Chapter 12: Managing Virtual File Systems and Core Dumps

Overview

The processes running on the system use physical memory for their code instructions and data. Because a system may be running many processes (applications) concurrently, it may run low on memory because memory is a finite resource. In this case, a predetermined space on the disk, configured as swap space, is used as memory. Because the space being used is not real physical memory, it is called virtual memory. The file system that manages the swap space is called a swap file system (SWAPFS) and is a virtual file system. Another related virtual file system is TMPFS, which uses the memory for file read and write and is a default file system for the /tmp directory. This file system is used to improve the performance of the applications, and uses the space resources from the swap space.

Virtual file systems make the storage space transparent to the processes. The network file system (NFS) makes the location of the machine on which the file systems exist transparent to the users and the applications. Using NFS, a user can mount a file system that exists on a remote NFS server and use it as though it existed on the local machine.

Both the processes and the system can crash from time to time; when they do, they leave some information behind in the form of core files and crash dump files that you can use to diagnose the problem. The core issue to think about in this chapter is: how are the memory location and the file location made transparent for the processes and how do we manage the crash information of a process or the system? In search of an answer, we will explore three thought streams: virtual memory, network file system, and process and system crashes.