kolla-ansible/contrib/dev/vagrant/Vagrantfile.custom.example

103 lines
2.5 KiB
Plaintext
Raw Normal View History

# -*- mode: ruby -*-
# vi: set ft=ruby :
# This file is an example of Vagrant configuration.
# Copy it to Vagrantfile.custom and configure it to your liking to customize
# the Vagrant deployment. The Vagrantfile.custom file is sourced by the
# Vagrantfile, it has to be valid ruby code.
# Either libvirt or virtualbox
# PROVIDER = "libvirt"
# Either centos or ubuntu
# DISTRO = "centos"
# The libvirt graphics_ip used for each guest. Only applies if PROVIDER
# is libvirt.
# GRAPHICSIP = "127.0.0.1"
# Provisioning other boxes than the default ones may therefore
# require changes to bootstrap.sh.
# PROVISION_SCRIPT = "bootstrap.sh"
# The bootstrap.sh provisioning script requires CentOS or Ubuntu; see below
# for the supported versions for each provider.
#
# kolla and kolla-ansible will be mounted in $HOME and the username depends
# on the image.
#
# PROVIDER_DEFAULTS = {
# libvirt: {
# centos: {
# base_image: "centos/7",
# bridge_interface: "virbr0",
# vagrant_shared_folder: "/home/vagrant/sync",
# sync_method: "nfs",
# username: "vagrant"
# }
# },
# virtualbox: {
# centos: {
# base_image: "centos/7",
# bridge_interface: "wlp3s0b1",
# vagrant_shared_folder: "/home/vagrant/sync",
# sync_method: "virtualbox",
# username: "vagrant"
# },
# ubuntu: {
# base_image: "ubuntu/xenial64",
# bridge_interface: "wlp3s0b1",
# vagrant_shared_folder: "/home/vagrant/sync",
# sync_method: "virtualbox",
# username: "ubuntu"
# }
# }
# }
Attach external NIC to a NAT-Network if on Wi-Fi On computers with wi-fi adapters, promiscuous mode on the VirtualBox (or maybe other hypervisors as well) NICs does not work, which means the default way of connecting the Neutron external interface to a bridged adapter, will not allow communication to and from the Nova VMs over floating IPs with any computer on the external network (except the host computer) or with the wi-fi router. This means no ability to connect to the Nova VMs and no internet access inside the Nova VMs. According to VirtualBox documentation (excerpt): "Bridging to a wireless interface is done differently from bridging to a wired interface, because most wireless adapters do not support promiscuous mode. All traffic has to use the MAC address of the host’s wireless adapter, and therefore VirtualBox needs to replace the source MAC address in the Ethernet header of an outgoing packet to make sure the reply will be sent to the host interface. When VirtualBox sees an incoming packet with a destination IP address that belongs to one of the virtual machine adapters it replaces the destination MAC address in the Ethernet header with the VM adapter’s MAC address and passes it on. VirtualBox examines ARP and DHCP packets in order to learn the IP addresses of virtual machines." To fix this issue, a new flag has been introduced: WIFI. If true, the default Vagrant public network is not created anymore. Instead, the 3rd NIC will be connected to a NAT-Network named OSNetwork. The NAT-Network has a virtual gateway, which will be used to communicate with the external physical wi-fi router. Since Vagrant does not have a high-level mechanism to attach an adapter to a NAT-Network, the code uses the low-level Vagrant construct vm.customize which makes it provider specific. Promiscuous mode is now activated by default on the 3rd NIC. The WIFI flag is false by default. This commit only addresses VirtualBox, and it is currently unknown if the problem described and fixed in this commit is present in other hypervisors. DocImpact Closes-Bug: #1558766 Change-Id: I0b4dbbc562d87191b2179f47b634cdd6f6361a5e Signed-off-by: Andrei-Lucian Șerb <lucian.serb@icloud.com>
2016-03-17 23:30:32 +00:00
# Whether the host network adapter is Wi-Fi.
# On VirtualBox, the user must first manually create a NAT-Network
# named OSNetwork. The default network CIDR must be changed.
# The Neutron external interface will be connected to this Network.
# WIFI = false
# Whether to do Multi-node or All-in-One deployment
# MULTINODE = false
# The following is only used when deploying in Multi-nodes
# NUMBER_OF_CONTROL_NODES = 3
# NUMBER_OF_COMPUTE_NODES = 1
# NUMBER_OF_STORAGE_NODES = 1
# NUMBER_OF_NETWORK_NODES = 1
# NUMBER_OF_MONITOR_NODES = 1
# NODE_SETTINGS = {
# aio: {
# cpus: 4,
# memory: 4096
# },
# operator: {
# cpus: 1,
# memory: 1024
# },
# control: {
# cpus: 1,
# memory: 2048
# },
# compute: {
# cpus: 1,
# memory: 1024
# },
# storage: {
# cpus: 1,
# memory: 1024
# },
# network: {
# cpus: 1,
# memory: 1024
# },
# monitor: {
# cpus: 1,
# memory: 1024
# }
# }