The watchdogd utility interfaces with the kernels watchdog facility to ensure that the system is in a working state. If watchdogd is unable to interface with the kernel over a specific timeout, the kernel will take actions to assist in debugging or restarting the computer.
If -e cmd is specified, watchdogd will attempt to execute this command with system(3), and only if the command returns with a zero exit code will the watchdog be reset. If -e cmd is not specified, the daemon will perform a trivial file system check instead.
The -s sleep argument can be used to control the sleep period between each execution of the check and defaults to one second.
The -t timeout specifies the desired timeout period in seconds.
One possible circumstance which will cause a watchdog timeout is an interrupt storm. If this occurs, watchdogd will no longer execute and thus the kernels watchdog routines will take action after a configurable timeout.
Upon receiving the SIGTERM or SIGINT signals, watchdogd will first instruct the kernel to no longer perform watchdog checks and then will terminate.
The watchdogd utility recognizes the following runtime options: