A django based system that provides a stable URL for every item in the libraries' catalog. Various discovery services will link to these URLs. The page for each item will in turn link out to various other resources that provide methods for accessing the content of the items.

This software should be runnable on any kind of operating system. However, these installation instructions are tailored to a Linux server, and have only been tested on ubuntu 14.04 LTS.

Part I - Basic server requirements

1. Install Apache and other dependencies:

    sudo apt-get install -y libxml2-dev libxslt1-dev zlib1g-dev
    sudo apt-get install apache2 libapache2-mod-wsgi libaio-dev python-dev python-profiler memcached libmemcached-dev  libxslt-dev sqlite3
    sudo apt-get install supervisor
   

2. Install git

    sudo apt-get install git
   

3. Download Oracle Instant Client 11.2.0.3 for Linux x86-64 or the platform that matches your system from http://www.oracle.com/technetwork/topics/linuxx86-64soft-092277.html (accept terms and login to download the files)

    * instant-client-basic
    * instant-client-devel
    * instant-client-sqlplus
   

   Or find a copy of the 11.2.0.3.0-1.x86_64 rpms in:

    /vol/backup/dependencies
   

4. Install alien in order to install the RPM packages on Debian or ubuntu Linux Distributions. This is not required if you are using RedHat or similar distribution.

    sudo apt-get install alien
   

5. Install all three download packages of oracle using the following commands (for Debian and Ubuntu distros). For RedHat and similar Distros use rpm -ivh packagename command.

   Note: if you have previous versions of these packages installed, you will likely have to remove them first; updates to the sqlplus client (at least) conflict with previous installations.

    sudo alien --install oracle-instantclient11.2-basic-11.2.0.3.0-1.x86_64.rpm
    sudo alien --install oracle-instantclient11.2-devel-11.2.0.3.0-1.x86_64.rpm
    sudo alien --install oracle-instantclient11.2-sqlplus-11.2.0.3.0-1.x86_64.rpm
   

6. Setup ORACLE_HOME and LD_LIBRARY_PATH environment variables. Change the path according to your installation of oracle.

    export ORACLE_HOME=/usr/lib/oracle/11.2/client64/
    export LD_LIBRARY_PATH=/usr/lib/oracle/11.2/client64/lib
   

7. Set environment variables in oracle.conf file (create the file if it does not exist)

    sudo vim /etc/ld.so.conf.d/oracle-instantclient11.2-basic.conf
   

   Add the following lines

    /lib
    /usr/lib/oracle/11.2/client64/lib
   

   Remove any lines that refer to previous versions of the library.

   Remove any files in /etc/ld.so.conf.d that refer to previous versions of the library

   Now run

    sudo ldconfig  
   

Part II - Setting up the project environment

1. Install virtualenv

    sudo apt-get install python-setuptools
    sudo easy_install virtualenv
   

2. Create directory for your projects (replace LPHOME with your root dir)

    mkdir /LPHOME
    cd /LPHOME
   

3. Pull down the project from github

    (GW staff only)
    git clone git@github.com:gwu-libraries/launchpad.git
   
    (everyone else)
    git clone https://github.com/gwu-libraries/launchpad.git
   

4. Create virtual Python environment for the project

    cd /LPHOME/launchpad
    virtualenv --no-site-packages ENV
   

5. Activate your virtual environment

    source ENV/bin/activate
   

6. Install django, cx_Oracle, and other python dependencies

    pip install -r requirements.txt
   

7. Note: Voyager's Oracle implementation requires ASCII encoding on database connections. Django strictly mandates UTF8 encodings on connections, so we have to override this by setting 'NLS_LANG' to '.US7ASCII' in our wsgi.py, overriding Django's strong preference (and that of its community). See the "Release: All" section of the release notes on our project wiki for examples to test.

    https://github.com/gwu-libraries/launchpad/wiki
   

   See https://code.djangoproject.com/ticket/15313#comment:4 for more details and a helpful response by a core Django developer advocating this approach over our previous Django-patching madness.

   If Oracle throws a access denied or Apache displays permissions error set directory permissions for launchpad/ and subdirectories as required.

8. Configure a database using sqlite in the appropriate directory.

    cd LPHOME/launchpad/lp/lp
    sqlite3 lp.db
   

Part III - Configuring your installation

Configure database and other settings in a local_settings file:

    cd lp/lp
    cp local_settings.py.template local_settings.py
    vim local_settings.py

• Set EMAIL_SUBJECT_PREFIX and SERVER_EMAIL to appropriate values.
• Change database login and password and any other parameters you wish to change.
• Define one or more Z39.50 servers if needed under Z3950_SERVERS.
• If you want to use memcache, define CACHES and ITEM_PAGE_CACHE_SECONDS.
  For development or testing, set ITEM_PAGE_CACHE_SECONDS to something low.
• NOTE: If you are deploying to production, set DEBUG = False. Also, set GOOGLE_ANALYTICS_UA to your UA to enable google analytics in production.
• Comment out CACHES if you are not using memcached

Edit wsgi file:

    cp lp/wsgi.py.template lp/wsgi.py
    vim lp/wsgi.py

• Change parameter for site.addsitedir() to your local path. You will need to change the path and possibly the Python version number.

If you want to use memcached, configure and ensure it has started:

    sudo vim /etc/memcached.conf
    sudo /etc/init.d/memcached start

Run the database migrations command.

    ./manage.py migrate

At this point, you should be able to run the app and view it working, even without apache configured. This might be sufficient for dev/test.

    python manage.py runserver 0.0.0.0:8080
    (visit http://your-server:8080/item/198738 to test)

• Performing this step will result in a new local file PyZ3950_parsetab.py generated by the Z39.50 support library. If you follow the next steps to enable apache integration without doing this first, apache might not be able to write this file, depending on whether you've made changes to permissions, groups, or the system user apache uses. It's easy just to do this by hand first. :)

If you want to use apache, add apache config file to sites-enabled and edit it

    sudo cp ../apache/lp.conf /etc/apache2/sites-available/lp.conf
    sudo vim /etc/apache2/sites-available/lp.conf

The apache config assumes you are using SSL and have set up certs as usual for GW Libraries applications. This includes renaming the file appropriately.

• Change the values of LPHOME, server, user, and python version in the document as appropriate.

Enable these apache modules:

    sudo a2enmod expires
    sudo a2enmod headers
    sudo a2enmod proxy
    sudo a2enmod proxy_http

Enable the app in apache and bounce apache2 to start it up

    sudo a2ensite lp
    sudo /etc/init.d/apache2 restart

If you are in production mode, and want to provide sitemaps, set launchpad's ENABLE_SITEMAPS, SITEMAPS_DIR, and SITEMAPS_BASE_URL in local_settings.py. Make sure SITEMAPS_DIR exists, then run the management command:

    manage.py make_sitemap

If you are in production mode, be sure to set DEBUG = False and the appropriate ALLOWED_HOSTS in lp/local_settings.py.

• Set up logsitter.

Logsitter is a script intended to be a workaround for #1004. Logsitter monitors the apache log and as soon as it sees the error message referenced in #1004, it automatically restarts apache. To ensure that it stays running, logsitter is deployed with and managed by supervisor.

If the LPHOME/scripts directory is not owned by root, change its ownership to root:

    sudo chown -R root:root LPHOME/scripts/

To set up logsitter, copy the logsitter supervisor configuration file, then restart supervisor:

    # Before copying, first edit the `command` parameter
    # in LPHOME/scripts/logsitter.conf to use the correct path.
    sudo cp LPHOME/scripts/logsitter.conf /etc/supervisor/conf.d/
    sudo service supervisor restart