Open System Services NFS Overview

Pathnames
To specify a file in the current OSS, DOS, or UNIX directory, you provide the name of a file. To
specify a file in some other directory, you specify a pathname either from the current directory or
from a parent or root directory.
You first mount the portion of the OSS NFS file hierarchy you want to use by specifying one of the
mount points (directory pathnames) provided by the system manager. Then you can use a command
on your local system to list the names of all directories and files you have mounted. For example,
you might use the ls command on a UNIX system or the dir command on a DOS system. You can
determine the file hierarchy and how to specify the NFS pathnames to the files you want to use.
NFS Pathname Syntax
As a DOS or UNIX user, you are familiar with the syntax of pathnames on your operating system.
OSS NFS pathnames generally follow the syntax rules for UNIX pathnames modified by NFS
standards. The general rules are as follows:
• A pathname consists of directory names separated by slashes (/) followed by a filename, for
example: /usr1/source/progabc.
• A directory name or filename can contain any character except the slash (/) or ASCII NUL.
The names are case-sensitive; that is, the names myfile, MYFILE, and Myfile are distinct.
• The maximum length of a directory name or filename is 255 bytes.
• The maximum length of a pathname is 1024 bytes.
The names you use for files managed by NFS servers can follow the general NFS syntax rules.
This means that your existing UNIX or DOS files can be stored on a NonStop system through an
NFS server using the name that the file has on your local system.
NOTE: Your NFS client program probably has its own mechanisms for mapping names, which
you must also consider. The information given here describes only how the OSS NFS servers
process the names.
NFS Example
Before examining how OSS NFS servers manage files, consider the way directories are mounted
locally on a UNIX system to create a file hierarchy. Figure 1 (page 11) shows an example of a
structure created by mounting subtrees from four devices: sd0, sp1, sd1, and sd2. The following
commands create the file hierarchy:
makefs
mount /dev/sd0 /
mkdir /etc
mkdir /usr
mount /dev/sp1 /etc
mount /dev/sd1 /usr
mkdir /usr/lib
mount /dev/sd2 /usr/lib
The mount commands associate physical devices (/dev/sd2) with directories in the UNIX file
hierarchy (/usr/lib).
10 Introduction to NFS for Open System Services