nagios4/base/utils.c
2019-08-03 18:28:19 +02:00

3804 lines
108 KiB
C

/*****************************************************************************
*
* UTILS.C - Miscellaneous utility functions for Nagios
*
*
* License:
*
* 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., 675 Mass Ave, Cambridge, MA 02139, USA.
*
*****************************************************************************/
#include "../include/config.h"
#include "../include/common.h"
#include "../include/objects.h"
#include "../include/statusdata.h"
#include "../include/comments.h"
#include "../include/downtime.h"
#include "../include/macros.h"
#include "../include/nagios.h"
#include "../include/netutils.h"
#include "../include/perfdata.h"
#include "../include/broker.h"
#include "../include/nebmods.h"
#include "../include/nebmodules.h"
#include "../include/workers.h"
#include "../xdata/xodtemplate.h"
#include <sys/time.h>
#include <sys/resource.h>
/* global variables only used by the daemon */
char *nagios_binary_path = NULL;
char *config_file = NULL;
char *command_file;
char *temp_file;
char *temp_path;
char *check_result_path;
char *lock_file;
int num_check_workers;
char *qh_socket_path;
char *nagios_user;
char *nagios_group;
char *ocsp_command;
char *ochp_command;
command *ocsp_command_ptr = NULL;
command *ochp_command_ptr = NULL;
int ocsp_timeout;
int ochp_timeout;
char *illegal_object_chars;
int use_regexp_matches;
int use_true_regexp_matching;
int use_syslog;
char *log_file;
char *log_archive_path;
int log_notifications;
int log_service_retries;
int log_host_retries;
int log_event_handlers;
int log_initial_states;
int log_current_states;
int log_external_commands;
int log_passive_checks;
unsigned long logging_options = 0;
unsigned long syslog_options = 0;
int service_check_timeout;
int service_check_timeout_state;
int host_check_timeout;
int event_handler_timeout;
int notification_timeout;
char *object_precache_file;
char *global_host_event_handler;
char *global_service_event_handler;
command *global_host_event_handler_ptr = NULL;
command *global_service_event_handler_ptr = NULL;
int service_inter_check_delay_method;
int host_inter_check_delay_method;
int service_interleave_factor_method;
int max_host_check_spread;
int max_service_check_spread;
int check_reaper_interval;
int max_check_reaper_time;
int service_freshness_check_interval;
int host_freshness_check_interval;
int auto_rescheduling_interval;
struct load_control loadctl;
int check_orphaned_services;
int check_orphaned_hosts;
int check_service_freshness;
int check_host_freshness;
int auto_reschedule_checks;
int auto_rescheduling_window;
int additional_freshness_latency;
int check_for_updates;
int bare_update_check;
time_t last_update_check = 0L;
unsigned long update_uid = 0L;
int update_available = FALSE;
char *last_program_version = NULL;
char *new_program_version = NULL;
time_t last_program_stop = 0L;
int use_aggressive_host_checking;
time_t cached_host_check_horizon;
time_t cached_service_check_horizon;
int enable_predictive_host_dependency_checks;
int enable_predictive_service_dependency_checks;
int soft_state_dependencies;
int retain_state_information;
int retention_update_interval;
int use_retained_program_state;
int use_retained_scheduling_info;
int retention_scheduling_horizon;
char *retention_file;
unsigned long modified_process_attributes = MODATTR_NONE;
unsigned long modified_host_process_attributes = MODATTR_NONE;
unsigned long modified_service_process_attributes = MODATTR_NONE;
unsigned long retained_host_attribute_mask;
unsigned long retained_service_attribute_mask;
unsigned long retained_contact_host_attribute_mask;
unsigned long retained_contact_service_attribute_mask;
unsigned long retained_process_host_attribute_mask;
unsigned long retained_process_service_attribute_mask;
unsigned long next_event_id = 0L;
unsigned long next_problem_id = 1L;
unsigned long next_comment_id = 0L;
unsigned long next_notification_id = 0L;
int verify_config = FALSE;
int test_scheduling = FALSE;
int precache_objects = FALSE;
int use_precached_objects = FALSE;
int sigshutdown = FALSE;
int sigrestart = FALSE;
int caught_signal = FALSE;
int sig_id = 0;
int daemon_dumps_core;
int max_parallel_service_checks;
int currently_running_service_checks = 0;
int currently_running_host_checks = 0;
time_t event_start = 0L;
int translate_passive_host_checks;
int passive_host_checks_are_soft;
int status_update_interval;
int time_change_threshold;
unsigned long event_broker_options;
double low_service_flap_threshold;
double high_service_flap_threshold;
double low_host_flap_threshold;
double high_host_flap_threshold;
int use_large_installation_tweaks;
int enable_environment_macros;
int free_child_process_memory;
int child_processes_fork_twice;
char *use_timezone;
char *website_url;
int allow_empty_hostgroup_assignment;
int host_down_disable_service_checks;
int service_skip_check_dependency_status;
int service_skip_check_parent_status;
int service_skip_check_host_down_status;
int host_skip_check_dependency_status;
/*** perfdata variables ***/
int perfdata_timeout;
char *host_perfdata_command;
char *service_perfdata_command;
char *host_perfdata_file_template;
char *service_perfdata_file_template;
char *host_perfdata_file;
char *service_perfdata_file;
int host_perfdata_file_append;
int service_perfdata_file_append;
int host_perfdata_file_pipe;
int service_perfdata_file_pipe;
unsigned long host_perfdata_file_processing_interval;
unsigned long service_perfdata_file_processing_interval;
char *host_perfdata_file_processing_command;
char *service_perfdata_file_processing_command;
int host_perfdata_process_empty_results;
int service_perfdata_process_empty_results;
/*** end perfdata variables */
/* Filename variables used by handle_sigxfsz */
extern char *status_file;
static long long check_file_size(char *, unsigned long, struct rlimit);
notification *notification_list;
time_t max_check_result_file_age;
check_stats check_statistics[MAX_CHECK_STATS_TYPES];
char *debug_file;
int debug_level;
int debug_verbosity;
unsigned long max_debug_file_size;
iobroker_set *nagios_iobs = NULL;
squeue_t *nagios_squeue = NULL; /* our scheduling queue */
sched_info scheduling_info;
/* from GNU defines errno as a macro, since it's a per-thread variable */
#ifndef errno
extern int errno;
#endif
/* Initialize the non-shared main configuration variables */
void init_main_cfg_vars(int first_time) {
/* Initialize only the first time */
if(first_time) {
num_check_workers = 0; /* auto-decide */
qh_socket_path = NULL; /* disabled */
}
log_file = NULL;
temp_file = NULL;
temp_path = NULL;
check_result_path = NULL;
command_file = NULL;
lock_file = NULL;
log_archive_path = NULL;
debug_file = NULL;
object_precache_file = (char *)strdup(DEFAULT_PRECACHED_OBJECT_FILE);
nagios_user = NULL;
nagios_group = NULL;
website_url = NULL;
use_regexp_matches = FALSE;
use_true_regexp_matching = FALSE;
use_syslog = DEFAULT_USE_SYSLOG;
log_service_retries = DEFAULT_LOG_SERVICE_RETRIES;
log_host_retries = DEFAULT_LOG_HOST_RETRIES;
log_initial_states = DEFAULT_LOG_INITIAL_STATES;
if(first_time) {
/* Not sure why this is not reset in reset_variables() */
log_current_states = DEFAULT_LOG_CURRENT_STATES;
}
log_notifications = DEFAULT_NOTIFICATION_LOGGING;
log_event_handlers = DEFAULT_LOG_EVENT_HANDLERS;
log_external_commands = DEFAULT_LOG_EXTERNAL_COMMANDS;
log_passive_checks = DEFAULT_LOG_PASSIVE_CHECKS;
service_check_timeout = DEFAULT_SERVICE_CHECK_TIMEOUT;
if(first_time) {
/* Not sure why this is not reset in reset_variables() */
service_check_timeout_state = STATE_CRITICAL;
}
host_check_timeout = DEFAULT_HOST_CHECK_TIMEOUT;
event_handler_timeout = DEFAULT_EVENT_HANDLER_TIMEOUT;
notification_timeout = DEFAULT_NOTIFICATION_TIMEOUT;
ocsp_timeout = DEFAULT_OCSP_TIMEOUT;
ochp_timeout = DEFAULT_OCHP_TIMEOUT;
if(first_time) {
/* Not sure why this is not reset in reset_variables() */
illegal_object_chars = NULL;
}
service_inter_check_delay_method = ICD_SMART;
host_inter_check_delay_method = ICD_SMART;
service_interleave_factor_method = ILF_SMART;
max_service_check_spread = DEFAULT_SERVICE_CHECK_SPREAD;
max_host_check_spread = DEFAULT_HOST_CHECK_SPREAD;
use_aggressive_host_checking = DEFAULT_AGGRESSIVE_HOST_CHECKING;
cached_host_check_horizon = DEFAULT_CACHED_HOST_CHECK_HORIZON;
cached_service_check_horizon = DEFAULT_CACHED_SERVICE_CHECK_HORIZON;
enable_predictive_host_dependency_checks =
DEFAULT_ENABLE_PREDICTIVE_HOST_DEPENDENCY_CHECKS;
enable_predictive_service_dependency_checks =
DEFAULT_ENABLE_PREDICTIVE_SERVICE_DEPENDENCY_CHECKS;
soft_state_dependencies = FALSE;
retain_state_information = FALSE;
retention_update_interval = DEFAULT_RETENTION_UPDATE_INTERVAL;
use_retained_program_state = TRUE;
use_retained_scheduling_info = FALSE;
retention_scheduling_horizon = DEFAULT_RETENTION_SCHEDULING_HORIZON;
if(first_time) {
/* Not sure why this is not reset in reset_variables() */
retention_file = NULL;
}
retained_host_attribute_mask = 0L;
retained_service_attribute_mask = 0L;
retained_process_host_attribute_mask = 0L;
retained_process_service_attribute_mask = 0L;
retained_contact_host_attribute_mask = 0L;
retained_contact_service_attribute_mask = 0L;
check_reaper_interval = DEFAULT_CHECK_REAPER_INTERVAL;
max_check_reaper_time = DEFAULT_MAX_REAPER_TIME;
max_check_result_file_age = DEFAULT_MAX_CHECK_RESULT_AGE;
service_freshness_check_interval = DEFAULT_FRESHNESS_CHECK_INTERVAL;
host_freshness_check_interval = DEFAULT_FRESHNESS_CHECK_INTERVAL;
auto_rescheduling_interval = DEFAULT_AUTO_RESCHEDULING_INTERVAL;
auto_rescheduling_window = DEFAULT_AUTO_RESCHEDULING_WINDOW;
check_orphaned_services = DEFAULT_CHECK_ORPHANED_SERVICES;
check_orphaned_hosts = DEFAULT_CHECK_ORPHANED_HOSTS;
check_service_freshness = DEFAULT_CHECK_SERVICE_FRESHNESS;
check_host_freshness = DEFAULT_CHECK_HOST_FRESHNESS;
auto_reschedule_checks = DEFAULT_AUTO_RESCHEDULE_CHECKS;
if(first_time) {
/* Not sure why this is not reset in reset_variables() */
daemon_dumps_core = TRUE;
}
max_parallel_service_checks = DEFAULT_MAX_PARALLEL_SERVICE_CHECKS;
status_update_interval = DEFAULT_STATUS_UPDATE_INTERVAL;
event_broker_options = BROKER_NOTHING;
time_change_threshold = DEFAULT_TIME_CHANGE_THRESHOLD;
low_service_flap_threshold = DEFAULT_LOW_SERVICE_FLAP_THRESHOLD;
high_service_flap_threshold = DEFAULT_HIGH_SERVICE_FLAP_THRESHOLD;
low_host_flap_threshold = DEFAULT_LOW_HOST_FLAP_THRESHOLD;
high_host_flap_threshold = DEFAULT_HIGH_HOST_FLAP_THRESHOLD;
translate_passive_host_checks = DEFAULT_TRANSLATE_PASSIVE_HOST_CHECKS;
passive_host_checks_are_soft = DEFAULT_PASSIVE_HOST_CHECKS_SOFT;
use_large_installation_tweaks = DEFAULT_USE_LARGE_INSTALLATION_TWEAKS;
enable_environment_macros = FALSE;
free_child_process_memory = -1;
child_processes_fork_twice = -1;
if(first_time) {
/* Not sure why these are not reset in reset_variables() */
use_timezone = NULL;
allow_empty_hostgroup_assignment =
DEFAULT_ALLOW_EMPTY_HOSTGROUP_ASSIGNMENT;
host_down_disable_service_checks = FALSE;
