As part of a data analytics course with Udacity I recently took, I had the chance to play a bit with Hadoop. It seems like I never want to take the easy path and just do the assignment, as this seemed liked the perfect chance to also experiment with Amazon’s Elastic MapReduce (EMR). And of course, why not also experiment with using node.js as well. It turns out after a bit of tweaking command line paramaters, it is actually relatively easy to spin up a EMR cluster, send it some jobs, and then shut it down.

There are many reasons someone might need to create a custom authentication method in meteor. For our project it was the use of a custom CAS authentication server for which there was no available module since it was a completely custom fork of CAS.

I took the plunge and decided to migrate from Aperture to Photos. I really like taking photos, and like to pretend that I’m a better photographer than I am. To this end, for many years I have used Aperture for Mac OS X to grab photos of my camera then tag, rate, organize, and do some basic brightness/contrast adjustments to the ones that really needed it.

I have been recently working quite a bit with single page applications, and I love putting the same technology to use for personal projects, that I use on a daily basis for professional projects.

A big fan of meteor right now. Making apps in meteor can be quite a pleasure. Here are the modules I almost always end up using.

I have run a personal server for as long as I can remember. It started with 10BASE2 connections across my machines that I put together in middle school with the help of a neighbor, and has been going strong ever since. It evolved into hosting sites over a cable modem (one of which maxed out the bandwidth for a while), then to a machine running on a static IP in a dorm room at college. Things got serious after that, with co-location fees starting to come in for my personal server. Finally, the co-location evolved away and I had a virtual servers–switching several times between AWS and RackSpace.

Setting up a new installation of Drupal 7 is really quite easy on Fedora. All the software you need is available with a simple yum install command, so it is pretty much just a matter of installing the necessary services, and downloading the code. This video goes through setting up a brand new Rackspace cloud server running Fedora 16 so that it will serve drupal pages.

Using replication is a fast and easy way to keep a continuous backup of data in MYSQL. It cannot guard against user error and accidental deletes, and therefore should not be the only tool in use for backups; however, when recovering from a server outage it is nice to know you will not loose more than a second or two of data.