int main(int argc, char **argv)
{
simple_worker *wp;
int i;
#ifdef HAVE_SIGACTION
struct sigaction sig_action;
sig_action.sa_sigaction = NULL;
sigfillset(&sig_action.sa_mask);
sig_action.sa_flags=SA_NOCLDSTOP;
sig_action.sa_handler = child_exited;
sigaction(SIGCHLD, &sig_action, NULL);
sig_action.sa_flags = SA_NODEFER|SA_RESTART;
sig_action.sa_handler = sighandler;
sigfillset(&sig_action.sa_mask);
sigaction(SIGINT, &sig_action, NULL);
sigaction(SIGPIPE, &sig_action, NULL);
#else /* HAVE_SIGACTION */
signal(SIGINT, sighandler);
signal(SIGPIPE, sighandler);
signal(SIGCHLD, child_exited);
#endif /* HAVE_SIGACTION */
iobs = iobroker_create();
if (!iobs)
die("Failed to create io broker set");
for (i = 0; i < NWPS; i++) {
wp = spawn_worker(print_some_crap, "lalala");
if (!wp) {
die("Failed to spawn worker(s)\n");
}
wps[i] = wp;
printf("Registering worker sd %d with io broker\n", wp->sd);
iobroker_register(iobs, wp->sd, wp, print_input);
}
iobroker_register(iobs, fileno(stdin), NULL, send_command);
/* get to work */
while (!sigreceived && iobroker_get_num_fds(iobs)) {
iobroker_poll(iobs, -1);
}
for (i = 0; i < NWPS; i++) {
kill(wps[i]->pid, SIGKILL);
}
return 0;
}
/* reap 'jobs' jobs or 'secs' seconds, whichever comes first */
void wproc_reap(int jobs, int msecs)
{
time_t start, now;
start = time(NULL);
/* one input equals one job (or close enough to it anyway) */
do {
int inputs;
now = time(NULL);
inputs = iobroker_poll(nagios_iobs, (now - start) * 1000);
jobs -= inputs;
} while (jobs > 0 && start + (msecs * 1000) <= now);
}
int main(int argc, char **argv)
{
int listen_fd, flags, sockopt = 1;
struct sockaddr_in sain;
int error;
const char *err_msg;
t_set_colors(0);
t_start("iobroker ipc test");
error = iobroker_get_max_fds(NULL);
ok_int(error, IOBROKER_ENOSET, "test errors when passing null");
err_msg = iobroker_strerror(error);
test(err_msg && !strcmp(err_msg, iobroker_errors[(~error) + 1].string), "iobroker_strerror() returns the right string");
iobs = iobroker_create();
error = iobroker_get_max_fds(iobs);
test(iobs && error >= 0, "max fd's for real iobroker set must be > 0");
listen_fd = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
flags = fcntl(listen_fd, F_GETFD);
flags |= FD_CLOEXEC;
fcntl(listen_fd, F_SETFD, flags);
(void)setsockopt(listen_fd, SOL_SOCKET, SO_REUSEADDR, &sockopt, sizeof(sockopt));
memset(&sain, 0, sizeof(sain));
sain.sin_port = ntohs(9123);
sain.sin_family = AF_INET;
bind(listen_fd, (struct sockaddr *)&sain, sizeof(sain));
listen(listen_fd, 128);
iobroker_register(iobs, listen_fd, iobs, listen_handler);
if (argc == 1)
conn_spam(&sain);
for (;;) {
iobroker_poll(iobs, -1);
if (iobroker_get_num_fds(iobs) <= 1) {
break;
}
}
iobroker_close(iobs, listen_fd);
iobroker_destroy(iobs, 0);
t_end();
return 0;
}
/* Try to reap 'jobs' jobs for 'msecs' milliseconds. Return early on error. */
void wproc_reap(int jobs, int msecs)
{
struct timeval start;
gettimeofday(&start, NULL);
while (jobs > 0 && msecs > 0) {
int inputs = iobroker_poll(nagios_iobs, msecs);
if (inputs < 0) return;
jobs -= inputs; /* One input is roughly equivalent to one job. */
struct timeval now;
gettimeofday(&now, NULL);
msecs -= tv_delta_msec(&start, &now);
start = now;
}
}
static int conn_spam(struct sockaddr_in *sain)
{
int i;
signal(SIGALRM, (void (*)(int))sighandler);
signal(SIGINT, (void (*)(int))sighandler);
signal(SIGPIPE, SIG_IGN);
alarm(20);
for (i = 0; i < NUM_PROCS; i++) {
