Hadoop: code and fault injection

I’ve put this latest write up about injection framework I’ve put together into Hadoop.

This page will be getting more details about Hadoop development and testing using AspectJ based Injection Framework.


The idea of code injection is fairly simple: it is an infusion of new or modification of existing behavior into a code of a software application. Fault Injection, which will be discussed later, is a similar mechanism for adding errors and exceptions into an application’s logic to achieve a higher coverage and fault tolerance of the system. Different implementations of this idea are available today. Hadoop’s inject framework is built on top of Aspect Oriented Paradigm (AOP) implemented on top of AspectJ toolkit.

The document below relates to injection technique with regards to code/behavior modifications and to injection of programmatic faults (fault-injection). I’ll refer to both injection and fault injection aspects of the framework unless narrower cases are explicitly mentioned…

You can read the rest of it from Apache Hadoop Wiki site

US national debt is outgrow $12T by 11/10/09

Basically, all people leaving in this country and being mugged by its tax collection, are indebted for an extra $44K each second. Doesn’t sounds like much? Well…

According to my calculations the national debt of US will outgrow $12T (yes, it is trillions!) by midnight of 11/10/2009 (Tuesday of coming week). Take a look for yourself – Debt Clock

Let’s celebrate, folks! It’s going to be a true fest in a midst of a plague 😦

Hadoop and ApacheCon’09

I’ve started working on Hadoop at Yahoo! Great place and very interesting software project. By the way, we are hiring big time! Speak Java concurrently? Can explain how locking on object is different from class’ locking? Looking for great place to work? Leave me a comment here w/ your email and I’ll get back to you in no time.

That said, I’ve went to ApacheCon’09 last night (10 years anniversary of Apache Foundation) and happen to a group picture with The elephant – the actual Hadoop 🙂 – feels me with pride.