scheduler.c
No OneTemporary
Actions

Size

15 KB

Subscribers

None

scheduler.c
View Options

	/* **************************************************************
	Copyright (C) 2010 Hewlett-Packard Development Company, L.P.

	This program is free software; you can redistribute it and/or
	modify it under the terms of the GNU General Public License
	version 2 as published by the Free Software Foundation.

	This program 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 General Public License for more details.

	You should have received a copy of the GNU General Public License along
	with this program; if not, write to the Free Software Foundation, Inc.,
	51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
	************************************************************** */

	/* local includes */
	#include <libfossrepo.h>
	#include <agent.h>
	#include <database.h>
	#include <event.h>
	#include <host.h>
	#include <interface.h>
	#include <logging.h>
	#include <scheduler.h>

	#include <fossconfig.h>

	/* std library includes */
	#include <stdio.h>
	#include <stdlib.h>
	#include <string.h>

	/* unix system includes */
	#include <dirent.h>
	#include <fcntl.h>
	#include <signal.h>
	#include <sys/mman.h>
	#include <sys/types.h>
	#include <sys/wait.h>
	#include <unistd.h>
	#include <pwd.h>
	#include <grp.h>

	/* glib includes */
	#include <glib.h>
	#include <gio/gio.h>

	#define AGENT_CONF "mods-enabled"
	#ifndef PROCESS_NAME
	#define PROCESS_NAME "fo_scheduler"
	#endif

	#define TEST_ERROR(...) \
	if(error) \
	{ \
	lprintf("ERROR %s.%d: %s\n", \
	__FILE__, __LINE__, error->message); \
	lprintf("ERROR %s.%d: ", __FILE__, __LINE__); \
	lprintf(__VA_ARGS__); \
	lprintf("\n"); \
	error = NULL; \
	continue; \
	} \

	/* global flags */
	int verbose = 0;
	int closing = 0;
	int startup = 0;
	int pause_f = 1;
	int s_pid;
	int s_daemon;
	int s_port;
	char* sysconfigdir;
	fo_conf* sysconfig;
	char* logdir;

	/* ************************************************************************** */
	/* ** signals and events ************************************************ */
	/* ************************************************************************** */

	/**
	* handle signals from the child process. This will only be called on a SIGCHLD
	* and will handle the effects of the death of the child process.
	*
	* @param signo
	* @param INFO
	* @param context
	*/
	void chld_sig(int signo)
	{
	pid_t n; // the next pid that has died
	pid_t* pass;
	int status; // status returned by waitpit()

	/* get all of the dead children's pids */
	while((n = waitpid(-1, &status, WNOHANG)) > 0)
	{
	if(TVERB_SCHED)
	clprintf("SIGNALS: received sigchld for pid %d\n", n);
	pass = g_new0(pid_t, 1);
	*pass = n;
	event_signal(agent_death_event, pass);
	}
	}

	/**
	* Handles any signals sent to the scheduler that are not SIGCHLD.
	*
	* Currently Handles:
	* SIGALRM: scheduler will run agent updates and database updates
	* SIGTERM: scheduler will gracefully shut down
	* SIGQUIT: scheduler will gracefully shut down
	* SIGINT: scheduler will gracefully shut down
	* SIGHIP: scheduler will reload configuration data
	*
	* @param signo the number of the signal that was sent
	*/
	void prnt_sig(int signo)
	{
	switch(signo)
	{
	case SIGALRM:
	V_SCHED("SIGNALS: Scheduler received alarm signal, checking job states\n");
	event_signal(agent_update_event, NULL);
	event_signal(database_update_event, NULL);
	alarm(CHECK_TIME);
	break;
	case SIGTERM:
	V_SCHED("SIGNALS: Scheduler received terminate signal, shutting down scheduler\n");
	event_signal(scheduler_close_event, NULL);
	break;
	case SIGQUIT:
	V_SCHED("SIGNALS: Scheduler received quit signal, shutting down scheduler\n");
	event_signal(scheduler_close_event, NULL);
	break;
	case SIGINT:
	V_SCHED("SIGNALS: Scheduler received interrupt signal, shutting down scheduler\n");
	event_signal(scheduler_close_event, NULL);
	break;
	case SIGHUP:
	V_SCHED("SIGNALS: Scheduler received SGIHUP, reloading configuration data\n");
	load_config(NULL);
	break;
	case SIGSEGV:
	FATAL("scheduler received a segmentation fault (i.e. the scheduler crashed");
	break;
	}
	}