int fd, sockopt = 1;
fd = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
(void)setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &sockopt, sizeof(sockopt));
proc_counter[i] = 0;
iobroker_register(iobs, fd, (void *)&proc_counter[i], connected_handler);
if (connect(fd, (struct sockaddr *)sain, sizeof(*sain))) {
perror("connect");
}
iobroker_poll(iobs, -1);
}
return 0;
}
/* this is the main event handler loop */
int event_execution_loop(void)
{
timed_event *temp_event, *last_event = NULL;
time_t last_time = 0L;
time_t current_time = 0L;
time_t last_status_update = 0L;
int poll_time_ms;
log_debug_info(DEBUGL_FUNCTIONS, 0, "event_execution_loop() start\n");
time(&last_time);
while (1) {
struct timeval now;
const struct timeval *event_runtime;
int inputs;
/* super-priority (hardcoded) events come first */
/* see if we should exit or restart (a signal was encountered) */
if (sigshutdown == TRUE || sigrestart == TRUE)
break;
/* get the current time */
time(¤t_time);
if (sigrotate == TRUE) {
rotate_log_file(current_time);
update_program_status(FALSE);
}
/* hey, wait a second... we traveled back in time! */
if (current_time < last_time)
compensate_for_system_time_change((unsigned long)last_time, (unsigned long)current_time);
/* else if the time advanced over the specified threshold, try and compensate... */
else if ((current_time - last_time) >= time_change_threshold)
compensate_for_system_time_change((unsigned long)last_time, (unsigned long)current_time);
/* get next scheduled event */
current_event = temp_event = (timed_event *)squeue_peek(nagios_squeue);
/* if we don't have any events to handle, exit */
if (!temp_event) {
log_debug_info(DEBUGL_EVENTS, 0, "There aren't any events that need to be handled! Exiting...\n");
break;
}
/* keep track of the last time */
last_time = current_time;
/* update status information occassionally - NagVis watches the NDOUtils DB to see if Nagios is alive */
if ((unsigned long)(current_time - last_status_update) > 5) {
last_status_update = current_time;
update_program_status(FALSE);
}
event_runtime = squeue_event_runtime(temp_event->sq_event);
if (temp_event != last_event) {
log_debug_info(DEBUGL_EVENTS, 1, "** Event Check Loop\n");
log_debug_info(DEBUGL_EVENTS, 1, "Next Event Time: %s", ctime(&temp_event->run_time));
log_debug_info(DEBUGL_EVENTS, 1, "Current/Max Service Checks: %d/%d (%.3lf%% saturation)\n",
currently_running_service_checks, max_parallel_service_checks,
((float)currently_running_service_checks / (float)max_parallel_service_checks) * 100);
}
last_event = temp_event;
gettimeofday(&now, NULL);
poll_time_ms = tv_delta_msec(&now, event_runtime);
if (poll_time_ms < 0)
poll_time_ms = 0;
else if (poll_time_ms >= 1500)
poll_time_ms = 1500;
log_debug_info(DEBUGL_SCHEDULING, 2, "## Polling %dms; sockets=%d; events=%u; iobs=%p\n",
poll_time_ms, iobroker_get_num_fds(nagios_iobs),
squeue_size(nagios_squeue), nagios_iobs);
inputs = iobroker_poll(nagios_iobs, poll_time_ms);
if (inputs < 0 && errno != EINTR) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Polling for input on %p failed: %s", nagios_iobs, iobroker_strerror(inputs));
break;
}
log_debug_info(DEBUGL_IPC, 2, "## %d descriptors had input\n", inputs);
/*
* if the event we peaked was removed from the queue from
* one of the I/O operations, we must take care not to
* try to run at, as we're (almost) sure to access free'd
* or invalid memory if we do.
