infrastructure is now ready for the K8s deployment.

K8s Cluster Deployment and Configuration

Kubespray Deployment and Configuration

In this solution, the Kubernetes (K8s) cluster is deployed using a modified Kubespray (based on release v2.28.1) with a non-root depuser account from the Jump Node. The modifications in Kubespray are designed to meet the DPF prerequisites as described in the User Manual and facilitate cluster deployment and scaling.

Download the modified Kubespray archive: modified_kubespray_v2.28.1.tar.gz.

Extract the contents and navigate to the extracted directory:

Jump Node Console

$ tar -xzf /home/depuser/modified_kubespray_v2.28.1.tar.gz
$ cd kubespray/
depuser@jump:~/kubespray$

Set the K8s API VIP address and DNS record. Replace it with your own IP address and DNS record if different:

Jump Node Console

depuser@jump:~/kubespray$ sed -i '/# kube_vip_address:/s/.*/kube_vip_address: 10.0.110.10/' inventory/mycluster/group_vars/k8s_cluster/addons.yml
depuser@jump:~/kubespray$ sed -i '/apiserver_loadbalancer_domain_name:/s/.*/apiserver_loadbalancer_domain_name: "kube-vip.dpf.rdg.local.domain"/' roles/kubespray_defaults/defaults/main/main.yml

Install the necessary dependencies and set up the Python virtual environment:

Jump Node Console

depuser@jump:~/kubespray$ sudo apt -y install python3-pip jq python3.12-venv
depuser@jump:~/kubespray$ python3 -m venv .venv
depuser@jump:~/kubespray$ source .venv/bin/activate
(.venv) depuser@jump:~/kubespray$ python3 -m pip install --upgrade pip
(.venv) depuser@jump:~/kubespray$ pip install -U -r requirements.txt

Review and edit the inventory/mycluster/hosts.yaml file to define the cluster nodes. The following is the configuration for this deployment:

All of the nodes are already labeled and annotated as per DPF user manual prerequisites.
The worker nodes include additional kubelet configuration which will be applied during their deployment to achieve best performance, allowing:
- Containers in Guaranteed pods with integer CPU requests access to exclusive CPUs on the node.
- Reserve some cores for the system using the reservedSystemCPUs option (kubelet requires a CPU reservation greater than zero to be made when the static policy is enabled), and make sure they belong to NUMA 0 (because the NIC in the example is wired to NUMA node 1, use cores from NUMA 1 if the NIC is wired to NUMA node 0).
- Define the topology to be single-numa-node so it only allows a pod to be admitted if all requested CPUs and devices can be allocated from exactly one NUMA node.
The kube_node hosts worker1 and worker2 are marked with # to only deploy the cluster with control plane nodes at the beginning (worker nodes will be added later on after the various components that are necessary for the DPF system are installed).

inventory/mycluster/hosts.yaml

YAML

all:
  hosts:
    master1:
      ansible_host: 10.0.110.1
      ip: 10.0.110.1
      access_ip: 10.0.110.1
      node_labels:
        "k8s.ovn.org/zone-name": "master1"
    master2:
      ansible_host: 10.0.110.2
      ip: 10.0.110.2
      access_ip: 10.0.110.2
      node_labels:
        "k8s.ovn.org/zone-name": "master2"
    master3:
      ansible_host: 10.0.110.3
      ip: 10.0.110.3
      access_ip: 10.0.110.3
      node_labels:
        "k8s.ovn.org/zone-name": "master3"
    worker1:
      ansible_host: 10.0.110.21
      ip: 10.0.110.21
      access_ip: 10.0.110.21
      node_labels:
        "node-role.kubernetes.io/worker": ""
        "k8s.ovn.org/dpu-host": ""
        "k8s.ovn.org/zone-name": "worker1"
      node_annotations:
        "k8s.ovn.org/remote-zone-migrated": "worker1"
      kubelet_cpu_manager_policy: static
      kubelet_topology_manager_policy: single-numa-node
      kubelet_reservedSystemCPUs: 0-7
    worker2:
      ansible_host: 10.0.110.22
      ip: 10.0.110.22
      access_ip: 10.0.110.22
      node_labels:
        "node-role.kubernetes.io/worker": ""
        "k8s.ovn.org/dpu-host": ""
        "k8s.ovn.org/zone-name": "worker2"
      node_annotations:
        "k8s.ovn.org/remote-zone-migrated": "worker2"
      kubelet_cpu_manager_policy: static
      kubelet_topology_manager_policy: single-numa-node
      kubelet_reservedSystemCPUs: 0-7
  children:
    kube_control_plane:
      hosts:
        master1:
        master2:
        master3:
    kube_node:
      hosts:
#        worker1:
#        worker2:
    etcd:
      hosts:
        master1:
        master2:
        master3:
    k8s_cluster:
      children:
        kube_control_plane:
        kube_node:

Deploying Cluster Using Kubespray Ansible Playbook

Run the following command from the Jump Node to initiate the deployment process:

Ensure you are in the Python virtual environment (.venv) when running the command.

Jump Node Console
```
(.venv) depuser@jump:~/kubespray$ ansible-playbook -i inventory/mycluster/hosts.yaml --become --become-user=root cluster.yml
```
It takes a while for this deployment to complete. Make sure there are no errors. Successful result example:

It is recommended to keep the shell from which Kubespray has been running open, later on it will be useful when performing cluster scale out to add the worker nodes.

K8s Deployment Verification

To simplify managing the K8s cluster from the Jump Host, set up kubectl with bash auto-completion.

Copy kubectl and the kubeconfig file from master1 to the Jump Host:

Jump Node Console

## Connect to master1
depuser@jump:~$ ssh master1
depuser@master1:~$ cp /usr/local/bin/kubectl /tmp/
depuser@master1:~$ sudo cp /root/.kube/config /tmp/kube-config
depuser@master1:~$ sudo chmod 644 /tmp/kube-config

In another terminal tab, copy the files to the Jump Host:

Jump Node Console

depuser@jump:~$ scp master1:/tmp/kubectl /tmp/
depuser@jump:~$ sudo chown root:root /tmp/kubectl
depuser@jump:~$ sudo mv /tmp/kubectl /usr/local/bin/
depuser@jump:~$ mkdir -p ~/.kube
depuser@jump:~$ scp master1:/tmp/kube-config ~/.kube/config
depuser@jump:~$ chmod 600 ~/.kube/config

Enable bash auto-completion for kubectl:
1. Verify if bash-completion is installed:
  
  Jump Node Console
  depuser@jump:~$ type _init_completion
  If installed, the output will include:
  
  Jump Node Console
  _init_completion is a function
2. If not installed, install it:
  
  Jump Node Console
  depuser@jump:~$ sudo apt install -y bash-completion
3. Set up the kubectl completion script:
  
  Jump Node Console
  depuser@jump:~$ kubectl completion bash | sudo tee /etc/bash_completion.d/kubectl > /dev/null depuser@jump:~$ bash

Check the status of the nodes in the cluster:

Jump Node Console

depuser@jump:~$ kubectl get nodes

Expected output:

Nodes will be in the NotReady state because the deployment did not include CNI components.

Jump Node Console

NAME      STATUS     ROLES           AGE   VERSION
master1   NotReady   control-plane   12m   v1.32.8
master2   NotReady   control-plane   12m   v1.32.8
master3   NotReady   control-plane   12m   v1.32.8

Check the pods in all namespaces:

Jump Node Console

depuser@jump:~$ kubectl get pods -A

Expected output:

coredns and dns-autoscaler pods will be in the Pending state due to the absence of CNI components.

Jump Node Console

NAMESPACE     NAME                              READY   STATUS    RESTARTS   AGE
kube-system   coredns-5c54f84c97-7fl6m          0/1     Pending   0          12m
kube-system   dns-autoscaler-56cb45595c-mkdjq   0/1     Pending   0          12m
kube-system   kube-apiserver-master1            1/1     Running   0          13m
kube-system   kube-apiserver-master2            1/1     Running   0          12m
kube-system   kube-apiserver-master3            1/1     Running   0          12m
kube-system   kube-controller-manager-master1   1/1     Running   1          13m
kube-system   kube-controller-manager-master2   1/1     Running   1          12m
kube-system   kube-controller-manager-master3   1/1     Running   1          12m
kube-system   kube-scheduler-master1            1/1     Running   1          13m
kube-system   kube-scheduler-master2            1/1     Running   1          12m
kube-system   kube-scheduler-master3            1/1     Running   1          12m
kube-system   kube-vip-master1                  1/1     Running   0          13m
kube-system   kube-vip-master2                  1/1     Running   0          12m
kube-system   kube-vip-master3                  1/1     Running   0          12m

DPF Installation

Software Prerequisites and Required Variables

Start by installing the remaining software perquisites.

