Chapter 13: The Cluster File System (CFS)


Overview

In this chapter we discuss one of the truly outstanding features of Tru64 UNIX clustering: the Cluster File System. You may have heard the term "single-system image (SSI)", and while we disagree with the term "image" because it suggests using one kernel for all cluster members, the CFS does provide a common, transparent view of all file systems to all cluster members. This ability, along with the Cluster Alias subsystem (covered in chapter 16), makes up the heart and soul of Tru64 UNIX clustering.

You may be thinking, "What's so special? I can do that with NFS!" No, actually you can't. Can NFS enable every member in a cluster to share the root (/) file system? No. Moreover, the CFS does this transparently. No fancy mount commands, no additional configuration files.0 If you want to mount a file system on a cluster, you log in to any member in the cluster, issue a normal mount(8) command, and all members in the cluster see the same file system at the same mount point. And what happens to an NFS file system if the NFS server crashes? Bye-bye file system. The CFS does not suffer from this problem because another cluster member will transparently take over the job of serving the file system to the other cluster members. The CFS does all this and still manages to adhere to the X-Open and POSIX standards. This is really cool stuff!

The net result is that in a TruCluster Server environment you can log in to any member of the cluster and have virtually an identical view of the files on the cluster mounted file systems. We say "virtually" because some files will be member-specific, implemented as CDSLs (see discussions in chapters 6 and 12).




TruCluster Server Handbook
TruCluster Server Handbook (HP Technologies)
ISBN: 1555582591
EAN: 2147483647
Year: 2005
Pages: 273

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net