service_skip_check_dependency_status = -1;
service_skip_check_parent_status = -1;
service_skip_check_host_down_status = -1;
host_skip_check_dependency_status = -1;
perfdata_timeout = 0;
host_perfdata_command = NULL;
service_perfdata_command = NULL;
host_perfdata_file_template = NULL;
service_perfdata_file_template = NULL;
host_perfdata_file = NULL;
service_perfdata_file = NULL;
host_perfdata_file_pipe = FALSE;
host_perfdata_file_append = TRUE;
service_perfdata_file_pipe = FALSE;
service_perfdata_file_append = TRUE;
host_perfdata_file_processing_interval = 0L;
service_perfdata_file_processing_interval = 0L;
host_perfdata_file_processing_command = NULL;
service_perfdata_file_processing_command = NULL;
host_perfdata_process_empty_results =
DEFAULT_HOST_PERFDATA_PROCESS_EMPTY_RESULTS;
service_perfdata_process_empty_results =
DEFAULT_SERVICE_PERFDATA_PROCESS_EMPTY_RESULTS;
}
additional_freshness_latency = DEFAULT_ADDITIONAL_FRESHNESS_LATENCY;
if(first_time) {
/* Not sure why these are not reset in reset_variables() */
check_for_updates = DEFAULT_CHECK_FOR_UPDATES;
bare_update_check = DEFAULT_BARE_UPDATE_CHECK;
debug_level = DEFAULT_DEBUG_LEVEL;
debug_verbosity = DEFAULT_DEBUG_VERBOSITY;
max_debug_file_size = DEFAULT_MAX_DEBUG_FILE_SIZE;
}
global_host_event_handler = NULL;
global_service_event_handler = NULL;
ocsp_command = NULL;
ochp_command = NULL;
return;
}
static const char *worker_source_name(const void *source) {
return source ? (const char *)source : "unknown internal source (voodoo, perhaps?)";
}
static const char *spool_file_source_name(const void *source) {
return "check result spool dir";
}
struct check_engine nagios_check_engine = {
"Nagios Core",
worker_source_name,
NULL,
};
static struct check_engine nagios_spool_check_engine = {
"Spooled checkresult file",
spool_file_source_name,
NULL,
};
const char *check_result_source(check_result *cr) {
if(cr->engine)
return cr->engine->source_name(cr->source);
return cr->source ? (const char *)cr->source : "(unknown engine)";
}
int set_loadctl_options(char *opts, unsigned int len)
{
struct kvvec *kvv;
int i;
kvv = buf2kvvec(opts, len, '=', ';', 0);
for (i = 0; i < kvv->kv_pairs; i++) {
struct key_value *kv = &kvv->kv[i];
if (!strcmp(kv->key, "enabled")) {
if (*kv->value == '1') {
if (!(loadctl.options & LOADCTL_ENABLED))
logit(0, 0, "Warning: Enabling experimental load control\n");
loadctl.options |= LOADCTL_ENABLED;
}
else {
if (loadctl.options & LOADCTL_ENABLED)
logit(0, 0, "Warning: Disabling experimental load control\n");
loadctl.options &= (~LOADCTL_ENABLED);
}
} else if (!strcmp(kv->key, "jobs_max")) {
loadctl.jobs_max = atoi(kv->value);
} else if (!strcmp(kv->key, "jobs_min")) {
loadctl.jobs_min = atoi(kv->value);
} else if (!strcmp(kv->key, "jobs_limit")) {
loadctl.jobs_limit = atoi(kv->value);
} else if (!strcmp(kv->key, "check_interval")) {
loadctl.check_interval = strtoul(kv->value, NULL, 10);
} else if (!strcmp(kv->key, "backoff_limit")) {
loadctl.backoff_limit = strtod(kv->value, NULL);
} else if (!strcmp(kv->key, "rampup_limit")) {
loadctl.rampup_limit = strtod(kv->value, NULL);
} else if (!strcmp(kv->key, "backoff_change")) {
loadctl.backoff_change = atoi(kv->value);
} else if (!strcmp(kv->key, "rampup_change")) {
loadctl.rampup_change = atoi(kv->value);
} else {
logit(NSLOG_CONFIG_ERROR, TRUE, "Error: Bad loadctl option; %s = %s\n", kv->key, kv->value);
return 400;
}
}
/* precedence order is "jobs_min -> jobs_max -> jobs_limit" */
if (loadctl.jobs_max < loadctl.jobs_min)
loadctl.jobs_max = loadctl.jobs_min;
if (loadctl.jobs_limit > loadctl.jobs_max)
loadctl.jobs_limit = loadctl.jobs_max;
if (loadctl.jobs_limit < loadctl.jobs_min)
loadctl.jobs_limit = loadctl.jobs_min;
kvvec_destroy(kvv, 0);
return 0;
}
/******************************************************************/
/******************** SYSTEM COMMAND FUNCTIONS ********************/
/******************************************************************/
/* executes a system command - used for notifications, event handlers, etc. */
int my_system_r(nagios_macros *mac, char *cmd, int timeout, int *early_timeout, double *exectime, char **output, int max_output_length) {
pid_t pid = 0;
int status = 0;
int result = 0;
char buffer[MAX_INPUT_BUFFER] = "";
int fd[2];
FILE *fp = NULL;
int bytes_read = 0;
struct timeval start_time, end_time;
dbuf output_dbuf;
int dbuf_chunk = 1024;
int flags;
#ifdef HAVE_SIGACTION
struct sigaction sig_action;
#endif
log_debug_info(DEBUGL_FUNCTIONS, 0, "my_system_r()\n");
/* initialize return variables */
if(output != NULL)
*output = NULL;
*early_timeout = FALSE;
*exectime = 0.0;
/* if no command was passed, return with no error */
if(cmd == NULL)
return STATE_OK;
log_debug_info(DEBUGL_COMMANDS, 1, "Running command '%s'...\n", cmd);
/* create a pipe */
pipe(fd);
/* make the pipe non-blocking */
fcntl(fd[0], F_SETFL, O_NONBLOCK);
fcntl(fd[1], F_SETFL, O_NONBLOCK);
/* get the command start time */
gettimeofday(&start_time, NULL);
#ifdef USE_EVENT_BROKER
/* send data to event broker */
end_time.tv_sec = 0L;
end_time.tv_usec = 0L;
broker_system_command(NEBTYPE_SYSTEM_COMMAND_START, NEBFLAG_NONE, NEBATTR_NONE, start_time, end_time, *exectime, timeout, *early_timeout, result, cmd, NULL, NULL);
#endif
/* fork */
pid = fork();
/* return an error if we couldn't fork */
if(pid == -1) {
logit(NSLOG_RUNTIME_WARNING, TRUE, "Warning: fork() in my_system_r() failed for command \"%s\" - errno: %s\n", cmd, strerror(errno));
/* close both ends of the pipe */
close(fd[0]);
close(fd[1]);
return STATE_UNKNOWN;
}
/* execute the command in the child process */
if(pid == 0) {
/* become process group leader */
setpgid(0, 0);
/* set environment variables */
set_all_macro_environment_vars_r(mac, TRUE);
/* ADDED 11/12/07 EG */
/* close external command file and shut down worker thread */
close_command_file();
/* reset signal handling */
reset_sighandler();
/* close pipe for reading */
close(fd[0]);
/* prevent fd from being inherited by child processed */
flags = fcntl(fd[1], F_GETFD, 0);
flags |= FD_CLOEXEC;
fcntl(fd[1], F_SETFD, flags);
/* trap commands that timeout */
#ifdef HAVE_SIGACTION
sig_action.sa_sigaction = NULL;
sig_action.sa_handler = my_system_sighandler;
sigfillset(&sig_action.sa_mask);
sig_action.sa_flags = SA_NODEFER|SA_RESTART;
sigaction(SIGALRM, &sig_action, NULL);
#else
signal(SIGALRM, my_system_sighandler);
#endif /* HAVE_SIGACTION */
alarm(timeout);
/* run the command */
fp = (FILE *)popen(cmd, "r");
/* report an error if we couldn't run the command */
if(fp == NULL) {
strncpy(buffer, "(Error: Could not execute command)\n", sizeof(buffer) - 1);
buffer[sizeof(buffer) - 1] = '\x0';
/* write the error back to the parent process */
write(fd[1], buffer, strlen(buffer) + 1);
result = STATE_CRITICAL;
}
else {
/* write all the lines of output back to the parent process */
while(fgets(buffer, sizeof(buffer) - 1, fp))
write(fd[1], buffer, strlen(buffer));
/* close the command and get termination status */
status = pclose(fp);
/* report an error if we couldn't close the command */
if(status == -1)
result = STATE_CRITICAL;
else {
if(WEXITSTATUS(status) == 0 && WIFSIGNALED(status))
result = 128 + WTERMSIG(status);
result = WEXITSTATUS(status);
}
}
/* close pipe for writing */
close(fd[1]);
/* reset the alarm */
alarm(0);
/* clear environment variables */
set_all_macro_environment_vars_r(mac, FALSE);
#ifndef DONT_USE_MEMORY_PERFORMANCE_TWEAKS
/* free allocated memory */
/* this needs to be done last, so we don't free memory for variables before they're used above */
if(free_child_process_memory == TRUE)
free_memory(mac);
#endif
_exit(result);
}
/* parent waits for child to finish executing command */
else {
/* close pipe for writing */
close(fd[1]);
/* wait for child to exit */
waitpid(pid, &status, 0);
/* get the end time for running the command */
gettimeofday(&end_time, NULL);
/* return execution time in milliseconds */
*exectime = (double)((double)(end_time.tv_sec - start_time.tv_sec) + (double)((end_time.tv_usec - start_time.tv_usec) / 1000) / 1000.0);
if(*exectime < 0.0)
*exectime = 0.0;
/* get the exit code returned from the program */
result = WEXITSTATUS(status);
/* check for possibly missing scripts/binaries/etc */
if(result == 126 || result == 127) {
logit(NSLOG_RUNTIME_WARNING, TRUE, "Warning: Attempting to execute the command \"%s\" resulted in a return code of %d. Make sure the script or binary you are trying to execute actually exists...\n", cmd, result);
}
/* check bounds on the return value */
if(result < -1 || result > 3)
result = STATE_UNKNOWN;
/* initialize dynamic buffer */
dbuf_init(&output_dbuf, dbuf_chunk);
/* Opsera patch to check timeout before attempting to read output via pipe. Originally by Sven Nierlein */
/* if there was a critical return code AND the command time exceeded the timeout thresholds, assume a timeout */
if(result == STATE_CRITICAL && (end_time.tv_sec - start_time.tv_sec) >= timeout) {
/* set the early timeout flag */
*early_timeout = TRUE;
/* try to kill the command that timed out by sending termination signal to child process group */
kill((pid_t)(-pid), SIGTERM);
sleep(1);
kill((pid_t)(-pid), SIGKILL);
}
/* read output if timeout has not occurred */
else {
/* initialize output */
strcpy(buffer, "");
/* try and read the results from the command output (retry if we encountered a signal) */
do {
bytes_read = read(fd[0], buffer, sizeof(buffer) - 1);
/* append data we just read to dynamic buffer */
if(bytes_read > 0) {
buffer[bytes_read] = '\x0';
dbuf_strcat(&output_dbuf, buffer);
}
/* handle errors */
if(bytes_read == -1) {
/* we encountered a recoverable error, so try again */
if(errno == EINTR)
continue;
/* patch by Henning Brauer to prevent CPU hogging */
else if(errno == EAGAIN) {
struct pollfd pfd;
pfd.fd = fd[0];
pfd.events = POLLIN;
poll(&pfd, 1, -1);
continue;
}
else
break;
}
/* we're done */
if(bytes_read == 0)
break;
}
while(1);
/* cap output length - this isn't necessary, but it keeps runaway plugin output from causing problems */
if(max_output_length > 0 && (int)output_dbuf.used_size > max_output_length)
output_dbuf.buf[max_output_length] = '\x0';
if(output != NULL && output_dbuf.buf)
*output = (char *)strdup(output_dbuf.buf);
}
log_debug_info(DEBUGL_COMMANDS, 1, "Execution time=%.3f sec, early timeout=%d, result=%d, output=%s\n", *exectime, *early_timeout, result, (output_dbuf.buf == NULL) ? "(null)" : output_dbuf.buf);
#ifdef USE_EVENT_BROKER
/* send data to event broker */
broker_system_command(NEBTYPE_SYSTEM_COMMAND_END, NEBFLAG_NONE, NEBATTR_NONE, start_time, end_time, *exectime, timeout, *early_timeout, result, cmd, (output_dbuf.buf == NULL) ? NULL : output_dbuf.buf, NULL);
#endif
/* free memory */
dbuf_free(&output_dbuf);
/* close the pipe for reading */
close(fd[0]);
}
return result;
}
/*
* For API compatibility, we must include a my_system() whose
* signature doesn't include the nagios_macros variable.
* NDOUtils uses this. Possibly other modules as well.
