QuickStart OpenStack, Mitaka Deployment (1)--pre-environment set-up preparation __openstack

Reference: Official configuration document (http://docs.openstack.org/mitaka/install-guide-rdo/overview.html) Preface

What OpenStack is. Official explanation: OpenStack is a cloud operating system this controls large pools of compute, storage, and networking resources Throug Hout a datacenter, all managed through a dashboard this gives administrators control while empowering their users to provi Sion resources through a web interface. The general meaning is that OpenStack is a cloud operating system with a lot of computing power, storage, and even the entire network of data centers. It can be managed by dashboard this web management interface.
I am also just a openstack beginner, write this sequence of blog is the most important purpose is to share the record of their learning growth process, and secondly, I hope that we can learn in the least resource planning to achieve a openstack without brain, And then through the deployment process to preliminary understanding of OpenStack.
This article takes the Unistedstack public cloud as an example, chooses other cloud server or the virtual machine The child also not to be too tangled, the principle is same.
This article is only the beginning, more simple and rough. But if the early preparation is sufficient, the later will save a lot of trouble. A lot of people start to encounter a variety of problems, such as network impassability, service has been started normally but can not grasp the package. So in order to avoid these problems, we need to start by emptying the firewall off. OpenStack Overall Architecture

We can get an overview of OpenStack's overall architecture and related services through the official overview. Official Overview:

Overall Architecture:

OpenStack Components:

Follow-up will be introduced for each component, now only need to preliminary understand the functions of each component can be.
Compute (Nova) Computing Services
Identity Service (KEYSTONE) certification services
Image Service (glance) mirroring services
Networking (neutron/quantum) network services
Dashboard (Horizon) instrument panel
Object Storage (Swift) objects Store
Block Storage (cinder) chunk storage
Orchestration (Heat) choreography
Telemetry (ceilometer) monitoring
DB Service (TROVE) database services
Processing (Sahara) data processing network extension complement map

Although most environments include Identity, Image service, Compute, at least one networking service, and the dashboard, T He Object Storage service can operate independently. If Your use case only involves Object Storage, can skip to Object Storage service after configuring the appropriate No Des for it. However, the dashboard requires at least the Image service, Compute, and networking.
According to the official advice, I feel that if you do not consider the performance and security, the entire network extension map can be designed as a solution, I intend to put neutron into the controller node (this is very bad, will lead to controller node directly exposed to the public online), But that only requires three nodes, and my goal is to implement OpenStack's basic functionality with minimal resources. If the computer resources are full of students I recommend the use of traditional two, because OpenStack is a coupled architecture, so the deployment approach is the same, just need to install the corresponding services to your extension map in the corresponding node. programme I:

As shown, I used three network segments, one is the management management network, one is the instance instance network, the other is connected with the external network uses the public network. As for what these networks are for, they will be introduced to neutron. So at this stage everyone needs to prepare three network segment can. Where the controller node requires three network cards, compute nodes need two network cards, storage node only need a network card. Programme II:

This is a more classic collocation scheme, interested friends can try it on their own. Environmental Preparedness because of the company's free, I'm using the Uos cloud:
Unistedstack Console Portal: https://console.ustack.com
The following steps can also be manipulated on VMware. To create three virtual machines, it is recommended to use the official recommended configuration. 4 kernel +8g memory +20g hard disk, dual core can also, but not enough memory, you can add swap virtual partition, will be used.
According to the official documentation requirements, the system uses 64-bit centos7.0 uniformly. Say you've seen 32-bit centos7. Cold face ... Create network, NIC, bind static IP, change host name. I configure the following
managesubnet:192.168.0.0/24:
Controller.example.com:192.168.0.17/24
Compute.example.com:192.168.0.18/24
Compute.example.com:192.168.0.19/24
instancesubnet:172.16.0.0/24:
Controller.example.com:172.16.0.17/24
Compute.example.com:172.16.0.18/24
publicsubnet:10.10.10.0/24:
Controller.example.com:10.10.10.17/24

Configuring an NTP time server
To install on each of the machines:
Yum Install Chrony-y
Modify the/etc/chrony.conf configuration file on the controller node, comment out the other server, reassign the NTP server IP, and allow host access to the 192.168.0.0/24 network segment. Save the exit, restart the service, and set the boot up.
Vim/etc/chrony.conf

Server 192.168.0.17 iburst
 allow 192.168.0.0/24
 systemctl enable Chronyd.service
 systemctl start Chronyd.service

Do the same with other nodes
Vim/etc/chrony.conf

Server 192.168.0.17 iburst
systemctl enable Chronyd.service
systemctl start Chronyd.service

To configure DNS domain name resolution servers:
Lazy, so I used a static, modified/etc/hosts for each node
Vim/etc/hosts

192.168.0.17 controller.example.com
192.168.0.18 compute.example.com
192.168.0.19 storage.example.com
172.16.0.17 controller.example.com
172.16.0.18 compute.example.com

You can ping each other's host name to do the test if the ping does not pass, the configuration file is not wrong, you can check your firewall. The official suggestion is to turn off the firewall and SELinux.

Systemctl Mask Iptables.service
Systemctl Mask Ip6tables.service
Systemctl Mask Ebtables.service
Systemctl Mask Firewalld.service
Setenforce 0

Edit SELinux file, set to boot default shutdown
Vim/etc/selinux/config
selinux=disabled install Openstackpackage, each machine is executed once:
Install yum-plugin-priorities packs to prevent high-priority software from being overwritten by Low-priority software.

Install Epel extension Yum Source, this package will be updated to use the previous address to find the latest package:

Installing the Openstackyum source will also update:

Update system

Yum Upgrade-y

Reboot system

Reboot

Install Openstackclient

Yum Install Python-openstackclient-y

Installation of OpenStack SELinux

Yum Install Openstack-selinux

installing. This article concludes
have a nice day!