This book describes Symantec’s Veritas Cluster File System (CFS), an implementation of the cluster architectural model for sharing data among multiple computers. Based on Symantec’s long-proven Storage Foundation File System (commonly known as VxFS), CFS is exceptionally scalable, robust, and high-performing. Building on a solid base of VxFS technology, CFS adds clusterwide cache coherency, distributed resource control and file system transactions, and other features that enable scaling and enhance performance and load balancing across a cluster of application or database servers. An astute reviewer noted that the book seems to address different audiences, and that is indeed the case. The Introduction sets the stage by defining the problem of sharing data among multiple application servers and contrasts the CFS approach with other commonly encountered solutions.
Part I presents an overview of CFS, with emphasis on the features that make it unique and the capabilities they enable, along with some examples of scenarios in which CFS is particularly suitable. It should be of interest to new application designers who require an appreciation of online data management in the data center environment.
Part II describes the CFS “secret sauce”-the internal architecture that gives CFS its unique combination of scalability, flexibility, robustness, and performance. It should be of interest to experienced developers and administrators who need to understand what’s going on “under the hood.”
Part III is a guide to installing CFS and tuning file systems. Multi-server cluster environments are inherently complex, as are their file I/O requirements. This part should be of interest to system administrators who are either installing CFS for the first time or are tasked with monitoring and managing the performance of a CFS cluster.
Most of the material is available in some form from other sources. The unique contribution of this book is to “pull it all together,” and to answer the question, “why” in addition to describing the “what.”
Books tend to live longer than software product versions. While every effort has been made to ensure the accuracy of the material, you should consult current product documentation and support sources before putting the principles and techniques described herein into practice.
This book is a collaborative effort of many experts in CFS, and indeed, in the field of computer file systems at large. The following are the principal technical contributors:
Jobi Ariyamannil, Brad Boyer, Sai Chivukula, Bala Kumaresan, David Noy, Pramodh Pisupati, Hal Prince, Mike Root, Meher Shah