Version 2.1 ist ein veraltete Veröffentlichung, diese Dokumente werden nicht mehr gewartet.

Linux

We use and recommend Ubuntu Linux as an easy to use Linux distro with a large and lively community behind it - however, we’ve obtained satisfactory results with a number of other distributions like Debian, CentOS or Fedora. Most of the configuration steps described in this document apply to any modern Linux environment, however some of them will apply only to Ubuntu and likely to any Ubuntu-based distribution.

To be specific, this document is based in Ubuntu 12.04 LTS (Precise Pangolin) and Ubuntu 14.04 LTS (Trusty Tahr). Once you have installed it, you should be able to follow the instructions described below. In particular, we are going to use Ubuntu packages that can be found under the repositories main and universe.

Important

Please make sure you have reviewed the requirements before proceeding.

Install the dependencies

MySQL

We strongly recommend using MySQL 5.5 as it’s much better than its previous major release in terms of speed, scalability and user-friendliness. Also, we’ve experienced very good results using forks like Percona Server or MariaDB, so don’t be afraid and use them if you want!

Let’s install MySQL using apt-get:

sudo apt-get install mysql-server-5.5

During the installation, you will be prompted to set a password for the default administrator user (root). We strongly recommend that you use a strong password and please write it down as you are going to need it later.

Elasticsearch

A relatively new search server based on Apache Lucene and developed in Java that has brought AtoM a lot of advanced features, performance and scalability. This is probably the biggest change introduced in AtoM 2.1 and we are pleased with the results.

Ubuntu doesn’t provide a package but you can download it directly from the Elasticsearch site if you are unable to download it using the method that follows.

First, make sure that Java is installed. Elasticsearch is compatible with OpenJDK (package openjdk-7-jre-headless) but we are going to use Oracle’s JVM distributed by Web Upd8.

sudo add-apt-repository ppa:webupd8team/java
sudo apt-get update
sudo apt-get install oracle-java8-installer

After successfully installing Java, proceed to install Elasticsearch. Download and install the public signing key used in their repository:

wget -qO - http://packages.elasticsearch.org/GPG-KEY-elasticsearch | sudo apt-key add -

Add the following to your /etc/apt/sources.list to enable the repository:

deb http://packages.elasticsearch.org/elasticsearch/1.3/debian stable main

Ready to be installed. Run:

sudo apt-get update
sudo apt-get install elasticsearch

Start the service and configure it to start when the system is booted.

sudo update-rc.d elasticsearch defaults 95 10
sudo /etc/init.d/elasticsearch start

Nginx

There are many web servers out there capable of working well with PHP. Apache is probably the most popular and we like it, but we’ve found that Nginx adapts itself much better to limited resource environments while it also scales better and more predictably under high loads. You are welcome to try other solutions, but the following documentation will focus upon Nginx, our preferred web server solution.

The installation in Ubuntu is simple:

sudo apt-get install nginx

Ubuntu 12.04 uses Nginx 1.1. However, the team behind Nginx provides an official PPA (Personal Package Archive) channel for Ubuntu users that supports more recent stable packages for the different releases of Ubuntu, including 12.04. This could be a good choice if you want to enjoy some of the latest features and improvements added to Nginx while taking minimal risks in your production environments. If you are interested, run the following commands:

sudo add-apt-repository ppa:nginx/stable
sudo apt-get update
sudo apt-get install nginx

Nginx deploys a default server (aka VirtualHost, for Apache users) called default and you can find it in /etc/nginx/sites-available/default. In order to install AtoM you could edit the existing server block or add a new one. We are going to you show you how to do the latter:

sudo touch /etc/nginx/sites-available/atom
sudo ln -sf /etc/nginx/sites-available/atom /etc/nginx/sites-enabled/atom
sudo rm /etc/nginx/sites-enabled/default

We have now created the configuration file and linked it from sites-enabled/, which is the directory that Nginx will look for. This means that you could disable a site by removing its symlink from sites-enabled/ while keeping the original one under sites-available/, in case that you want to re-use it in the future. You can do this with the Nginx default server.