*/
if (!current_event) {
log_debug_info(DEBUGL_EVENTS, 0, "Event was cancelled by iobroker input\n");
continue;
}
gettimeofday(&now, NULL);
if (tv_delta_msec(&now, event_runtime) >= 0)
continue;
/* move on if we shouldn't run this event */
if (should_run_event(temp_event) == FALSE)
continue;
/* handle the event */
handle_timed_event(temp_event);
/*
* we must remove the entry we've peeked, or
* we'll keep getting the same one over and over.
* This also maintains sync with broker modules.
*/
remove_event(nagios_squeue, temp_event);
/* reschedule the event if necessary */
if (temp_event->recurring == TRUE)
reschedule_event(nagios_squeue, temp_event);
/* else free memory associated with the event */
else
my_free(temp_event);
}
log_debug_info(DEBUGL_FUNCTIONS, 0, "event_execution_loop() end\n");
return OK;
}
static void enter_worker(int sd)
{
/* created with socketpair(), usually */
master_sd = sd;
parent_pid = getppid();
(void)chdir("/tmp");
(void)chdir("nagios-workers");
if (setpgid(0, 0)) {
/* XXX: handle error somehow, or maybe just ignore it */
}
/* we need to catch child signals the default way */
signal(SIGCHLD, SIG_DFL);
fcntl(fileno(stdout), F_SETFD, FD_CLOEXEC);
fcntl(fileno(stderr), F_SETFD, FD_CLOEXEC);
fcntl(master_sd, F_SETFD, FD_CLOEXEC);
iobs = iobroker_create();
if (!iobs) {
/* XXX: handle this a bit better */
worker_die("Worker failed to create io broker socket set");
}
/*
* Create a modest scheduling queue that will be
* more than enough for our needs
*/
sq = squeue_create(1024);
set_socket_options(master_sd, 256 * 1024);
iobroker_register(iobs, master_sd, NULL, receive_command);
while (iobroker_get_num_fds(iobs) > 0) {
int poll_time = -1;
/* check for timed out jobs */
for (;;) {
child_process *cp;
struct timeval now, tmo;
/* stop when scheduling queue is empty */
cp = (child_process *)squeue_peek(sq);
if (!cp)
break;
tmo.tv_usec = cp->start.tv_usec;
tmo.tv_sec = cp->start.tv_sec + cp->timeout;
gettimeofday(&now, NULL);
poll_time = tv_delta_msec(&now, &tmo);
/*
* A little extra takes care of rounding errors and
* ensures we never kill a job before it times out.
* 5 milliseconds is enough to take care of that.
*/
poll_time += 5;
if (poll_time > 0)
break;
/* this job timed out, so kill it */
wlog("job with pid %d timed out. Killing it", cp->pid);
kill_job(cp, ETIME);
}
iobroker_poll(iobs, poll_time);
/*
* if our parent goes away we can't really do anything
* sensible at all, so let's just break out and exit
*/
if (kill(parent_pid, 0) < 0 && errno == ESRCH) {
break;
}
}
/* we exit when the master shuts us down */
exit(EXIT_SUCCESS);
}
int main(int argc, char **argv)
{
int result;
int error = FALSE;
int display_license = FALSE;
int display_help = FALSE;
int c = 0;
struct tm *tm, tm_s;
time_t now;
char datestring[256];
nagios_macros *mac;
const char *worker_socket = NULL;
int i;
#ifdef HAVE_GETOPT_H
int option_index = 0;
static struct option long_options[] = {
{"help", no_argument, 0, 'h'},
{"version", no_argument, 0, 'V'},
{"license", no_argument, 0, 'V'},
{"verify-config", no_argument, 0, 'v'},
{"daemon", no_argument, 0, 'd'},
{"test-scheduling", no_argument, 0, 's'},
{"precache-objects", no_argument, 0, 'p'},
{"use-precached-objects", no_argument, 0, 'u'},
{"enable-timing-point", no_argument, 0, 'T'},
{"worker", required_argument, 0, 'W'},
{0, 0, 0, 0}
};
#define getopt(argc, argv, o) getopt_long(argc, argv, o, long_options, &option_index)
#endif
memset(&loadctl, 0, sizeof(loadctl));
mac = get_global_macros();
