This is Part 1 in a series of blog posts describing my move of ipythonblogs.org from Rackspace to Heroku. In this first post I’ll describe the existing deployment and what I intend to change during the migration. Other posts are:

• Part 1: Introduction and Architecture
• Part 2: Data Migration
• Part 3: Database Interface
• Part 4: Application Updates

As a side project I maintain a Python library called ipythonblocks that displays colored grids in a Jupyter Notebook. (See also this intro blog post.) It can be useful for teaching or for a bit of fun art. I also maintain the website ipythonblocks.org that allows users to post their ipythonblocks grids on the internet to be shared. ipythonblocks.org has been hosted on Rackspace since I first launched it, but now I’m migrating the site to Heroku for easier maintenance and deployment.

Rackspace Deployment

ipythonblocks.org has been running on a single Rackspace computer with 2 GB of RAM. I deployed it by copying code up over SSH then connecting over SSH and running a shell script. Very low tech! There were two Docker containers running, one for the Python app that served the site and one for memcached (used to cache the results of database lookups). The grid data sent to ipythonblocks.org was stored in a sqlite database on disk.

The application itself uses Tornado to build and serve HTML pages, there’s no client side JavaScript. You can read more about the libraries used in the app in this previous blog post. You can browse the application code on GitHub in its state prior to the migration here.

Why Change Anything?

The site periodically goes down for unknown reasons and I’m over SSH-ing into the computer to restart things until it works again. I’ve been hosting Days Since it Rained on Heroku for years and never had to think about it and I’d like that same level of reliability for ipythonblocks.org. The site gets practically no traffic so I can easily use Heroku’s free tier.

Planned Changes

The biggest change will be not using Docker anymore when Heroku is handling all the server stuff for me. I’ll be able to stop using all code related to deployment and focus only on application code. I’m also planning to make some other changes:

Update to Python 3.6

I mean, of course.

SQLite to Postgres

Heroku doesn’t have a persistent filesystem so I can’t use SQLite, but Heroku does offer a Postgres add-on with a free tier. As an added bonus I can use Postgres’ JSON types, which are a good fit for some of the data I store. (To store JSON-like data in SQLite I wrote some wrapper code to convert it to/from strings.)

Dataset to SQLAlchemy

I originally used dataset as a SQL abstraction for ipythonblocks.org. I’m switching to SQLAlchemy so I can use Postgres’ JSON data type and so I can use SQLAlchemy’s ORM, which is a good fit for a web app.

No memcached

Caching is a really great idea in web apps, but ipythonblocks.org sees so little traffic that I think I’ll be able to read from the DB for every request without problem and simplify the app logic.

What’s Next

In my next post I’ll describe the grid data model and how I migrated the existing grid data from SQLite to Postgres.