This section describes the internal operations when NFS is started and when a remote resource is mounted.
Starting NFS
The NFS daemons are started resp. restarted by running the script "/etc/rc2.d/S20nfs start" or by sending the signal SIGTRAP to the init process (PID 1). This is done automatically at POSIX subsystem start or at NFS (update) installation.
By default following daemons are started:
1 mountd 4 nfsd 4 biod 1 pcnfsd 1 rpcbind
Mounting a remote resource
As an example for all mount operations involving remote resources, a description follows of what happens when the command mount -F nfs ... is entered.
You enter the following command:
mount -F nfs -o intr,rw,soft tanz:/usr/src /usr/src/tanz.src
A check is made whether in the directory /etc/fs/nfs the mount command for mounting remote resources is present and whether the entered command is syntactically correct.
The mount command checks in the file /etc/mnttab whether the resource is already mounted.
Via the rpcbind daemon, the mount command requests the port number which the mountd daemon has on the computer tanz.
The mount command passes the path name of the resource (/usr/src) to the mountd daemon on the computer tanz.
The mountd daemon on the computer tanz processes the mount request and checks whether the resource has been shared and with what permissions.
The mount command receives a positive response from the mountd daemon on the computer tanz and adds an entry for the mounted resource to the file /etc/mnttab.
When a file from the mounted resource is to be processed, the request is processed by the nfsd daemon on the computer tanz. Using the options specified when sharing the resource, this daemon checks whether the request can be satisfied or must be rejected.