INSTALL.md

Install

# check for hardware support
LC_ALL=C lscpu | grep Virtualization
# check kernel support
zgrep CONFIG_KVM /proc/config.gz
# Debian packages
sudo apt -y install \
       bridge-utils \
       clustershell \
       libguestfs-tools \
       libnss-libvirt \
       libosinfo-bin \
       libvirt-clients \
       libvirt-daemon-system \
       libvirt-dev \
       ovmf \
       qemu-system-x86 \
       qemu-utils \
       ruby \
       virt-manager \
       virt-viewer \
       virtinst
# Fedora/CentOS packages
sudo dnf -y install \
      @virtualization \
      clustershell
# ArchLinux packages
sudo pacman -Sy --noconfirm \
       bridge-utils \
       cpio \
       dnsmasq \
       ebtables \
       iptables \
       libvirt \
       openbsd-netcat \
       python-pip \
       ruby \
       vde2 \
       virt-install \
       virt-manager \
       virt-viewer \
       qemu \
       qemu-block-rbd
sudo pip install ClusterShell

Enable your user to manage virtual machines, re-login to activate these group rights:

>>> sudo usermod -a -G libvirt,kvm `id -un`      

Configure the libvirt service to run with your user ID (here illustrated with ID jdow):

>>> sudo grep -e '^user' -e '^group' /etc/libvirt/qemu.conf
user = "jdow"
group = "jdow"
# restart the Libvirt service daemon
>>> sudo systemctl restart libvirtd

Libvirt NSS Module

For host access to guests on non-isolated, bridged networks, enable the libvirt NSS module:

>>> grep hosts /etc/nsswitch.conf
hosts: files libvirt libvirt_guest ...
>>> getent hosts lxdev01
10.1.1.30       lxdev01

Commands like ssh should work with VM instance names now:

>>> vm shadow ${image} lxdev01
>>> ssh -i $(vm path lxdev01)/keys/id_rsa devops@lxdev01
Last login: Tue Jul  9 11:42:57 2019 from gateway
[devops@lxdev01 ~]$