Jump Node Console

## Connect to master1 to copy helm client utility that was installed during kubespray deployment
$ depuser@jump:~$ ssh master1
depuser@master1:~$ cp /usr/local/bin/helm /tmp/

## In another tab 
depuser@jump:~$ scp master1:/tmp/helm /tmp/
depuser@jump:~$ sudo chown root:root /tmp/helm
depuser@jump:~$ sudo mv /tmp/helm /usr/local/bin/

## Verify that envsubst utility is installed 
depuser@jump:~$ which envsubst
/usr/bin/envsubst

Proceed to clone the doca-platform Git repository:

Jump Node Console
```
git clone https://github.com/NVIDIA/doca-platform.git
```
Change directory to doca-platform and checkout to tag v25.10.0:

Jump Node Console
```
cd doca-platform/
git checkout v25.10.0
```
Change directory to readme.md from where all the commands will be run:

Jump Node Console
```
cd docs/public/user-guides/host-trusted/use-cases/hbn-ovnk/
```

Use the following file to define the required variables for the installation:

Replace the values for the variables in the following file with the values that fit your setup. Specifically, pay attention to DPU_P0 , DPU_P0_VF1 and DPUCLUSTER_INTERFACE.

manifests/00-env-vars/envvars.env

Bash

## IP Address for the Kubernetes API server of the target cluster on which DPF is installed.
## This should never include a scheme or a port.
## e.g. 10.10.10.10
export TARGETCLUSTER_API_SERVER_HOST=10.0.110.10

## Port for the Kubernetes API server of the target cluster on which DPF is installed.
export TARGETCLUSTER_API_SERVER_PORT=6443

## IP address range for hosts in the target cluster on which DPF is installed.
## This is a CIDR in the form e.g. 10.10.10.0/24
export TARGETCLUSTER_NODE_CIDR=10.0.110.0/24

## Virtual IP used by the load balancer for the DPU Cluster. Must be a reserved IP from the management subnet and not allocated by DHCP.
export DPUCLUSTER_VIP=10.0.110.200

## DPU_P0 is the name of the first port of the DPU. This name must be the same on all worker nodes.
export DPU_P0=ens5f0np0

## DPU_P0_VF1 is the name of the second Virtual Function (VF) of the first port of the DPU. This name must be the same on all worker nodes.
## Note: The VF will be created after the DPU is provisioned and the phase "Host Network Configuration" is completed.
export DPU_P0_VF1=ens5f0v1

## Interface on which the DPUCluster load balancer will listen. Should be the management interface of the control plane node.
export DPUCLUSTER_INTERFACE=brenp1s0

## IP address to the NFS server used as storage for the BFB.
export NFS_SERVER_IP=10.0.110.253

## The repository URL for the NVIDIA Helm chart registry.
## Usually this is the NVIDIA Helm NGC registry. For development purposes, this can be set to a different repository.
export HELM_REGISTRY_REPO_URL=https://helm.ngc.nvidia.com/nvidia/doca

## The repository URL for the HBN container image.
## Usually this is the NVIDIA NGC registry. For development purposes, this can be set to a different repository.
export HBN_NGC_IMAGE_URL=nvcr.io/nvidia/doca/doca_hbn

## The repository URL for the OVN-Kubernetes Helm chart.
## Usually this is the NVIDIA GHCR repository. For development purposes, this can be set to a different repository.
export OVN_KUBERNETES_REPO_URL=oci://ghcr.io/nvidia

## POD_CIDR is the CIDR used for pods in the target Kubernetes cluster.
export POD_CIDR=10.233.64.0/18

## SERVICE_CIDR is the CIDR used for services in the target Kubernetes cluster.
## This is a CIDR in the form e.g. 10.10.10.0/24
export SERVICE_CIDR=10.233.0.0/18

## The DPF REGISTRY is the Helm repository URL where the DPF Operator Chart resides.
## Usually this is the NVIDIA Helm NGC registry. For development purposes, this can be set to a different repository.
export REGISTRY=https://helm.ngc.nvidia.com/nvidia/doca

## The DPF TAG is the version of the DPF components which will be deployed in this guide.
export TAG=v25.10.0

## URL to the BFB used in the `bfb.yaml` and linked by the DPUSet.
export BFB_URL="https://content.mellanox.com/BlueField/BFBs/Ubuntu24.04/bf-bundle-3.2.1-34_25.11_ubuntu-24.04_64k_prod.bfb"

Export environment variables for the installation:

Jump Node Console
```
source manifests/00-env-vars/envvars.env
```

CNI Installation

OVN Kubernetes is used as the primary CNI for the cluster. On worker nodes, the primary CNI will be accelerated by offloading work to the DPU. On control plane nodes, OVN Kubernetes will run without offloading.

Create the NS for the CNI:

Jump Node Console
```
kubectl create ns ovn-kubernetes
```

Install the OVN Kubernetes CNI components from the helm chart substituting the environment variables with the ones we defined before.

Note that MTU field with value of 8940 has been added to the yaml to override the default value and to be able to achieve better performance results.

YAML

commonManifests:
  enabled: true
nodeWithoutDPUManifests:
  enabled: true
controlPlaneManifests:
  enabled: true
nodeWithDPUManifests:
  enabled: true
  nodeMgmtPortNetdev: $DPU_P0_VF1
  dpuServiceAccountNamespace: dpf-operator-system
gatewayOpts: --gateway-interface=$DPU_P0
## Note this CIDR is followed by a trailing /24 which informs OVN Kubernetes on how to split the CIDR per node.
podNetwork: $POD_CIDR/24
serviceNetwork: $SERVICE_CIDR
k8sAPIServer: https://$TARGETCLUSTER_API_SERVER_HOST:$TARGETCLUSTER_API_SERVER_PORT
mtu: 8940

Run the following command:

Jump Node Console

envsubst < manifests/01-cni-installation/helm-values/ovn-kubernetes.yml | helm upgrade --install -n ovn-kubernetes ovn-kubernetes ${OVN_KUBERNETES_REPO_URL}/ovn-kubernetes-chart --version $TAG --values -

Verify the CNI installation:

The following verification commands may need to be run multiple times to ensure the condition is met.

Jump Node Console

$ kubectl wait --for=condition=ready --namespace ovn-kubernetes pods --all --timeout=300s
pod/ovn-kubernetes-cluster-manager-7df49fcffb-dx9zs condition met
pod/ovn-kubernetes-node-gh7gc condition met
pod/ovn-kubernetes-node-vzqkx condition met
pod/ovn-kubernetes-node-zktqt condition met

$ kubectl wait --for=condition=ready nodes --all
node/master1 condition met
node/master2 condition met
node/master3 condition met

$ kubectl wait --for=condition=ready --namespace kube-system pods --all
pod/coredns-5c54f84c97-7fl6m condition met
pod/coredns-5c54f84c97-vgs2l condition met
pod/dns-autoscaler-56cb45595c-mkdjq condition met
pod/kube-apiserver-master1 condition met
pod/kube-apiserver-master2 condition met
pod/kube-apiserver-master3 condition met
pod/kube-controller-manager-master1 condition met
pod/kube-controller-manager-master2 condition met
pod/kube-controller-manager-master3 condition met
pod/kube-scheduler-master1 condition met
pod/kube-scheduler-master2 condition met
pod/kube-scheduler-master3 condition met
pod/kube-vip-master1 condition met
pod/kube-vip-master2 condition met
pod/kube-vip-master3 condition met

DPF Operator Installation

Create Storage Required by the DPF Operator

Create the NS for the operator:

Jump Node Console
```
kubectl create ns dpf-operator-system
```

The following YAML file defines storage (for the BFB image) that is required by the DPF operator.

YAML

---
apiVersion: v1
kind: PersistentVolume
metadata:
  name: bfb-pv
spec:
  capacity:
    storage: 10Gi
  volumeMode: Filesystem
  accessModes:
    - ReadWriteMany
  nfs:
    path: /mnt/dpf_share/bfb
    server: $NFS_SERVER_IP
  persistentVolumeReclaimPolicy: Delete
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: bfb-pvc
  namespace: dpf-operator-system
spec:
  accessModes:
  - ReadWriteMany
  resources:
    requests:
      storage: 10Gi
  volumeMode: Filesystem
  storageClassName: ""

Run the following command to substitute the environment variables using envsubst and apply the yaml file:

Jump Node Console
```
cat manifests/02-dpf-operator-installation/*.yaml | envsubst | kubectl apply -f -
```

Additional Dependencies

The DPF Operator requires several prerequisite components to function properly in a Kubernetes environment. Starting with DPF v25.7, all Helm dependencies have been removed from the DPF chart. This means that all dependencies must be installed manually before installing the DPF chart itself. The following commands describe an opiniated approach to install those dependencies (for more information, check: Helm Prerequisites - NVIDIA Docs).
1. Install helmfile binary:
  
  Jump Node Console
  wget https://github.com/helmfile/helmfile/releases/download/v1.1.2/helmfile_1.1.2_linux_amd64.tar.gz tar -xvf helmfile_1.1.2_linux_amd64.tar.gz sudo mv ./helmfile /usr/local/bin/
2. Change directory to doca-platform:
  
  Use another shell from the one where you run all the other installation commands for DPF.
  