	/* ************************************************************************** */
	/* ** The actual scheduler ********************************************** */
	/* ************************************************************************** */

	/**
	* The heart of the scheduler, the actual scheduling algorithm. This will be
	* passed to the event loop as a call back and will be called every time an event
	* is executed. Therefore the code should be light weight since it will be run
	* very frequently.
	*
	* TODO:
	* currently this will only grab a job and create a single agent to execute
	* the job.
	*
	* TODO: allow for runonpfile jobs to have multiple agents based on size
	* TODO: allow for job preemption. The scheduler can pause jobs, allow it
	* TODO: allow for specific hosts to be chossen.
	*/
	void update_scheduler()
	{
	/* queue used to hold jobs if an exclusive job enters the system */
	static job j = NULL;
	static int lockout = 0;

	/* locals */
	host machine = NULL;
	int n_agents = num_agents();
	int n_jobs = active_jobs();

	if(startup && n_agents == 0)
	{
	event_signal(database_update_event, NULL);
	startup = 0;
	}

	if(closing && n_agents == 0 && n_jobs == 0)
	{
	event_loop_terminate();
	return;
	}

	if(lockout && n_agents == 0 && n_jobs == 0)
	lockout = 0;

	if(j == NULL && !lockout)
	{
	while(peek_job() != NULL)
	{
	if((machine = get_host(1)) == NULL)
	break;

	j = next_job();
	if(is_special(job_type(j), SAG_EXCLUSIVE))
	break;

	V_SCHED("Starting JOB[%d].%s\n", job_id(j), job_type(j));
	agent_init(machine, j);
	j = NULL;
	}
	}

	if(j != NULL && n_agents == 0 && n_jobs == 0)
	{
	agent_init(get_host(1), j);
	lockout = 1;
	j = NULL;
	}

	if(pause_f)
	{
	startup = 1;
	pause_f = 0;
	}
	}

	/* ************************************************************************** */
	/* ** main utility functions ******************************************** */
	/* ************************************************************************** */

	#define GU_HEADER "DIRECTORIES"
	#define GU_GROUP "PROJECTGROUP"
	#define GU_USER "PROJECTUSER"

	/**
	* Correctly set the project user and group. The fossology scheduler must run as
	* the user specified by PROJECT_USER and PROJECT_GROUP since the agents must be
	* able to connect to the database. This ensures that that happens correctly.
	*/
	void set_usr_grp()
	{
	/* locals */
	struct group* grp;
	struct passwd* pwd;

	char* group =
	fo_config_has_key(sysconfig, GU_HEADER, GU_GROUP) ?
	fo_config_get (sysconfig, GU_HEADER, GU_GROUP, NULL) : PROJECT_GROUP;
	char* user =
	fo_config_has_key(sysconfig, GU_HEADER, GU_USER) ?
	fo_config_get (sysconfig, GU_HEADER, GU_USER, NULL) : PROJECT_USER;

	/* make sure group exists */
	grp = getgrnam(group);
	if(!grp)
	{
	// TODO error message
	}

	/* set the project group */
	setgroups(1, &(grp->gr_gid));
	if((setgid(grp->gr_gid) != 0) \|\| (setegid(grp->gr_gid) != 0))
	{
	fprintf(stderr, "FATAL %s.%d: %s must be run as root or %s\n",
	__FILE__, __LINE__, PROCESS_NAME, user);
	fprintf(stderr, "FATAL Set group '%s' aborting due to error: %s\n",
	group, strerror(errno));
	exit(-1);
	}

	/* run as project user */
	pwd = getpwnam(user);
	if(!pwd)
	{
	fprintf(stderr, "FATAL %s.%d: user '%s' not found\n",
	__FILE__, __LINE__, user);
	exit(-1);
	}

	/* run as correct user, not as root or any other user */
	if((setuid(pwd->pw_uid) != 0) \|\| (seteuid(pwd->pw_uid) != 0))
	{
	fprintf(stderr, "FATAL %s.%d: %s must run this as %s\n",
	__FILE__, __LINE__, PROCESS_NAME, user);
	fprintf(stderr, "FATAL SETUID aborting due to error: %s\n",
	strerror(errno));
	exit(-1);
	}
	}