*/
int my_system(char *cmd, int timeout, int *early_timeout, double *exectime, char **output, int max_output_length) {
return my_system_r(get_global_macros(), cmd, timeout, early_timeout, exectime, output, max_output_length);
}
/* given a "raw" command, return the "expanded" or "whole" command line */
int get_raw_command_line_r(nagios_macros *mac, command *cmd_ptr, char *cmd, char **full_command, int macro_options) {
char temp_arg[MAX_COMMAND_BUFFER] = "";
char *arg_buffer = NULL;
register int x = 0;
register int y = 0;
register int arg_index = 0;
log_debug_info(DEBUGL_FUNCTIONS, 0, "get_raw_command_line_r()\n");
/* clear the argv macros */
clear_argv_macros_r(mac);
/* make sure we've got all the requirements */
if(cmd_ptr == NULL || full_command == NULL)
return ERROR;
log_debug_info(DEBUGL_COMMANDS | DEBUGL_CHECKS | DEBUGL_MACROS, 2, "Raw Command Input: %s\n", cmd_ptr->command_line);
/* get the full command line */
*full_command = (char *)strdup((cmd_ptr->command_line == NULL) ? "" : cmd_ptr->command_line);
/* XXX: Crazy indent */
/* get the command arguments */
if(cmd != NULL) {
/* skip the command name (we're about to get the arguments)... */
for(arg_index = 0;; arg_index++) {
if(cmd[arg_index] == '!' || cmd[arg_index] == '\x0')
break;
}
/* get each command argument */
for(x = 0; x < MAX_COMMAND_ARGUMENTS; x++) {
/* we reached the end of the arguments... */
if(cmd[arg_index] == '\x0')
break;
/* get the next argument */
/* can't use strtok(), as that's used in process_macros... */
for(arg_index++, y = 0; y < (int)sizeof(temp_arg) - 1; arg_index++) {
/* handle escaped argument delimiters */
if(cmd[arg_index] == '\\' && cmd[arg_index+1] == '!') {
arg_index++;
} else if(cmd[arg_index] == '!' || cmd[arg_index] == '\x0') {
/* end of argument */
break;
}
/* copy the character */
temp_arg[y] = cmd[arg_index];
y++;
}
temp_arg[y] = '\x0';
/* ADDED 01/29/04 EG */
/* process any macros we find in the argument */
process_macros_r(mac, temp_arg, &arg_buffer, macro_options);
mac->argv[x] = arg_buffer;
}
}
log_debug_info(DEBUGL_COMMANDS | DEBUGL_CHECKS | DEBUGL_MACROS, 2, "Expanded Command Output: %s\n", *full_command);
return OK;
}
/*
* This function modifies the global macro struct and is thus not
* threadsafe
*/
int get_raw_command_line(command *cmd_ptr, char *cmd, char **full_command, int macro_options) {
nagios_macros *mac;
mac = get_global_macros();
return get_raw_command_line_r(mac, cmd_ptr, cmd, full_command, macro_options);
}
/******************************************************************/
/******************** ENVIRONMENT FUNCTIONS ***********************/
/******************************************************************/
/* sets or unsets an environment variable */
int set_environment_var(char *name, char *value, int set) {
#ifndef HAVE_SETENV
char *env_string = NULL;
#endif
/* we won't mess with null variable names */
if(name == NULL)
return ERROR;
/* set the environment variable */
if(set == TRUE) {
#ifdef HAVE_SETENV
setenv(name, (value == NULL) ? "" : value, 1);
#else
/* needed for Solaris and systems that don't have setenv() */
/* this will leak memory, but in a "controlled" way, since lost memory should be freed when the child process exits */
asprintf(&env_string, "%s=%s", name, (value == NULL) ? "" : value);
if(env_string)
putenv(env_string);
#endif
}
/* clear the variable */
else {
#ifdef HAVE_UNSETENV
unsetenv(name);
#endif
}
return OK;
}
/******************************************************************/
/************************* TIME FUNCTIONS *************************/
/******************************************************************/
/* Checks if the given time is in daylight time saving period */
static int is_dst_time(time_t *time) {
struct tm *bt = localtime(time);
return bt->tm_isdst;
}
/* Returns the shift in seconds if the given times are across the daylight time saving period change */
static int get_dst_shift(time_t *start, time_t *end) {
int shift = 0, dst_end, dst_start;
dst_start = is_dst_time(start);
dst_end = is_dst_time(end);
if(dst_start < dst_end) {
shift = 3600;
}
else if(dst_start > dst_end) {
shift = -3600;
}
return shift;
}
/*#define TEST_TIMEPERIODS_A 1*/
static timerange* _get_matching_timerange(time_t test_time, timeperiod *tperiod) {
daterange *temp_daterange = NULL;
time_t start_time = (time_t)0L;
time_t end_time = (time_t)0L;
unsigned long days = 0L;
int year = 0;
int shift = 0;
time_t midnight = (time_t)0L;
struct tm *t, tm_s;
int daterange_type = 0;
int test_time_year = 0;
int test_time_mon = 0;
int test_time_wday = 0;
log_debug_info(DEBUGL_FUNCTIONS, 0, "_get_matching_timerange()\n");
if(tperiod == NULL)
return NULL;
t = localtime_r((time_t *)&test_time, &tm_s);
test_time_year = t->tm_year;
test_time_mon = t->tm_mon;
test_time_wday = t->tm_wday;
/* calculate the start of the day (midnight, 00:00 hours) when the specified test time occurs */
t->tm_sec = 0;
t->tm_min = 0;
t->tm_hour = 0;
midnight = mktime(t);
/**** check exceptions first ****/
for(daterange_type = 0; daterange_type < DATERANGE_TYPES; daterange_type++) {
for(temp_daterange = tperiod->exceptions[daterange_type]; temp_daterange != NULL; temp_daterange = temp_daterange->next) {
#ifdef TEST_TIMEPERIODS_A
printf("TYPE: %d\n", daterange_type);
printf("TEST: %llu = %s", (unsigned long long)test_time, ctime(&test_time));
printf("MIDNIGHT: %llu = %s", (unsigned long long)midnight, ctime(&midnight));
#endif
/* get the start time */
switch(daterange_type) {
case DATERANGE_CALENDAR_DATE:
t->tm_sec = 0;
t->tm_min = 0;
t->tm_hour = 0;
t->tm_wday = 0;
t->tm_mday = temp_daterange->smday;
t->tm_mon = temp_daterange->smon;
t->tm_year = (temp_daterange->syear - 1900);
t->tm_isdst = -1;
start_time = mktime(t);
break;
case DATERANGE_MONTH_DATE:
start_time = calculate_time_from_day_of_month(test_time_year, temp_daterange->smon, temp_daterange->smday);
break;
case DATERANGE_MONTH_DAY:
start_time = calculate_time_from_day_of_month(test_time_year, test_time_mon, temp_daterange->smday);
break;
case DATERANGE_MONTH_WEEK_DAY:
start_time = calculate_time_from_weekday_of_month(test_time_year, temp_daterange->smon, temp_daterange->swday, temp_daterange->swday_offset);
break;
case DATERANGE_WEEK_DAY:
start_time = calculate_time_from_weekday_of_month(test_time_year, test_time_mon, temp_daterange->swday, temp_daterange->swday_offset);
break;
default:
continue;
break;
}
/* get the end time */
switch(daterange_type) {
case DATERANGE_CALENDAR_DATE:
t->tm_sec = 0;
t->tm_min = 0;
t->tm_hour = 0;
t->tm_wday = 0;
t->tm_mday = temp_daterange->emday;
t->tm_mon = temp_daterange->emon;
t->tm_year = (temp_daterange->eyear - 1900);
t->tm_isdst = -1;
end_time = mktime(t);
break;
case DATERANGE_MONTH_DATE:
year = test_time_year;
end_time = calculate_time_from_day_of_month(year, temp_daterange->emon, temp_daterange->emday);
/* advance a year if necessary: august 2 - february 5 */
if(end_time < start_time) {
year++;
end_time = calculate_time_from_day_of_month(year, temp_daterange->emon, temp_daterange->emday);
}
break;
case DATERANGE_MONTH_DAY:
end_time = calculate_time_from_day_of_month(test_time_year, test_time_mon, temp_daterange->emday);
break;
case DATERANGE_MONTH_WEEK_DAY:
year = test_time_year;
end_time = calculate_time_from_weekday_of_month(year, temp_daterange->emon, temp_daterange->ewday, temp_daterange->ewday_offset);
/* advance a year if necessary: thursday 2 august - monday 3 february */
if(end_time < start_time) {
year++;
end_time = calculate_time_from_weekday_of_month(year, temp_daterange->emon, temp_daterange->ewday, temp_daterange->ewday_offset);
}
break;
case DATERANGE_WEEK_DAY:
end_time = calculate_time_from_weekday_of_month(test_time_year, test_time_mon, temp_daterange->ewday, temp_daterange->ewday_offset);
break;
default:
continue;
break;
}
#ifdef TEST_TIMEPERIODS_A
printf("START: %llu = %s", (unsigned long long)start_time, ctime(&start_time));
printf("END: %llu = %s", (unsigned long long)end_time, ctime(&end_time));
#endif
/* start date was bad, so skip this date range */
if((unsigned long)start_time == 0L)
continue;
/* end date was bad - see if we can handle the error */
if((unsigned long)end_time == 0L) {
switch(daterange_type) {
case DATERANGE_CALENDAR_DATE:
continue;
break;
case DATERANGE_MONTH_DATE:
/* end date can't be helped, so skip it */
if(temp_daterange->emday < 0)
continue;
/* else end date slipped past end of month, so use last day of month as end date */
/* use same year calculated above */
end_time = calculate_time_from_day_of_month(year, temp_daterange->emon, -1);
break;
case DATERANGE_MONTH_DAY:
/* end date can't be helped, so skip it */
if(temp_daterange->emday < 0)
continue;
/* else end date slipped past end of month, so use last day of month as end date */
end_time = calculate_time_from_day_of_month(test_time_year, test_time_mon, -1);
break;
case DATERANGE_MONTH_WEEK_DAY:
/* end date can't be helped, so skip it */
if(temp_daterange->ewday_offset < 0)
continue;
/* else end date slipped past end of month, so use last day of month as end date */
/* use same year calculated above */
end_time = calculate_time_from_day_of_month(year, test_time_mon, -1);
break;
case DATERANGE_WEEK_DAY:
/* end date can't be helped, so skip it */
if(temp_daterange->ewday_offset < 0)
continue;
/* else end date slipped past end of month, so use last day of month as end date */
end_time = calculate_time_from_day_of_month(test_time_year, test_time_mon, -1);
break;
default:
continue;
break;
}
}
/* calculate skip date start (and end) */
if(temp_daterange->skip_interval > 1) {
/* skip start date must be before test time */
if(start_time > test_time)
continue;
/* check if interval is across dlst change and gets the compensation */
shift = get_dst_shift(&start_time, &midnight);
/* how many days have passed between skip start date and test time? */
days = (shift + (unsigned long)midnight - (unsigned long)start_time) / (3600 * 24);
/* if test date doesn't fall on a skip interval day, bail out early */
if((days % temp_daterange->skip_interval) != 0)
continue;
/* use midnight of test date as start time */
else
start_time = midnight;
/* if skipping range has no end, use test date as end */
if((daterange_type == DATERANGE_CALENDAR_DATE) && (is_daterange_single_day(temp_daterange) == TRUE))
end_time = midnight;
}
#ifdef TEST_TIMEPERIODS_A
printf("NEW START: %llu = %s", (unsigned long long)start_time, ctime(&start_time));
printf("NEW END: %llu = %s", (unsigned long long)end_time, ctime(&end_time));
printf("%d DAYS PASSED\n", days);
printf("DLST SHIFT: %i\n", shift);
#endif
/* time falls inside the range of days
* end time < start_time when range covers end-of-$unit
* (fe. end-of-month) */
if((midnight + 84800UL >= start_time && (midnight <= end_time || start_time > end_time)) || (midnight <= end_time && start_time > end_time)) {
#ifdef TEST_TIMEPERIODS_A
printf("(MATCH)\n");
#endif
return temp_daterange->times;
}
}
}
return tperiod->days[test_time_wday];
}
/* see if the specified time falls into a valid time range in the given time period */
int check_time_against_period(time_t test_time, timeperiod *tperiod) {
timerange *temp_timerange = NULL;
timeperiodexclusion *temp_timeperiodexclusion = NULL;
struct tm *t, tm_s;
time_t midnight = (time_t)0L;
time_t day_range_start = (time_t)0L;
time_t day_range_end = (time_t)0L;
log_debug_info(DEBUGL_FUNCTIONS, 0, "check_time_against_period()\n");
t = localtime_r((time_t *)&test_time, &tm_s);
t->tm_sec = 0;
t->tm_min = 0;
t->tm_hour = 0;
midnight = mktime(t);
/* if no period was specified, assume the time is good */
if(tperiod == NULL)
return OK;
for(temp_timeperiodexclusion = tperiod->exclusions; temp_timeperiodexclusion != NULL; temp_timeperiodexclusion = temp_timeperiodexclusion->next) {
if(check_time_against_period(test_time, temp_timeperiodexclusion->timeperiod_ptr) == OK) {
return ERROR;
}
}
for(temp_timerange = _get_matching_timerange(test_time, tperiod); temp_timerange != NULL; temp_timerange = temp_timerange->next) {
day_range_start = (time_t)(midnight + temp_timerange->range_start);
day_range_end = (time_t)(midnight + temp_timerange->range_end);
if(test_time >= day_range_start && test_time <= day_range_end)
return OK;
}
return ERROR;
}
/*#define TEST_TIMEPERIODS_B 1*/
void _get_next_valid_time(time_t pref_time, time_t *valid_time, timeperiod *tperiod);
static void _get_next_invalid_time(time_t pref_time, time_t *invalid_time, timeperiod *tperiod) {
timeperiodexclusion *temp_timeperiodexclusion = NULL;
int depth = 0;
int max_depth = 300; // commonly roughly equal to "days in the future"
struct tm *t, tm_s;
time_t earliest_time = pref_time;
time_t last_earliest_time = 0;
time_t midnight = (time_t)0L;
time_t day_range_start = (time_t)0L;
time_t day_range_end = (time_t)0L;
/* if no period was specified, assume the time is good */
if(tperiod == NULL) {
*invalid_time = pref_time;
return;
}
while (earliest_time != last_earliest_time && depth < max_depth) {