  Jump Node Console
  cd doca-platform/
3. Install Helm dependencies using the following command:
  
  Jump Node Console
  make HELMFILE_FILE=deploy/helmfiles/prereqs.yaml test-deploy-helmfile
Ensure that the KUBERNETES_SERVICE_HOST and KUBERNETES_SERVICE_PORT environment variables are set in the node-feature-discovery-worker DaemonSet:

Run this command from the previous shell where the environment variables were exported.

Jump Node Console
```
kubectl -n dpf-operator-system set env daemonset/node-feature-discovery-worker \
KUBERNETES_SERVICE_HOST=$TARGETCLUSTER_API_SERVER_HOST \
KUBERNETES_SERVICE_PORT=$TARGETCLUSTER_API_SERVER_PORT
```

DPF Operator Deployment

Run the following commands to install the DPF Operator:

Jump Node Console

helm repo add --force-update dpf-repository ${REGISTRY}
helm repo update
helm upgrade --install -n dpf-operator-system dpf-operator dpf-repository/dpf-operator --version=$TAG

Verify the DPF Operator installation by ensuring the deployment is available and all the pods are ready:

The following verification commands may need to be run multiple times to ensure the conditions are met.

Jump Node Console

$ kubectl rollout status deployment --namespace dpf-operator-system dpf-operator-controller-manager
deployment "dpf-operator-controller-manager" successfully rolled out

$ kubectl wait --for=condition=ready --namespace dpf-operator-system pods --all
pod/argo-cd-argocd-application-controller-0 condition met
pod/argo-cd-argocd-redis-77dfd8fcb4-fhn27 condition met
pod/argo-cd-argocd-repo-server-7b6c5b8cdb-t8rkb condition met
pod/argo-cd-argocd-server-744d5f9c7c-sxptv condition met
pod/dpf-operator-controller-manager-645467745b-xhtfv condition met
pod/kamaji-556cb86895-mbjgh condition met
pod/kamaji-etcd-0 condition met
pod/kamaji-etcd-1 condition met
pod/kamaji-etcd-2 condition met
pod/maintenance-operator-585767f779-gttvx condition met
pod/node-feature-discovery-gc-7f64f764f8-hbmv8 condition met
pod/node-feature-discovery-master-6fbc95665c-l8zrc condition met

DPF System Installation

This section involves creating the DPF system components and some basic infrastructure required for a functioning DPF-enabled cluster.

The following YAML files define the DPFOperatorConfig to install the DPF System components and the DPUCluster to serve as Kubernetes control plane for DPU nodes.

Note that to achieve high performance results you need to adjust the operatorconfig.yaml to support MTU 9000.

YAML

---
apiVersion: operator.dpu.nvidia.com/v1alpha1
kind: DPFOperatorConfig
metadata:
  name: dpfoperatorconfig
  namespace: dpf-operator-system
spec:
  overrides:
    kubernetesAPIServerVIP: $TARGETCLUSTER_API_SERVER_HOST
    kubernetesAPIServerPort: $TARGETCLUSTER_API_SERVER_PORT
  provisioningController:
    bfbPVCName: "bfb-pvc"
    dmsTimeout: 900
  kamajiClusterManager:
    disable: false
  networking:
    controlPlaneMTU: 9000
    highSpeedMTU: 9000

YAML

---
apiVersion: provisioning.dpu.nvidia.com/v1alpha1
kind: DPUCluster
metadata:
  name: dpu-cplane-tenant1
  namespace: dpu-cplane-tenant1
spec:
  type: kamaji
  maxNodes: 10
  clusterEndpoint:
    # deploy keepalived instances on the nodes that match the given nodeSelector.
    keepalived:
      # interface on which keepalived will listen. Should be the oob interface of the control plane node.
      interface: $DPUCLUSTER_INTERFACE
      # Virtual IP reserved for the DPU Cluster load balancer. Must not be allocatable by DHCP.
      vip: $DPUCLUSTER_VIP
      # virtualRouterID must be in range [1,255], make sure the given virtualRouterID does not duplicate with any existing keepalived process running on the host
      virtualRouterID: 126
      nodeSelector:
        node-role.kubernetes.io/control-plane: ""

Create NS for the Kubernetes control plane of the DPU nodes:

Jump Node Console
```
kubectl create ns dpu-cplane-tenant1
```

Apply the previous YAML files:

Jump Node Console

cat manifests/03-dpf-system-installation/*.yaml | envsubst | kubectl apply -f -

Verify the DPF system by ensuring that the provisioning and DPUService controller manager deployments are available, that all other deployments in the DPF Operator system are available, and that the DPUCluster is ready for nodes to join.

Jump Node Console

$ kubectl rollout status deployment --namespace dpf-operator-system dpf-provisioning-controller-manager dpuservice-controller-manager
deployment "dpf-provisioning-controller-manager" successfully rolled out
deployment "dpuservice-controller-manager" successfully rolled out

$ kubectl rollout status deployment --namespace dpf-operator-system
deployment "argo-cd-argocd-applicationset-controller" successfully rolled out
deployment "argo-cd-argocd-redis" successfully rolled out
deployment "argo-cd-argocd-repo-server" successfully rolled out
deployment "argo-cd-argocd-server" successfully rolled out
deployment "dpf-operator-controller-manager" successfully rolled out
deployment "dpf-provisioning-controller-manager" successfully rolled out
deployment "dpuservice-controller-manager" successfully rolled out
deployment "kamaji" successfully rolled out
deployment "kamaji-cm-controller-manager" successfully rolled out
deployment "maintenance-operator" successfully rolled out
deployment "node-feature-discovery-gc" successfully rolled out
deployment "node-feature-discovery-master" successfully rolled out
deployment "servicechainset-controller-manager" successfully rolled out

$ kubectl wait --for=condition=ready --namespace dpu-cplane-tenant1 dpucluster --all
dpucluster.provisioning.dpu.nvidia.com/dpu-cplane-tenant1 condition met

Install Components to Enable Accelerated CNI Nodes

OVN Kubernetes accelerates traffic by attaching a VF to each pod using the primary CNI. This VF is used to offload flows to the DPU. This section details the components needed to connect pods to the offloaded OVN Kubernetes CNI.

Install Multus and SRIOV Network Operator using NVIDIA Network Operator

Add the NVIDIA Network Operator Helm repository:

Jump Node Console

helm repo add nvidia https://helm.ngc.nvidia.com/nvidia --force-update

The following network-operator.yaml values file will be applied:

YAML

nfd:
  enabled: false
  deployNodeFeatureRules: false
sriovNetworkOperator:
  enabled: true
sriov-network-operator:
  operator:
    affinity:
      nodeAffinity:
        requiredDuringSchedulingIgnoredDuringExecution:
          nodeSelectorTerms:
            - matchExpressions:
                - key: node-role.kubernetes.io/master
                  operator: Exists
            - matchExpressions:
                - key: node-role.kubernetes.io/control-plane
                  operator: Exists
  crds:
    enabled: true
  sriovOperatorConfig:
    deploy: true
    configDaemonNodeSelector: null
operator:
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
          - matchExpressions:
              - key: node-role.kubernetes.io/master
                operator: Exists
          - matchExpressions:
              - key: node-role.kubernetes.io/control-plane
                operator: Exists

Deploy the operator:

Jump Node Console

helm upgrade --no-hooks --install --create-namespace --namespace nvidia-network-operator network-operator nvidia/network-operator --version 25.7.0 -f ./manifests/04-enable-accelerated-cni/helm-values/network-operator.yml

Ensure all the pods in nvidia-network-operator namespace are ready:

Jump Node Console

$ kubectl wait --for=condition=ready --namespace nvidia-network-operator pods --all
pod/network-operator-5bc6758dbd-g5slv condition met
pod/network-operator-sriov-network-operator-6779b646b7-5fq4m condition met

Install OVN Kubernetes resource injection webhook

The OVN Kubernetes resource injection webhook is injected into each pod scheduled to a worker node with a request for a VF and a Network Attachment Definition. This webhook is part of the same helm chart as the other components of the OVN Kubernetes CNI. Here it is installed by adjusting the existing helm installation to add the webhook component to the installation.

