The official location for Hadoop is the Apache SVN repository; Git is also supported, and useful if you want to make lots of local changes -and keep those changes under some form or private or public revision control.
The official location is now the Apache Git repository.
A lot of people use Git with Hadoop because they have their own patches to make to Hadoop, and Git helps them manage it. Apache serves up read-only Git versions of their source at http://git.apache.org/. People cannot commit changes with Git; for that the patches need to be applied to the SVN repositories.