	/**
	* kills all other running schedulers. This is accomplished by reading from the
	* /proc/ file system provided by the linux kernel.
	*/
	void kill_scheduler()
	{
	char f_name[FILENAME_MAX];
	struct dirent* ep;
	DIR* dp;
	FILE* file;

	if((dp = opendir("/proc/")) == NULL)
	{
	fprintf(stderr, "ERROR %s.%d: Could not open /proc/ file system\n",
	__FILE__, __LINE__);
	exit(-1);
	}

	while((ep = readdir(dp)) != NULL)
	{
	if(string_is_num(ep->d_name))
	{
	snprintf(f_name, sizeof(f_name), "/proc/%s/cmdline", ep->d_name);
	if((file = fopen(f_name, "rt")))
	{
	V_SCHED("KILL: found \"%s\" in cmdline file\n", f_name);
	if(fgets(f_name, sizeof(f_name), file) != NULL &&
	strstr(f_name, "fo_scheduler") && s_pid != atoi(ep->d_name))
	{
	lprintf("KILL: sending kill signal to pid %s\n", ep->d_name);
	kill(atoi(ep->d_name), SIGKILL);
	}
	}
	}
	}
	}

	/**
	* TODO
	*/
	void load_agent_config()
	{
	DIR* dp; // directory pointer used to load meta agents;
	struct dirent* ep; // information about directory
	char addbuf[512]; // standard string buffer
	int max = -1; // the number of agents to a host or number of one type running
	int special = 0; // anything that is special about the agent (EXCLUSIVE)
	int i;
	char* name;
	char* cmd;
	char* tmp;
	GError* error = NULL;
	fo_conf* config;

	/* clear previous configurations */
	agent_list_clean();

	snprintf(addbuf, sizeof(addbuf), "%s/%s/", sysconfigdir, AGENT_CONF);
	if((dp = opendir(addbuf)) == NULL)
	{
	FATAL("Could not open agent config directory: %s", addbuf);
	return;
	}

	/* load the configuration for the agents */
	while((ep = readdir(dp)) != NULL)
	{
	if(ep->d_name[0] != '.')
	{
	snprintf(addbuf, sizeof(addbuf), "%s/%s/%s/%s.conf",
	sysconfigdir, AGENT_CONF, ep->d_name, ep->d_name);

	config = fo_config_load(addbuf, &error);
	if(error && error->code == fo_missing_file)
	{
	V_SCHED("CONFIG: Could not find %s\n", addbuf);
	g_error_free(error);
	error = NULL;
	continue;
	}
	TEST_ERROR("no additional info");
	V_SCHED("CONFIG: loading config file %s\n", addbuf);

	if(!fo_config_has_group(config, "default"))
	{
	lprintf("ERROR: %s must have a \"default\" group\n", addbuf);
	lprintf("ERROR: cause by %s.%d\n", __FILE__, __LINE__);
	error = NULL;
	continue;
	}

	special = 0;
	name = ep->d_name;
	max = fo_config_list_length(config, "default", "special", &error);
	TEST_ERROR("the special key should be of type list");
	for(i = 0; i < max; i++)
	{
	cmd = fo_config_get_list(config, "default", "special", i, &error);
	TEST_ERROR("failed to load element %d of special list", i)
	if(strncmp(cmd, "EXCLUSIVE", 9) == 0)
	special \|= SAG_EXCLUSIVE;
	else if(strncmp(cmd, "NOEMAIL", 7) == 0)
	special \|= SAG_NOEMAIL;
	else if(strlen(cmd) != 0)
	WARNING("Invalid special type for agent %s: %s", name, cmd);
	}

	cmd = fo_config_get(config, "default", "command", &error);
	TEST_ERROR("the default group must have a command key");
	tmp = fo_config_get(config, "default", "max", &error);
	TEST_ERROR("the default group must have a max key");