The following ovn-kubernetes.yaml values file will be applied:

YAML

ovn-kubernetes-resource-injector:
  ## Enable the ovn-kubernetes-resource-injector
  enabled: true

Run the following command:

Jump Node Console

envsubst < manifests/04-enable-accelerated-cni/helm-values/ovn-kubernetes.yml | helm upgrade --install -n ovn-kubernetes ovn-kubernetes-resource-injector ${OVN_KUBERNETES_REPO_URL}/ovn-kubernetes-chart --version $TAG --values -

Verify that the resource injector deployment has been successfully rolled out.

Jump Node Console

$ kubectl rollout status deployment --namespace ovn-kubernetes ovn-kubernetes-resource-injector
deployment "ovn-kubernetes-resource-injector" successfully rolled out

Apply NicClusterPolicy and SriovNetworkNodePolicy

The following NicClusterPolicy and SriovNetworkNodePolicy configuration files should be applied.

manifests/04-enable-accelerated-cni/nic_cluster_policy.yaml

YAML

---
apiVersion: mellanox.com/v1alpha1
kind: NicClusterPolicy
metadata:
  name: nic-cluster-policy
spec:
  secondaryNetwork:
    multus:
      image: multus-cni
      imagePullSecrets: []
      repository: ghcr.io/k8snetworkplumbingwg
      version: v3.9.3

manifests/04-enable-accelerated-cni/sriov_network_operator_policy.yaml

YAML

---
apiVersion: sriovnetwork.openshift.io/v1
kind: SriovNetworkNodePolicy
metadata:
  name: bf3-p0-vfs
  namespace: nvidia-network-operator
spec:
  nicSelector:
    deviceID: "a2dc"
    vendor: "15b3"
    pfNames:
    - $DPU_P0#2-45
  nodeSelector:
    node-role.kubernetes.io/worker: ""
  numVfs: 46
  resourceName: bf3-p0-vfs
  isRdma: true
  externallyManaged: true
  deviceType: netdevice
  linkType: eth

Apply those configuration files:

Jump Node Console

cat manifests/04-enable-accelerated-cni/*.yaml | envsubst | kubectl apply -f -

Verify the DPF system by ensuring that the following DaemonSets were successfully rolled out:

Jump Node Console

$ kubectl rollout status daemonset --namespace nvidia-network-operator kube-multus-ds sriov-network-config-daemon sriov-device-plugin
daemon set "kube-multus-ds" successfully rolled out
daemon set "sriov-network-config-daemon" successfully rolled out
daemon set "sriov-device-plugin" successfully rolled out

DPU Provisioning and Service Installation

Before deploying the objects under manifests/05-dpudeployment-installationdirectory, few adjustments need to be made to later achieve better performance results.

Create a new DPUFlavor using the following YAML:

The parameter NUM_VF_MSIX is configured to be 48 in the provided example, which is suited for the HP servers that were used in this RDG.
Set it to the physical number of cores in the NUMA node the NIC is located in.

YAML

---
apiVersion: provisioning.dpu.nvidia.com/v1alpha1
kind: DPUFlavor
metadata:
  name: hbn-ovnk-$TAG-performance
  namespace: dpf-operator-system
spec:
  grub:
    kernelParameters:
      - console=hvc0
      - console=ttyAMA0
      - earlycon=pl011,0x13010000
      - fixrttc
      - net.ifnames=0
      - biosdevname=0
      - iommu.passthrough=1
      - cgroup_no_v1=net_prio,net_cls
      - hugepagesz=2048kB
      - hugepages=8072
  nvconfig:
    - device: "*"
      parameters:
        - PF_BAR2_ENABLE=0
        - PER_PF_NUM_SF=1
        - PF_TOTAL_SF=20
        - PF_SF_BAR_SIZE=10
        - NUM_PF_MSIX_VALID=0
        - PF_NUM_PF_MSIX_VALID=1
        - PF_NUM_PF_MSIX=228
        - INTERNAL_CPU_MODEL=1
        - INTERNAL_CPU_OFFLOAD_ENGINE=0
        - SRIOV_EN=1
        - NUM_OF_VFS=46
        - LAG_RESOURCE_ALLOCATION=1
        - LINK_TYPE_P1=ETH
        - LINK_TYPE_P2=ETH
        - NUM_VF_MSIX=48
  ovs:
    rawConfigScript: |
      _ovs-vsctl() {
        ovs-vsctl --no-wait --timeout 15 "$@"
      }

      _ovs-vsctl set Open_vSwitch . other_config:doca-init=true
      _ovs-vsctl set Open_vSwitch . other_config:dpdk-max-memzones=50000
      _ovs-vsctl set Open_vSwitch . other_config:hw-offload=true
      _ovs-vsctl set Open_vSwitch . other_config:pmd-quiet-idle=true
      _ovs-vsctl set Open_vSwitch . other_config:max-idle=20000
      _ovs-vsctl set Open_vSwitch . other_config:max-revalidator=5000
      _ovs-vsctl set Open_vSwitch . other_config:ctl-pipe-size=1024
      _ovs-vsctl --if-exists del-br ovsbr1
      _ovs-vsctl --if-exists del-br ovsbr2
      _ovs-vsctl --may-exist add-br br-sfc
      _ovs-vsctl set bridge br-sfc datapath_type=netdev
      _ovs-vsctl set bridge br-sfc fail_mode=secure
      _ovs-vsctl --may-exist add-br br-hbn
      _ovs-vsctl set bridge br-hbn datapath_type=netdev
      _ovs-vsctl set bridge br-hbn fail_mode=secure
      _ovs-vsctl --may-exist add-port br-sfc p0
      _ovs-vsctl set Interface p0 type=dpdk
      _ovs-vsctl set Interface p0 mtu_request=9216
      _ovs-vsctl set Port p0 external_ids:dpf-type=physical
      _ovs-vsctl --may-exist add-port br-sfc p1
      _ovs-vsctl set Interface p1 type=dpdk
      _ovs-vsctl set Interface p1 mtu_request=9216
      _ovs-vsctl set Port p1 external_ids:dpf-type=physical

      _ovs-vsctl set Open_vSwitch . external-ids:ovn-bridge-datapath-type=netdev
      _ovs-vsctl --may-exist add-br br-ovn
      _ovs-vsctl set bridge br-ovn datapath_type=netdev
      _ovs-vsctl br-set-external-id br-ovn bridge-id br-ovn
      _ovs-vsctl br-set-external-id br-ovn bridge-uplink puplinkbrovntobrsfc
      _ovs-vsctl set Interface br-ovn mtu_request=9216
      _ovs-vsctl --may-exist add-port br-ovn pf0hpf
      _ovs-vsctl set Interface pf0hpf type=dpdk
      _ovs-vsctl set Interface pf0hpf mtu_request=9216

      cat <<EOT > /etc/netplan/99-dpf-comm-ch.yaml
      network:
        renderer: networkd
        version: 2
        ethernets:
          pf0vf0:
            mtu: 9000
            dhcp4: no
        bridges:
          br-comm-ch:
            dhcp4: yes
            interfaces:
              - pf0vf0
      EOT

  bfcfgParameters:
    - UPDATE_ATF_UEFI=yes
    - UPDATE_DPU_OS=yes
    - WITH_NIC_FW_UPDATE=yes

  hostNetworkInterfaceConfigs:
    - portNumber: 0
      dhcp: true
      mtu: 9000

  configFiles:
  - path: /etc/mellanox/mlnx-bf.conf
    operation: override
    raw: |
        ALLOW_SHARED_RQ="no"
        IPSEC_FULL_OFFLOAD="no"
        ENABLE_ESWITCH_MULTIPORT="yes"
    permissions: "0644"
  - path: /etc/mellanox/mlnx-ovs.conf
    operation: override
    raw: |
        CREATE_OVS_BRIDGES="no"
        OVS_DOCA="yes"
    permissions: "0644"
  - path: /etc/mellanox/mlnx-sf.conf
    operation: override
    raw: ""
    permissions: "0644"

Adjust dpudeployment.yaml to reference the DPUFlavor suited for performance (This component provisions DPUs on the worker nodes and describes a set of DPUServices and DPUServiceChain that run on those DPUs):

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUDeployment
metadata:
  name: ovn-hbn
  namespace: dpf-operator-system
spec:
  dpus:
    bfb: bf-bundle-$TAG
    flavor: hbn-ovnk-$TAG-performance
    dpuSets:
    - nameSuffix: "dpuset1"
      nodeSelector:
        matchLabels:
          feature.node.kubernetes.io/dpu-enabled: "true"
      dpuSelector:
        provisioning.dpu.nvidia.com/dpudevice-pf0-name: $DPU_P0
  services:
    ovn:
      serviceTemplate: ovn
      serviceConfiguration: ovn
    hbn:
      serviceTemplate: hbn
      serviceConfiguration: hbn
    dts:
      serviceTemplate: dts
      serviceConfiguration: dts
    blueman:
      serviceTemplate: blueman
      serviceConfiguration: blueman
  serviceChains:
    switches:
      - ports:
        - serviceInterface:
            matchLabels:
              uplink: p0
        - service:
            name: hbn
            interface: p0_if
      - ports:
        - serviceInterface:
            matchLabels:
              uplink: p1
        - service:
            name: hbn
            interface: p1_if
      - ports:
        - serviceInterface:
            matchLabels:
              port: ovn
        - service:
            name: hbn
            interface: pf2dpu2_if

