driver-installation.yaml

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: &name install-mlx-driver
  labels:
    app: *name
spec:
  selector:
    matchLabels:
      app: *name
  template:
    metadata:
      labels:
        app: *name
    spec:
      affinity:
        nodeAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
            nodeSelectorTerms:
            - matchExpressions:
              - key: node.kubernetes.io/instance-type
                operator: In
                values:
                - Standard_H16mr
                - Standard_H16mr_Promo
                - Standard_H16r
                - Standard_H16r_Promo
                - Standard_H8_Promo
                - Standard_HB120-16rs_v2
                - Standard_HB120-32rs_v2
                - Standard_HB120-64rs_v2
                - Standard_HB120-96rs_v2
                - Standard_HB120rs_v2
                - Standard_HB120-16rs_v3
                - Standard_HB120-32rs_v3
                - Standard_HB120-64rs_v3
                - Standard_HB120-96rs_v3
                - Standard_HB120rs_v3
                - Standard_HB60-15rs
                - Standard_HB60-30rs
                - Standard_HB60-45rs
                - Standard_HB60rs
                - Standard_NC24rs_v2
                - Standard_ND96amsr_A100_v4
                - Standard_ND96asr_v4
                - Standard_HC44rs
                - standard_hb120-96rs_v3

      hostNetwork: true
      hostPID: true
      containers:
      - image: ghcr.io/converged-computing/aks-infiniband-install:ubuntu-22.04
        imagePullPolicy: Always
        name: *name
        command: ["/entrypoint.sh"] 
        args: ["installonly"]
        resources:
          requests:
            {}
          limits:
            {}
        securityContext:
          privileged: true
        volumeMounts:
        - name: actions
          mountPath: "/opt/actions"
        - name: hostmount
          mountPath: "/mnt/actions"
        - name: debs
          mountPath: "/mnt/debs"
      volumes:
      - name: debs
        hostPath:
          path: /opt/debs
          type: DirectoryOrCreate
      - name: hostmount
        hostPath:
          path: /opt/actions
          type: DirectoryOrCreate
      - name: actions
        configMap:
          name: install-driver-actions
---
apiVersion: v1
kind: ConfigMap
metadata:
  name: install-driver-actions
  labels:
    app: install-mlx-driver
data:
  installonly: |
    #!/usr/bin/env bash
    set -eo pipefail

    umount /mnt/install || true
    cd /opt/debs
    mkdir -p /mnt/install
    mount -o ro,loop ./MLNX_OFED_LINUX-24.04-0.7.0.0-ubuntu22.04-x86_64.iso /mnt/install
    /mnt/install/mlnxofedinstall --force
    /etc/init.d/openibd restart
    
    # Read in ip links, rename to ib0, and start
    # ip link set ibP257s75549 name ib0    
    # ip link set ib0 up
    wget https://raw.githubusercontent.com/converged-computing/aks-infiniband-install/main/ubuntu22.04/parse-links.py
    python3 parse-links.py
    ip link
    check_ib="$(ip link | grep ib0)"
    if [ -z "$check_ib" ];
    then
        echo "driver installation failed"
        exit 1
    fi
    echo "installation completed."
    exit 0