The T Mahout example package provides classes to generate a sample dataset.

For K-means, DisplayKmeans is the class that displays the cluster. You can directly run the class. As per the code in the class, points are generated as follows:

generateSamples(500, 1, 1, 3); // 500 samples of sd 3
generateSamples(300, 1, 0, 0.5); //300 sample of sd 0.5
generateSamples(300, 0, 2, 0.1); //300 sample of sd 0.1

Data is a set of randomly-generated 2D data points, and the points are generated using a normal distribution centered at a mean location with a constant standard deviation.

Once you run this class, you will view the clusters, as shown here:

The final clustering done by the algorithm is shown using a bold red colored circle. In the console, you can find the output related to points generation and cluster formation.