/* make sure we have the correct number of command line arguments */
if (argc < 2)
error = TRUE;
/* get all command line arguments */
while (1) {
c = getopt(argc, argv, "+hVvdspuxTW");
if (c == -1 || c == EOF)
break;
switch (c) {
case '?': /* usage */
case 'h':
display_help = TRUE;
break;
case 'V': /* version */
display_license = TRUE;
break;
case 'v': /* verify */
verify_config++;
break;
case 's': /* scheduling check */
test_scheduling = TRUE;
break;
case 'd': /* daemon mode */
daemon_mode = TRUE;
break;
case 'p': /* precache object config */
precache_objects = TRUE;
break;
case 'u': /* use precached object config */
use_precached_objects = TRUE;
break;
case 'T':
enable_timing_point = TRUE;
break;
case 'W':
worker_socket = optarg;
break;
case 'x':
printf("Warning: -x is deprecated and will be removed\n");
break;
default:
break;
}
}
/* if we're a worker we can skip everything below */
if (worker_socket) {
exit(nagios_core_worker(worker_socket));
}
if (daemon_mode == FALSE) {
printf("\nNaemon Core %s\n", PROGRAM_VERSION);
printf("Copyright (c) 2013-present Naemon Core Development Team and Community Contributors\n");
printf("Copyright (c) 2009-2013 Nagios Core Development Team and Community Contributors\n");
printf("Copyright (c) 1999-2009 Ethan Galstad\n");
printf("Last Modified: %s\n", PROGRAM_MODIFICATION_DATE);
printf("License: GPL\n\n");
printf("Website: http://www.naemon.org\n");
}
/* just display the license */
if (display_license == TRUE) {
printf("This program is free software; you can redistribute it and/or modify\n");
printf("it under the terms of the GNU General Public License version 2 as\n");
printf("published by the Free Software Foundation.\n\n");
printf("This program is distributed in the hope that it will be useful,\n");
printf("but WITHOUT ANY WARRANTY; without even the implied warranty of\n");
printf("MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n");
printf("GNU General Public License for more details.\n\n");
printf("You should have received a copy of the GNU General Public License\n");
printf("along with this program; if not, write to the Free Software\n");
printf("Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.\n\n");
exit(OK);
}
/* make sure we got the main config file on the command line... */
if (optind >= argc)
error = TRUE;
/* if there are no command line options (or if we encountered an error), print usage */
if (error == TRUE || display_help == TRUE) {
printf("Usage: %s [options] <main_config_file>\n", argv[0]);
printf("\n");
printf("Options:\n");
printf("\n");
printf(" -v, --verify-config Verify all configuration data (-v -v for more info)\n");
printf(" -s, --test-scheduling Shows projected/recommended check scheduling and other\n");
printf(" diagnostic info based on the current configuration files.\n");
printf(" -T, --enable-timing-point Enable timed commentary on initialization\n");
printf(" -x, --dont-verify-paths Deprecated (Don't check for circular object paths)\n");
printf(" -p, --precache-objects Precache object configuration\n");
printf(" -u, --use-precached-objects Use precached object config file\n");
printf(" -d, --daemon Starts Naemon in daemon mode, instead of as a foreground process\n");
printf(" -W, --worker /path/to/socket Act as a worker for an already running daemon\n");
printf("\n");
printf("Visit the Naemon website at http://www.naemon.org/ for bug fixes, new\n");
printf("releases, online documentation, FAQs and more...\n");
printf("\n");
exit(ERROR);
}
/*
* config file is last argument specified.