time_t potential_time = 0;
depth++;
last_earliest_time = earliest_time;
t = localtime_r((time_t *)&earliest_time, &tm_s);
t->tm_sec = 0;
t->tm_min = 0;
t->tm_hour = 0;
midnight = mktime(t);
timerange *temp_timerange = _get_matching_timerange(earliest_time, tperiod);
for(; temp_timerange != NULL; temp_timerange = temp_timerange->next) {
/* ranges with start/end of zero mean exclude this day */
if(temp_timerange->range_start == 0 && temp_timerange->range_end == 0)
continue;
day_range_start = (time_t)(midnight + temp_timerange->range_start);
day_range_end = (time_t)(midnight + temp_timerange->range_end);
#ifdef TEST_TIMEPERIODS_B
printf(" INVALID RANGE START: %lu (%lu) = %s", temp_timerange->range_start, (unsigned long)day_range_start, ctime(&day_range_start));
printf(" INVALID RANGE END: %lu (%lu) = %s", temp_timerange->range_end, (unsigned long)day_range_end, ctime(&day_range_end));
#endif
if(day_range_start <= earliest_time && day_range_end > earliest_time)
potential_time = day_range_end + 60;
else
potential_time = earliest_time;
if(potential_time > earliest_time) {
earliest_time = potential_time;
#ifdef TEST_TIMEPERIODS_B
printf(" EARLIEST INVALID TIME: %llu = %s", (unsigned long long)earliest_time, ctime(&earliest_time));
#endif
}
}
for(temp_timeperiodexclusion = tperiod->exclusions; temp_timeperiodexclusion != NULL; temp_timeperiodexclusion = temp_timeperiodexclusion->next) {
_get_next_valid_time(last_earliest_time, &potential_time, temp_timeperiodexclusion->timeperiod_ptr);
if (potential_time + 60 < earliest_time)
earliest_time = potential_time + 60;
}
}
#ifdef TEST_TIMEPERIODS_B
printf(" FINAL EARLIEST INVALID TIME: %llu = %s", (unsigned long long)earliest_time, ctime(&earliest_time));
#endif
if (depth == max_depth)
*invalid_time = pref_time;
else
*invalid_time = earliest_time;
}
/* Separate this out from public get_next_valid_time for testing */
void _get_next_valid_time(time_t pref_time, time_t *valid_time, timeperiod *tperiod) {
timeperiodexclusion *temp_timeperiodexclusion = NULL;
int depth = 0;
int max_depth = 300; // commonly roughly equal to "days in the future"
time_t earliest_time = pref_time;
time_t last_earliest_time = 0;
struct tm *t, tm_s;
time_t midnight = (time_t)0L;
time_t day_range_start = (time_t)0L;
time_t day_range_end = (time_t)0L;
int have_earliest_time = FALSE;
/* if no period was specified, assume the time is good */
if(tperiod == NULL) {
*valid_time = pref_time;
return;
}
while (earliest_time != last_earliest_time && depth < max_depth) {
time_t potential_time = 0;
have_earliest_time = FALSE;
depth++;
last_earliest_time = earliest_time;
t = localtime_r((time_t *)&earliest_time, &tm_s);
t->tm_sec = 0;
t->tm_min = 0;
t->tm_hour = 0;
midnight = mktime(t);
timerange *temp_timerange = _get_matching_timerange(earliest_time, tperiod);
#ifdef TEST_TIMEPERIODS_B
printf(" RANGE START: %lu\n", temp_timerange ? temp_timerange->range_start : 0);
printf(" RANGE END: %lu\n", temp_timerange ? temp_timerange->range_end : 0);
#endif
for(; temp_timerange != NULL; temp_timerange = temp_timerange->next) {
/* ranges with start/end of zero mean exclude this day */
if(temp_timerange->range_start == 0 && temp_timerange->range_end == 0)
continue;
day_range_start = (time_t)(midnight + temp_timerange->range_start);
day_range_end = (time_t)(midnight + temp_timerange->range_end);
#ifdef TEST_TIMEPERIODS_B
printf(" RANGE START: %lu (%lu) = %s", temp_timerange->range_start, (unsigned long)day_range_start, ctime(&day_range_start));
printf(" RANGE END: %lu (%lu) = %s", temp_timerange->range_end, (unsigned long)day_range_end, ctime(&day_range_end));
#endif
/* range is out of bounds */
if(day_range_end < last_earliest_time)
continue;
/* preferred time occurs before range start, so use range start time as earliest potential time */
if(day_range_start >= last_earliest_time)
potential_time = day_range_start;
/* preferred time occurs between range start/end, so use preferred time as earliest potential time */
else if(day_range_end >= last_earliest_time)
potential_time = last_earliest_time;
/* is this the earliest time found thus far? */
if(have_earliest_time == FALSE || potential_time < earliest_time) {
earliest_time = potential_time;
#ifdef TEST_TIMEPERIODS_B
printf(" EARLIEST TIME: %llu = %s", (unsigned long long)earliest_time, ctime(&earliest_time));
#endif
}
have_earliest_time = TRUE;
}
if (have_earliest_time == FALSE) {
earliest_time = midnight + 86400;
} else {
for(temp_timeperiodexclusion = tperiod->exclusions; temp_timeperiodexclusion != NULL; temp_timeperiodexclusion = temp_timeperiodexclusion->next) {
_get_next_invalid_time(earliest_time, &earliest_time, temp_timeperiodexclusion->timeperiod_ptr);
#ifdef TEST_TIMEPERIODS_B
printf(" FINAL EARLIEST TIME: %llu = %s", (unsigned long long)earliest_time, ctime(&earliest_time));
#endif
}
}
}
if (depth == max_depth)
*valid_time = pref_time;
else
*valid_time = earliest_time;
}
/* given a preferred time, get the next valid time within a time period */
void get_next_valid_time(time_t pref_time, time_t *valid_time, timeperiod *tperiod) {
time_t current_time = (time_t)0L;
log_debug_info(DEBUGL_FUNCTIONS, 0, "get_next_valid_time()\n");
/* get time right now, preferred time must be now or in the future */
time(&current_time);
pref_time = (pref_time < current_time) ? current_time : pref_time;
_get_next_valid_time(pref_time, valid_time, tperiod);
}
/* Given the next valid time in a timeperiod, the timeperiod itself, and the normal rescheduling window, */
/* return the next check time */
time_t reschedule_within_timeperiod(time_t starting_valid_time, timeperiod* check_period_ptr, time_t check_window) {
log_debug_info(DEBUGL_FUNCTIONS, 0, "reschedule_within_timeperiod");
/* First, find the next time that is outside the timeperiod */
time_t ending_valid_time;
_get_next_invalid_time(starting_valid_time, &ending_valid_time, check_period_ptr);
/* _get_next_invalid_time returns the first invalid minute. The maximum allowable should be a minute earlier */
ending_valid_time -= 60;
/* Determine whether the next invalid time or the outside of the check_window is closer */
time_t max_nudge = ending_valid_time - starting_valid_time;
/* max_nudge will be less than zero when there's no 'invalid' time */
/* Otherwise, use the closest of the two times to reschedule the check */
if (max_nudge <= 0 || max_nudge > check_window) {
log_debug_info(DEBUGL_CHECKS, 0, "Using raw check_window instead of timeperiod for scheduling \n");
max_nudge = check_window;
}
/* Reschedule within the smaller range */
return starting_valid_time + ranged_urand(0, max_nudge);
}
/* tests if a date range covers just a single day */
int is_daterange_single_day(daterange *dr) {
if(dr == NULL)
return FALSE;
if(dr->syear != dr->eyear)
return FALSE;
if(dr->smon != dr->emon)
return FALSE;
if(dr->smday != dr->emday)
return FALSE;
if(dr->swday != dr->ewday)
return FALSE;
if(dr->swday_offset != dr->ewday_offset)
return FALSE;
return TRUE;
}
/* returns a time (midnight) of particular (3rd, last) day in a given month */
time_t calculate_time_from_day_of_month(int year, int month, int monthday) {
time_t midnight;
int day = 0;
struct tm t;
#ifdef TEST_TIMEPERIODS
printf("YEAR: %d, MON: %d, MDAY: %d\n", year, month, monthday);
#endif
/* positive day (3rd day) */
if(monthday > 0) {
t.tm_sec = 0;
t.tm_min = 0;
t.tm_hour = 0;
t.tm_year = year;
t.tm_mon = month;
t.tm_mday = monthday;
t.tm_isdst = -1;
midnight = mktime(&t);
#ifdef TEST_TIMEPERIODS
printf("MIDNIGHT CALC: %s", ctime(&midnight));
#endif
/* if we rolled over to the next month, time is invalid */
/* assume the user's intention is to keep it in the current month */
if(t.tm_mon != month)
midnight = (time_t)0L;
}
/* negative offset (last day, 3rd to last day) */
else {
/* find last day in the month */
day = 32;
do {
/* back up a day */
day--;
/* make the new time */
t.tm_sec = 0;
t.tm_min = 0;
t.tm_hour = 0;
t.tm_mon = month;
t.tm_year = year;
t.tm_mday = day;
t.tm_isdst = -1;
midnight = mktime(&t);
}
while(t.tm_mon != month);
/* now that we know the last day, back up more */
/* make the new time */
t.tm_mon = month;
t.tm_year = year;
/* -1 means last day of month, so add one to to make this correct - Mike Bird */
t.tm_mday += (monthday < -30) ? -30 : monthday + 1;
t.tm_isdst = -1;
midnight = mktime(&t);
/* if we rolled over to the previous month, time is invalid */
/* assume the user's intention is to keep it in the current month */
if(t.tm_mon != month)
midnight = (time_t)0L;
}
return midnight;
}
/* returns a time (midnight) of particular (3rd, last) weekday in a given month */
time_t calculate_time_from_weekday_of_month(int year, int month, int weekday, int weekday_offset) {
time_t midnight;
int days = 0;
int weeks = 0;
struct tm t;
t.tm_sec = 0;
t.tm_min = 0;
t.tm_hour = 0;
t.tm_year = year;
t.tm_mon = month;
t.tm_mday = 1;
t.tm_isdst = -1;
midnight = mktime(&t);
/* how many days must we advance to reach the first instance of the weekday this month? */
days = weekday - (t.tm_wday);
if(days < 0)
days += 7;
/* positive offset (3rd thursday) */
if(weekday_offset > 0) {
/* how many weeks must we advance (no more than 5 possible) */
weeks = (weekday_offset > 5) ? 5 : weekday_offset;
days += ((weeks - 1) * 7);
/* make the new time */
t.tm_mon = month;
t.tm_year = year;
t.tm_mday = days + 1;
t.tm_isdst = -1;
midnight = mktime(&t);
/* if we rolled over to the next month, time is invalid */
/* assume the user's intention is to keep it in the current month */
if(t.tm_mon != month)
midnight = (time_t)0L;
}
/* negative offset (last thursday, 3rd to last tuesday) */
else {
/* find last instance of weekday in the month */
days += (5 * 7);
do {
/* back up a week */
days -= 7;
/* make the new time */
t.tm_mon = month;
t.tm_year = year;
t.tm_mday = days + 1;
t.tm_isdst = -1;
midnight = mktime(&t);
}
while(t.tm_mon != month);
/* now that we know the last instance of the weekday, back up more */
weeks = (weekday_offset < -5) ? -5 : weekday_offset;
days = ((weeks + 1) * 7);
/* make the new time */
t.tm_mon = month;
t.tm_year = year;
t.tm_mday += days;
t.tm_isdst = -1;
midnight = mktime(&t);
/* if we rolled over to the previous month, time is invalid */
/* assume the user's intention is to keep it in the current month */
if(t.tm_mon != month)
midnight = (time_t)0L;
}
return midnight;
}
/* get the next time to schedule a log rotation */
time_t get_next_log_rotation_time(void) {
time_t current_time;
struct tm *t, tm_s;
int is_dst_now = FALSE;
time_t run_time;
int expected_mday;
time(&current_time);
t = localtime_r(&current_time, &tm_s);
t->tm_min = 0;
t->tm_sec = 0;
is_dst_now = (t->tm_isdst > 0) ? TRUE : FALSE;
switch(log_rotation_method) {
case LOG_ROTATION_HOURLY:
t->tm_hour++;
run_time = mktime(t);
break;
case LOG_ROTATION_DAILY:
t->tm_mday++;
t->tm_hour = 0;
run_time = mktime(t);
break;
case LOG_ROTATION_WEEKLY:
t->tm_mday += (7 - t->tm_wday);
t->tm_hour = 0;
run_time = mktime(t);
break;
case LOG_ROTATION_MONTHLY:
default:
t->tm_mon++;
t->tm_mday = 1;
t->tm_hour = 0;
run_time = mktime(t);
break;
}
if(is_dst_now == TRUE && t->tm_isdst == 0)
run_time += 3600;
return run_time;
}
/******************************************************************/
/******************** SIGNAL HANDLER FUNCTIONS ********************/
/******************************************************************/
/* trap signals so we can exit gracefully */
void setup_sighandler(void) {
#ifdef HAVE_SIGACTION
struct sigaction sig_action;
#endif
/* reset the shutdown flag */
sigshutdown = FALSE;
/* remove buffering from stderr, stdin, and stdout */
setbuf(stdin, (char *)NULL);
setbuf(stdout, (char *)NULL);
setbuf(stderr, (char *)NULL);
/* initialize signal handling */
#ifdef HAVE_SIGACTION
sig_action.sa_sigaction = NULL;
sig_action.sa_handler = SIG_IGN;
sigemptyset(&sig_action.sa_mask);
sig_action.sa_flags = 0;
sigaction(SIGPIPE, &sig_action, NULL);
sig_action.sa_handler = sighandler;
sigfillset(&sig_action.sa_mask);
sig_action.sa_flags = SA_NODEFER|SA_RESTART;
sigaction(SIGQUIT, &sig_action, NULL);
sigaction(SIGTERM, &sig_action, NULL);
sigaction(SIGHUP, &sig_action, NULL);
if(daemon_dumps_core == FALSE && daemon_mode == TRUE)
sigaction(SIGSEGV, &sig_action, NULL);
sig_action.sa_flags = SA_NOCLDWAIT;
#else /* HAVE_SIGACTION */
signal(SIGPIPE, SIG_IGN);
signal(SIGQUIT, sighandler);
signal(SIGTERM, sighandler);
signal(SIGHUP, sighandler);
if(daemon_dumps_core == FALSE && daemon_mode == TRUE)
signal(SIGSEGV, sighandler);
#endif /* HAVE_SIGACTION */
return;
}
/* reset signal handling... */
void reset_sighandler(void) {
/* set signal handling to default actions */
#ifdef HAVE_SIGACTION
struct sigaction sig_action;
sig_action.sa_sigaction = NULL;
sig_action.sa_handler = SIG_DFL;
sigemptyset(&sig_action.sa_mask);
sig_action.sa_flags = 0;
sigaction(SIGQUIT, &sig_action, NULL);
sigaction(SIGTERM, &sig_action, NULL);
sigaction(SIGHUP, &sig_action, NULL);
