With remote file locking, only one user or process can use a file at a time. Remote file locking allows for file locking and file sharing by NFS clients, and for the automatic maintenance of those locks in the event of a client or server crash. The following sections describe how remote file locking works, how to configure it for your network, and how to troubleshoot it.
Two RPC-based daemons interact to effect file locks: lockd and statd. lockd is the Network Lock Manager and statd is the Network Status Monitor. The following sections describe how they interact to lock files.
The Network Lock Manager provides file and record locking and DOS-compatible file sharing and locking in an NFS environment. The SCO NFS lock manager is the lockd daemon. lockd runs on both client and server to process lock requests and lock releases, and to handle crash recovery operations. For more information, refer to the lockd(NADM) manual page.
The Network Status Monitor provides application processes with host status information. The SCO NFS Network Status Monitor is the daemon statd. statd runs on both client and server in such a way that client and server monitor their own status and that of each other.
To effect properly monitored locks, lockd relies upon notification of host state changes from the local statd. When the client statd detects a client crash, it notifies the server statd of the state change so that the server can release any locks it is holding for the client. When the server crashes, its local statd notifies the client statd of the state change; the client statd informs the local lockd and lockd reclaims any existing locks established on the server.
For a detailed description of the sequence of events involved in the
interaction of the Network Lock Manager and the Network Status Monitor
during locking, crash, recovery, and unlocking, see the standards
outlined in the Common Applications Environment (CAE)
specification Protocols for X/Open Internetworking: XNFS.
For further information
on the daemon statd, see the