Skip to content

Latest commit

 

History

History
111 lines (69 loc) · 5.18 KB

INSTALLING-web.md

File metadata and controls

111 lines (69 loc) · 5.18 KB
Raw

Installing the web app

Fetching the code

To install the web system, you should first clone the repo somewhere in your server.

git clone https://github.com/JoseTomasTocino/pfc-ii.git siteup

We'll be working in the web folder from now on.

cd siteup/web

Installing Python environment and dependencies

In order to install the proper requirements you need to create a virtualenv. I highly recommend using virtualenvwrapper alongside virtualenv.

pip install virtualenv
pip install virtualenvwrapper

Make sure you check virtualenvwrapper's documentation and follow the steps to fulfill the installation.

With virtualenvwrapper installed, create a new virtualenv and install the requirements using pip. Note though that some requirements may need some header files, located in the python-dev package in debian-based distributions.

mkvirtualenv siteup
sudo apt-get install python-dev
pip install -r web/requirements.txt

Building the database

Next step is to set up the database. Use the following commands to build the database and the tables. Currently the project uses sqlite for both development and production environments, but this will change in the future.

./manage.py syncdb
./manage.py migrate siteup_api

Installing required services

Any web project built using a Python web framework, such as Django, needs a WSGI http server, that will run the python files within the project, and a HTTP Proxy server, that will receive the actual requests from the users and route them to the WSGI server. For the first, SiteUp uses gunicorn. It gets installed via pip along with the other dependencies in the previous steps.

For the second, I recommend using nginx. It's a powerful HTTP server and reverse proxy, and you can install it easily using the package manager of your distribution

sudo apt-get install nginx

Additionally, SiteUp needs a way of running tasks asynchronously. For that task we use celery, a the task queue that triggers the checks periodically. Celery is also installed automatically along the other python dependencies.

Finally, Siteup needs a AMQP server to store the tasks. Recommended option is RabbitMQ:

sudo apt-get install rabbitmq-server

Running the services

Usually, both nginx and rabbitmq will run when you the computer starts as they've been installed system-wide. On the other hand, it's necessary to have a way of running gunicorn and celery. The best option I've found is using supervisord, it's pretty simple and easier to use than traditional init.d scripts.

SiteUp includes a management task that automatically creates the configuration file for supervisord.

./manage.py supervisor_conf
sudo mv supervisor-siteup.conf /etc/supervisord/conf.d

That task will ask for some variables that SiteUp reads from the environment.

  • GMAIL_PASS, used in siteup/settings/base.py. Password for the gmail account used to send emails.
  • GCM_API_KEY, used in siteup_checker/task_notification.py. Google Cloud Messaging api key.

In development you won't be using Supervisord, so you should set them up appropiately. A good option is placing them in your virtualenv's postactivate script, like this:

cdvirtualenv
cd bin
cat >> postactivate
export GMAIL_PASS=yourpass
export GCM_API_KEY=yourkey

Now you need to configure your proxy server. I'm using nginx because it's very easy to configure as a proxy. A possible configuration could be this (change your paths accordingly):

server {
    listen 80;
    server_name siteup.josetomastocino.com;

    # Static files
    location /static/ {
        root /srv/siteup.josetomastocino.com/siteup/web/siteup_frontend;
        try_files $uri $uri/ =404;
    }


    location / {
        proxy_set_header X-Real-IP        $remote_addr;
        proxy_set_header X-Forwarded-For  $proxy_add_x_forwarded_for;
        proxy_set_header Host             $http_host;
        proxy_redirect off;
        if (!-f $request_filename) {
            proxy_pass       http://127.0.0.1:8000;
            break;
        }
    }
}

After that you should just restart both supervisord and nginx and you should be good to go:

sudo service nginx restart
sudo service supervisord restart

Performance issues

Rabbitmq default configuration makes it consume a lot of memory and can become a problem in systems with limited resources, such as low-end vps. In order to fix that you can add the following to /etc/rabbitmq/rabbitmq.config to limit the max memory to 15% of the total.

[
    {rabbit, [{vm_memory_high_watermark, 0.15}]}
].

If the number of checks in your system is going to be low, you can decrease the number of celery workers by modifying the sup_celery.conf file generated before, modifying the -c 16 parameter in the command attribute.