* Make sure it uses an absolute path
*/
config_file = nspath_absolute(argv[optind], NULL);
if (config_file == NULL) {
printf("Error allocating memory.\n");
exit(ERROR);
}
config_file_dir = nspath_absolute_dirname(config_file, NULL);
/*
* Set the signal handler for the SIGXFSZ signal here because
* we may encounter this signal before the other signal handlers
* are set.
*/
signal(SIGXFSZ, handle_sigxfsz);
/*
* let's go to town. We'll be noisy if we're verifying config
* or running scheduling tests.
*/
if (verify_config || test_scheduling || precache_objects) {
reset_variables();
/*
* if we don't beef up our resource limits as much as
* we can, it's quite possible we'll run headlong into
* EAGAIN due to too many processes when we try to
* drop privileges later.
*/
set_loadctl_defaults();
if (verify_config)
printf("Reading configuration data...\n");
/* read our config file */
result = read_main_config_file(config_file);
if (result != OK) {
printf(" Error processing main config file!\n\n");
exit(EXIT_FAILURE);
}
if (verify_config)
printf(" Read main config file okay...\n");
/* drop privileges */
if ((result = drop_privileges(naemon_user, naemon_group)) == ERROR) {
printf(" Failed to drop privileges. Aborting.");
exit(EXIT_FAILURE);
}
/*
* this must come after dropping privileges, so we make
* sure to test access permissions as the right user.
*/
if (test_configured_paths() == ERROR) {
printf(" One or more path problems detected. Aborting.\n");
exit(EXIT_FAILURE);
}
/* read object config files */
result = read_all_object_data(config_file);
if (result != OK) {
printf(" Error processing object config files!\n\n");
/* if the config filename looks fishy, warn the user */
if (!strstr(config_file, "naemon.cfg")) {
printf("\n***> The name of the main configuration file looks suspicious...\n");
printf("\n");
printf(" Make sure you are specifying the name of the MAIN configuration file on\n");
printf(" the command line and not the name of another configuration file. The\n");
printf(" main configuration file is typically '%s'\n", DEFAULT_CONFIG_FILE);
}
printf("\n***> One or more problems was encountered while processing the config files...\n");
printf("\n");
printf(" Check your configuration file(s) to ensure that they contain valid\n");
printf(" directives and data defintions. If you are upgrading from a previous\n");
printf(" version of Naemon, you should be aware that some variables/definitions\n");
printf(" may have been removed or modified in this version. Make sure to read\n");
printf(" the HTML documentation regarding the config files, as well as the\n");
printf(" 'Whats New' section to find out what has changed.\n\n");
exit(EXIT_FAILURE);
}
if (verify_config) {
printf(" Read object config files okay...\n\n");
printf("Running pre-flight check on configuration data...\n\n");
}
/* run the pre-flight check to make sure things look okay... */
result = pre_flight_check();
if (result != OK) {
printf("\n***> One or more problems was encountered while running the pre-flight check...\n");
printf("\n");
printf(" Check your configuration file(s) to ensure that they contain valid\n");
printf(" directives and data defintions. If you are upgrading from a previous\n");
printf(" version of Naemon, you should be aware that some variables/definitions\n");
printf(" may have been removed or modified in this version. Make sure to read\n");
printf(" the HTML documentation regarding the config files, as well as the\n");
printf(" 'Whats New' section to find out what has changed.\n\n");
exit(EXIT_FAILURE);
}
if (verify_config) {
printf("\nThings look okay - No serious problems were detected during the pre-flight check\n");
}
/* scheduling tests need a bit more than config verifications */
if (test_scheduling == TRUE) {
/* we'll need the event queue here so we can time insertions */
init_event_queue();
timing_point("Done initializing event queue\n");
/* read initial service and host state information */
initialize_retention_data(config_file);
read_initial_state_information();
timing_point("Retention data and initial state parsed\n");
/* initialize the event timing loop */
init_timing_loop();
timing_point("Timing loop initialized\n");
/* display scheduling information */
display_scheduling_info();
}
if (precache_objects) {
result = fcache_objects(object_precache_file);
timing_point("Done precaching objects\n");
if (result == OK) {
printf("Object precache file created:\n%s\n", object_precache_file);
} else {
printf("Failed to precache objects to '%s': %s\n", object_precache_file, strerror(errno));
}
}
/* clean up after ourselves */
cleanup();
/* exit */
timing_point("Exiting\n");
/* make valgrind shut up about still reachable memory */
neb_free_module_list();
free(config_file_dir);
free(config_file);
exit(result);
}
/* start to monitor things... */
/*
* if we're called with a relative path we must make
* it absolute so we can launch our workers.
