Preparing for Installation

KVM CPU/IO Requirements

OS and Hardware Requirements

Component Requirement
Supported OS versions Debian 11, Rocky Linux 8
CPU Intel i7 or higher, with a minimum of 4 physical cores per instance
Memory Minimum 8GB per instance
Disk 32 GB per instance
NICs
  • 1Gbps: Intel I350
  • 10Gbps: Intel X520, X550, X710, XL710, and X722
Optic Connectors Intel-compatible
Core switch (for certain topologies) RUCKUS ICX switch family or a third-party switch

Required Software Packages

Ensure the latest software packages are installed for upgrade and bug fixes.

OS Versions

OS Toolkit
Debian 11 libvirt-daemon-system, ovmf, qemu-system-x86, virtinst
Rocky Linux 8 libvirt, edk2-ovmf, qemu-kvm, virt-install

Install Required Packages

OS Command
Debian 11
sudo apt-get update
sudo apt-get install <PKG_NAME> [<PKG_NAME> ...]
Rocky Linux 8
sudo dnf update
sudo dnf install <PKG_NAME> [<PKG_NAME> ...]

libvirt

  1. Check if libvirtd is running:
    pgrep libvirtd

    If nothing is returned by pgrep, start the libvirtd service manually:

    sudo systemctl start libvirtd.service

  2. Make the libvirtd service automatically start at host boot:
    sudo systemctl enable libvirtd.service
  3. Disable libvirtd AppArmor profile.

    Disabling the Profile

    OS Command
    Debian 11

    By default, AppArmor is enabled for Debian 11.

    Unload libvirtd AppArmor profile.

    sudo apparmor_parser -R /etc/apparmor.d/usr.sbin.libvirtd
    sudo ln -s /etc/apparmor.d/usr.sbin.libvirtd /etc/apparmor.d/disable/
    Rocky Linux 8 NA

Networking

RUCKUS Virtual Edge needs at least one network interface to function, either a PCI passthrough or a virtio interface.

PCI Passthrough Support

Skip this section if there is no requirement to use NIC PCI passthrough.

  1. Enable IOMMU support.
    Note: You can check this by checking the status of the /sys/kernel/iommu_groups directory. If this directory exists and has content, IOMMU is likely enabled.

    Enabling IOMMU Support

    Debian 11 Rocky Linux 8
    Debian 11 Linux kernel has
    CONFIG_INTEL_IOMMU_DEFAULT_ON_INTGPU_OFF=y
    Note: No manual enabling required.
    If IOMMU is not enabled or you need to adjust its settings, you will usually need to modify your GRUB (or GRUB2) configuration, as follows: . Here’s how you can do it:
    • Make a backup of the original grub.cfg file.sudo grub2-mkconfig -o <GRUB_CFG_PATH>
      Note: The <GRUB_CFG_PATH> depends on the host OS installation type: legacy BIOS boot or UEFI boot. Find the original file grub.cfg in /boot and make a backup.
    • Edit the GRUB configuration file using a text editor (such as nano or vi):
      sudo vi /etc/default/grub
    • Locate the GRUB_CMDLINE_LINUX parameter and append intel_iommu=on.
      GRUB_CMDLINE_LINUX="... intel_iommu=on"
    • Save the file and exit the text editor.
    • After making changes to the GRUB configuration file, update GRUB to apply the changes.
  2. Auto-insert the vfio-pci driver at boot.
    echo 'vfio-pci' | sudo tee /etc/modules-load.d/vfio-pci.conf
  3. Enable user access to the vfio device files.
    Note: Ensure you log in using the user account that is going to execute the installation process and not the 'root' account.

    For example, if the user $USER is the one that is going to execute the .bin file.

    echo 'SUBSYSTEM=="vfio", OWNER="root", GROUP="kvm"' | sudo tee /etc/udev/rules.d/10-qemu-hw-users.rules
    sudo usermod -aG kvm "$USER"

  4. Edit /etc/security/limits.conf, for example, sudo vi /etc/security/limits.conf and add the two lines for the normal user that is going to execute the .bin file to unlimit memlock.
    foo soft memlock unlimited
    foo hard memlock unlimited

    Note: Replace foo with your account name (not the root) that is going to execute the installation process.

  5. Reboot the host after completing the preparation for PCI passthrough and verify:
    • If IOMMU is enabled by checking the /sys/kernel/iommu_groups directory again.
    • If vfio-pci is loaded by entering the: lsmod | grep vfio_pci command again.

Virtual Network

  1. Skip this section if there is no requirement to use virtio virtual network interfaces.

    Use one of the following methods to create bridges for RUCKUS Virtual Edge.

    • Using libvirt commands and XML-based configurations.

      Example: libvirt default network

      Debian 11 Rocky Linux 8
      sudo virsh net-define /etc/libvirt/qemu/networks/default.xml
      sudo virsh net-start default
      sudo virsh net-autostart default

      A bridge, virbr0, with NAT/DHCP services is created and is ready for virtio interfaces to attach.

      The configuration is enabled by default when libvirt is installed and started.

      A bridge, virbr0, with NAT/DHCP services is created and ready for virtio interfaces to attach.

    • Manually create a bridge using the following commands:
      sudo ip link add name vse-br0 type bridge
      sudo ip link set dev vse-br0 up

      Add some host network interfaces, for example, eth0 to the bridge

      sudo ip link set dev eth0 master vse-br0

      Add NAT/DHCP services for devices that connect to the bridge.

      A bridge named vse-br0 is created and ready for virtio interfaces to attach.

  2. Enable permission to access bridges.

    This applies to both methods listed in the previous step.

    Permission to Access Bridges

    Debian 11 Rocky Linux 8
    sudo mkdir -p /etc/qemu
    sudo vi /etc/qemu/bridge.conf
    sudo vi /etc/qemu-kvm/bridge.conf
  3. Provide permission to create network tunnel devices.

    Permission to Create Network Tunnel Devices

    Debian 11 Rocky Linux 8
    sudo chmodu+s /usr/lib/qemu/qemu-bridge-helper By default, /usr/libexec/qemu-bridge-helper has already been set with setuid.