Set the mtu to 8940 for the OVN DPUServiceConfig (to deploy the OVN Kubernetes workloads on the DPU with the same MTU as in the host):

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceConfiguration
metadata:
  name: ovn
  namespace: dpf-operator-system
spec:
  deploymentServiceName: "ovn"
  serviceConfiguration:
    helmChart:
      values:
        k8sAPIServer: https://$TARGETCLUSTER_API_SERVER_HOST:$TARGETCLUSTER_API_SERVER_PORT
        podNetwork: $POD_CIDR/24
        serviceNetwork: $SERVICE_CIDR
        mtu: 8940
        dpuManifests:
          kubernetesSecretName: "ovn-dpu" # user needs to populate based on DPUServiceCredentialRequest
          vtepCIDR: "10.0.120.0/22" # user needs to populate based on DPUServiceIPAM
          hostCIDR: $TARGETCLUSTER_NODE_CIDR # user needs to populate
          ipamPool: "pool1" # user needs to populate based on DPUServiceIPAM
          ipamPoolType: "cidrpool" # user needs to populate based on DPUServiceIPAM
          ipamVTEPIPIndex: 0
          ipamPFIPIndex: 1

Use the following dpuserviceconfig_hbn.yaml so it will calculate the BGP ASN (Autonomous System Number) based on the loopback IPAM and not using static values:

manifests/05-dpudeployment-installation/dpuserviceconfig_hbn.yaml

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceConfiguration
metadata:
  name: hbn
  namespace: dpf-operator-system
spec:
  deploymentServiceName: "hbn"
  serviceConfiguration:
    serviceDaemonSet:
      annotations:
        k8s.v1.cni.cncf.io/networks: |-
          [
          {"name": "iprequest", "interface": "ip_lo", "cni-args": {"poolNames": ["loopback"], "poolType": "cidrpool"}},
          {"name": "iprequest", "interface": "ip_pf2dpu2", "cni-args": {"poolNames": ["pool1"], "poolType": "cidrpool", "allocateDefaultGateway": true}}
          ]
    helmChart:
      values:
        configuration:
          perDPUValuesYAML: |
            - hostnamePattern: "*"
              values:
                bgp_peer_group: hbn
          startupYAMLJ2: |
            - header:
                model: BLUEFIELD
                nvue-api-version: nvue_v1
                rev-id: 1.0
                version: HBN 2.4.0
            - set:
                interface:
                  lo:
                    ip:
                      address:
                        {{ ipaddresses.ip_lo.ip }}/32: {}
                    type: loopback
                  p0_if,p1_if:
                    type: swp
                    link:
                      mtu: 9000
                  pf2dpu2_if:
                    ip:
                      address:
                        {{ ipaddresses.ip_pf2dpu2.cidr }}: {}
                    type: swp
                    link:
                      mtu: 9000
                router:
                  bgp:
                    autonomous-system: {{ ( ipaddresses.ip_lo.ip.split(".")[3] | int ) + 65101 }}
                    enable: on
                    graceful-restart:
                      mode: full
                    router-id: {{ ipaddresses.ip_lo.ip }}
                vrf:
                  default:
                    router:
                      bgp:
                        address-family:
                          ipv4-unicast:
                            enable: on
                            redistribute:
                              connected:
                                enable: on
                          ipv6-unicast:
                            enable: on
                            redistribute:
                              connected:
                                enable: on
                        enable: on
                        neighbor:
                          p0_if:
                            peer-group: {{ config.bgp_peer_group }}
                            type: unnumbered
                          p1_if:
                            peer-group: {{ config.bgp_peer_group }}
                            type: unnumbered
                        path-selection:
                          multipath:
                            aspath-ignore: on
                        peer-group:
                          {{ config.bgp_peer_group }}:
                            remote-as: external

  interfaces:
    ## NOTE: Interfaces inside the HBN pod must have the `_if` suffix due to a naming convention in HBN.
  - name: p0_if
    network: mybrhbn
  - name: p1_if
    network: mybrhbn
  - name: pf2dpu2_if
    network: mybrhbn

The rest of the configuration files remain the same, including:

BFB to download BlueField Bitstream to a shared volume.

YAML

---
apiVersion: provisioning.dpu.nvidia.com/v1alpha1
kind: BFB
metadata:
  name: bf-bundle-$TAG
  namespace: dpf-operator-system
spec:
  url: $BFB_URL

OVN DPUServiceTemplate to deploy OVN Kubernetes workloads to the DPUs.

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceTemplate
metadata:
  name: ovn
  namespace: dpf-operator-system
spec:
  deploymentServiceName: "ovn"
  helmChart:
    source:
      repoURL: $OVN_KUBERNETES_REPO_URL
      chart: ovn-kubernetes-chart
      version: $TAG
    values:
      commonManifests:
        enabled: true
      dpuManifests:
        enabled: true
      leaseNamespace: "ovn-kubernetes"
      gatewayOpts: "--gateway-interface=br-ovn"

HBN DPUServiceTemplate to deploy HBN workloads to the DPUs.

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceTemplate
metadata:
  name: hbn
  namespace: dpf-operator-system
spec:
  deploymentServiceName: "hbn"
  helmChart:
    source:
      repoURL: $HELM_REGISTRY_REPO_URL
      version: 1.0.5
      chart: doca-hbn
    values:
      image:
        repository: $HBN_NGC_IMAGE_URL
        tag: 3.2.1-doca3.2.1
      resources:
        memory: 6Gi
        nvidia.com/bf_sf: 3

DOCA Telemetry Service (DTS) DPUServiceConfig and DPUServiceTemplate to deploy DTS to the DPUs.

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceConfiguration
metadata:
  name: dts
  namespace: dpf-operator-system
spec:
  deploymentServiceName: "dts"

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceTemplate
metadata:
  name: dts
  namespace: dpf-operator-system
spec:
  deploymentServiceName: "dts"
  helmChart:
    source:
      repoURL: $HELM_REGISTRY_REPO_URL
      version: 1.23.4
      chart: doca-telemetry

Blueman DPUServiceConfig and DPUServiceTemplate to deploy Blueman to the DPUs.

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceConfiguration
metadata:
  name: blueman
  namespace: dpf-operator-system
spec:
  deploymentServiceName: "blueman"

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceTemplate
metadata:
  name: blueman
  namespace: dpf-operator-system
spec:
  deploymentServiceName: "blueman"
  helmChart:
    source:
      repoURL: $HELM_REGISTRY_REPO_URL
      version: 1.0.8
      chart: doca-blueman

OVN DPUServiceCredentialRequest to allow cross cluster communication.

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceCredentialRequest
metadata:
  name: ovn-dpu
  namespace: dpf-operator-system
spec:
  serviceAccount:
    name: ovn-dpu
    namespace: dpf-operator-system
  duration: 24h
  type: tokenFile
  secret:
    name: ovn-dpu
    namespace: dpf-operator-system
  metadata:
    labels:
      dpu.nvidia.com/image-pull-secret: ""

DPUServiceInterfaces for physical ports on the DPU.

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceInterface
metadata:
  name: p0
  namespace: dpf-operator-system
spec:
  template:
    spec:
      template:
        metadata:
          labels:
            uplink: "p0"
        spec:
          interfaceType: physical
          physical:
            interfaceName: p0
---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceInterface
metadata:
  name: p1
  namespace: dpf-operator-system
spec:
  template:
    spec:
      template:
        metadata:
          labels:
            uplink: "p1"
        spec:
          interfaceType: physical
          physical:
            interfaceName: p1

OVN DPUServiceInterface to define the ports attached to OVN workloads on the DPU.

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceInterface
metadata:
  name: ovn
  namespace: dpf-operator-system
spec:
  template:
    spec:
      template:
        metadata:
          labels:
            port: ovn
        spec:
          interfaceType: ovn

DPUServiceIPAM to set up IP Address Management on the DPUCluster.

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceIPAM
metadata:
  name: pool1
  namespace: dpf-operator-system
spec:
  ipv4Network:
    network: "10.0.120.0/22"
    gatewayIndex: 3
    prefixSize: 29

DPUServiceIPAM for the loopback interface in HBN.

