RFInd Monitor

For deployment skip below to deployment.

Suggested installation steps

Setup a venv and source it

python3 -m venv venv
source venv/bin/activate

update pip

(venv)$ pip install --upgrade pip

Install this package and it's requirements with

(venv)$ pip install .

Add a -e flag to make it editable if you're changing the code or regularly updating the repo

Install Redis

wget http://download.redis.io/redis-stable.tar.gz
tar xvzf redis-stable.tar.gz
cd redis-stable
make
make test
sudo make install

Test the app with simulated data

Build the simulation data

(venv)$ python -m rfind_monitor.sim.data_gen

To use the local zmq simulation, move the new data.h5 file into /rfind_monitor/sim/.

Run the app.

(venv)$ python -m rfind_monitor.frontend.dash

In a separate terminal (sourcing the same venv) run the Redis server

(venv)$ python -m rfind_monitor.backend.redis_store

In a separate terminal (sourcing the same venv) run the middle man.

(venv)$ python -m rfind_monitor.backend.middle_man

In a separate terminal (sourcing the same venv) run the data server

(venv)$ python -m rfind_monitor.sim.zmq_pusher

The middle man handles the zmq binding that has data pushed to it and writes that to the Redis storage.

Deployment

1. Spin up CC cloud instance (Ubuntu 20.04) using the quick start guide.

2. SSH into the VM.

3. Update the VM with sudo apt update and sudo apt upgrade

4. Install some dependencies with sudo apt install apache2 libapache2-mod-wsgi-py3 python3-venv (which we'll use to serve both the application and the data)

5. Clone this repo into /home/ubuntu/ then install the venv and this package as described above. Make sure to install redis as well as decribed above.

6. Make a folder in the home directory to host the server

    cd
    mkdir public_html
    mkdir public_html/data
    mkdir public_html/wsgi

7. Create a file public_html/wsgi/dash.wsgi that contains

   #!/usr/bin/python3
   import sys
   sys.path.insert(0, "/home/ubuntu/rfind-monitor/venv/lib/python3.8/site-packages")
   from rfind_monitor.frontend.dash import server as application

8. Create a file /etc/apache2/sites-available/dash.conf that contains

   WSGIDaemonProcess dash user=ubuntu group=ubuntu home=/home/ubuntu threads=5 socket-user=#33
   WSGIScriptAlias /live /home/ubuntu/public_html/wsgi/dash.wsgi
   
   WSGIProcessGroup dash
   WSGIApplicationGroup %{GLOBAL}

9. Edit /etc/apache2/sites-available/000-default.conf:

   • Change the line DocumentRoot /var/www to DocumentRoot /home/ubuntu/public_html

10. Edit /etc/apache2/apache2.conf

    • Change the line <Directory /var/www/> to <Directory /home/ubuntu/public_html/>.
    • Add RedirectMatch ^/$ /live to the bottom of the file.

11. Enable the new wsgi site using sudo /usr/sbin/a2ensite dash.conf

12. Restart the apache server with sudo systemctl reload apache2.

13. Move the redis service file into its location and start the service.

    sudo cp /home/ubuntu/rfind-monitor/conf/redis.service /etc/systemd/system/redis.service
    sudo systemctl enable /etc/systemd/system/redis.service
    sudo systemctl start redis.service

    There are also a few lines you need to add to the redis.conf file - ask Nick for details.

14. Move the middle man service file into its location and start the service.

    sudo cp /home/ubuntu/rfind-monitor/conf/middleman.service /etc/systemd/system/middleman.service
    sudo systemctl enable /etc/systemd/system/middleman.service
    sudo systemctl start middleman.service

Refs

DASH

1 Intro to client-side callbacks and the extendData method.

2 Seems like someone was having this exact problem with large images. Will need to look into shaders.

DEPLOYMENT

3 Not much in here honestly. The user was on the wrong py version.

4 Explains wsgi conf file.

5 An example.

6 I always forgot this. Don't forget to a2ensuite!