* If not, we needn't bother, as we're using execvp()
*/
if (strchr(argv[0], '/'))
naemon_binary_path = nspath_absolute(argv[0], NULL);
else
naemon_binary_path = strdup(argv[0]);
if (!naemon_binary_path) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Unable to allocate memory for naemon_binary_path\n");
exit(EXIT_FAILURE);
}
if (!(nagios_iobs = iobroker_create())) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Failed to create IO broker set: %s\n",
strerror(errno));
exit(EXIT_FAILURE);
}
/* keep monitoring things until we get a shutdown command */
do {
/* reset internal book-keeping (in case we're restarting) */
wproc_num_workers_spawned = wproc_num_workers_online = 0;
caught_signal = sigshutdown = sigrestart = FALSE;
sig_id = 0;
/* reset program variables */
reset_variables();
timing_point("Variables reset\n");
/* get PID */
nagios_pid = (int)getpid();
/* read in the configuration files (main and resource config files) */
result = read_main_config_file(config_file);
if (result != OK) {
logit(NSLOG_CONFIG_ERROR, TRUE, "Error: Failed to process config file '%s'. Aborting\n", config_file);
exit(EXIT_FAILURE);
}
timing_point("Main config file read\n");
/* NOTE 11/06/07 EG moved to after we read config files, as user may have overridden timezone offset */
/* get program (re)start time and save as macro */
program_start = time(NULL);
my_free(mac->x[MACRO_PROCESSSTARTTIME]);
asprintf(&mac->x[MACRO_PROCESSSTARTTIME], "%lu", (unsigned long)program_start);
/* drop privileges */
if (drop_privileges(naemon_user, naemon_group) == ERROR) {
logit(NSLOG_PROCESS_INFO | NSLOG_RUNTIME_ERROR | NSLOG_CONFIG_ERROR, TRUE, "Failed to drop privileges. Aborting.");
cleanup();
exit(ERROR);
}
if (test_path_access(naemon_binary_path, X_OK)) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: failed to access() %s: %s\n", naemon_binary_path, strerror(errno));
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Spawning workers will be impossible. Aborting.\n");
exit(EXIT_FAILURE);
}
if (test_configured_paths() == ERROR) {
/* error has already been logged */
exit(EXIT_FAILURE);
}
/* enter daemon mode (unless we're restarting...) */
if (daemon_mode == TRUE && sigrestart == FALSE) {
result = daemon_init();
/* we had an error daemonizing, so bail... */
if (result == ERROR) {
logit(NSLOG_PROCESS_INFO | NSLOG_RUNTIME_ERROR, TRUE, "Bailing out due to failure to daemonize. (PID=%d)", (int)getpid());
cleanup();
exit(EXIT_FAILURE);
}
/* get new PID */
nagios_pid = (int)getpid();
}
/* this must be logged after we read config data, as user may have changed location of main log file */
logit(NSLOG_PROCESS_INFO, TRUE, "Naemon %s starting... (PID=%d)\n", PROGRAM_VERSION, (int)getpid());
/* log the local time - may be different than clock time due to timezone offset */
now = time(NULL);
tm = localtime_r(&now, &tm_s);
strftime(datestring, sizeof(datestring), "%a %b %d %H:%M:%S %Z %Y", tm);
logit(NSLOG_PROCESS_INFO, TRUE, "Local time is %s", datestring);
/* write log version/info */
write_log_file_info(NULL);
/* open debug log now that we're the right user */
open_debug_log();
#ifdef USE_EVENT_BROKER
/* initialize modules */
neb_init_modules();
neb_init_callback_list();
#endif
timing_point("NEB module API initialized\n");
/* handle signals (interrupts) before we do any socket I/O */
setup_sighandler();
/*
* Initialize query handler and event subscription service.