YAML

---
apiVersion: svc.dpu.nvidia.com/v1alpha1
kind: DPUServiceIPAM
metadata:
  name: loopback
  namespace: dpf-operator-system
spec:
  ipv4Network:
    network: "11.0.0.0/24"
    prefixSize: 32

Apply all of the YAML files mentioned above using the following command:

Jump Node Console
```
cat manifests/05-dpudeployment-installation/*.yaml | envsubst | kubectl apply -f - 
```

Verify the DPU and Service installation by ensuring the DPUServices are created and have been reconciled, that the DPUServiceIPAMs have been reconciled, that the DPUServiceInterfaces have been reconciled, and that the DPUServiceChain have been reconciled:

Notes

These verification commands may need to be run multiple times to ensure the conditions are met.

Jump Node Console

$ kubectl wait --for=condition=ApplicationsReconciled --namespace dpf-operator-system dpuservices -l svc.dpu.nvidia.com/owned-by-dpudeployment=dpf-operator-system_ovn-hbn
dpuservice.svc.dpu.nvidia.com/blueman-vd2z4 condition met
dpuservice.svc.dpu.nvidia.com/dts-zsf4n condition met
dpuservice.svc.dpu.nvidia.com/hbn-vtm7n condition met
dpuservice.svc.dpu.nvidia.com/ovn-b5z57 condition met

$ kubectl wait --for=condition=DPUIPAMObjectReconciled --namespace dpf-operator-system dpuserviceipam --all
dpuserviceipam.svc.dpu.nvidia.com/loopback condition met
dpuserviceipam.svc.dpu.nvidia.com/pool1 condition met

$ kubectl wait --for=condition=ServiceInterfaceSetReconciled --namespace dpf-operator-system dpuserviceinterface --all
dpuserviceinterface.svc.dpu.nvidia.com/hbn-p0-if-4lckn condition met
dpuserviceinterface.svc.dpu.nvidia.com/hbn-p1-if-8m45f condition met
dpuserviceinterface.svc.dpu.nvidia.com/hbn-pf2dpu2-if-zj2sv condition met
dpuserviceinterface.svc.dpu.nvidia.com/ovn condition met
dpuserviceinterface.svc.dpu.nvidia.com/p0 condition met
dpuserviceinterface.svc.dpu.nvidia.com/p1 condition met

$ kubectl wait --for=condition=ServiceChainSetReconciled --namespace dpf-operator-system dpuservicechain --all
dpuservicechain.svc.dpu.nvidia.com/ovn-hbn-prwdc condition met

K8s Cluster Scale-out

Add Worker Nodes to the Cluster

At this point workers should be added to the cluster. As workers are added to the cluster, DPUs will be provisioned and DPUServices will begin to be spun up.

Return to the shell where Kubespray was previously run to deploy the cluster, unmark worker1 and worker2 under group kube_node in the hosts.yaml file, and add the worker nodes to the cluster:

Ensure you are in the Python virtual environment (.venv) when running the command.

Jump Node Console
```
(.venv) depuser@jump:~/kubespray$ cat inventory/mycluster/hosts.yaml
...
   kube_node:
     hosts:
       worker1:
       worker2:
...

(.venv) depuser@jump:~/kubespray$ ansible-playbook -i inventory/mycluster/hosts.yaml --become --become-user=root scale.yml
```
The scale-out shouldn't take a long time, and a successful run should look similar to the following output:

Verification

To follow the progress of the DPU provisioning, run the following command to check in which phase it currently is:

Jump Node Console

$ watch -n10 "kubectl describe dpu -n dpf-operator-system | grep 'Node Name\|Type\|Last\|Phase'"
Every 10.0s: kubectl describe dpu -n dpf-operator-system | grep 'Node Name\|Type\|Last\|Phase'

  Dpu Node Name:                                      worker1
    Last Transition Time:  2025-12-24T12:48:56Z
    Type:                  BFBPrepared
    Last Transition Time:  2025-12-24T12:48:30Z
    Type:                  BFBReady
    Last Transition Time:  2025-12-24T12:48:30Z
    Type:                  Initialized
    Last Transition Time:  2025-12-24T12:48:55Z
    Type:                  NodeEffectReady
    Last Transition Time:  2025-12-24T12:48:56Z
    Type:                  FWConfigured
    Last Transition Time:  2025-12-24T12:48:55Z
    Type:                  InterfaceInitialized
    Last Transition Time:  2025-12-24T12:48:57Z
    Type:                  OSInstalled
  Phase:                OS Installing
  Dpu Node Name:                                      worker2
    Last Transition Time:  2025-12-24T12:52:25Z
    Type:                  BFBPrepared
    Last Transition Time:  2025-12-24T12:51:59Z
    Type:                  BFBReady
    Last Transition Time:  2025-12-24T12:51:58Z
    Type:                  Initialized
    Last Transition Time:  2025-12-24T12:52:24Z
    Type:                  NodeEffectReady
    Last Transition Time:  2025-12-24T12:52:25Z
    Type:                  FWConfigured
    Last Transition Time:  2025-12-24T12:52:24Z
    Type:                  InterfaceInitialized
    Last Transition Time:  2025-12-24T12:52:25Z
    Type:                  OSInstalled
  Phase:                OS Installing

Validate that the DPUs have been provisioned successfully by ensuring they're in ready state:

Jump Node Console

$ kubectl wait --for=condition=ready --namespace dpf-operator-system dpu --all
dpu.provisioning.dpu.nvidia.com/worker1-mt2404xz0c97 condition met
dpu.provisioning.dpu.nvidia.com/worker2-mt2333xz0xvb condition met

Ensure that the following DaemonSets have 2 ready replicas:

Jump Node Console

$ kubectl wait ds --for=jsonpath='{.status.numberReady}'=2 --namespace nvidia-network-operator kube-multus-ds sriov-network-config-daemon sriov-device-plugin
daemonset.apps/kube-multus-ds condition met
daemonset.apps/sriov-network-config-daemon condition met
daemonset.apps/sriov-device-plugin condition met

$ kubectl wait ds --for=jsonpath='{.status.numberReady}'=2 --namespace ovn-kubernetes ovn-kubernetes-node-dpu-host
daemonset.apps/ovn-kubernetes-node-dpu-host condition met

Validate that all the different DPUServices, DPUServiceIPAMs, DPUServiceInterfaces and DPUServiceChain objects are now in ready state:

Jump Node Console

$ kubectl wait --for=condition=ApplicationsReady --namespace dpf-operator-system dpuservices -l svc.dpu.nvidia.com/owned-by-dpudeployment=dpf-operator-system_ovn-hbn
dpuservice.svc.dpu.nvidia.com/blueman-vd2z4 condition met
dpuservice.svc.dpu.nvidia.com/dts-zsf4n condition met
dpuservice.svc.dpu.nvidia.com/hbn-vtm7n condition met
dpuservice.svc.dpu.nvidia.com/ovn-b5z57 condition met

$ kubectl wait --for=condition=DPUIPAMObjectReady --namespace dpf-operator-system dpuserviceipam --all
dpuserviceipam.svc.dpu.nvidia.com/loopback condition met
dpuserviceipam.svc.dpu.nvidia.com/pool1 condition met

$ kubectl wait --for=condition=ServiceInterfaceSetReady --namespace dpf-operator-system dpuserviceinterface --all
dpuserviceinterface.svc.dpu.nvidia.com/hbn-p0-if-4lckn condition met
dpuserviceinterface.svc.dpu.nvidia.com/hbn-p1-if-8m45f condition met
dpuserviceinterface.svc.dpu.nvidia.com/hbn-pf2dpu2-if-zj2sv condition met
dpuserviceinterface.svc.dpu.nvidia.com/ovn condition met
dpuserviceinterface.svc.dpu.nvidia.com/p0 condition met
dpuserviceinterface.svc.dpu.nvidia.com/p1 condition met

$ kubectl wait --for=condition=ServiceChainSetReady --namespace dpf-operator-system dpuservicechain --all
dpuservicechain.svc.dpu.nvidia.com/ovn-hbn-prwdc condition met

Congratulations, the DPF system has been successfully installed!

Infrastructure Latency & Bandwidth Validation

Verify the deployment and that you can reach link-speed performance and good latency results on the DPF system by using various tests:

RDMA - for latency measurements
Iperf TCP - for bandwidth measurements

Each of the tests is described thoroughly. At the end of each test, you'll see the achieved performance.

Make sure that the servers are tuned for maximum performance (not covered in this document).

