NFS Performance Tuning for HP-UX 11.0 and 11i Systems

nfs performance tuning for hp-ux 11.0 and 11i systems page 60

Notes:

Page 60July 22, 2002

rpc.lockd

rpc.statd

What are rpc.lockd and rpc.statd?

• Implement the Network Lock Manager (NLM)

Protocol, providing NFS file locking semantics

• rpc.lockd handles lock requests

• rpc.statd monitors the systems involved in NFS file

locking and is an integral component in recovering

locks after system failures

• NLM is required to maintain some “state” information as

opposed to NFS which is a “stateless” protocol

The NFS protocol is commonly described as being “stateless” because the NFS

server maintains no state information about the clients. Every NFS request is

treated individually by the server and no assumptions are made based on previous

requests. This design works fine, except when it comes to NFS file locking.

In the file locking case, both the client and the server need to maintain a list of

which files are locked, by which processes, and on which NFS clients. Some of

this information must be stored on local disk in order for it to survive a client or

server reboot (since anything in memory is lost during a reboot). In order for NFS

to maintain this “state” information on top of a “stateless” protocol, a new

mechanism was created to manage NFS file locks – the Network Lock Manager

(NLM) protocol.

On HP-UX systems, the rpc.lockd and rpc.statd daemons (commonly referred to as

simply “lockd” and “statd”) implement the Network Lock Manager protocol.