Installation instructions

OpenREM can be installed with a single command; however, there are two prerequisites that need to be installed first – python and pip – and RabbitMQ needs to be installed for exports and patient size imports to work.

Once installed, there are a few configuration choices that need to be made, and finally a couple of services that need to be started. Then you are ready to go!

Install the prerequisites

Install Python 2.7.x

Add Python and the scripts folder to the path

Windows only – this is usually automatic in linux

Add the following to the end of the path environment variable (to see how to edit the environment variables, see


Setuptools and pip

Install setuptools and pip – for details go to The quick version is as follows:


Download the latest version using the same method as for Windows, or get the version in your package manager, for example:

sudo apt-get install python-pip


Download the installer script and save it locally – right click and Save link as... or equivalent.

Open a command window (Start menu, cmd.exe) and navigate to the place you saved the get‑ file:


Quick check of python and pip

To check everything is installed correctly so far, type the following in a command window/shell. You should have the version number of pip returned to you:

pip -V

Install RabbitMQ

(New for version 0.4.3)

For either install, just follow the defaults – no special configurations required.


Before continuing, consider virtualenv

Install and configure OpenREM

Install latest beta version

pip install --pre openrem==0.5.0b1

Will need ``sudo`` or equivalent if installing on linux without using a virtualenv


Locate install location

  • Linux: /usr/local/lib/python2.7/dist-packages/openrem/ or /usr/lib/python2.7/site-packages/openrem/
  • Windows: C:\Python27\Lib\site-packages\openrem\

There are two files that need renaming:

  • openremproject/ to openremproject/
  • openremproject/ to openremproject/

In the file, set the database details, the MEDIA_ROOT path, the secret key and the ALLOWED_HOSTS.


Windows notepad will not recognise the Unix style line endings. Please use an editor such as Notepad++ or Notepad2 if you can, else use WordPad – on the View tab you may wish to set the Word wrap to ‘No wrap’

Database settings

For testing you can use the SQLite3 database

'ENGINE': 'django.db.backends.sqlite3',
  • Linux example: 'NAME': '/home/user/openrem/openrem.db',
  • Windows example: 'NAME': 'C:/Users/myusername/Documents/OpenREM/openrem.db', Note use of forward slash in configuration files

For production use, see Database options below

Location setting for imports and exports

Csv and xlsx study information exports and patient size csv imports are written to disk at a location defined by MEDIA_ROOT.

The path set for MEDIA_ROOT is up to you, but the user that runs the webserver must have read/write access to the location specified because it is the webserver than reads and writes the files. In a debian linux, this is likely to be www-data for a production install. Remember to use forward slashes for the config file, even for Windows.

Linux example:

MEDIA_ROOT = "/var/openrem/media/"

Windows example:

MEDIA_ROOT = "C:/Users/myusername/Documents/OpenREM/media/"

Secret key

Generate a new secret key and replace the one in the file. You can use for this.

Allowed hosts

The ALLOWED_HOSTS needs to be defined, as the DEBUG mode is now set to False. This needs to contain the server name or IP address that will be used in the URL in the web browser. For example:


A dot before a hostname allows for subdomains (eg www.doseserver), a dot after a hostname allows for FQDNs (eg Alternatively, a single '*' allows any host, but removes the security the feature gives you.

Create the database


python /usr/local/lib/python2.7/dist-packages/openrem/ syncdb


python C:\Python27\Lib\site-packages\openrem\ syncdb

Answer each question as it is asked, do setup a superuser. This username and password wil be used to log into the admin interface to create the usernames for using the web interface. See the Start using it! section below.

Help! I get a value too long for type character varying(50) error!
This error with part of the Django auth_permissions system that we are not using, and can safely be ignored. This is being tracked as Issue 62

For production installs, convert to South

(What is south?)


python /usr/local/lib/python2.7/dist-packages/openrem/ convert_to_south remapp


python C:\Python27\Lib\site-packages\openrem\ convert_to_south remapp

Start all the services

Start test web server


python /usr/local/lib/python2.7/dist-packages/openrem/ runserver --insecure


python C:\Python27\Lib\site-packages\openrem\ runserver --insecure

If you are using a headless server and need to be able to see the web interface from another machine, use python /usr/lib/python2.7/dist-packages/openrem/ runserver x.x.x.x:8000 --insecure (or Windows equivalent) replacing the x with the IP address of the server and 8000 with the port you wish to use.

Open the web addesss given, appending /openrem (http://localhost:8000/openrem)


Why are we using the --insecure option? With DEBUG mode set to True the test web server would serve up the static files. In this release, DEBUG mode is set to False, which prevents the test web server serving those files. The --insecure option allows them to be served again.

Start the Celery task queue

(New for version 0.4.3)

Celery will have been automatically installed with OpenREM, and along with RabbitMQ allows for asynchronous task processing for imports and exports.


The webserver and Celery both need to be able to read and write to the MEDIA_ROOT location. Therefore you might wish to consider starting Celery using the same user or group as the webserver, and setting the file permissions accordingly.

In a new shell:


cd /usr/local/lib/python2.7/dist-packages/openrem/
celery -A openremproject worker -l info


cd C:\Python27\Lib\site-packages\openrem\
celery -A openremproject worker -l info

For production use, see Daemonising Celery below

Start using it!

Add some data! rdsrfile.dcm

Add some users (New in version 0.4.0)

  • Go to the admin interface (eg http://localhost:8000/admin) and log in with the user created when you created the database (syncdb)

  • Create some users and add them to the appropriate groups (if there are no groups, go to the OpenREM homepage and they should be created).

    • viewgroup can browse the data only
    • exportgroup can do as view group plus export data to a spreadsheet
    • admingroup can delete studies and import height and weight data in addition to anything the export group can do
  • Return to the OpenREM interface (eg http://localhost:8000/openrem) and log out of the superuser in the top right corner and log in again using one of the new users you have just created.

Further instructions

Database options

SQLite is great for getting things running quickly and testing if the setup works, but is really not recommended for production use on any scale. Therefore it is recommended to use a different database such as PostgreSQL or MySQL.

Here are instructions for installing PostgreSQL on linux and on Windows:

Database migrations

South is a django application to manage database migrations. Using South means that future changes to the database model can be calculated and executed automatically with simple commands when OpenREM is upgraded.

Production webservers

Unlike the database, the production webserver can be left till later and can be changed again at any time.

For performance it is recommended that a production webserver is used instead of the inbuilt ‘runserver’. Popular choices would be either Apache or you can do as the cool kids do and use Gunicorn with nginx.

The django website has instructions and links to get you set up with Apache.

Daemonising Celery

(New for version 0.4.3)

In a production environment, Celery will need to start automatically and not depend on a particular user being logged in. Therefore, much like the webserver, it will need to be daemonised. For now, please refer to the instructions and links at

Virtualenv and virtualenvwrapper

If the server is to be used for more than one python application, or you wish to be able to test different versions of OpenREM or do any development, it is highly recommended that you use virtualenv or maybe virtualenvwrapper

Virtualenv sets up an isolated python environment and is relatively easy to use.

If you do use virtualenv, all the paths referred to in the documentation will be changed to:

  • Linux: lib/python2.7/site-packages/openrem/
  • Windows: Lib\site-packages\openrem

In Windows, even when the virtualenv is activated you will need to call python and provide the full path to script in the Scripts folder. If you call the script (such as without prefixing it with python, the system wide Python will be used instead. This doesn’t apply to Linux, where once activated, the scripts can be called without a python prefix from anywhere.