Using a Network File System (NFS) directory for document storage

Network File System (NFS) is a distributed file system implementation providing remote, transparent access to files and directories. OnDemand can use an NFS exported directory as a disk pool for document storage.

A few key points regarding NFS include the following:
  • You make a directory or an object available to NFS clients by exporting it. Therefore, you have very specific control over which parts of your system you will make available to NFS clients in your network.
  • When you export, you can specify which clients have access to the objects. You identify a client by system name or IP address.
  • You make a directory or an object available on your NFS client by mounting it.
  • When you mount remote server file systems over local client directories, you allow IBM® i servers to work with file systems that have been exported from a remote server. The mounted file systems will act and perform as if they exist on the local server.
  • The NFS does not provide password protection. It is designed and intended for data sharing within a trusted community of systems. When a user requests access, the server receives the user's User ID number (UID). The UID is used to determine the permissions of the user.
  • The use of synchronous writes is recommended. The use of asynchronous writes can result in data loss in the event of a system failure on either the archive or storage system.
  • Each NFS mount (across all instances and logical partitions (LPARs)) must use a unique NFS directory.
  • When using NFS with OnDemand, note that directories located in an Independent ASP (IASP) cannot be mounted over an NFS exported directory. This limitation affects only installations that locate a complete instance, including the instance library, within an IASP.

More information on NFS can be found within the IBM i information center at http://www.ibm.com/systems/i/infocenter/