* This must be done before modules are initialized, so
* the modules can use our in-core stuff properly
*/
if (qh_init(qh_socket_path ? qh_socket_path : DEFAULT_QUERY_SOCKET) != OK) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Failed to initialize query handler. Aborting\n");
exit(EXIT_FAILURE);
}
timing_point("Query handler initialized\n");
nerd_init();
timing_point("NERD initialized\n");
/* initialize check workers */
if (init_workers(num_check_workers) < 0) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Failed to spawn workers. Aborting\n");
exit(EXIT_FAILURE);
}
timing_point("%u workers spawned\n", wproc_num_workers_spawned);
i = 0;
while (i < 50 && wproc_num_workers_online < wproc_num_workers_spawned) {
iobroker_poll(nagios_iobs, 50);
i++;
}
timing_point("%u workers connected\n", wproc_num_workers_online);
/* now that workers have arrived we can set the defaults */
set_loadctl_defaults();
/* read in all object config data */
if (result == OK)
result = read_all_object_data(config_file);
/*
* the queue has to be initialized befor loading the neb modules
* to give them the chance to register user events.
* (initializing event queue requires number of objects, so do
* this after parsing the objects)
*/
init_event_queue();
timing_point("Event queue initialized\n");
#ifdef USE_EVENT_BROKER
/* load modules */
if (neb_load_all_modules() != OK) {
logit(NSLOG_CONFIG_ERROR, ERROR, "Error: Module loading failed. Aborting.\n");
/* if we're dumping core, we must remove all dl-files */
if (daemon_dumps_core)
neb_unload_all_modules(NEBMODULE_FORCE_UNLOAD, NEBMODULE_NEB_SHUTDOWN);
exit(EXIT_FAILURE);
}
timing_point("Modules loaded\n");
/* send program data to broker */
broker_program_state(NEBTYPE_PROCESS_PRELAUNCH, NEBFLAG_NONE, NEBATTR_NONE, NULL);
timing_point("First callback made\n");
#endif
/* there was a problem reading the config files */
if (result != OK)
logit(NSLOG_PROCESS_INFO | NSLOG_RUNTIME_ERROR | NSLOG_CONFIG_ERROR, TRUE, "Bailing out due to one or more errors encountered in the configuration files. Run Naemon from the command line with the -v option to verify your config before restarting. (PID=%d)", (int)getpid());
else {
/* run the pre-flight check to make sure everything looks okay*/
if ((result = pre_flight_check()) != OK)
logit(NSLOG_PROCESS_INFO | NSLOG_RUNTIME_ERROR | NSLOG_VERIFICATION_ERROR, TRUE, "Bailing out due to errors encountered while running the pre-flight check. Run Naemon from the command line with the -v option to verify your config before restarting. (PID=%d)\n", (int)getpid());
}
/* an error occurred that prevented us from (re)starting */
if (result != OK) {
/* if we were restarting, we need to cleanup from the previous run */
if (sigrestart == TRUE) {
/* clean up the status data */
cleanup_status_data(TRUE);
}
#ifdef USE_EVENT_BROKER
/* send program data to broker */
broker_program_state(NEBTYPE_PROCESS_SHUTDOWN, NEBFLAG_PROCESS_INITIATED, NEBATTR_SHUTDOWN_ABNORMAL, NULL);
#endif
cleanup();
exit(ERROR);
}
timing_point("Object configuration parsed and understood\n");
/* write the objects.cache file */
fcache_objects(object_cache_file);
timing_point("Objects cached\n");
#ifdef USE_EVENT_BROKER
/* send program data to broker */
broker_program_state(NEBTYPE_PROCESS_START, NEBFLAG_NONE, NEBATTR_NONE, NULL);
#endif
/* initialize status data unless we're starting */
if (sigrestart == FALSE) {
initialize_status_data(config_file);
timing_point("Status data initialized\n");
}
/* initialize scheduled downtime data */
initialize_downtime_data();