sigaction(SIGSEGV, &sig_action, NULL);
sigaction(SIGPIPE, &sig_action, NULL);
sigaction(SIGXFSZ, &sig_action, NULL);
#else /* HAVE_SIGACTION */
signal(SIGQUIT, SIG_DFL);
signal(SIGTERM, SIG_DFL);
signal(SIGHUP, SIG_DFL);
signal(SIGSEGV, SIG_DFL);
signal(SIGPIPE, SIG_DFL);
signal(SIGXFSZ, SIG_DFL);
#endif /* HAVE_SIGACTION */
return;
}
/* handle signals */
void sighandler(int sig) {
const char *sigs[35] = {"EXIT", "HUP", "INT", "QUIT", "ILL", "TRAP", "ABRT", "BUS", "FPE", "KILL", "USR1", "SEGV", "USR2", "PIPE", "ALRM", "TERM", "STKFLT", "CHLD", "CONT", "STOP", "TSTP", "TTIN", "TTOU", "URG", "XCPU", "XFSZ", "VTALRM", "PROF", "WINCH", "IO", "PWR", "UNUSED", "ZERR", "DEBUG", (char *)NULL};
int x = 0;
/* if shutdown is already true, we're in a signal trap loop! */
/* changed 09/07/06 to only exit on segfaults */
if(sigshutdown == TRUE && sig == SIGSEGV)
exit(ERROR);
caught_signal = TRUE;
if(sig < 0)
sig = -sig;
for(x = 0; sigs[x] != (char *)NULL; x++);
sig %= x;
sig_id = sig;
/* we received a SIGHUP, so restart... */
if(sig == SIGHUP)
sigrestart = TRUE;
/* else begin shutting down... */
else if(sig < 16) {
logit(NSLOG_PROCESS_INFO, TRUE, "Caught SIG%s, shutting down...\n", sigs[sig]);
sigshutdown = TRUE;
}
return;
}
/* handle timeouts when executing commands via my_system_r() */
void my_system_sighandler(int sig) {
/* force the child process to exit... */
_exit(STATE_CRITICAL);
}
/* Handle the SIGXFSZ signal. A SIGXFSZ signal is received when a file exceeds
the maximum allowable size either as dictated by the fsize parameter in
/etc/security/limits.conf (ulimit -f) or by the maximum size allowed by
the filesystem */
void handle_sigxfsz(int sig) {
static time_t lastlog_time = (time_t)0; /* Save the last log time so we
don't log too often. */
unsigned long log_interval = 300; /* How frequently to log messages
about receiving the signal */
struct rlimit rlim;
time_t now;
char *files[] = {
log_file,
debug_file,
host_perfdata_file,
service_perfdata_file,
object_cache_file,
object_precache_file,
status_file,
retention_file,
};
int x;
char **filep;
long long size;
long long max_size = 0LL;
char *max_name = NULL;
if(SIGXFSZ == sig) { /* Make sure we're handling the correct signal */
/* Check the current time and if less time has passed since the last
time the signal was received, ignore it */
time(&now);
if((unsigned long)(now - lastlog_time) < log_interval) return;
/* Get the current file size limit */
if(getrlimit(RLIMIT_FSIZE, &rlim) != 0) {
/* Attempt to log the error, realizing that the logging may fail
if it is the log file that is over the size limit. */
logit(NSLOG_RUNTIME_ERROR, TRUE,
"Unable to determine current resource limits: %s\n",
strerror(errno));
}
/* Try to figure out which file caused the signal and react
appropriately */
for(x = 0, filep = files; x < (sizeof(files) / sizeof(files[0]));
x++, filep++) {
if((*filep != NULL) && strcmp(*filep, "/dev/null")) {
if((size = check_file_size(*filep, 1024, rlim)) == -1) {
lastlog_time = now;
return;
}
else if(size > max_size) {
max_size = size;
max_name = log_file;
}
}
}
if((max_size > 0) && (max_name != NULL)) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "SIGXFSZ received because a "
"file's size may have exceeded the file size limits of "
"the filesystem. The largest file checked, '%s', has a "
"size of %lld bytes", max_name, max_size);
}
else {
logit(NSLOG_RUNTIME_ERROR, TRUE, "SIGXFSZ received but unable to "
"determine which file may have caused it.");
}
}
return;
}
/* Checks a file to determine whether it exceeds resource limit imposed
limits. Returns the file size if file is OK, 0 if it's status could not
be determined, or -1 if not OK. fudge is the fudge factor (in bytes) for
checking the file size */
static long long check_file_size(char *path, unsigned long fudge,
struct rlimit rlim) {
struct stat status;
/* Make sure we were passed a legitimate file path */
if(NULL == path) {
return 0;
}
/* Get the status of the file */
if(stat(path, &status) == 0) {
/* Make sure it is a file */
if(S_ISREG(status.st_mode)) {
/* If the file size plus the fudge factor exceeds the
current resource limit imposed size limit, log an error */
if(status.st_size + fudge > rlim.rlim_cur) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Size of file '%s' (%llu) "
"exceeds (or nearly exceeds) size imposed by resource "
"limits (%llu). Consider increasing limits with "
"ulimit(1).\n", path,
(unsigned long long)status.st_size,
(unsigned long long)rlim.rlim_cur);
return -1;
}
else {
return status.st_size;
}
}
else {
return 0;
}
}
else {
/* If we could not determine the file status, log an error message */
logit(NSLOG_RUNTIME_ERROR, TRUE,
"Unable to determine status of file %s: %s\n",
log_file, strerror(errno));
return 0;
}
}
/******************************************************************/
/************************ DAEMON FUNCTIONS ************************/
/******************************************************************/
int daemon_init(void)
{
pid_t pid = -1;
int pidno = 0;
int lockfile = 0;
int val = 0;
char buf[256] = { 0 };
char *homedir = NULL;
char *cp = NULL;
struct flock lock;
#ifdef RLIMIT_CORE
struct rlimit limit;
#endif
/* change working directory. scuttle home if we're dumping core */
if (daemon_dumps_core == TRUE) {
homedir = getenv("HOME");
if (homedir && *homedir) {
chdir(homedir);
}
else if (log_file && *log_file) {
homedir = strdup(log_file);
cp = strrchr(homedir, '/');
if (cp) {
*cp = '\0';
}
else {
strcpy(homedir, "/");
}
chdir(homedir);
free(homedir);
} else {
chdir("/");
}
}
umask(S_IWGRP | S_IWOTH);
/* close existing stdin, stdout, stderr */
close(0);
close(1);
close(2);
/* THIS HAS TO BE DONE TO AVOID PROBLEMS WITH STDERR BEING REDIRECTED TO SERVICE MESSAGE PIPE! */
/* re-open stdin, stdout, stderr with known values */
open("/dev/null", O_RDONLY);
open("/dev/null", O_WRONLY);
open("/dev/null", O_WRONLY);
lockfile = open(lock_file, O_RDWR | O_CREAT, S_IWUSR | S_IRUSR | S_IRGRP | S_IROTH);
if (lockfile < 0) {
logit(NSLOG_RUNTIME_ERROR, TRUE,
"Failed to obtain lock on file %s: %s\n", lock_file, strerror(errno));
logit(NSLOG_PROCESS_INFO | NSLOG_RUNTIME_ERROR, TRUE,
"Bailing out due to errors encountered while attempting to daemonize... (PID=%d)", (int)getpid());
cleanup();
exit(ERROR);
}
/* see if we can read the contents of the lockfile */
val = read(lockfile, buf, (size_t)10);
if (val < 0) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Lockfile exists but cannot be read");
cleanup();
exit(ERROR);
}
/* we read something - check the PID */
if (val > 0) {
val = sscanf(buf, "%d", &pidno);
if (val < 1) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Lockfile '%s' does not contain a valid PID (%s)", lock_file, buf);
cleanup();
exit(ERROR);
}
}
pid = (pid_t)pidno;
if (val == 1) {
/* check for SIGHUP */
if (pid == getpid()) {
close(lockfile);
return OK;
}
/* send a signal to see if pid alive */
val = kill(pid, 0);
/* is this process alive? */
if (val == 0) {
logit(NSLOG_RUNTIME_ERROR, TRUE,
"Lockfile '%s' contains PID of running process (%d)", lock_file, pidno);
cleanup();
exit(ERROR);
}
}
/* exit on errors... */
pid = fork();
if (pid < 0) {
return(ERROR);
}
/* parent process goes away.. */
else if (pid != 0) {
iobroker_destroy(nagios_iobs, IOBROKER_CLOSE_SOCKETS);
cleanup();
cleanup_performance_data();
cleanup_downtime_data();
my_free(lock_file);
my_free(config_file);
my_free(config_file_dir);
my_free(nagios_binary_path);
exit(OK);
}
/* child continues... */
/* child becomes session leader... */
setsid();
/* place a file lock on the lock file */
lock.l_type = F_WRLCK;
lock.l_start = 0;
lock.l_whence = SEEK_SET;
lock.l_len = 0;
val = fcntl(lockfile, F_SETLK, &lock);
if (val < 0) {
if (errno == EACCES || errno == EAGAIN) {
fcntl(lockfile, F_GETLK, &lock);
logit(NSLOG_RUNTIME_ERROR, TRUE,
"Lockfile '%s' looks like its already held by another instance of Nagios (PID %d). Bailing out...",
lock_file, (int)lock.l_pid);
}
else {
logit(NSLOG_RUNTIME_ERROR, TRUE,
"Cannot lock lockfile '%s': %s. Bailing out...",
lock_file, strerror(errno));
}
cleanup();
exit(ERROR);
}
/* prevent daemon from dumping a core file... */
#ifdef RLIMIT_CORE
if(daemon_dumps_core == FALSE) {
getrlimit(RLIMIT_CORE, &limit);
limit.rlim_cur = 0;
setrlimit(RLIMIT_CORE, &limit);
}
#endif
/* write PID to lockfile... */
lseek(lockfile, 0, SEEK_SET);
ftruncate(lockfile, 0);
sprintf(buf, "%d\n", (int)getpid());
write(lockfile, buf, strlen(buf));
/* make sure lock file stays open while program is executing... */
val = fcntl(lockfile, F_GETFD, 0);
val |= FD_CLOEXEC;
fcntl(lockfile, F_SETFD, val);
return OK;
}
/******************************************************************/
/*********************** SECURITY FUNCTIONS ***********************/
/******************************************************************/
/* drops privileges */
int drop_privileges(char *user, char *group) {
uid_t uid = -1;
gid_t gid = -1;
struct group *grp = NULL;
struct passwd *pw = NULL;
int result = OK;
#ifdef HAVE_SYS_RESOURCE_H
struct rlimit rl;
#endif
/* only drop privileges if we're running as root, so we don't interfere with being debugged while running as some random user */
if(getuid() != 0)
return OK;
/* set effective group ID */
if(group != NULL) {
/* see if this is a group name */
if(strspn(group, "0123456789") < strlen(group)) {
grp = (struct group *)getgrnam(group);
if(grp != NULL)
gid = (gid_t)(grp->gr_gid);
else
logit(NSLOG_RUNTIME_WARNING, TRUE, "Warning: Could not get group entry for '%s'", group);
}
/* else we were passed the GID */
else
gid = (gid_t)atoi(group);
}
/* set effective user ID */
if(user != NULL) {
/* see if this is a user name */
if(strspn(user, "0123456789") < strlen(user)) {
pw = (struct passwd *)getpwnam(user);
if(pw != NULL)
uid = (uid_t)(pw->pw_uid);
else
logit(NSLOG_RUNTIME_WARNING, TRUE, "Warning: Could not get passwd entry for '%s'", user);
}
/* else we were passed the UID */
else
uid = (uid_t)atoi(user);
}
/* now that we know what to change to, we fix log file permissions */
fix_log_file_owner(uid, gid);
/* set effective group ID if other than current EGID */
if(gid != getegid()) {
if(setgid(gid) == -1) {
logit(NSLOG_RUNTIME_WARNING, TRUE, "Warning: Could not set effective GID=%d", (int)gid);
result = ERROR;
}
}
#ifdef HAVE_INITGROUPS
if(uid != geteuid()) {
/* initialize supplementary groups */
if(initgroups(user, gid) == -1) {
if(errno == EPERM)
logit(NSLOG_RUNTIME_WARNING, TRUE, "Warning: Unable to change supplementary groups using initgroups() -- I hope you know what you're doing");
else {
logit(NSLOG_RUNTIME_WARNING, TRUE, "Warning: Possibly root user failed dropping privileges with initgroups()");
return ERROR;
}
}
}
#endif
#ifdef HAVE_SYS_RESOURCE_H
if (result == OK && daemon_dumps_core == TRUE) {
rl.rlim_cur = RLIM_INFINITY;
rl.rlim_max = RLIM_INFINITY;
setrlimit(RLIMIT_CORE, &rl);
}
#endif
if(setuid(uid) == -1) {
logit(NSLOG_RUNTIME_WARNING, TRUE, "Warning: Could not set effective UID=%d", (int)uid);
result = ERROR;
}
#ifdef HAVE_SYS_PRCTL_H
if (result == OK && daemon_dumps_core == TRUE) {
if (prctl(PR_SET_DUMPABLE, 1))
logit(NSLOG_RUNTIME_WARNING, TRUE, "Warning: Call to prctl(PR_SET_DUMPABLE, 1) failed with error %d", errno);
}
#endif
return result;
}
/******************************************************************/
/************************* IPC FUNCTIONS **************************/
/******************************************************************/
/* processes files in the check result queue directory */
int process_check_result_queue(char *dirname) {
char file[MAX_FILENAME_LENGTH];
DIR *dirp = NULL;
struct dirent *dirfile = NULL;
register int x = 0;
struct stat stat_buf;
struct stat ok_stat_buf;
char *temp_buffer = NULL;
int result = OK, check_result_files = 0;
time_t start;
/* make sure we have what we need */
if(dirname == NULL) {
logit(NSLOG_CONFIG_ERROR, TRUE, "Error: No check result queue directory specified.\n");
return ERROR;
}
/* open the directory for reading */
if((dirp = opendir(dirname)) == NULL) {
logit(NSLOG_CONFIG_ERROR, TRUE, "Error: Could not open check result queue directory '%s' for reading.\n", dirname);
return ERROR;
}
log_debug_info(DEBUGL_CHECKS, 0, "Starting to read check result queue '%s'...\n", dirname);
start = time(NULL);
/* process all files in the directory... */
while((dirfile = readdir(dirp)) != NULL) {
/* bail out if we encountered a signal */
if (sigshutdown == TRUE || sigrestart == TRUE) {
log_debug_info(DEBUGL_CHECKS, 0, "Breaking out of check result reaper: signal encountered\n");
break;
}
/* break out if we've been here too long */
if (start + max_check_reaper_time < time(NULL)) {
log_debug_info(DEBUGL_CHECKS, 0, "Breaking out of check result reaper: max time (%ds) exceeded\n", max_check_reaper_time);
break;
}
/* create /path/to/file */
snprintf(file, sizeof(file), "%s/%s", dirname, dirfile->d_name);
file[sizeof(file) - 1] = '\x0';
/* process this if it's a check result file...