The following diagram illustrates the test environment and how the network traffic is redirected via the accelerated OVN-Kubernetes and HBN services using SFC:

Performance Tests

RoCE Latency Test

Apply the following NetworkPolicy to enable stateless traffic:

stateless_netpolicy.yaml

YAML

apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
  name: multi-port-egress
  namespace: default
  annotations:
    k8s.ovn.org/acl-stateless: "true"
spec:
  podSelector: {}
  policyTypes:
  - Egress
  - Ingress
  egress:
   - {}
  ingress:
   - {}

Jump Node Console

kubectl apply -f stateless_netpolicy.yaml

Create a test Deployment using the following YAML to create 2 replicas on 2 different worker nodes:

The container image specified below must include NVIDIA user space drivers and perftest

testapp-performance-test-deployment.yaml

YAML

---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: testapp-performance
  labels:
    app: testapp-performance
spec:
  replicas: 2
  selector:
    matchLabels:
      app: testapp-performance
  template:
    metadata:
      labels:
        app: testapp-performance
    spec:
      topologySpreadConstraints:
      - maxSkew: 1
        topologyKey: kubernetes.io/hostname
        whenUnsatisfiable: DoNotSchedule
        labelSelector:
          matchLabels:
            app: testapp-performance
      containers:
      - name: testapp-pod
        image: <container_image>
        imagePullPolicy: Always
        command: ['sh', '-c', 'trap : TERM INT; sleep infinity & wait']
        securityContext:
          capabilities:
            add: [ "IPC_LOCK" ]
        resources:
          requests:
            cpu: '24'
            memory: '8Gi'
          limits:
            cpu: '24'
            memory: '8Gi'

Apply the resource:

Jump Node Console

kubectl apply -f testapp-performance-test-deployment.yaml

Validate that the deployment is running successfully:

Jump Node Console

$ kubectl get pods -o wide
NAME                                   READY   STATUS    RESTARTS   AGE   IP            NODE      NOMINATED NODE   READINESS GATES
testapp-performance-567cfdbd4b-7xzxm   1/1     Running   0          15s   10.233.68.3   worker2   <none>           <none>
testapp-performance-567cfdbd4b-m9vzc   1/1     Running   0          15s   10.233.67.3   worker1   <none>           <none>

Connect to one of the pods in the Deployment:

Jump Node Console

kubectl exec -it testapp-performance-567cfdbd4b-7xzxm -- bash

From within the container, check its IP address on its interface and see that it is recognizable as an RDMA device:

First Pod Console

root@testapp-performance-567cfdbd4b-7xzxm:/# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
134: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 8940 qdisc mq state UP group default qlen 1000
    link/ether 0a:58:0a:e9:44:03 brd ff:ff:ff:ff:ff:ff permaddr fa:85:f3:5f:a9:f7
    altname enp137s0f0v4
    inet 10.233.68.3/24 brd 10.233.68.255 scope global eth0
       valid_lft forever preferred_lft forever
    inet6 fe80::f885:f3ff:fe5f:a9f7/64 scope link
       valid_lft forever preferred_lft forever

root@testapp-performance-567cfdbd4b-7xzxm:/# rdma link | grep eth0
link mlx5_6/1 state ACTIVE physical_state LINK_UP netdev eth0

Start the ib_read_lat server side:

First Pod Console

root@testapp-performance-567cfdbd4b-7xzxm:/# ib_read_lat -F -n 20000

************************************
* Waiting for client to connect... *
************************************

Using another console window, reconnect to the jump node and connect to the second pod in the deployment.

Jump Node Console
```
kubectl exec -it testapp-performance-567cfdbd4b-m9vzc -- bash
```

From within the container, start the ib_read_lat client (use the IP address from the server-side container) and check the latency results:

First Pod Console

root@testapp-performance-567cfdbd4b-m9vzc:/# ib_read_lat -F -n 20000 10.233.68.3
---------------------------------------------------------------------------------------
                    RDMA_Read Latency Test
 Dual-port       : OFF          Device         : mlx5_14
 Number of qps   : 1            Transport type : IB
 Connection type : RC           Using SRQ      : OFF
 PCIe relax order: ON
 ibv_wr* API     : ON
 TX depth        : 1
 Mtu             : 4096[B]
 Link type       : Ethernet
 GID index       : 5
 Outstand reads  : 16
 rdma_cm QPs     : OFF
 Data ex. method : Ethernet
---------------------------------------------------------------------------------------
 local address: LID 0000 QPN 0x015f PSN 0xa2d513 OUT 0x10 RKey 0x044500 VAddr 0x00652c78a07000
 GID: 00:00:00:00:00:00:00:00:00:00:255:255:10:233:67:03
 remote address: LID 0000 QPN 0x016e PSN 0xe1afa9 OUT 0x10 RKey 0x048500 VAddr 0x00599c97d3c000
 GID: 00:00:00:00:00:00:00:00:00:00:255:255:10:233:68:03
---------------------------------------------------------------------------------------
 #bytes #iterations    t_min[usec]    t_max[usec]  t_typical[usec]    t_avg[usec]    t_stdev[usec]   99% percentile[usec]   99.9% percentile[usec]
 2       20000          3.94           8.33         4.04               4.68             0.72            7.60                    7.75
---------------------------------------------------------------------------------------

iPerf TCP Bandwidth Test

Create a test Deployment using the YAML from the previous example to create a pod on each worker that you can use to test TCP connectivity and performance.

The container image specified in the test must include iperf.

Connect to one of the pods in the deployment:

Jump Node Console

kubectl exec -it testapp-performance-567cfdbd4b-7xzxm -- bash

Before starting the iperf3 server listeners, and to be able to achieve good results, check in another tab the cores the pod is currently running on:

To be able to bind to specific cores, make sure to schedule a pod in Guaranteed QoS class.

Check on which worker node the pod is running on:

Jump Node Console

$ kubectl get pods -o wide | grep 7xzxm
testapp-performance-567cfdbd4b-7xzxm   1/1     Running   0          10m   10.233.68.3   worker2   <none>           <none>

SSH to the worker:

Jump Node Console

depuser@jump:~$ ssh worker2  
depuser@worker2:~$ sudo -i
root@worker2:~#

Inspect the pod current cores:

Worker2 Console

root@worker2:~# crictl ps | grep testapp
a7f2268086471       032269a586520       10 minutes ago      Running             testapp-pod                   0                   89e06306373c1       testapp-performance-567cfdbd4b-7xzxm   default
root@worker2:~# crictl inspect a7f2268086471 | jq '.status.resources.linux.cpusetCpus'

Output example:

Worker2 Console
```
"28-51"
```

Back within the container of the pod, use the following script to start multiple iperf3 servers (1 for each core) on different ports:

iperf_server.sh

Bash

#!/bin/bash

# Cores to bind the iperf3 server processes to
CORES=$1

# Calculate the first_core and last_core to provide the CPU range
first_core=$(echo $CORES | cut -d "-" -f1)
last_core=$(echo $CORES | cut -d "-" -f2)

# Loop over the ports (5201 + i*2) for i in the given CPU range and run iperf3 servers
for i in $(seq $first_core $last_core); do
   echo "Running iperf3 server on core $i"
   taskset -c $i iperf3 -s -p $((5201 + i * 2)) > /dev/null 2>&1 &
done

Start the script using the previous CPU range (leave 1 core as a buffer):

First Pod Console

root@testapp-performance-567cfdbd4b-7xzxm:/# chmod +x iperf_server.sh
root@testapp-performance-567cfdbd4b-7xzxm:/# ./iperf_server.sh 28-50
Running iperf3 server on core 28
Running iperf3 server on core 29

...
...
Running iperf3 server on core 49
Running iperf3 server on core 50

root@testapp-performance-567cfdbd4b-7xzxm:/# ps -ef | grep iperf3
root          39       1  0 13:57 pts/1    00:00:00 iperf3 -s -p 5257
root          40       1  0 13:57 pts/1    00:00:00 iperf3 -s -p 5259
...
...
root          60       1  0 13:57 pts/1    00:00:00 iperf3 -s -p 5299
root          61       1  0 13:57 pts/1    00:00:00 iperf3 -s -p 5301

Connect to the second pod:

Jump Node Console

kubectl exec -it testapp-performance-567cfdbd4b-m9vzc -- bash

Follow the previously displayed method to identify the CPU cores the second pod is running on.