timing_point("Downtime data initialized\n");
/* read initial service and host state information */
initialize_retention_data(config_file);
timing_point("Retention data initialized\n");
read_initial_state_information();
timing_point("Initial state information read\n");
/* initialize comment data */
initialize_comment_data();
timing_point("Comment data initialized\n");
/* initialize performance data */
initialize_performance_data(config_file);
timing_point("Performance data initialized\n");
/* initialize the event timing loop */
init_timing_loop();
timing_point("Event timing loop initialized\n");
/* initialize check statistics */
init_check_stats();
timing_point("check stats initialized\n");
/* update all status data (with retained information) */
update_all_status_data();
timing_point("Status data updated\n");
/* log initial host and service state */
log_host_states(INITIAL_STATES, NULL);
log_service_states(INITIAL_STATES, NULL);
timing_point("Initial states logged\n");
/* reset the restart flag */
sigrestart = FALSE;
/* fire up command file worker */
launch_command_file_worker();
timing_point("Command file worker launched\n");
#ifdef USE_EVENT_BROKER
/* send program data to broker */
broker_program_state(NEBTYPE_PROCESS_EVENTLOOPSTART, NEBFLAG_NONE, NEBATTR_NONE, NULL);
#endif
/* get event start time and save as macro */
event_start = time(NULL);
my_free(mac->x[MACRO_EVENTSTARTTIME]);
asprintf(&mac->x[MACRO_EVENTSTARTTIME], "%lu", (unsigned long)event_start);
timing_point("Entering event execution loop\n");
/***** start monitoring all services *****/
/* (doesn't return until a restart or shutdown signal is encountered) */
event_execution_loop();
/*
* immediately deinitialize the query handler so it
* can remove modules that have stashed data with it
*/
qh_deinit(qh_socket_path ? qh_socket_path : DEFAULT_QUERY_SOCKET);
/* 03/01/2007 EG Moved from sighandler() to prevent FUTEX locking problems under NPTL */
/* 03/21/2007 EG SIGSEGV signals are still logged in sighandler() so we don't loose them */
/* did we catch a signal? */
if (caught_signal == TRUE) {
if (sig_id == SIGHUP)
logit(NSLOG_PROCESS_INFO, TRUE, "Caught SIGHUP, restarting...\n");
}
#ifdef USE_EVENT_BROKER
/* send program data to broker */
broker_program_state(NEBTYPE_PROCESS_EVENTLOOPEND, NEBFLAG_NONE, NEBATTR_NONE, NULL);
if (sigshutdown == TRUE)
broker_program_state(NEBTYPE_PROCESS_SHUTDOWN, NEBFLAG_USER_INITIATED, NEBATTR_SHUTDOWN_NORMAL, NULL);
else if (sigrestart == TRUE)
broker_program_state(NEBTYPE_PROCESS_RESTART, NEBFLAG_USER_INITIATED, NEBATTR_RESTART_NORMAL, NULL);
#endif
disconnect_command_file_worker();
/* save service and host state information */
save_state_information(FALSE);
cleanup_retention_data();
/* clean up performance data */
cleanup_performance_data();
/* clean up the scheduled downtime data */
cleanup_downtime_data();
/* clean up the status data unless we're restarting */
if (sigrestart == FALSE) {
cleanup_status_data(TRUE);
}
free_worker_memory(WPROC_FORCE);
/* shutdown stuff... */
if (sigshutdown == TRUE) {
iobroker_destroy(nagios_iobs, IOBROKER_CLOSE_SOCKETS);
nagios_iobs = NULL;
/* log a shutdown message */
logit(NSLOG_PROCESS_INFO, TRUE, "Successfully shutdown... (PID=%d)\n", (int)getpid());
}
/* clean up after ourselves */
cleanup();
/* close debug log */
close_debug_log();
} while (sigrestart == TRUE && sigshutdown == FALSE);
if (daemon_mode == TRUE)
unlink(lock_file);
/* free misc memory */
my_free(lock_file);
my_free(config_file);
my_free(config_file_dir);
my_free(naemon_binary_path);
return OK;
}