PyFCGI/src/main.c

/*
 * Copyright (C) 2019 Weber Yann
 * 
 * This file is part of PyFCGI.
 * 
 * PyFCGI is free software: you can redistribute it and/or modify
 * it under the terms of the GNU Affero General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * any later version.
 * 
 * PyFCGI is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU Affero General Public License for more details.
 * 
 * You should have received a copy of the GNU Affero General Public License
 * along with PyFCGI.  If not, see <http://www.gnu.org/licenses/>.
 */

/**@defgroup processes Process organisation of PyFCGI
 *
 * PyFCGI is organised in three layer : 
 * A @ref main_proc : simple, that keep running and spawn a
 * @ref work_master_proc . This process handles @ref worker_process creation
 * and try to maintain a pool able to reply efficiently to CGI requests.
 */
/**@defgroup main_proc Main process
 * @brief The main process in the @ref main() function
 * @ingroup processes
 */

#include <fcgi_stdio.h> /* fcgi library; put it first*/

#include <unistd.h>
#include <stdlib.h>
#include <syslog.h>
#include <string.h>
#include <errno.h>
#include <signal.h>
#include <time.h>
#include <sys/types.h>
#include <sys/wait.h>

#include "conf.h"
#include "logger.h"
#include "responder.h"
#include "ipc.h"

#define IDENT_FMT "pyfcgi[%d]"
#define MAX_REQS 1024

#define EARLY_ERR(err_str) write(2, err_str, strlen(err_str))

extern pyfcgi_conf_t PyFCGI_conf;

pid_t pool_handler_pid;
pid_t monitor_serv_pid;

void sighandler(int signum)
{
	int status, ret, i;
	struct timespec req;

	req.tv_sec = 0;
	req.tv_nsec = 100000000; //0.1s

	if(signum == SIGALRM)
	{
		pyfcgi_log(LOG_WARNING, "Master process received SIGALRM !");
		return;
	}
	else if(signum == SIGINT)
	{
		pyfcgi_log(LOG_INFO,
			"Master process received ctrl+c, exiting...");
	}
	else
	{
		pyfcgi_log(LOG_INFO,
			"Master process received signal %s(%d), exiting...",
			strsignal(signum), signum);
	}
	if(pool_handler_pid)
	{
		pyfcgi_log(LOG_INFO, "Killing pool_handler(%d)",
			pool_handler_pid);
		kill(pool_handler_pid, SIGTERM);
		for(i=0; i<5; i++)
		{

			ret = waitpid(pool_handler_pid, &status, WNOHANG);
			if(ret <= 0)
			{
				nanosleep(&req, NULL);
				continue;
			}
			pool_handler_pid = 0;
			break;
		}
	}
	if(monitor_serv_pid)
	{
		pyfcgi_log(LOG_INFO, "Killing monitor(%d)",
			monitor_serv_pid);
		kill(monitor_serv_pid, SIGTERM);
		for(i=0; i<5; i++)
		{

			ret = waitpid(monitor_serv_pid, &status, WNOHANG);
			if(ret <= 0)
			{
				nanosleep(&req, NULL);
				continue;
			}
			monitor_serv_pid = 0;
			break;
		}
	}
	if(pool_handler_pid)
	{
		pyfcgi_log(LOG_WARNING,
			"pool_handler(%d) seems to freeze, sending SIGKILL",
			pool_handler_pid);
		kill(pool_handler_pid, SIGKILL);
	}
	if(monitor_serv_pid)
	{
		pyfcgi_log(LOG_WARNING,
			"pool_handler(%d) seems to freeze, sending SIGKILL",
			monitor_serv_pid);
		kill(monitor_serv_pid, SIGKILL);
	}
	pyfcgi_log(LOG_INFO,
		"Master process exiting.");
	exit(0);
}

void debug_sighandler(int signum)
{
	pyfcgi_log(LOG_WARNING, "Master process received signal %s(%d)",
		strsignal(signum), signum);
	return;
}

int main(int argc, char **argv)
{
	int emerg_sleep = 3;
	int child_ret;
	pid_t rpid;
	struct sigaction act;
	short fails, need_wait;
	//Sleeping on waitpid WNOHANG
	struct timespec tsleep;
	char *child_name;

	act.sa_handler = sighandler;
	sigemptyset(&act.sa_mask);
	act.sa_flags = 0;
	act.sa_restorer = NULL;

	if(sigaction(SIGTERM, &act, NULL))
	{
		perror("Sigaction error");
		exit(4);
	}
	if(sigaction(SIGINT, &act, NULL))
	{
		perror("Sigaction error");
		exit(4);
	}
	act.sa_handler = debug_sighandler;
	if(sigaction(SIGALRM, &act, NULL))
	{
		perror("Sigaction2 error");
		exit(4);
	}