Use the following script to start multiple iperf3 clients that will connect to each iperf3 server in the first pod:

The script receives 3 parameters: server IP to connect to, the cores it will spawn the iperf3 processes to, and the duration the iperf3 test will run. Make sure to pass all 3 when initiating the script and providing the CPU cores as a range (28-50 in this example).
jq and bc should be installed on the pod to properly run it.

iperf_client.sh

Bash

#!/bin/bash

# IP address of the server where iperf3 servers are running
SERVER_IP=$1  # Change to your server's IP

# Cores to bind the iperf3 client processes to
CORES=$2

# Duration to run the iperf3 test
DUR=$3

# Variable to accumulate the total bandwidth in Gbit/sec
total_bandwidth_Gbit=0

# Calculate the first_core and last_core to provide the CPU range
first_core=$(echo $CORES | cut -d "-" -f1)
last_core=$(echo $CORES | cut -d "-" -f2)

# Array to store the PIDs of background tasks
pids=()

# Loop over the ports (5201 + i*2) for i in the given CPU range
for i in $(seq $first_core $last_core); do
    port=$((5201 + i * 2))
    cpu_core=$i  # Assign CPU core based on the value of i
    output_file="iperf3_client_results_$port.log"

    # Run the iperf3 client in the background with CPU core binding
    timeout $(( DUR +5 )) taskset -c $cpu_core iperf3 -c $SERVER_IP -p $port -t $DUR -Z -J > $output_file &
    pid=$!
    pids+=("$pid")
done

# Wait for all background tasks to complete and check their status
for pid in "${pids[@]}"; do
    wait $pid
    if [[ $? -ne 0 ]]; then
        echo "Process with PID $pid failed or timed out."
    fi
done

# Summarize the results from each log file
echo "Summary of iperf3 client results:"
for i in $(seq $first_core $last_core); do
    port=$((5201 + i * 2))
    output_file="iperf3_client_results_$port.log"

    if [[ -f $output_file ]]; then
        echo "Results for port $port:"

        # Parse the results and print a summary
        bandwidth_bps=$(jq '.end.sum_received.bits_per_second' $output_file)

        if [[ -n $bandwidth_bps ]]; then
           # Convert bandwidth from bps to Gbit/sec
           bandwidth_Gbit=$(echo "scale=3; $bandwidth_bps / 1000000000" | bc)
           echo "  Bandwidth: $bandwidth_Gbit Gbit/sec"

           # Accumulate the bandwidth for the total summary
           total_bandwidth_Gbit=$(echo "scale=3; $total_bandwidth_Gbit + $bandwidth_Gbit" | bc)

           # Delete current log file
           rm $output_file
        else
           echo "No bandwidth data found in $output_file"
        fi

    else
        echo "No results found for port $port"
    fi
done

# Print the total bandwidth summary
echo "Total Bandwidth across all streams: $total_bandwidth_Gbit Gbit/sec"

Run the script and check the performance results:

Second Pod Console

root@testapp-performance-567cfdbd4b-m9vzc:/# chmod +x iperf_client.sh
root@testapp-performance-567cfdbd4b-m9vzc:/# ./iperf_client.sh 10.233.68.3 28-50 30
Summary of iperf3 client results:
Results for port 5257:
  Bandwidth: 10.299 Gbit/sec
Results for port 5259:
  Bandwidth: 14.417 Gbit/sec
Results for port 5261:
  Bandwidth: 26.517 Gbit/sec
Results for port 5263:
  Bandwidth: 14.869 Gbit/sec
Results for port 5265:
  Bandwidth: 6.053 Gbit/sec
Results for port 5267:
  Bandwidth: 29.648 Gbit/sec
Results for port 5269:
  Bandwidth: 16.708 Gbit/sec
Results for port 5271:
  Bandwidth: 5.970 Gbit/sec
Results for port 5273:
  Bandwidth: 10.411 Gbit/sec
Results for port 5275:
  Bandwidth: 31.203 Gbit/sec
Results for port 5277:
  Bandwidth: 14.025 Gbit/sec
Results for port 5279:
  Bandwidth: 30.534 Gbit/sec
Results for port 5281:
  Bandwidth: 13.452 Gbit/sec
Results for port 5283:
  Bandwidth: 6.014 Gbit/sec
Results for port 5285:
  Bandwidth: 25.819 Gbit/sec
Results for port 5287:
  Bandwidth: 26.472 Gbit/sec
Results for port 5289:
  Bandwidth: 5.940 Gbit/sec
Results for port 5291:
  Bandwidth: 10.068 Gbit/sec
Results for port 5293:
  Bandwidth: 5.981 Gbit/sec
Results for port 5295:
  Bandwidth: 13.352 Gbit/sec
Results for port 5297:
  Bandwidth: 29.973 Gbit/sec
Results for port 5299:
  Bandwidth: 13.464 Gbit/sec
Results for port 5301:
  Bandwidth: 31.478 Gbit/sec
Total Bandwidth across all streams: 392.667 Gbit/sec

Connecting to BlueMan Web Interface

As part of the DPF system installation, DTS and Blueman DPUServices were deployed.

DOCA Telemetry Service (DTS) collects data from built-in providers (data providers such as sysfs, ethtool and tc, and aggregation providers such as fluent_aggr and prometheus_aggr), and from external telemetry applications.

DOCA BlueMan runs in the DPU as a standalone web dashboard and consolidates all the basic information, health, and telemetry counters into a single interface.
All the information that BlueMan provides is gathered from the DOCA Telemetry Service (DTS).

To be able to log into BlueMan and view the local DTS instance data in a convenient way, the management IP address of the DPU should be entered to a web browser located in the same network as the DPU. In this RDG, it will be demonstrated by using RDP to connect to the jump node and opening a web browser in it (same as with MaaS, Firewall).

To find out the DPU management IP address in the 10.0.110.0/24 subnet, obtain the DPU names:

Jump Node Console

$ kubectl get dpus -n dpf-operator-system
NAME                   READY   PHASE   AGE
worker1-mt2404xz0c97   True    Ready   74m
worker2-mt2333xz0xvb   True    Ready   71m

Obtain the DPU management IPs:

Jump Node Console

$ kubectl get dpus -n dpf-operator-system -o json | jq '.items[].status.addresses[0].address' | cut -d '"' -f2
10.0.110.88
10.0.110.89

In the RDP session, open a web browser and enter https://<DPU_INTERNAL_IP>. A warning of self-signed certificate should appear; click accept the risk and proceed.

Afterwards it will open the login page:

The login credentials to use are the same pair used for the SSH connection to the DPU (ubuntu/ubuntu). However, login straight away won't work and an additional certificate exception in the browser has to be made.
Open another tab in the browser and enter https://<DPU_INTERNAL_IP>:10000. It will again prompt a warning of self-signed certificate; click accept the risk to add it to your browser exception list.
Return to the BlueMan login page, enter the credentials, and you should be able to login.

Authors

Guy Zilberman

Guy Zilberman is a solution architect at NVIDIA's Networking Solutions Labs, bringing extensive experience from several leadership roles in cloud computing. He specializes in designing and implementing solutions for cloud and containerized workloads, leveraging NVIDIA's advanced networking technologies. His work primarily focuses on open-source cloud infrastructure, with expertise in platforms such as Kubernetes (K8s) and OpenStack.

Last updated: May 31, 2026

Scope

Abbreviations and Acronyms

Introduction

References

Solution Architecture

Key Components and Technologies

Solution Design

Solution Logical Design

K8s Cluster Logical Design

Firewall Design

Software Stack Components

Bill of Materials

Deployment and Configuration

Node and Switch Definitions

Wiring

Hypervisor Node

K8s Worker Node

Fabric Configuration

Updating Cumulus Linux

Configuring the Cumulus Linux Switch

Host Configuration

Hypervisor Installation and Configuration

Hypervisor netplan configuration

Hypervisor Console

Prepare Infrastructure Servers

Firewall VM - pfSense Installation and Interface Configuration

Jump VM

Firewall VM – Web Configuration

MaaS VM

K8s Master VMs

Provision Master VMs and Worker Nodes Using MaaS

Master VMs

Install virsh and Set Up SSH Access

Get MAC Addresses of the Master VMs

MaaS Console

MaaS Console

Add Master VMs to MaaS

Configure OVS Bridges on Master VMs

Deploy Master VMs Using Cloud-Init

Verify Deployment

Finalize Setup

Master1 Console

Worker Nodes

Create Worker Machines in MaaS

Create a Tag for Kernel Parameters

Adjust Network Settings

Deploy Worker Nodes Using Cloud-Init

Verify Deployment

Finalize Deployment

Jump Node Console

K8s Cluster Deployment and Configuration

Kubespray Deployment and Configuration

Deploying Cluster Using Kubespray Ansible Playbook

K8s Deployment Verification

DPF Installation

Software Prerequisites and Required Variables

CNI Installation

DPF Operator Installation

Create Storage Required by the DPF Operator

Additional Dependencies

DPF Operator Deployment

DPF System Installation

Install Components to Enable Accelerated CNI Nodes

Install Multus and SRIOV Network Operator using NVIDIA Network Operator

Install OVN Kubernetes resource injection webhook

Apply NicClusterPolicy and SriovNetworkNodePolicy

DPU Provisioning and Service Installation

K8s Cluster Scale-out

Add Worker Nodes to the Cluster

Verification

Infrastructure Latency & Bandwidth Validation

Performance Tests

RoCE Latency Test

iPerf TCP Bandwidth Test

Connecting to BlueMan Web Interface

Authors

Install `virsh` and Set Up SSH Access