	if(!add_meta_agent(name, cmd, atoi(tmp), special))
	{
	V_SCHED("CONFIG: could not create meta agent using %s\n", ep->d_name);
	}
	else if(TVERB_SCHED)
	{
	lprintf("CONFIG: added new agent\n");
	lprintf(" name = %s\n", name);
	lprintf(" command = %s\n", cmd);
	lprintf(" max = %d\n", max);
	lprintf(" special = %d\n", special);
	lprintf("\"\n");
	}

	fo_config_free(config);
	}
	}
	closedir(dp);
	for_each_host(test_agents);
	}

	/**
	* TODO
	*/
	void load_foss_config()
	{
	char* tmp; // pointer into a string
	char** keys;
	int max = -1; // the number of agents to a host or number of one type running
	int special = 0; // anything that is special about the agent (EXCLUSIVE)
	char addbuf[512]; // standard string buffer
	char dirbuf[512]; // standard string buffer
	GError* error = NULL;
	int i;

	/* clear all previous configurations */
	host_list_clean();

	/* parse the config file */
	snprintf(addbuf, sizeof(addbuf), "%s/fossology.conf", sysconfigdir);
	sysconfig = fo_config_load(addbuf, &error);
	if(error)
	FATAL("%s", error->message);

	/* load the port setting */
	if(s_port < 0)
	s_port = atoi(fo_config_get(sysconfig, "FOSSOLOGY", "port", &error));
	set_port(s_port);

	/* load the host settings */
	keys = fo_config_key_set(sysconfig, "HOSTS", &special);
	for(i = 0; i < special; i++)
	{
	tmp = fo_config_get(sysconfig, "HOSTS", keys[i], &error);
	if(error)
	{
	lprintf(error->message);
	error = NULL;
	continue;
	}

	sscanf(tmp, "%s %s %d", addbuf, dirbuf, &max);
	if(strcmp(addbuf, "localhost") == 0) strcpy(dirbuf, sysconfigdir);

	host_init(keys[i], addbuf, dirbuf, max);
	if(TVERB_SCHED)
	{
	lprintf("CONFIG: added new host\n");
	lprintf(" name = %s\n", keys[i]);
	lprintf(" address = %s\n", addbuf);
	lprintf(" directory = %s\n", dirbuf);
	lprintf(" max = %d\n", max);
	}
	}
	}

	/**
	* TODO
	*
	* @param unused
	*/
	void load_config(void* unused)
	{
	load_foss_config();
	load_agent_config();
	}

	/**
	* TODO
	*
	* @param unused
	*/
	void scheduler_close_event(void* unused)
	{
	closing = 1;
	kill_agents();
	}

	/**
	* TODO
	*
	* @return
	*/
	int close_scheduler()
	{
	job_list_clean();
	host_list_clean();
	agent_list_clean();
	interface_destroy();
	database_destroy();
	event_loop_destroy();
	fo_RepClose();
	return 0;
	}

	/**
	* Checks if a string is entirely composed of numeric characters
	*
	* @param str the string to test
	* @return TRUE if the string is entirely numeric, FALSE otherwise
	*/
	gint string_is_num(gchar* str)
	{
	int len = strlen(str);
	int i;

	for(i = 0; i < len; i++)
	if(!isdigit(str[i]))
	return FALSE;
	return TRUE;
	}

	/**
	* Utility function that enables the use of the strcmp function with a GTree.
	*
	* @param a The first string
	* @param b The second string
	* @param user_data unused in this function
	* @return integral value idicating the relatioship between the two strings
	*/
	gint string_compare(gconstpointer a, gconstpointer b, gpointer user_data)
	{
	return strcmp((char)a, (char)b);
	}

	/**
	* Utility function that enable the agents to be stored in a GTree using
	* the PID of the associated process.
	*
	* @param a The pid of the first process
	* @param b The pid of the second process
	* @param user_data unused in this function
	* @return integral value idicating the relationship between the two pids
	*/
	gint int_compare(gconstpointer a, gconstpointer b, gpointer user_data)
	{
	return (int)a - (int)b;
	}

File Metadata

Mime Type: text/x-c
Expires: Wed, Jun 4, 7:15 PM (1 w, 12 h ago)
Storage Engine: blob
Storage Format: Raw Data
Storage Handle: 3399014

scheduler.cNo OneTemporaryActions

scheduler.cView Options

File Metadata

Event Timeline

scheduler.c
No OneTemporary
Actions

scheduler.c
View Options