Preliminary check

  1. Make sure that the server is physical and supports virtualization at the CPU level. To do this, boot Linux on the server and run the command:

    grep -P 'vmx|svm' /proc/cpuinfo
    BASH

     If the reply to the command is not empty, then the CPU supports virtualization. Also make sure that the server BIOS has options enabled to allow hardware virtualization.

  2. Make sure that the server hardware meets the requirements. A server being connected to a cluster with the Second network interface option requires at least two physical network adapters.

Operating system

  1. Install one of the supported operating systems on the server. When installing, use the official unmodified distribution of the OS in the minimal edition without third-party repositories. Do not install any additional software or system services on the node server.

    Note

    Do not install the VMmanager platform on the node server. It must be installed on a separate server.

  2. When partitioning the disk, allocate the main part of the disk space for the root file system. If you plan to use LVM virtual disk storage on the server, the part of the server disk space intended for storage should be left free so that it can be used to create a group of LVM volumes in the future.
  3. For servers located in OVH data centers: When installing the OS through the OVH control panel, enable the Install original kernel option. Use the original OS kernel for proper operation of the cluster servers in an OVH data center.
  4. Install OS updates:

    1. Alma Linux 8:

      dnf update
      BASH

    2. CentOS 7:

      yum update
      BASH

    3. Ubuntu 20.04:

      apt update && apt dist-upgrade
      BASH

Network settings

  1. Make sure that the server has a unique hostname.
  2. For clusters with "Switching" or "Routing" network configuration: configure a public IP address on the server and make sure it is properly routed (has access to the Internet). Make sure that the address is configured statically (without using DHCP) on the physical or VLAN interface of the server. An attempt to use an address assigned dynamically by DHCP will result in errors when connecting the server to the VMmanager cluster.
  3. Make sure that incoming network connections on the following ports are allowed:
    1. SSH: port 22/tcp (standard) or another port to which the sshd service is bound.
    2. If you plan to use virtual networks (VxLANs) in the cluster: 179/tcp, 4789/udp.
    3. VNC: range 5900-6900/tcp.  If the VNC is accessed only through the VMmanager web interface, this port range must be open to the network that joins the cluster nodes.
    4. TLS port of the libvirtd service: 16514/tcp.
    5. Ports used by libvirtd when migrating virtual machines between cluster nodes: range 49152-49261/tcp.