Layout

The website is split two independent parts. (Which if needed could run on two different machines.)

The Frontend is the UI which provides the tool and feel of the system, this system requires no databases.

The Tracker is the backend which keeps track of stats about the system. It is only accessed through a JSON API.

For how this website fits into the overall streaming system look at the following diagram;

For how the website fits into the overall TimVideos.us project, look at the following diagram:

Initial Configuration

To get the code and dependencies, as a non-root user:

git clone git@github.com:timvideos/streaming-system.git
cd streaming-system/website

cp private/settings.py.example private/settings.py
vi private/settings.py

sudo apt-get install python-dev python-pip || sudo yum install python-devel python-pip
sudo pip install virtualenv

Running a test server

Run make serve; this will configure a virtualenv, download and install dependencies (inside the virtualenv; your system will not be touched); and a test server will be started.

If this is your first time running make serve you'll be prompted to provide a username and password for an admin account.

This test server should not be used in production (see relevant section below).

Requests are served at http://127.0.0.1:8000 by default. If you need to change the listening IP or port, try ARGS="192.168.1.1:8000" make serve where ARGS represents your desired listening IP and port.

This test server should NOT be used in production (see relevant section below).

Registering a (fake) flumotion encoder server onto tracker

The frontend will only show rooms which are available on the tracking module.

The register.py command is normally used on a flumotion encoder machine to register the encoder's existence. If you are not running an encoder machine locally, you can use the fake_register command to pretend you have one.

# cd ../tools/register
# ../../website/bin/python fake_register.py --ip 127.0.0.1

Sample output:

Registered example at 2013-04-24 09:32:43.961403 result OK
Registered example at 2013-04-24 09:32:45.130072 result OK
...

This keeps on going.

Open http://127.0.0.1:8000/example to view the page

Note: DO NOT run make serve as root. If you do, virtualenv folders won't have write permissions, so the tracker won't be able to write to sqlite database.

Setting up groups

Groups in the streaming system are one per "room" where you are streaming from.

The following options are available to each group.

• channel: The IRC channel to use.
• irclog: A URL to a HTML log of the IRC channel, to be shown on the overview.
• preview: A URL to an image containing a recent frame of the video stream.
• twitter: Twitter search parameters to use for the Twitter feed.
• logo: Logo to be displayed in some context
• title: Title of the stream.

Schedule options:

• link: Link to a HTML version of the schedule for the conference.
• schedule: Reference to a computer-readable version of the conference programme. Supports a few different format, default is eventfeed (Zookeepr).
• schedule-timezone: Olsen timezone identifier to interpret the conference programme with.

Flumotion (stream) options:

• flumotion-mixer:
• flumotion-encoder:
• flumotion-collector:

YouTube options:

• youtube: Video ID of where the stream is being broadcast. For the URI http://www.youtube.com/watch?v=dQw4w9WgXcQ, you should enter dQw4w9WgXcQ here. If not specified, YouTube playback will be disabled.

Setting up a new instance of the system

Copy config.private.json.example to config.private.json.

cd ..
cp config.private.json.example config.private.json
vi config.private.json

You will then need to update the following files;

• streaming-system/config.private.json
  • Update the default object with the event details (where it says timvideos you will need to add your events id).
  • The parent object should have a child object for each room with the room_id matching the event feeds (used by eventfeed2internal.py) room_ids
• streaming-system/website/Makefile#L53
  • You may need to update the name of eventfeed2internal.py if you are creating a new parser.
• streaming-system/website/frontend/eventfeed2internal.py
  • This should generate a output that resembles https://gist.github.com/lukejohnosmahi/7938504
• streaming-system/website/frontend/static/logos/*.png
  • Remove any old logos and add the events logo (158x217 is a known to work dimension).
• streaming-system/website/frontend/static/img/favicon.ico
  • Update the favicon.ico with your events logo.

Production Deployment

• To account for differences between the dev and prod infrastructure a config-private.json can be provided at the base level directory.

• The deployment uses one user to deploy the code, and another user to run the code:

  tim@website:~$ grep website /etc/passwd /etc/group
  /etc/passwd:website:x:100:100::/home/website:/bin/bash
  /etc/passwd:website-run:x:101:101::/home/website-run:/bin/false
  /etc/group:website:x:101:tim
  /etc/group:website-run:x:105:website,tim
  

• Usually, the run user should only need read access to the files you've checked out. If there are specific files or directories that the run user needs to write to, simply use chgrp website-run $FILE; chmod g+w $FILE to make them accessible to both the deploy and run users. If this needs to be persisted across deployments, you may have to take care of this in your deploy script.

• To do the deployment, I use the setup.sh script in conf directory. It takes a copy of the current code in ~website/timvideos and puts it in ~website/$VERSION-$DATE-$TIME directory and then links ~website/current to that.

  This makes it relatively easy to revert to an earlier version of the code.

• The app runs inside Green Unicorn, and is started by upstart. Take a look at conf/init.conf

• The app uses nginx as a frontend, and to serve static files. Only a few changes from the default config are needed to accomplish this.

  Take a look at conf/nginx.conf