Tag/tag.png

Unsupported Version
This article applies to an unsupported version of Ubuntu. More info...

Tag/tag.png

Candidate for Deletion
This article may not be appropriate for this wiki, and may be deleted. More info...

Introduction

Warning /!\ This is instruction of Ubuntu 11.10. For Ubuntu 12.04 LTS, please see UbuntuCloudInfrastructure.

Ubuntu Cloud Infrastructure, which is the successor of Ubuntu Enterprise Cloud, is a ready to deploy Infrastructure-as-a-Service (IaaS) based on OpenStack. It is making it's official entry into Ubuntu 11.10.

Installation

There are multiple ways to install Ubuntu Cloud Infrastructure.

  • Ubuntu Cloud Live Image: a single node all in one cloud booting from USB
  • Ubuntu Cloud Infrastructure with Orchestra and Juju: our recommended method
  • Install Ubuntu Cloud Infrastructure from packages: the hard way

Ubuntu Cloud Live Image

  1. Download the image from http://cdimage.ubuntu.com/ubuntu-cloud-live/releases/11.10/

  2. Create a USB stick using USB Creator or some other tools as described here

  3. Boot from the created USB stick

Your Ubuntu Cloud Infrastructure is ready to be used. See the section Use your cloud for more details.

Ubuntu Cloud Infrastructure with Orchestra and Juju

This install is done in 3 steps:

  1. Install the Orchestra server
  2. Install Juju
  3. Deploy Ubuntu Cloud Infrastructure with Juju

Which are described below.

Requirements

  • It should be assumed that the user deploying this setup has all nodes with two network interfaces linked to two seperate physical networks. Deploying with a single network is possible with some hacking and workarounds, see original document for that.
  • It's also assumed the user has sufficient number of nodes available in Orchestra. The minimum deployment outlined below requires a minimum of 6 machines including the juju bootstrap node.

Installing the Orchestra server

  1. Follow the usual steps to install Ubuntu Server

  2. Once done, install the package named "ubuntu-orchestra-server"
    sudo apt-get update
    sudo apt-get install ubuntu-orchestra-server -y
  3. Answer some questions about your networking setup

Your basic Orchestra setup is now done. To learn more about Orchestra see Dustin Kirland's excellent blog post on the subject.

Install Juju

This can be done on the same server you have deployed Orchestra onto.

1. Installing juju

Starting Oneiric 11.10, juju is available in universe

sudo apt-get install juju

to install a daily build from PPA or to install from source code, check out the juju documentation

2. Configuring the environment

sudo mkdir -p ~/.juju
sudo vim ~/.juju/environments.yaml

And we copy the following

juju: environments
environments:
  orchestra:
    type: orchestra
    # Specify the orchestra server's IP address
    orchestra-server: 10.1.2.3
    # Specify storage. In this case we are using webdav installed by orchestra.
    storage-url: http://10.1.2.3/webdav
    # Specify cobbler's usr/pass
    orchestra-user: cobbler
    orchestra-pass: cobbler
    admin-secret: fooooo
    # Branch from where we will install ensemble
    # juju-branch: lp:juju
    # Mangement classes
    acquired-mgmt-class: orchestra-juju-acquired
    available-mgmt-class: orchestra-juju-available

3. Creating SSH keys

Ensemble requires SSH keys to be able to access the deployed nodes. In case those keys do not exist, then we have to create them before we bootstrap our environment:

ssh-keygen -t rsa

4. Bootstrapping the environment

After the configuration done above, we bootstrapped the environment:

juju bootstrap

This process will select any of the machines available to orchestra to be the zookeeper node. In this example, the machine's name is mabolo

ubuntu@orchtrasrvr:~$ juju status
2011-09-08 18:02:22,148 INFO Connecting to environment.
machines:
  0: {dns-name: mabolo.lan, instance-id: MTMxNDMwNzI0OS40ODkyODE2MzAuODE5ODY}

NOTE: Note that after executing the command we had to manually start/restart the system to PXE boot it. This can be automated if cobbler is configured to power-on the servers

Deploying Ubuntu Cloud Infrastructure with Juju

1. Clone Charms

Create local branches of the following Juju charms in ~/charms

ubuntu@orchestrasrvr:~$ mkdir -p charms/oneiric
ubuntu@orchestrasrvr:~$ cd charms/
ubuntu@orchestrasrvr:~/charms$ bzr branch lp:charm/rabbitmq-server oneiric/rabbitmq-server
ubuntu@orchestrasrvr:~/charms$ bzr branch lp:charm/mysql oneiric/mysql
ubuntu@orchestrasrvr:~/charms$ bzr branch lp:charm/nova-compute oneiric/nova-compute
ubuntu@orchestrasrvr:~/charms$ bzr branch lp:charm/nova-cloud-controller oneiric/nova-cloud-controller
ubuntu@orchestrasrvr:~/charms$ bzr branch lp:charm/glance oneiric/glance

