PIDFile

A pidfile is a small, plain-text file used in Unix-like systems to record the process ID (PID) of a running daemon or service. The file, commonly named program.pid and located in /var/run or /run, provides a simple reference that administrators and management scripts can use to identify or interact with the process, such as sending signals, checking status, or coordinating restarts. While not mandatory, a pidfile acts as a lightweight mechanism to indicate that an instance is already running.

Contents and semantics: A pidfile typically contains a single decimal number representing the PID of the running process. Some implementations may include extra metadata, but the PID is the essential content. The existence of the file can be used as a basic lock to prevent multiple simultaneous instances, though robust solutions employ explicit locking and startup checks to avoid race conditions.

Lifecycle and best practices: On startup, the daemon should check for an existing pidfile, verify that the referenced PID corresponds to a live process that belongs to the same program, and handle stale files by removing them. The pidfile should be created atomically, for example with an exclusive create operation, and removed on normal shutdown or after abnormal termination when feasible. The process should write its own PID and ensure proper cleanup via exit handlers. Access permissions should be restricted to appropriate users to prevent tampering.

Limitations and modern usage: Contemporary init systems and service managers, such as systemd, may manage process lifecycles differently, and pidfiles can be optional or discouraged in that context. Nevertheless, pidfiles remain common in traditional daemons and in scripts that manage processes outside supervisor control.

See also: daemon, /var/run, /run, systemd, process management.