You should clone this repository and put it as you own private project in your Git version control system. For now, let's say you fork the project on Github, and that your Github username is toto. Then the repository should be available on:

https://github.com/toto/demo_celery

If the machine is running Window, it may then be simpler to use the first option with Docker (see below). Also, do not forget to make sure the port 6379 is open.

Install Docker: https://docs.docker.com/engine/installation/

Then just run this (do not forget to change the password):

sudo docker run -d -p 0.0.0.0:6379:6379 --name redis redis redis-server --requirepass 1234

On Linux it is quite simple but you need to compile it:

sudo apt-get install wget make g++ screen
wget http://download.redis.io/redis-stable.tar.gz
tar xvzf redis-stable.tar.gz
cd redis-stable
make
sudo make install

Then you can open a screen to launch Redis in a persitant manner. Again, do not forget to change the password:

screen
redis-server --requirepass 1234

Here you clone the repo using SSH because you want to be able to push easily.

git clone git@github.com:toto/demo_celery.git

Then you just have to install the Python libs:

cd demo_celery
sudo pip install -r requirements.txt

Also you need to change the Redis config in redis_config.py to indicate the IP adress of the machine running Redis and the password you chose.

Here you clone using HTTPS because you may not have any SSH key setup:

git clone https://github.com/toto/demo_celery.git

Then you also just have to install the Python libs:

cd demo_celery
sudo pip install -r requirements.txt

Once you changed the Redis config in redis_config.py as you did previously on your local computer, you can start the worker with make worker or with:

celery multi start node1 -A distributed --loglevel=info --pidfile=distributed.pid
celery multi start node2 -A updater --loglevel=info --pidfile=updater.pid

It should create two files node1.log and node2.log. Take a look at their content and make sure the workers have no trouble connecting.

On your local computer, running python main.py should return 1129.

# What does it do?

If you commit changes, the worker pulls it when you call updater.update() and updates itself (actually only the file distributed.py will be updated, not updater.py). Then the sample code just sends a file (config.json) to the worker, read a dummy value from it and returns it.

You should put all your code in distributed.py.

Have fun !!