remember it needs to be in the format of
filename = cXXXXXX
where X is any integer
there must also be a filename present
okfile = cXXXXXX.ok
where the XXXXXX is the same as in the filename */
x = strlen(dirfile->d_name);
if(x == 7 && dirfile->d_name[0] == 'c') {
if(stat(file, &stat_buf) == -1) {
logit(NSLOG_RUNTIME_WARNING, TRUE, "Warning: Could not stat() check result file '%s'.\n", file);
continue;
}
/* we only care about real files */
if (!S_ISREG(stat_buf.st_mode))
continue;
/* at this point we have a regular file... */
/* if the file is too old, we delete it */
if (stat_buf.st_mtime + max_check_result_file_age < time(NULL)) {
if (delete_check_result_file(file) != OK
&& delete_check_result_file(dirfile->d_name) != OK)
logit(NSLOG_RUNTIME_WARNING, TRUE, "Error: Unable to delete '%s' or '%s'!", file, dirfile->d_name);
continue;
}
/* can we find the associated ok-to-go file ? */
asprintf(&temp_buffer, "%s.ok", file);
result = stat(temp_buffer, &ok_stat_buf);
my_free(temp_buffer);
if(result == -1)
continue;
/* process the file */
result = process_check_result_file(file);
/* break out if we encountered an error */
if(result == ERROR) {
log_debug_info(DEBUGL_CHECKS, 0, "Encountered an error processing the check result file\n");
break;
}
check_result_files++;
}
}
closedir(dirp);
log_debug_info(DEBUGL_CHECKS, 0, "Finished reaping %d check results\n", check_result_files);
return check_result_files;
}
int process_check_result(check_result *cr)
{
const char *source_name;
if (!cr)
return ERROR;
source_name = check_result_source(cr);
if (cr->object_check_type == SERVICE_CHECK) {
service *svc;
svc = find_service(cr->host_name, cr->service_description);
if (!svc) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Got check result for service '%s' on host '%s'. Unable to find service\n",
cr->service_description, cr->host_name);
return ERROR;
}
log_debug_info(DEBUGL_CHECKS, 2, "Processing check result for service '%s' on host '%s'\n",
svc->description, svc->host_name);
svc->check_source = source_name;
return handle_async_service_check_result(svc, cr);
}
if (cr->object_check_type == HOST_CHECK) {
host *hst;
hst = find_host(cr->host_name);
if (!hst) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Got host checkresult for '%s', but no such host can be found\n", cr->host_name);
return ERROR;
}
log_debug_info(DEBUGL_CHECKS, 2, "Processing check result for host '%s'\n", hst->name);
hst->check_source = source_name;
return handle_async_host_check_result(hst, cr);
}
/* We should never end up here */
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Unknown object check type for checkresult: %d; (host_name: %s; service_description: %s)\n",
cr->object_check_type,
cr->host_name ? cr->host_name : "(null)",
cr->service_description ? cr->service_description : "(null)");
return ERROR;
}
/* Unescapes newlines in a string. Declared here for now as it's not used
* elsewhere. */
/* static char *unescape_check_result_file_output(char*); */
/* reads check result(s) from a file */
int process_check_result_file(char *fname)
{
mmapfile *thefile = NULL;
char *input = NULL;
char *var = NULL;
char *val = NULL;
char *vartok = NULL;
char *valtok = NULL;
char *v1 = NULL, *v2 = NULL;
time_t current_time;
check_result cr;
if (fname == NULL) {
return ERROR;
}
init_check_result(&cr);
cr.engine = &nagios_spool_check_engine;
time(&current_time);
log_debug_info(DEBUGL_CHECKS, 1, "Processing check result file: '%s'\n", fname);
/* open the file for reading */
thefile = mmap_fopen(fname);
if (thefile == NULL) {
/* try removing the file - zero length files can't be mmap()'ed, so it might exist */
log_debug_info(DEBUGL_CHECKS, 1, "Failed to open check result file for reading: '%s'\n", fname);
delete_check_result_file(fname);
return ERROR;
}
/* read in all lines from the file */
while(1) {
/* free memory */
my_free(input);
my_free(var);
my_free(val);
/* read the next line */
input = mmap_fgets_multiline(thefile);
if (input == NULL) {
break;
}
/* skip comments */
if (input[0] == '#') {
continue;
}
/* empty line indicates end of record */
else if (input[0] == '\n') {
/* do we have the minimum amount of data? */
if (cr.host_name != NULL && cr.output != NULL) {
/* process the check result */
process_check_result(&cr);
}
/* cleanse for next check result */
free_check_result(&cr);
init_check_result(&cr);
cr.output_file = fname;
}
vartok = my_strtok_with_free(input, "=", FALSE);
if (vartok == NULL) {
continue;
}
valtok = my_strtok_with_free(NULL, "\n", FALSE);
if (valtok == NULL) {
vartok = my_strtok_with_free(NULL, NULL, TRUE);
continue;
}
/* clean up some memory before we go any further */
var = strdup(vartok);
val = strdup(valtok);
vartok = my_strtok_with_free(NULL, NULL, TRUE);
log_debug_info(DEBUGL_CHECKS, 2, " * %25s: %s\n", var, val);
/* found the file time */
if (!strcmp(var, "file_time")) {
/* file is too old - ignore check results it contains and delete it */
/* this will only work as intended if file_time comes before check results */
if ((max_check_result_file_age > 0)
&& (current_time - (strtoul(val, NULL, 0)) > max_check_result_file_age)) {
log_debug_info(DEBUGL_CHECKS, 1,
"Skipping check_result because file_time is %s and max cr file age is %lu",
val, max_check_result_file_age);
break;
}
}
/* else we have check result data */
else {
if (!strcmp(var, "host_name")) {
cr.host_name = (char *)strdup(val);
}
else if (!strcmp(var, "service_description")) {
cr.service_description = (char *)strdup(val);
cr.object_check_type = SERVICE_CHECK;
}
else if (!strcmp(var, "check_type")) {
cr.check_type = atoi(val);
}
else if (!strcmp(var, "check_options")) {
cr.check_options = atoi(val);
}
else if (!strcmp(var, "scheduled_check")) {
cr.scheduled_check = atoi(val);
}
else if (!strcmp(var, "reschedule_check")) {
cr.reschedule_check = atoi(val);
}
else if (!strcmp(var, "latency")) {
cr.latency = strtod(val, NULL);
}
else if (!strcmp(var, "start_time") || !strcmp(var, "finish_time")) {
v1 = strtok(val, ".");
if (v1 == NULL) {
continue;
}
v2 = strtok(NULL, "\n");
if (v2 == NULL) {
continue;
}
if (!strcmp(var, "start_time")) {
cr.start_time.tv_sec = strtoul(v1, NULL, 0);
cr.start_time.tv_usec = strtoul(v2, NULL, 0);
}
else {
cr.finish_time.tv_sec = strtoul(v1, NULL, 0);
cr.finish_time.tv_usec = strtoul(v2, NULL, 0);
}
}
else if (!strcmp(var, "early_timeout")) {
cr.early_timeout = atoi(val);
}
else if (!strcmp(var, "exited_ok")) {
cr.exited_ok = atoi(val);
}
else if (!strcmp(var, "return_code")) {
cr.return_code = atoi(val);
}
else if (!strcmp(var, "output")) {
/* Interpolate "\\\\" and "\\n" escape sequences to the literal
* characters they represent. This converts from the single line
* format used to store the output in a checkresult file, to the
* newline delimited format we use internally. By converting as
* soon as possible after reading from the file we don't have
* to worry about two different representations later. */
cr.output = unescape_check_result_output(val);
}
}
}
my_free(var);
my_free(val);
log_debug_info(DEBUGL_CHECKS, 2, " **************\n");
/* do we have the minimum amount of data? */
if (cr.host_name != NULL && cr.output != NULL) {
/* process check result */
process_check_result(&cr);
}
else {
/* log a debug message */
log_debug_info(DEBUGL_CHECKS, 1, "Minimum amount of data not present; Skipped check result file: '%s'\n", fname);
}
free_check_result(&cr);
/* free memory and close file */
my_free(input);
mmap_fclose(thefile);
/* delete the file (as well its ok-to-go file) */
delete_check_result_file(fname);
return OK;
}
/* deletes as check result file, as well as its ok-to-go file */
int delete_check_result_file(char *fname)
{
char *temp_buffer = NULL;
int result = OK;
/* delete the result file */
result = unlink(fname);
/* delete the ok-to-go file */
asprintf(&temp_buffer, "%s.ok", fname);
result |= unlink(temp_buffer);
my_free(temp_buffer);
return result;
}
/* initializes a host/service check result */
int init_check_result(check_result *info)
{
if (info == NULL) {
return ERROR;
}
/* reset vars */
info->object_check_type = HOST_CHECK;
info->host_name = NULL;
info->service_description = NULL;
info->check_type = CHECK_TYPE_ACTIVE;
info->check_options = CHECK_OPTION_NONE;
info->scheduled_check = FALSE;
info->reschedule_check = FALSE;
info->output_file_fp = NULL;
info->latency = 0.0;
info->start_time.tv_sec = 0;
info->start_time.tv_usec = 0;
info->finish_time.tv_sec = 0;
info->finish_time.tv_usec = 0;
info->early_timeout = FALSE;
info->exited_ok = TRUE;
info->return_code = 0;
info->output = NULL;
info->source = NULL;
info->engine = NULL;
return OK;
}
/* frees memory associated with a host/service check result */
int free_check_result(check_result *info)
{
if (info == NULL) {
return OK;
}
my_free(info->host_name);
my_free(info->service_description);
my_free(info->output);
return OK;
}
/******************************************************************/
/************************ STRING FUNCTIONS ************************/
/******************************************************************/
/* gets the next string from a buffer in memory - strings are terminated by newlines, which are removed */
char *get_next_string_from_buf(char *buf, int *start_index, int bufsize) {
char *sptr = NULL;
const char *nl = "\n";
int x;
if(buf == NULL || start_index == NULL)
return NULL;
if(bufsize < 0)
return NULL;
if(*start_index >= (bufsize - 1))
return NULL;
sptr = buf + *start_index;
/* end of buffer */
if(sptr[0] == '\x0')
return NULL;
x = strcspn(sptr, nl);
sptr[x] = '\x0';
*start_index += x + 1;
return sptr;
}
/* determines whether or not an object name (host, service, etc) contains illegal characters */
int contains_illegal_object_chars(char *name) {
register int x = 0;
register int y = 0;
if(name == NULL || illegal_object_chars == NULL)
return FALSE;
x = (int)strlen(name) - 1;
for(; x >= 0; x--) {
/* illegal user-specified characters */
if(illegal_object_chars != NULL)
for(y = 0; illegal_object_chars[y]; y++)
if(name[x] == illegal_object_chars[y])
return TRUE;
}
return FALSE;
}
/* Escapes newlines in a string. */
char *escape_newlines(char *rawbuf) {
char *newbuf = NULL;
int x;
int y;
if (rawbuf == NULL)
return NULL;
/* Count the escapes we need to make. */
for (x = 0, y = 0; rawbuf[x]; x++) {
if (rawbuf[x] == '\\' || rawbuf[x] == '\n')
y++;
}
/* Just duplicate the string if we have nothing to escape. */
if (y == 0)
return strdup(rawbuf);
/* Allocate memory for the new string with escapes. */
if ((newbuf = malloc(x + y + 1)) == NULL)
return NULL;
for (x = 0, y = 0; rawbuf[x]; x++) {
/* Escape backslashes. */
if (rawbuf[x] == '\\') {
newbuf[y++] = '\\';
newbuf[y++] = '\\';
}
/* Escape newlines. */
else if (rawbuf[x] == '\n') {
newbuf[y++] = '\\';
newbuf[y++] = 'n';
}
else
newbuf[y++] = rawbuf[x];
}
newbuf[y] = '\0';
return newbuf;
}
/* Unescapes newlines (and backslashes) in a string. */
char *unescape_check_result_output(const char *rawbuf) {
char *newbuf = NULL;
int x;
int y;
if (rawbuf == NULL)
return NULL;
/* Count the replacements we need to make. */
for (x = 0, y = 0; rawbuf[x]; x++) {
if (rawbuf[x] == '\\' && (rawbuf[x + 1] == '\\' || rawbuf[x + 1] == 'n'))
x++, y++; /* Consume one more char for each replacement. */
}
/* Just duplicate the string if we have nothing to replace. */
if (y == 0)
return strdup(rawbuf);
/* Allocate memory for the new string, with our escape sequences replaced. */
if ((newbuf = malloc(x - y + 1)) == NULL)
return NULL;
for (x = 0, y = 0; rawbuf[x]; x++) {
/* Unescape backslashes. */
if (rawbuf[x] == '\\' && rawbuf[x + 1] == '\\') {
x++;
newbuf[y++] = '\\';
}
/* Unescape newlines. */
else if (rawbuf[x] == '\\' && rawbuf[x + 1] == 'n') {
x++;
newbuf[y++] = '\n';
}
else
newbuf[y++] = rawbuf[x];
}
newbuf[y] = '\0';
return newbuf;
}
/* compares strings */
int compare_strings(char *val1a, char *val2a) {
/* use the compare_hashdata() function */
return compare_hashdata(val1a, NULL, val2a, NULL);
}
/******************************************************************/
/************************* FILE FUNCTIONS *************************/
/******************************************************************/
/* renames a file - works across filesystems (Mike Wiacek) */
int my_rename(char *source, char *dest) {
int rename_result = 0;
/* make sure we have something */
if(source == NULL || dest == NULL)
return -1;
/* first see if we can rename file with standard function */
rename_result = rename(source, dest);
/* handle any errors... */
if(rename_result == -1) {
/* an error occurred because the source and dest files are on different filesystems */
if(errno == EXDEV) {
/* try copying the file */
if(my_fcopy(source, dest) == ERROR) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Unable to rename file '%s' to '%s': %s\n", source, dest, strerror(errno));
return -1;
}
/* delete the original file */
unlink(source);
/* reset result since we successfully copied file */
rename_result = 0;
}
/* some other error occurred */
else {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Unable to rename file '%s' to '%s': %s\n", source, dest, strerror(errno));
return rename_result;
}
}
return rename_result;
}
/*
* copy a file from the path at source to the already opened
* destination file dest.