The following is a recommended server block for AtoM. Put the following contents in /etc/nginx/sites-enabled/atom.

upstream atom {
  server unix:/var/run/php5-fpm.atom.sock;
}

server {

  listen 80;
  root /usr/share/nginx/atom;

  # http://wiki.nginx.org/HttpCoreModule#server_name
  # _ means catch any, but it's better if you replace this with your server
  # name, e.g. archives.foobar.com
  server_name _;

  client_max_body_size 72M;

  # http://wiki.nginx.org/HttpCoreModule#try_files
  location / {
    try_files $uri /index.php?$args;
  }

  location ~ /\. {
    deny all;
    return 404;
  }

  location ~* (\.yml|\.ini|\.tmpl)$ {
    deny all;
    return 404;
  }

  location ~* /(?:uploads|files)/.*\.php$ {
    deny all;
    return 404;
  }

  # This is the most important part, as here we are redirecting some specific
  # requests to PHP-FPM so PHP can do its job passing data to and from the
  # web server.
  location ~ ^/(index|qubit_dev)\.php(/|$) {
    include /etc/nginx/fastcgi_params;
    fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
    fastcgi_split_path_info ^(.+\.php)(/.*)$;
    fastcgi_pass atom;
  }

  location ~* \.php$ {
    deny all;
    return 404;
  }

}

Now you need to restart Nginx:

sudo service nginx restart

PHP

Our favorite way to deploy AtoM is using PHP-FPM, a process manager that scales better than other solutions like FastCGI. The following command will install it along with the rest of PHP extensions required by AtoM:

sudo apt-get install php5-cli php5-fpm php5-curl php5-mysql php5-xsl php5-json php5-ldap php-apc

If you are using Ubuntu 14.04, the package php5-readline is also required.

sudo apt-get install php5-readline

Let’s add a new PHP pool for AtoM by adding the following contents in a new file called /etc/php5/fpm/pool.d/atom.conf:

[atom]

# The user running the application
user = www-data
group = www-data

# Use UNIX sockets if Nginx and PHP-FPM are running in the same machine
listen = /var/run/php5-fpm.atom.sock
listen.owner = www-data
listen.group = www-data
listen.mode = 0600

# The following directives should be tweaked based in your hardware resources
pm = dynamic
pm.max_children = 30
pm.start_servers = 10
pm.min_spare_servers = 10
pm.max_spare_servers = 10
pm.max_requests = 200

chdir = /

# Some defaults for your PHP production environment
# A full list here: http://www.php.net/manual/en/ini.list.php
php_admin_value[expose_php] = off
php_admin_value[allow_url_fopen] = on
php_admin_value[memory_limit] = 512M
php_admin_value[max_execution_time] = 120
php_admin_value[post_max_size] = 72M
php_admin_value[upload_max_filesize] = 64M
php_admin_value[max_file_uploads] = 10
php_admin_value[cgi.fix_pathinfo] = 0
php_admin_value[display_errors] = off
php_admin_value[display_startup_errors] = off
php_admin_value[html_errors] = off
php_admin_value[session.use_only_cookies] = 0

# APC, which is still used in PHP 5.5 for userland memory cache unless you
# are switching to something like sfMemcacheCache
php_admin_value[apc.enabled] = 1
php_admin_value[apc.shm_size] = 64M
php_admin_value[apc.num_files_hint] = 5000
php_admin_value[apc.stat] = 0

# Zend OPcache
# Only in Ubuntu 14.04 (PHP 5.5).
# Don't use this in Ubuntu 12.04, it won't work.
php_admin_value[opcache.enable] = 1
php_admin_value[opcache.enable_cli] = 0
php_admin_value[opcache.memory_consumption] = 192
php_admin_value[opcache.interned_strings_buffer] = 16
php_admin_value[opcache.max_accelerated_files] = 4000
php_admin_value[opcache.validate_timestamps] = 0
php_admin_value[opcache.fast_shutdown] = 1

# This is a good place to define some environment variables, e.g. use
# ATOM_DEBUG_IP to define a list of IP addresses with full access to the
# debug frontend or ATOM_READ_ONLY if you want AtoM to prevent
# authenticated users
env[ATOM_DEBUG_IP] = "10.10.10.10,127.0.0.1"
env[ATOM_READ_ONLY] = "off"

Note that the section “Zend OPcache” won’t work in Ubuntu 12.04. Comment it out or remove it unless you are using Ubuntu 14.04.

The process manager has to be restarted:

sudo service php5-fpm restart

If the service fails to start, make sure that the configuration file has been has been pasted properly. You can also check the syntax by running:

sudo php5-fpm --test

If you are not planning to use the default PHP pool (www), feel free to remove it:

sudo rm /etc/php5/fpm/pool.d/www.conf
sudo service php5-fpm restart

Other packages

If you want AtoM to be able to process digital objects in formats like JPEG or to extract the text from your PDF documents, there are certain packages that you need to install. They are not mandatory but if they are found in the system, AtoM will use them to produce digital object derivatives from your master objects. for more information on each, see: Requirements: other dependencies. The following will install all the recommended dependencies at once:

sudo apt-get install imagemagick ghostscript poppler-utils

Install ffmpeg from Archivematica’s PPA, which works for both Ubuntu 12.04 (precise) and Ubuntu 14.04 (trusty).

sudo add-apt-repository ppa:archivematica/externals
sudo apt-get update
sudo apt-get install ffmpeg

Download AtoM

Now that we have installed and configured all dependencies, we are ready to download and install AtoM itself. The safest way is to install AtoM from the tarball, which you can find in the download section. However, experienced users may prefer to check out the code from our public repository.