	pool_handler_pid = 0;
	monitor_serv_pid = 0;
	default_conf();
	pyfcgi_name_IPC(getpid()); //name semaphore using master proc PID

	pyfcgi_logger_init();
	pyfcgi_logger_set_ident("MainProc");
	if(parse_args(argc, argv))
	{
		return 1;
	}

	pyfcgi_log(LOG_INFO, "New server started");

	while(1)
	{
		if(!pool_handler_pid)
		{
			pool_handler_pid = spawn_pool_handler();
			if(pool_handler_pid < 0)
			{
				fails = 1;
				pool_handler_pid = 0;
			}
		}
		if(PyFCGI_conf.mon_socket && !monitor_serv_pid)
		{
			monitor_serv_pid = pyfcgi_spawn_monitor();
			if(monitor_serv_pid < 0)
			{
				fails = 1;
				monitor_serv_pid = 0;
			}
		}
		if(fails)
		{
			if(emerg_sleep > 600)
			{
				pyfcgi_log(LOG_EMERG, "Abording...");
				exit(PYFCGI_FATAL);
			}
			fails = 0;
			pyfcgi_log(LOG_WARNING, "Sleeping %ds",
				emerg_sleep);
			emerg_sleep *= 2;
			continue;
		}
		else
		{
			emerg_sleep = 3;
		}

		need_wait = ((!PyFCGI_conf.mon_socket || monitor_serv_pid)
			&& pool_handler_pid);
		rpid = waitpid(0, &child_ret, need_wait?0:WNOHANG);
		if(rpid == pool_handler_pid)
		{
			child_name = "Pool handler";
			pool_handler_pid = 0;
			if(!child_ret)
			{
				pyfcgi_log(LOG_NOTICE,
					"Restarting main process after %d requests",
					MAX_REQS);
				continue;
			}
		}
		else if(rpid == monitor_serv_pid)
		{
			child_name = "Monitor server";
			monitor_serv_pid = 0;
		}
		else if(rpid == 0 && !need_wait)
		{
			tsleep.tv_sec = 0;
			tsleep.tv_nsec = 100000000;
			nanosleep(&tsleep, NULL);
			continue;
		}
		else if(rpid < 0)
		{
			pyfcgi_log(LOG_EMERG, "Unable to waitpid : %s",
				strerror(errno));
			exit(PYFCGI_FATAL);
		}
		else
		{
			child_name = "Unknown child";
			pyfcgi_log(LOG_WARNING,
				"Unknown child (PID %d) exited...", rpid);
		}

		if(child_ret)
		{

			pyfcgi_log(LOG_ERR, "%s exits with error code : %s",
				child_name, status2str(WEXITSTATUS(child_ret)));
			if(WIFSIGNALED(child_ret))
			{
				pyfcgi_log(LOG_ERR, "%s terminated by sig %s(%d)",
					child_name,
					strsignal(WTERMSIG(child_ret)),
					WTERMSIG(child_ret));
			}
		}
		else
		{
			pyfcgi_log(LOG_WARNING,
				"%s exits with no error status 0",
				child_name);
		}
	}

	closelog();
}

/**@mainpage PyFCGI
 * @section main_what What is PyFCGI ?
 * PyFCGI is a simple python3 fastcgi runner.
 *
 * Usage : Usage : spawn-fcgi [OPTIONS] -- pyfcgi -e PYMODULE [-E PYFUN] [OPTIONS]
 *
 * To run foo.entrypoint() python function.
 *
 * @subsection main_how_use How to use it ?
 * 
 * @warning For the moment PyFCGI is under heavy developpement and in early
 * stage. Everything will change ;o)
 *
 * PyFCGI should be runned with spawn-fcgi (or something similar), allowing
 * to configure & forward environnement variables to libFCGI.
 *
 * For the moment no configuration files exists. You have to pass arguments
 * to pyfcgi using -- argument of spawnn-fcgi
 *
 * When called this function will have to send valid CGI data to the webserver
 * using sys.stdin (the print() function for exemple) like :
 *<code>Content-type: text/html\\r\\n\\r\\nHello world !\\n</code>
 *
 * The function will have access to updated CGI os.environ containing
 * all informations about a request
 *
 * @subsubsection main_how_use_syslog Logging, using syslog
 * 
 * Right now PyFCGI uses pyfcgi_log() to log stuff using a pyfcgi ident.
 * PyFCGI logs can be filtered using /etc/rsyslog.d/pyfcgi.conf :
 * 
<pre>
if ($programname contains 'pyfcgi') then {
	-/var/log/pyfcgi/pyfcgi.log
	stop
}
</pre>
 *
 * @subsubsection main_how_use_hardcode Hardcoded stuffs
 *
 * Right now there is a lot of hardcodd stuff. Fortunatly a vast majority
 * is in @ref main.c like the minimum and maximum number of workers, or
 * the number of requests before a worker restart. Some timers and stuff
 * are harcoded but should not in @ref pyworker.c & @ref responder.c
 *
 * @subsection main_how_works How it works ?
 *
 * - @ref processes
 *  - @ref main_proc
 *  - @ref work_master_proc
 *  - @ref worker_process
 *
 */

