The idea of changing root is that of running security-critical applications in a minimal environment so that an attacker who manages to subvert the application has limited access only to the chroot environment and can use only the tools that this environment supplies.

As with all ideas to improve security, chrooting presents its own share of security problems. If a user gains root privileges, she can break out of a chroot environment by either remounting the file system or chdir() out of the chroot until the original root is reached, and then chroot to this directory. So we should configure our Lighttpd to give up root privileges in addition to changing root, and also be sure not to have any User ID executable in our chroot environment.

Set User ID (SUID for short) is a method to run a process under a different user's privileges. UNIX file systems allow setting up a SetUID flag on an executable combined with a User ID. The program will be executed with the privileges of the User...