2. Set Global Configuration

We'll need to set some global configuration for the installation. Create a file /home/ubuntu/charms/openstack.yaml:

nova-cloud-controller:
  nova-release: distro
  nova-config: /etc/nova/nova.conf
  db-user: nova
  nova-db: nova
  rabbit-user: nova
  rabbit-vhost: nova
  network-manager: FlatDHCPManager
  bridge-interface: br100
nova-compute:
  nova-release: distro
  nova-config: /etc/nova/nova.conf
  db-user: nova
  nova-db: nova
  rabbit-user: nova
  rabbit-vhost: nova
  flat-interface: eth1
  virt-type: kvm
glance:
  glance-release: distro
  registry-config: /etc/glance/glance-registry.conf
  api-config: /etc/glance/glance-api.conf
  db-user: glance
  nova-db: nova

Note: Most of these are defaults. The only option specific to our example deployment is 'flat-interface: eth1'. This will configure the virtual network on eth1. Later, nova will create a virtual network on these bridges for instances.

4. Deploy Services

After branching the necessary charms to /home/ubuntu/charms/ and setting our configuration at /home/ubuntu/charms/openstack.yaml, we can begin deploying the services.

MySQL and RabbitMQ

First, deploy the MySQL and RabbitMQ services.

ubuntu@orchestrasrvr:~$ export REPO=~/charms
ubuntu@orchestrasrvr:~$ export CONFIG=~/charms/openstack.yaml
ubuntu@orchestrasrvr:~$ juju deploy --repository=$REPO local:mysql
ubuntu@orchestrasrvr:~$ juju deploy --repository=$REPO local:rabbitmq

Nova and Glance

Deploy nova-cloud-controller, nova-compute and glance to their own nodes:

ubuntu@orchestrasrvr:~$ juju deploy --repository=$REPO --config=$CONFIG local:nova-cloud-controller
ubuntu@orchestrasrvr:~$ juju deploy --repository=$REPO --config=$CONFIG local:nova-compute
ubuntu@orchestrasrvr:~$ juju deploy --repository=$REPO --config=$CONFIG local:glance

6. Add Relations

Now we can add the necessary relations between the services.

Give nova-cloud-controller, nova-compute and glance access to the database:

ubuntu@orchestrasrvr:~$ juju add-relation mysql nova-cloud-controller
ubuntu@orchestrasrvr:~$ juju add-relation mysql nova-compute
ubuntu@orchestrasrvr:~$ juju add-relation mysql glance

NOTE: The first of these 3 relations runs the initial database migration. It's a good idea to allow it time to complete entirely before adding the others.

Give the Nova components access to the messaging queue:

ubuntu@orchestrasrvr:~$ juju add-relation rabbitmq nova-compute
ubuntu@orchestrasrvr:~$ juju add-relation rabbitmq nova-cloud-controller

Configure the compute node for the network manager specified in our config.

ubuntu@orchestrasrvr:~$ juju add-relation nova-cloud-controller:nova-network nova-compute:nova-network

Let both the cloud controller and the compute nodes know where to fetch virtual machine images:

ubuntu@orchestrasrvr:~$ juju add-relation glance:image-service nova-cloud-controller:image-service
ubuntu@orchestrasrvr:~$ juju add-relation glance:image-service nova-compute:image-service

Your Ubuntu Cloud Infrastructure is ready to be used. See the section Use your cloud for more details.

Ubuntu Cloud Infrastructure from packages

This is a more complex process which is described in great details in the OpenStack Beginner’s Guide V2.0 for Ubuntu 11.10 written by CSS Corp.

Use Your Cloud

At this point, the Openstack cloud has been deployed and should be functioning. If you login to the nova-cloud-controller node (as reported by 'juju status'), you can create the first nova users, projects, networks and export credentials to be used with euca-tools or other API tools.

ubuntu@orchestrasrvr:~$ juju ssh 2
ubuntu@marula:~$ sudo nova-manage user admin admin
ubuntu@marula:~$ sudo nova-manage project create novaproject admin
ubuntu@marula:~$ sudo nova-manage network create \
                  --label=novanet \
                  --fixed_range_v4=10.0.0.0/24 \
                  --num_networks=1 \
                  --network_size=255 \
                  --bridge=br100 \
                  --bridge_interface=eth1 \
                  --multi_host=T \
                  --project_id=novaproject
ubuntu@marula:~$ sudo nova-manage project zipfile novaproject admin

Now we can unzip our credentials and begin using the cloud.

ubuntu@marula:~$ unzip nova.zip
ubuntu@marula:~$ . novarc
ubuntu@marula:~$ wget http://uec-images.ubuntu.com/server/server/releases/oneiric/beta-1/ubuntu-11.10-beta1-server-cloudimg-amd64.tar.gz
ubuntu@marula:~$ uec-publish-tarball ./ubuntu-11.10-beta1-server-cloudimg-amd64.tar.gz images
ubuntu@marula:~$ euca-add-keypair adam >adam.pk
ubuntu@marula:~$ euca-run-instances -k adam ami-00000002

We still have 1 extra machine. We can easily be added as a compute node through Juju to increase our resources:

ubuntu@orchestrasrvr:~$ juju add-unit nova-compute

See Also

UbuntuCloudInfrastructure11.10 (last edited 2017-09-20 01:25:56 by ckimes)