The following instructions assume that we are installing AtoM under /usr/share/nginx and that you are using AtoM 2.1.2.

Option 1: Download the tarball

wget https://storage.accesstomemory.org/releases/atom-2.1.2.tar.gz
sudo mkdir /usr/share/nginx/atom
sudo tar xzf atom-2.1.2.tar.gz -C /usr/share/nginx/atom --strip 1

Option 2: Check out the code from our git repository

Install git:

sudo apt-get install git
sudo mkdir /usr/share/nginx/atom
sudo git clone -b stable/2.1.x http://github.com/artefactual/atom.git /usr/share/nginx/atom
cd /usr/share/nginx/atom

If you are not interested in downloading all the history from git, you could also truncate it to a specific number of revisions, e.g.: just one revision

git clone --depth 1 http://github.com/artefactual/atom.git /usr/share/nginx/atom

Once you’ve cloned the code from our git repository, you’ll need to compile the CSS files:

sudo add-apt-repository ppa:chris-lea/node.js
sudo apt-get update
sudo apt-get install nodejs make
sudo npm install -g "less@<2.0.0"
cd /usr/share/nginx/atom/plugins/arDominionPlugin/
sudo make # At this point the files still belong to root

Filesystem permissions

By default, Nginx runs as the www-data user. There are a few directories under AtoM that must be writable by the web server. The easiest was to ensure this is to update the owner of the full directory and its contents by running:

sudo chown -R www-data:www-data /usr/share/nginx/atom

Create the database

Assuming that you are running MySQL in localhost, please create the database by running the following command using the password you created earlier:

mysql -h localhost -u root -p -e "CREATE DATABASE atom CHARACTER SET utf8 COLLATE utf8_unicode_ci;"

Notice that the database has been called atom. Feel free to change its name.

In case your MySQL server is not the same as your web server, replace “localhost” with the address of your MySQL server.

Warning

Plase make sure that you are using an empty database! Don’t reuse an old database unless it’s empty. You can always drop it by using the DROP DATABASE command and then create it again.

Additionally, it’s always a good idea to create a specific MySQL user for AtoM to keep things safer. This is how you can create an user called atom with password 12345 and the permissions needed for the database created above.

mysql -h localhost -u root -p -e "GRANT INDEX, CREATE, SELECT, INSERT, UPDATE, DELETE, ALTER, LOCK TABLES ON atom.* TO 'atom'@'localhost' IDENTIFIED BY '12345';"

Note that the INDEX, CREATE and ALTER privileges are only necessary during the installation process or when you are upgrading AtoM to a newer version. They can be removed from the user once you are finished with the installation or you can change the user used by AtoM in config.php.

Run the web installer

You should now be ready to run the installer. It’s a simple web interface that changes some internal configuration files according to your environment and adds the necessary tables and initial data to the database recently created.

Open your browser and type the URL in the address bar. The URL can greatly change depending on your web server configuration. The URL will usually be something like http://localhost. AtoM will redirect you to the installer automatically.

The installation process consists of a number of steps where you will be asked for configuration details such as the location of your database server. If you have followed this document to the letter, this is how you should fill the following fields:

  • Database name: atom
  • Database username: atom
  • Database password: 12345
  • Database host: localhost
  • Database port: 3306
  • Search host: localhost
  • Search port: 9200
  • Search index: atom

Of course, some of these fields will look very different if you are running AtoM in a distributed way, where your services like MySQL or Elasticsearch are running in separate machines.

The rest of the fields can be filled as you need:

  • Site title
  • Site description
  • Username
  • E-mail address
  • Password

Deployment of workers

Optionally, you can use Gearman to add support for asynchronous tasks like SWORD deposits. Check out the following page for further installation details: asynchronous jobs and worker management.

Security considerations

Now that AtoM is installed, please take a moment to read our security section where we will show you how to configure the firewall in Ubuntu and back up AtoM.

We strongly encourage our users to configure a firewall because some of the services configured should not be exposed in the wild, e.g. Elasticsearch was not designed to be accessible from untrusted networks and it’s a common attack vector.