/**@page pyfcgi
 * @brief Python Fast CGI runner
 * @section man_syn SYNOPSIS
 * spawn-fcgi [OPTS] -- pyfcgi [OPTIONS]
 * @section man_desc DESCRIPTION
 * Run WSGI python application in a pool of worker.
 *
 * @section man_opt OPTIONS
 * @subsection man_opt_spawn spawn-fcgi OPTIONS
 * OPTS are @ref spawn-fcgi(1) options.
 *
 * @subsection man_opt_gen General OPTIONS
 * @par -h --help
 * Display help and exit
 * @par -V --version
 * Display pyfcgi and Python version and exit
 * @par -c --config=FILE
 * load a configuration file
 * @par -e --pymodule=MODULE_NAME
 * python entrypoint module name
 * @par -E --pyapp=FUNC_NAME
 * python entrypoint function name
 * @par -A --alt-io
 * use stdout to communicate with web server instead of entrypoint return as
 * specified in PEP333
 * @par -P --pid-file=PATH
 * Create a pidfile with master process PID
 *
 * @subsection man_opt_pool Worker pool OPTIONS
 * @par -w --min-worker
 * minimum worker in the pool
 * @par -W --max-worker
 * maximum worker in the pool
 * @par -f --fast-spawn
 * If not given there is at least 1s between two child creation. When given
 * childs may be spawned in small burst.
 * @par -t --timeout=SECONDS
 * Request timeout. If the timeout expires the worker process is restarted.
 *
 * @subsection man_opt_log Logging & monitoring OPTIONS
 * @par -L --log=LOGGERSPEC
 * Add a logfile using syntax "LOGFILE[;FILT][;FMT]" See @ref man_logger_specs
 * @par -S --syslog
 * Use syslog for logging
 * @par -v --verbose
 * Send all loglines on STDERR
 * @par -s --socket-server=SOCKURL
 * Indicate a socket like "tcp://localhost:8765" to listen on, replying status
 * and statistics. See @ref man_url_spec
 *
 * @section man_logger_specs Logfile format
 * A logger specification use the following format : LOGILE[;FILT][;FMT] with
 * - LOGFILE the logfile name
 * - FILT a number (in decimal or hex 0xHH) indicating wich facility/level to
 * log
 * - FMT the logline format, following the markup format described bellow
 * @subsection man_logger_fmt Logline format
 * Logline's format is indicated using a simple format using fields between
 * '{' and '}'. Supported fields are : 
 * - {datetime} {datetime:SIZE} {datetime:SIZE:FMT} defines a format and a
 * constant length for a datetime field. Default : {datetime:25:%F %T%z}
 * - {level} the loglevel
 * - {facility} the log facility
 * - {pid} the process PID
 * - {ident} the process ident (friendly name)
 * - {msg} the log message (can appear only once)
 *
 * Specials chars '{' & '}' can be escpaed using "{{" and "}}", and all field
 * names can be abreviated to one character.
 * 
 * @section man_url_spec Statistics socket URL format
 * Statistics server listen on a socket specified by an URL like : 
 *
 * PROT://HOST:[PORT] with PROT one of :
 * - tcp with HOST a valid INET address
 * - udp with HOST a valid INET address
 * - unix with HOST a valid PATH
 *
 * @section EXAMPLES 
 * To run foo_pep333.entrypoint() PEP333 application :
 * 
 * spawn-fcgi -d . -n -p 9000 -a 127.0.0.1 -- src/pyfcgi -S
 * -e foo_pep333 -E entrypoint
 *
 * Logfile example :
 *
 * spawn-fcgi -d . -n -p 9000 -a 127.0.0.1 -- src/pyfcgi -S
 * -e foo_pep333 -E entrypoint
 * -L '/tmp/foo.log;0xff;{datetime} {msg} {ident}'
 *
 * @section AUTHOR
 * Written by Yann Weber &lt;yann.weber@member.fsf.org>
 *
 * @section COPYRIGHT
 * Copyright @ 2019 Yann Weber License GPLv3+: GNU GPL version 3 or later
 * &lt;http://gnu.org/licences/gpl.html>.
 *
 * This is free software: you are free to change and redistribute it.
 * There is NO WARRANTY, to extent permitted by law.
 *
 * @section man_see_also See Also
 * @ref spawn-fcgi(1)
 *
 * PyFCGI git repositorie &lt;https://git.yannweb.net/yannweb/pyfcgi>
 *
 */