* This is handy when creating tempfiles with mkstemp()
*/
int my_fdcopy(char *source, char *dest, int dest_fd) {
int source_fd, rd_result = 0, wr_result = 0;
int tot_written = 0, tot_read = 0, buf_size = 0;
struct stat st;
char *buf;
/* open source file for reading */
if((source_fd = open(source, O_RDONLY, 0644)) < 0) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Unable to open file '%s' for reading: %s\n", source, strerror(errno));
return ERROR;
}
/*
* find out how large the source-file is so we can be sure
* we've written all of it
*/
if(fstat(source_fd, &st) < 0) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Unable to stat source file '%s' for my_fcopy(): %s\n", source, strerror(errno));
close(source_fd);
return ERROR;
}
/*
* If the file is huge, read it and write it in chunks.
* This value (128K) is the result of "pick-one-at-random"
* with some minimal testing and may not be optimal for all
* hardware setups, but it should work ok for most. It's
* faster than 1K buffers and 1M buffers, so change at your
* own peril. Note that it's useful to make it fit in the L2
* cache, so larger isn't necessarily better.
*/
buf_size = st.st_size > 128 << 10 ? 128 << 10 : st.st_size;
buf = malloc(buf_size);
if(!buf) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Unable to malloc(%d) bytes: %s\n", buf_size, strerror(errno));
close(source_fd);
return ERROR;
}
/* most of the times, this loop will be gone through once */
while(tot_written < st.st_size) {
int loop_wr = 0;
rd_result = read(source_fd, buf, buf_size);
if(rd_result < 0) {
if(errno == EAGAIN || errno == EINTR)
continue;
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: my_fcopy() failed to read from '%s': %s\n", source, strerror(errno));
break;
}
tot_read += rd_result;
while(loop_wr < rd_result) {
wr_result = write(dest_fd, buf + loop_wr, rd_result - loop_wr);
if(wr_result < 0) {
if(errno == EAGAIN || errno == EINTR)
continue;
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: my_fcopy() failed to write to '%s': %s\n", dest, strerror(errno));
break;
}
loop_wr += wr_result;
}
if(wr_result < 0)
break;
tot_written += loop_wr;
}
/*
* clean up irregardless of how things went. dest_fd comes from
* our caller, so we mustn't close it.
*/
close(source_fd);
free(buf);
if(rd_result < 0 || wr_result < 0) {
/* don't leave half-written files around */
unlink(dest);
return ERROR;
}
return OK;
}
/* copies a file */
int my_fcopy(char *source, char *dest) {
int dest_fd, result;
/* make sure we have something */
if(source == NULL || dest == NULL)
return ERROR;
/* unlink destination file first (not doing so can cause problems on network file systems like CIFS) */
unlink(dest);
/* open destination file for writing */
if((dest_fd = open(dest, O_WRONLY | O_TRUNC | O_CREAT | O_APPEND, 0644)) < 0) {
logit(NSLOG_RUNTIME_ERROR, TRUE, "Error: Unable to open file '%s' for writing: %s\n", dest, strerror(errno));
return ERROR;
}
result = my_fdcopy(source, dest, dest_fd);
close(dest_fd);
return result;
}
/******************************************************************/
/******************** DYNAMIC BUFFER FUNCTIONS ********************/
/******************************************************************/
/* initializes a dynamic buffer */
int dbuf_init(dbuf *db, int chunk_size) {
if(db == NULL)
return ERROR;
db->buf = NULL;
db->used_size = 0L;
db->allocated_size = 0L;
db->chunk_size = chunk_size;
return OK;
}
/* frees a dynamic buffer */
int dbuf_free(dbuf *db) {
if(db == NULL)
return ERROR;
if(db->buf != NULL)
my_free(db->buf);
db->buf = NULL;
db->used_size = 0L;
db->allocated_size = 0L;
return OK;
}
/* dynamically expands a string */
int dbuf_strcat(dbuf *db, const char *buf) {
char *newbuf = NULL;
unsigned long buflen = 0L;
unsigned long new_size = 0L;
unsigned long memory_needed = 0L;
if(db == NULL || buf == NULL)
return ERROR;
/* how much memory should we allocate (if any)? */
buflen = strlen(buf);
new_size = db->used_size + buflen + 1;
/* we need more memory */
if(db->allocated_size < new_size) {
memory_needed = ((ceil(new_size / db->chunk_size) + 1) * db->chunk_size);
/* allocate memory to store old and new string */
if((newbuf = (char *)realloc((void *)db->buf, (size_t)memory_needed)) == NULL)
return ERROR;
/* update buffer pointer */
db->buf = newbuf;
/* update allocated size */
db->allocated_size = memory_needed;
/* terminate buffer */
db->buf[db->used_size] = '\x0';
}
/* append the new string */
strcat(db->buf, buf);
/* update size allocated */
db->used_size += buflen;
return OK;
}
/******************************************************************/
/********************** CHECK STATS FUNCTIONS *********************/
/******************************************************************/
/* initialize check statistics data structures */
int init_check_stats(void) {
int x = 0;
int y = 0;
for(x = 0; x < MAX_CHECK_STATS_TYPES; x++) {
check_statistics[x].current_bucket = 0;
for(y = 0; y < CHECK_STATS_BUCKETS; y++)
check_statistics[x].bucket[y] = 0;
check_statistics[x].overflow_bucket = 0;
for(y = 0; y < 3; y++)
check_statistics[x].minute_stats[y] = 0;
check_statistics[x].last_update = (time_t)0L;
}
return OK;
}
/* records stats for a given type of check */
int update_check_stats(int check_type, time_t check_time) {
time_t current_time;
unsigned long minutes = 0L;
int new_current_bucket = 0;
int this_bucket = 0;
int x = 0;
if(check_type < 0 || check_type >= MAX_CHECK_STATS_TYPES)
return ERROR;
time(&current_time);
if((unsigned long)check_time == 0L) {
#ifdef DEBUG_CHECK_STATS
printf("TYPE[%d] CHECK TIME==0!\n", check_type);
#endif
check_time = current_time;
}
/* do some sanity checks on the age of the stats data before we start... */
/* get the new current bucket number */
minutes = ((unsigned long)check_time - (unsigned long)program_start) / 60;
new_current_bucket = minutes % CHECK_STATS_BUCKETS;
/* its been more than 15 minutes since stats were updated, so clear the stats */
if((((unsigned long)current_time - (unsigned long)check_statistics[check_type].last_update) / 60) > CHECK_STATS_BUCKETS) {
for(x = 0; x < CHECK_STATS_BUCKETS; x++)
check_statistics[check_type].bucket[x] = 0;
check_statistics[check_type].overflow_bucket = 0;
#ifdef DEBUG_CHECK_STATS
printf("CLEARING ALL: TYPE[%d], CURRENT=%llu, LASTUPDATE=%llu\n", check_type, (unsigned long long)current_time, (unsigned long long)check_statistics[check_type].last_update);
#endif
}
/* different current bucket number than last time */
else if(new_current_bucket != check_statistics[check_type].current_bucket) {
/* clear stats in buckets between last current bucket and new current bucket - stats haven't been updated in a while */
for(x = check_statistics[check_type].current_bucket; x < (CHECK_STATS_BUCKETS * 2); x++) {
this_bucket = (x + CHECK_STATS_BUCKETS + 1) % CHECK_STATS_BUCKETS;
if(this_bucket == new_current_bucket)
break;
#ifdef DEBUG_CHECK_STATS
printf("CLEARING BUCKET %d, (NEW=%d, OLD=%d)\n", this_bucket, new_current_bucket, check_statistics[check_type].current_bucket);
#endif
/* clear old bucket value */
check_statistics[check_type].bucket[this_bucket] = 0;
}
/* update the current bucket number, push old value to overflow bucket */
check_statistics[check_type].overflow_bucket = check_statistics[check_type].bucket[new_current_bucket];
check_statistics[check_type].current_bucket = new_current_bucket;
check_statistics[check_type].bucket[new_current_bucket] = 0;
}
#ifdef DEBUG_CHECK_STATS
else
printf("NO CLEARING NEEDED\n");
#endif
/* increment the value of the current bucket */
check_statistics[check_type].bucket[new_current_bucket]++;
#ifdef DEBUG_CHECK_STATS
printf("TYPE[%d].BUCKET[%d]=%d\n", check_type, new_current_bucket, check_statistics[check_type].bucket[new_current_bucket]);
printf(" ");
for(x = 0; x < CHECK_STATS_BUCKETS; x++)
printf("[%d] ", check_statistics[check_type].bucket[x]);
printf(" (%d)\n", check_statistics[check_type].overflow_bucket);
#endif
/* record last update time */
check_statistics[check_type].last_update = current_time;
return OK;
}
/* generate 1/5/15 minute stats for a given type of check */
int generate_check_stats(void) {
time_t current_time;
int x = 0;
int new_current_bucket = 0;
int this_bucket = 0;
int last_bucket = 0;
int this_bucket_value = 0;
int last_bucket_value = 0;
int bucket_value = 0;
int seconds = 0;
int minutes = 0;
int check_type = 0;
float this_bucket_weight = 0.0;
float last_bucket_weight = 0.0;
time(&current_time);
/* do some sanity checks on the age of the stats data before we start... */
/* get the new current bucket number */
minutes = ((unsigned long)current_time - (unsigned long)program_start) / 60;
new_current_bucket = minutes % CHECK_STATS_BUCKETS;
for(check_type = 0; check_type < MAX_CHECK_STATS_TYPES; check_type++) {
/* its been more than 15 minutes since stats were updated, so clear the stats */
if((((unsigned long)current_time - (unsigned long)check_statistics[check_type].last_update) / 60) > CHECK_STATS_BUCKETS) {
for(x = 0; x < CHECK_STATS_BUCKETS; x++)
check_statistics[check_type].bucket[x] = 0;
check_statistics[check_type].overflow_bucket = 0;
#ifdef DEBUG_CHECK_STATS
printf("GEN CLEARING ALL: TYPE[%d], CURRENT=%llu, LASTUPDATE=%llu\n", check_type, (unsigned long long)current_time, (unsigned long long)check_statistics[check_type].last_update);
#endif
}
/* different current bucket number than last time */
else if(new_current_bucket != check_statistics[check_type].current_bucket) {
/* clear stats in buckets between last current bucket and new current bucket - stats haven't been updated in a while */
for(x = check_statistics[check_type].current_bucket; x < (CHECK_STATS_BUCKETS * 2); x++) {
this_bucket = (x + CHECK_STATS_BUCKETS + 1) % CHECK_STATS_BUCKETS;
if(this_bucket == new_current_bucket)
break;
#ifdef DEBUG_CHECK_STATS
printf("GEN CLEARING BUCKET %d, (NEW=%d, OLD=%d), CURRENT=%llu, LASTUPDATE=%llu\n", this_bucket, new_current_bucket, check_statistics[check_type].current_bucket, (unsigned long long)current_time, (unsigned long long)check_statistics[check_type].last_update);
#endif
/* clear old bucket value */
check_statistics[check_type].bucket[this_bucket] = 0;
}
/* update the current bucket number, push old value to overflow bucket */
check_statistics[check_type].overflow_bucket = check_statistics[check_type].bucket[new_current_bucket];
check_statistics[check_type].current_bucket = new_current_bucket;
check_statistics[check_type].bucket[new_current_bucket] = 0;
}
#ifdef DEBUG_CHECK_STATS
else
printf("GEN NO CLEARING NEEDED: TYPE[%d], CURRENT=%llu, LASTUPDATE=%llu\n", check_type, (unsigned long long)current_time, (unsigned long long)check_statistics[check_type].last_update);
#endif
/* update last check time */
check_statistics[check_type].last_update = current_time;
}
/* determine weights to use for this/last buckets */
seconds = ((unsigned long)current_time - (unsigned long)program_start) % 60;
this_bucket_weight = (seconds / 60.0);
last_bucket_weight = ((60 - seconds) / 60.0);
/* update statistics for all check types */
for(check_type = 0; check_type < MAX_CHECK_STATS_TYPES; check_type++) {
/* clear the old statistics */
for(x = 0; x < 3; x++)
check_statistics[check_type].minute_stats[x] = 0;
/* loop through each bucket */
for(x = 0; x < CHECK_STATS_BUCKETS; x++) {
/* which buckets should we use for this/last bucket? */
this_bucket = (check_statistics[check_type].current_bucket + CHECK_STATS_BUCKETS - x) % CHECK_STATS_BUCKETS;
last_bucket = (this_bucket + CHECK_STATS_BUCKETS - 1) % CHECK_STATS_BUCKETS;
/* raw/unweighted value for this bucket */
this_bucket_value = check_statistics[check_type].bucket[this_bucket];
/* raw/unweighted value for last bucket - use overflow bucket if last bucket is current bucket */
if(last_bucket == check_statistics[check_type].current_bucket)
last_bucket_value = check_statistics[check_type].overflow_bucket;
else
last_bucket_value = check_statistics[check_type].bucket[last_bucket];
/* determine value by weighting this/last buckets... */
/* if this is the current bucket, use its full value + weighted % of last bucket */
if(x == 0) {
bucket_value = (int)(this_bucket_value + floor(last_bucket_value * last_bucket_weight));
}
/* otherwise use weighted % of this and last bucket */
else {
bucket_value = (int)(ceil(this_bucket_value * this_bucket_weight) + floor(last_bucket_value * last_bucket_weight));
