When users access NFS volumes, simple commands, like ls, must appear to retrieve remotely stored information over the network. However, ls does not understand networks at all; it just asks the kernel for a directory's contents, formats the results, and displays them. There must be some transparent mechanism handling the network communications. Here, the kernel sees that the directory is mounted over NFS and makes a Remote Procedure Call (RPC).

When Linux needs a server to perform a service like accessing an NFS volume, it sends an RPC request packet to the server. The client then waits a predetermined amount of time before making the request again. Meanwhile, the server tries to perform the requested job and reply to the client with the results. Every single command does not need to know how to handle NFS; each command simply requests information from the kernel or appropriate daemon, which makes the RPC. The following SlideShow summarizes the process used to retrieve NFS information for the ls command from a server.

One caveat to the RPC process is that the server and client must be running their respective portmap daemons. The daemons set up the communication channels between the RPC client and server. The next lesson shows you how to use the mount command to access NFS shares.

Before moving on to the next lesson, click the Quiz link below to review the basics of RPC and NFS.
RPC NFS - Quiz