Scroll to navigation

KILLPROC(8) The SuSE boot concept KILLPROC(8)


Killproc - Send signals to processes by full path name


killproc [-vqLN] [-g|-G] [-x] [-p pid_file] [-i ingnore_file] [-c root] [-t <sec>] [-<SIG>] /full/path/to/executable

killproc -n [-vq] [-g|-G] [-t <sec>] [-<SIG>] name_of_kernel_thread

killproc    [-vq] [-g|-G] [-t <sec>] [-<SIG>] basename_of_executable

killproc -l


killproc [-p pid_file] /full/path/to/executable [-<SIG>]


killproc sends signals to all processes that use the specified executable. If no signal name is specified, the signal SIGTERM is sent. If this program is not called with the name killproc then SIGHUP is used. Note that if no signal is specified on the command line and the program killproc can not terminate a process with the default SIGTERM the signal SIGKILL is send after a few seconds (default is 5 seconds, see option -t). If a program has been terminated successfully and a verified pid file was found, this pid file will be removed if the terminated process didn't already do so.

killproc does not use the pid to send a signal to a process but the full path of the corresponding program which is used to identify the executable (see proc(5)). Only if inode number (/proc/<pid>/exe) and full path name of the executable (/proc/<pid>/cmdline) are unavailable or if the executable has changed its zeroth argument, killproc uses the base name (/proc/<pid>/stat) to identify the running program. Note that if the option -n for kernel thread is given, only (/proc/<pid>/stat) is used. For this case an existing symbolic link (/proc/<pid>/exe) indicates that the <pid> is not a kernel thread.

The last variant of the killproc call is only with the basename of the executable. This is required by the Linux Standard Base Specification (LSB). Note that this disables the comparision of the inodes of the executable and the information found within the proc table (see proc(5)). This implies that if no pid file is found it may happens that programs or script with the same base name will killed instead. killproc tries to determine the pid of its parent and pid of the parent of its parent to not to kill those two processes.

Extended functionality is provided by the -p pid_file option (former option -f changed due to the LSB specification). If this option is specified, killproc tries to send the signal to the pid read from this file instead of from the default pid file (/var/run/<basename>.pid). If the pid_file does not exist, killproc assumes that the daemon is not running. It is possible to use a process identity number instead of a pid file.

For the possibility of having two different sessions of one binary program, the option -i ignore_file allows to specify a pid file which pid number is used to ignore all processes of corresponding process session.

The option -v turns on verbosity about which signal is being sent. The option -q is ignored for compatibility with older versions. If the option -G is used, the signal will not only sent to the running executable, but also to all members of the session that the specified executable leads. Whereas the option -g the signal sends to the group including the executable. If unsure use -G.


/full/path/to/executable or name_of_kernel_thread
Specifies the executable to which the signal should be sent, or alternatively, if the option -n is used, the name of the kernel thread. This argument is always required.


Sends the signal to all session followers (children) of the identified process.
Sends the signal to all members of the session including the identified process. Note that usually the option -G should be used.
This option causes symlinks to be followed, as the like-named option in ls(1). Note: for the file name the original name of the program is used instead of the name of the symbolic link.
(Former option -f changed due to the LSB specification.) Use an alternate pid file instead of the default (/var/run/<basename>.pid). If the pid_file does not exist, killproc assumes that the daemon is not running. It is possible to use a process identity number instead of a pid file.
The pid found in this file is used as session id of the same binary program which should be ignored by killproc.
Change root directory to root for services which have been started with this option by startproc(8).
This option indicates that a kernel thread should be signaled. In this case not the executable with its full path name is required but the name of the kernel thread.
With this option the location of the executable is checked about NFS file system and if true the stat(2) system call is not applied on the exe symbolic link under /proc(5). Otherwise killproc could be locked if the corresponding NFS server is currently not online or available. This implies that the inode number check between the exectuable on the command line and the exectuable of the exec symbolic link will be skipped.
Signals can be specified either by name (e.g. -HUP, -SIGHUP) or by number (e.g. -1).
The number <sec> specifies the seconds to wait between the default signal SIGTERM and the subsequentially signal SIGKILL if the first SIGTERM does not show any result within the first few milli seconds. This timeout defaults to 5 seconds. On the other hand if the signal SIGTERM was explicitly used on the command line, the signal SIGKILL will be omitted, even if the process was not terminated after the timeout has expired.
This option is ignored.
Be more verbose.
This option list all available signals and some of their synonyms by their number and signal names to standard out. and exits.
Tells killproc to use the upto 15 byte long filename name found in /proc/<pid>/stat to be able to identify scripts overwriting their command line.


sends the signal SIGTERM to the running sendmail process. If a pid file exists in /var/run/ then the pid from this file is being used after verification to terminate the sendmail process. Other running processes are ignored.
sends the signal SIGTERM to the pid found in /var/myrun/ if and only if this pid belongs to /usr/sbin/lpd. If the named /var/myrun/ does not exist, killproc assumes that the daemon of /usr/sbin/lpd is not running. The exit status is set to 0 for successfully delivering the default signals SIGTERM and SIGKILL otherwise to 7 if the program was not running. It is also successful if no signal was specified and no program was there for Termination because it is already terminated.


The exit codes have the following LSB conform conditions:

Success or program was not running (no signal specified)
Generic or unspecified error
Invalid or excess argument(s)
Insufficient privilege(s)
Program is not installed
Program was not running to receive the specified signal

In some error cases, diagnostic output is sent to standard error, or, if standard error is not available, syslogd(8) is being used.


killproc is a replacment for the Bourne shell function killproc found in the widely used SysVinit package of Miquel van Smoorenburg, <>.


Identifying a process based on the executable file and the corresponding inode number only works if the process stays alive during killproc's execution. Impure executables like shell scripts (the inode number of the shell is not identical to that of the script) and programs rewriting their zeroth argument may not be identified by a file name.

Killproc isn't able to signal processes being in the zombie state. Zombies are processes which arn't alive but listed in the process table to have the exit status ready for the corresponding parent processes.


path to the proc file system (see proc(5)).
path to the SuSE boot concept script base directory as required by the Linux Standard Base Specification (LSB) (see init.d(7)).


startproc(8), checkproc(8), insserv(8), init.d(7), kill(1), skill(1), killall(8), killall5(8), signal(7), proc(5).


1994-2005 Werner Fink, 1996-2005 SuSE GmbH Nuernberg, Germany.


Werner Fink <>

November 10, 2000 3rd Berkeley Distribution