We use it extensively at LinkedIn. Many of them data-driven features you see on the site are powered by it. It's greatly improved not only the speed at which the data can be built but most importantly what can be done.

I've also introduced it to a start-up I worked for 1.5-2 years ago. I essentially did much of the optimization Todd described "in the dark"; this was before Cloudera was formed and when #hadoop on FreeNode only had a handful of Yahoo, Facebook and Rapleaf people.

Despite the amount of work involved (including introducing Java-based project to a LAMP-based start-up which was very ambivalent towards Java), it was a great productivity boost (over the mix of ad-hoc shell/Perl/PHP scripts a MySQL datawarehouse we started with).

MapReduce isn't just about scalability and performance. You don't need to have a "scalability crisis" to benefit from it. It's also about being able to do a great deal more by applying parallelizable algorithms (see for example, the Mahout or Katta projects).

I use it (and Hive) extensively at Facebook.

We have a vSphere installation, and have a hadoop VM on every physical machine, eating up spare cycles.

We're very happy with it.