Sending signals from DCL command line on OpenVMS - openvms

Question

I'm trying to send a signal via the command line on an OpenVMS server. Using Perl I have set up signal handlers between processes and Perl on VMS is able to send Posix signals. In addition, C++ programs are able to send and handle signals too. However, the problem I run into is that the processes could be running on another node in the cluster and I need to write a utility script to remotely send a signal to them.
I'm trying to avoid writing a new script and would rather simply execute a command remotely to send the signal from the command line. I need to send SIGUSR1, which translates to C$_SIGUSR1 for OpenVMS.
Thanks.

Answer 1

As far as I know, there is no supported command line interface to do this. But you can accomplish the task by calling an undocumented system service called SYS$SIGPRC(). This system service can deliver any condition value to the target process, not just POSIX signals. Here's the interface described in standard format:
FORMAT
SYS$SIGPRC process-id ,[process-name] ,condition-code
RETURNS
OpenVMS usage: cond_value
type: longword (unsigned)
access: write only
mechanism: by value
ARGUMENTS
process-id
OpenVMS usage: process_id
type: longword (unsigned)
access: modify
mechanism: by reference
Process identifier of the process for which is to receive the signal. The
process-id argument is the address of an unsigned longword containing the
process identifier. If you do not specify process-id, process-name is
used.
The process-id is updated to contain the process identifier actually
used, which may be different from what you originally requested if you
specified process-name.
process-name
OpenVMS usage: process_name
type: character string
access: read only
mechanism: by descriptor
A 1 to 15 character string specifying the name of the process for
which will receive the signal. The process-name argument is the
address of a descriptor pointing to the process name string. The name
must correspond exactly to the name of the process that is to receive
the signal; SYS$SIGPRC does not allow trailing blanks or abbreviations.
If you do not specify process-name, process-id is used. If you specify
neither process-name nor process-id, the caller's process is used.
Also, if you do not specify process-name and you specify zero for
process-id, the caller's process is used.
condition-value
OpenVMS usage: cond_value
type: longword (unsigned)
access: read only
mechanism: by value
OpenVMS 32-bit condition value. The condition-value argument is
an unsigned longword that contains the condition value delivered
to the process as a signal.
CONDITION VALUES RETURNED
SS$_NORMAL The service completed successfully
SS$_NONEXPR Specified process does not exist
SS$_NOPRIV The process does not have the privilege to signal
the specified process
SS$_IVLOGNAM The process name string has a length of 0 or has
more than 15 characters
(plus I suspect there are other possible returns having to do
with various cluster communications issues)
EXAMPLE CODE
#include <stdio.h>
#include <stdlib.h>
#include <ssdef.h>
#include <stsdef.h>
#include <descrip.h>
#include <errnodef.h>
#include <lib$routines.h>
int main (int argc, char *argv[]) {
/*
**
** To build:
**
** $ cc sigusr1
** $ link sigusr1
**
** Run example:
**
** $ sigusr1 := $dev:[dir]sigusr1.exe
** $ sigusr1 20206E53
**
*/
static unsigned int pid;
static unsigned int r0_status;
extern unsigned int sys$sigprc (unsigned int *,
struct dsc$descriptor_s *,
int);
if (argc < 2) {
(void)fprintf (stderr, "Usage: %s PID\n",
argv[0]);
exit (EXIT_SUCCESS);
}
sscanf (argv[1], "%x", &pid);
r0_status = sys$sigprc (&pid, 0, C$_SIGUSR1);
if (!$VMS_STATUS_SUCCESS (r0_status)) {
(void)lib$signal (r0_status);
}
}