In the clustered environment we have large number of independent components cooperating or collaborating on a computation. Any of this vast number of components can fail at any time, resulting in erroneous output. There are many techniques have been developed to resilience to these kinds of faults. This paper discuss various techniques of fault-tolerance. Success of cluster computing is based on the low cost, widely used commercial of-the-self(COTS) hardware and software. Copyright © 2012 ACM.