This is a blogpost about how I setup my test virtual machines. The seasoned sysadmin and DBA will notice that the techniques used here are perfectly usable for real production environments. The most important thing is there is no need to download or stage any software for installing the virtual machine, everything is downloaded when needed during installation. Obviously this works best when you have got reasonable bandwidth available for connecting to the internet.
The main infrastructure software components of this setup are:
– Virtualbox as the virtualisation software.
– Ansible as the configuration and management engine.
– Vagrant as the virtualisation manager.
Installation (Mac OSX specific).
– Virtual box is installed by downloading and installing the installation image in the normal way.
– Ansible requires a few steps. Ansible relies on python.
First install pip using easy_install:
$ sudo easy_install pip
Then install ansible for your current user:
$ sudo pip install ansible --user
Now to pick up the ansible local install, add the following to .bash_profile:
– Vagrant is installed by downloading and installing the installation image in the normal way.
$ VBoxManage -v
$ ansible --version
config file = /Users/fritshoogland/.ansible.cfg
configured module search path = Default w/o overrides
$ vagrant version
Installed Version: 1.8.5
Latest Version: 1.8.5
You're running an up-to-date version of Vagrant!
Okay! Now all the software is installed, up to the next step!
SSH keys issue for vagrant (OSX specific).
For ssh certificate based authentication to work, the private key file needs to protected by mode 600 (rw-|—|—). Vagrant uses a ssh key to access the virtual machine. I am running the vagrant directory from a share, which forces the mode of the files to be 700. In order to workaround that issue, I setup my own keys in ~/.vagrant_ssh:
$ mkdir .vagrant_ssh
$ cd .vagrant_ssh
$ ssh-keygen -q -N "" -f id_rsa
Later on some settings are made in the Vagrantfile to use the private key just generated.
SSH keys for normal authentication (OSX specific).
The Ansible scripts copy your ssh public key and put it in the authorised_keys file of both the oracle and the root user. By putting your public key in the authorized_keys file of a user, the authentication part of logging on is done via the public key.
However, for this to work, you first need to have a private and public key pair. There are many ways of doing that, this is an example of that:
Check if you have a dsa key pair:
$ cd ~
$ find .ssh
In this case, as you can see, there are two id_dsa files, one with the extension “.pub”, which is the public key, and the other one without an extension is the private key. The private key should be kept private (at all times). If you don’t have id_dsa keys, set them up in the following way:
$ cd ~
$ mkdir .ssh
$ chmod 700 .ssh
$ cd .ssh
$ ssh-keygen -q -N "" -t dsa -f id_dsa
Setting the Virtualbox images directory.
Virtualbox will put the virtual machines in it’s default machine folder. You can see where the default machine folder is set to using the VBoxManage utility:
$ VBoxManage list systemproperties | grep machine
Default machine folder: /Volumes/VirtualBox
Because the virtual machines are disk space intensive, you might want to change the location. At least, I later decided to dedicate an external disk to it. The way to get the disk images in another location, is change the virtualbox settings:
$ VBoxManage setproperty machinefolder /Volumes/Virtualbox
The vagrant ‘box’ location.
Vagrant uses operating system images that it is calling ‘boxes’. These ‘boxes’ are the base image that is copied when a virtual machine is created. It is stored to save it from downloading over and over. The location of these images or ‘boxes’ is ~/.vagrant.d/boxes. If you want to store the ‘boxes’ somewhere else, you can set the environment variable VAGRANT_HOME to make vagrant use a different location for storing its global state and the ‘boxes’.
Setting up the vagrant directory.
The next step is to set up the directory where the virtual machine will be run. The simplest way of doing so is cloning the files from github:
$ git clone https://github.com/FritsHoogland/Vagrant-oracle-setup.git
This will create a ‘Vagrant-oracle-setup’ directory.
Now you need to fill out the Vagrantfile in the directory:
a) line 12: set the hostname. (change optional)
b) line 14: set an IP address. (do not choose an already existing network, like the default virtualbox 10.0.2/24 network)
c) line 18/19: set the memory and CPU capacity for the VM to use.
d) line 41/42: set your My Oracle Support credentials to download the installation media.
e) line 43: set the IP address for ansible, as set at b)/line 14.
f) line 46: set the database name to create. (change optional)
Install the virtual machine with linux, the oracle database and setup a database.
Now the cool part: install linux, install the oracle database software and create a database without any manual intervention! For this you need to go into the Vagrant-oracle-setup directory, and execute:
$ vagrant up
You can follow the steps vagrant and ansible are executing. You’ll see vagrant first fetching (downloading) the OS image (which it will save to prevent from downloading it again), and then copy this image to be the image for the virtual machine to run. Vagrant then makes some settings, mostly network related, and hands over the control to ansible.
Ansible executes the playbook as set in the Vagrantfile, which I’ve put in the ‘ansible’ directory. Ansible sets up linux, configures the extra block devices using LVM for the software home and the database, gets the installation media from My Oracle Support, installs the software and then runs DBCA to create a database.
Handling vagrant to administer the virtual machine is easy. Go into the directory with the Vagrantfile and use:
– vagrant halt: stop the virtual machine.
– vagrant up: for a new machine, get the base image and provision the machine. For an existing machine: startup the virtual machine.
– vagrant destroy: stop the virtual machine and remove all the information it has used like configuration/provisioning status and the disks.
– vagrant provision: run the ansible playbook again.
Anything else? Well, yes, I didn’t setup a listener for example. It should be really simple and straightforward: just start the listener. The listener will start listening on all network devices, and the database will register itself. There might be much more, depending on your specific need.
Thanks to Maris Elsins for getMOSPatch.