}
/* 1 minute stats */
if(x == 0)
check_statistics[check_type].minute_stats[0] = bucket_value;
/* 5 minute stats */
if(x < 5)
check_statistics[check_type].minute_stats[1] += bucket_value;
/* 15 minute stats */
if(x < 15)
check_statistics[check_type].minute_stats[2] += bucket_value;
#ifdef DEBUG_CHECK_STATS2
printf("X=%d, THIS[%d]=%d, LAST[%d]=%d, 1/5/15=%d,%d,%d L=%d R=%d\n", x, this_bucket, this_bucket_value, last_bucket, last_bucket_value, check_statistics[check_type].minute_stats[0], check_statistics[check_type].minute_stats[1], check_statistics[check_type].minute_stats[2], left_value, right_value);
#endif
/* record last update time */
check_statistics[check_type].last_update = current_time;
}
#ifdef DEBUG_CHECK_STATS
printf("TYPE[%d] 1/5/15 = %d, %d, %d (seconds=%d, this_weight=%f, last_weight=%f)\n", check_type, check_statistics[check_type].minute_stats[0], check_statistics[check_type].minute_stats[1], check_statistics[check_type].minute_stats[2], seconds, this_bucket_weight, last_bucket_weight);
#endif
}
return OK;
}
/******************************************************************/
/************************ UPDATE FUNCTIONS ************************/
/******************************************************************/
/* check for new releases of Nagios */
int check_for_nagios_updates(int force, int reschedule) {
time_t current_time;
int result = OK;
int api_result = OK;
int do_check = TRUE;
time_t next_check = 0L;
unsigned int rand_seed = 0;
int randnum = 0;
time(&current_time);
/*
printf("NOW: %s",ctime(&current_time));
printf("LAST CHECK: %s",ctime(&last_update_check));
*/
/* seed the random generator */
rand_seed = (unsigned int)(current_time + nagios_pid);
srand(rand_seed);
/* generate a (probably) unique ID for this nagios install */
/* the server api currently sees thousands of nagios installs behind single ip addresses, so this help determine if there are really thousands of servers out there, or if some nagios installs are misbehaving */
if(update_uid == 0L)
update_uid = current_time;
/* update checks are disabled */
if(check_for_updates == FALSE)
do_check = FALSE;
/* we checked for updates recently, so don't do it again */
if((current_time - last_update_check) < MINIMUM_UPDATE_CHECK_INTERVAL)
do_check = FALSE;
/* the check is being forced */
if(force == TRUE)
do_check = TRUE;
/* do a check */
if(do_check == TRUE) {
/*printf("RUNNING QUERY...\n");*/
/* query api */
api_result = query_update_api();
}
/* should we reschedule the update check? */
if(reschedule == TRUE) {
/*printf("RESCHEDULING...\n");*/
randnum = rand();
/*
printf("RAND: %d\n",randnum);
printf("RANDMAX: %d\n",RAND_MAX);
printf("UCIW: %d\n",UPDATE_CHECK_INTERVAL_WOBBLE);
printf("MULT: %f\n",(float)randnum/RAND_MAX);
*/
/* we didn't do an update, so calculate next possible update time */
if(do_check == FALSE) {
next_check = last_update_check + BASE_UPDATE_CHECK_INTERVAL;
next_check = next_check + (unsigned long)(((float)randnum / RAND_MAX) * UPDATE_CHECK_INTERVAL_WOBBLE);
}
/* we tried to check for an update */
else {
/* api query was okay */
if(api_result == OK) {
next_check = current_time + BASE_UPDATE_CHECK_INTERVAL;
next_check += (unsigned long)(((float)randnum / RAND_MAX) * UPDATE_CHECK_INTERVAL_WOBBLE);
}
/* query resulted in an error - retry at a shorter interval */
else {
next_check = current_time + BASE_UPDATE_CHECK_RETRY_INTERVAL;
next_check += (unsigned long)(((float)randnum / RAND_MAX) * UPDATE_CHECK_RETRY_INTERVAL_WOBBLE);
}
}
/* make sure next check isn't in the past - if it is, schedule a check in 1 minute */
if(next_check < current_time)
next_check = current_time + 60;
/*printf("NEXT CHECK: %s",ctime(&next_check));*/
/* schedule the next update event */
schedule_new_event(EVENT_CHECK_PROGRAM_UPDATE, TRUE, next_check, FALSE, BASE_UPDATE_CHECK_INTERVAL, NULL, TRUE, NULL, NULL, 0);
}
return result;
}
/* checks for updates at api.nagios.org */
int query_update_api(void) {
const char *api_server = "api.nagios.org";
const char *api_path = "/versioncheck/";
char *api_query = NULL;
char *api_query_opts = NULL;
char *buf = NULL;
char recv_buf[1024];
int report_install = FALSE;
char *ptr = NULL;
int current_line = 0;
int buf_index = 0;
int in_header = TRUE;
char *var = NULL;
char *val = NULL;
int sd = 0;
int send_len = 0;
int recv_len = 0;
int update_check_succeeded = FALSE;
/* report a new install, upgrade, or rollback */
/* Nagios monitors the world and we monitor Nagios taking over the world. :-) */
if(last_update_check == (time_t)0L)
report_install = TRUE;
if(last_program_version == NULL || strcmp(PROGRAM_VERSION, last_program_version))
report_install = TRUE;
if(report_install == TRUE) {
asprintf(&api_query_opts, "&firstcheck=1");
if(last_program_version != NULL) {
char *qopts2 = NULL;
asprintf(&qopts2, "%s&last_version=%s", api_query_opts, last_program_version);
my_free(api_query_opts);
api_query_opts = qopts2;
}
}
/* generate the query */
asprintf(&api_query, "v=1&product=nagios&tinycheck=1&stableonly=1&uid=%lu", update_uid);
if(bare_update_check == FALSE) {
char *api_query2 = NULL;
asprintf(&api_query2, "%s&version=%s%s", api_query, PROGRAM_VERSION, (api_query_opts == NULL) ? "" : api_query_opts);
my_free(api_query);
api_query = api_query2;
}
/* generate the HTTP request */
asprintf(&buf,
"POST %s HTTP/1.0\r\nUser-Agent: Nagios/%s\r\n"
"Connection: close\r\nHost: %s\r\n"
"Content-Type: application/x-www-form-urlencoded\r\n"
"Content-Length: %lu\r\n\r\n%s",
api_path, PROGRAM_VERSION, api_server,
(unsigned long) strlen(api_query), api_query);
if (buf == NULL) {
abort();
}
my_tcp_connect(api_server, 80, &sd, 2);
if(sd > 0) {
/* send request */
send_len = strlen(buf);
my_sendall(sd, buf, &send_len, 2);
/* get response */
recv_len = sizeof(recv_buf);
my_recvall(sd, recv_buf, &recv_len, 2);
recv_buf[sizeof(recv_buf) - 1] = '\x0';
/* close connection */
close(sd);
/* parse the result */
in_header = TRUE;
while((ptr = get_next_string_from_buf(recv_buf, &buf_index, sizeof(recv_buf)))) {
strip(ptr);
current_line++;
if(!strcmp(ptr, "")) {
in_header = FALSE;
continue;
}
if(in_header == TRUE)
continue;
var = strtok(ptr, "=");
val = strtok(NULL, "\n");
if(!strcmp(var, "UPDATE_AVAILABLE")) {
update_available = atoi(val);
/* we were successful */
update_check_succeeded = TRUE;
}
else if(!strcmp(var, "UPDATE_VERSION")) {
if(new_program_version)
my_free(new_program_version);
new_program_version = strdup(val);
}
else if(!strcmp(var, "UPDATE_RELEASEDATE")) {
}
}
}
/* cleanup */
my_free(buf);
my_free(api_query);
my_free(api_query_opts);
/* we were successful! */
if(update_check_succeeded == TRUE) {
time(&last_update_check);
if(last_program_version)
free(last_program_version);
last_program_version = (char *)strdup(PROGRAM_VERSION);
}
return OK;
}
/******************************************************************/
/************************* MISC FUNCTIONS *************************/
/******************************************************************/
/* returns Nagios version */
char *get_program_version(void) {
return (char *)PROGRAM_VERSION;
}
/* returns Nagios modification date */
char *get_program_modification_date(void) {
return (char *)PROGRAM_MODIFICATION_DATE;
}
/******************************************************************/
/*********************** CLEANUP FUNCTIONS ************************/
/******************************************************************/
/* do some cleanup before we exit */
void cleanup(void) {
xodtemplate_free_memory();
#ifdef USE_EVENT_BROKER
/* unload modules */
if(test_scheduling == FALSE && verify_config == FALSE) {
neb_free_callback_list();
neb_unload_all_modules(NEBMODULE_FORCE_UNLOAD, (sigshutdown == TRUE) ? NEBMODULE_NEB_SHUTDOWN : NEBMODULE_NEB_RESTART);
neb_free_module_list();
neb_deinit_modules();
}
#endif
/* free all allocated memory - including macros */
free_memory(get_global_macros());
close_log_file();
return;
}
/* free the memory allocated to the linked lists */
void free_memory(nagios_macros *mac) {
int i;
/* free all allocated memory for the object definitions */
free_object_data();
free_comment_data();
/* free event queue data */
squeue_destroy(nagios_squeue, SQUEUE_FREE_DATA);
nagios_squeue = NULL;
/* free memory for global event handlers */
my_free(global_host_event_handler);
my_free(global_service_event_handler);
/* free any notification list that may have been overlooked */
free_notification_list();
/* free obsessive compulsive commands */
my_free(ocsp_command);
my_free(ochp_command);
my_free(object_cache_file);
my_free(object_precache_file);
/*
* free memory associated with macros.
* It's ok to only free the volatile ones, as the non-volatile
* are always free()'d before assignment if they're set.
* Doing a full free of them here means we'll wipe the constant
* macros when we get a reload or restart request through the
* command pipe, or when we receive a SIGHUP.
*/
clear_volatile_macros_r(mac);
free_macrox_names();
/* free illegal char strings */
my_free(illegal_object_chars);
my_free(illegal_output_chars);
/* free nagios user and group */
my_free(nagios_user);
my_free(nagios_group);
/* free version strings */
my_free(last_program_version);
my_free(new_program_version);
/* free file/path variables */
my_free(debug_file);
my_free(log_file);
mac->x[MACRO_LOGFILE] = NULL; /* assigned from 'log_file' */
my_free(temp_file);
mac->x[MACRO_TEMPFILE] = NULL; /* assigned from temp_file */
my_free(temp_path);
mac->x[MACRO_TEMPPATH] = NULL; /*assigned from temp_path */
my_free(command_file);
mac->x[MACRO_COMMANDFILE] = NULL; /* assigned from command_file */
my_free(check_result_path);
my_free(log_archive_path);
my_free(website_url);
my_free(status_file);
my_free(retention_file);
for (i = 0; i < MAX_USER_MACROS; i++) {
my_free(macro_user[i]);
}
/* these have no other reference */
my_free(mac->x[MACRO_PROCESSSTARTTIME]);
my_free(mac->x[MACRO_EVENTSTARTTIME]);
my_free(mac->x[MACRO_ADMINEMAIL]);
my_free(mac->x[MACRO_ADMINPAGER]);
my_free(mac->x[MACRO_RESOURCEFILE]);
my_free(mac->x[MACRO_OBJECTCACHEFILE]);
my_free(mac->x[MACRO_MAINCONFIGFILE]);
my_free(mac->x[MACRO_RETENTIONDATAFILE]);
my_free(mac->x[MACRO_HOSTPERFDATAFILE]);
my_free(mac->x[MACRO_STATUSDATAFILE]);
my_free(mac->x[MACRO_SERVICEPERFDATAFILE]);
return;
}
/* free a notification list that was created */
void free_notification_list(void) {
notification *temp_notification = NULL;
notification *next_notification = NULL;
temp_notification = notification_list;
while(temp_notification != NULL) {
next_notification = temp_notification->next;
my_free(temp_notification);
temp_notification = next_notification;
}
/* reset notification list pointer */
notification_list = NULL;
return;
}
/* reset all system-wide variables, so when we've receive a SIGHUP we can restart cleanly */
int reset_variables(void) {
/* First free any variables previously set */
my_free(log_file);
my_free(temp_file);
my_free(temp_path);
my_free(check_result_path);
my_free(command_file);
my_free(lock_file);
my_free(log_archive_path);
my_free(debug_file);
my_free(object_cache_file);
my_free(object_precache_file);
my_free(nagios_user);
my_free(nagios_group);
my_free(global_host_event_handler);
my_free(global_service_event_handler);
my_free(ocsp_command);
my_free(ochp_command);
my_free(website_url);
/* Next re-initialize configuration variables */
init_main_cfg_vars(0);
init_shared_cfg_vars(0);
/* Finally, set anything not set in the initialization routines */
log_file = (char *)strdup(DEFAULT_LOG_FILE);
temp_file = (char *)strdup(DEFAULT_TEMP_FILE);
temp_path = (char *)strdup(DEFAULT_TEMP_PATH);
check_result_path = (char *)strdup(DEFAULT_CHECK_RESULT_PATH);
command_file = (char *)strdup(DEFAULT_COMMAND_FILE);
lock_file = (char *)strdup(DEFAULT_LOCK_FILE); /* this is kept across restarts */
log_archive_path = (char *)strdup(DEFAULT_LOG_ARCHIVE_PATH);
debug_file = (char *)strdup(DEFAULT_DEBUG_FILE);
/* init_main_cfg_vars(0) sets object_precache_file = strdup(DEFAULT_PRECACHED_OBJECT_FILE); */
nagios_user = (char *)strdup(DEFAULT_NAGIOS_USER);
nagios_group = (char *)strdup(DEFAULT_NAGIOS_GROUP);
logging_options = NSLOG_RUNTIME_ERROR | NSLOG_RUNTIME_WARNING | NSLOG_VERIFICATION_ERROR | NSLOG_VERIFICATION_WARNING | NSLOG_CONFIG_ERROR | NSLOG_CONFIG_WARNING | NSLOG_PROCESS_INFO | NSLOG_HOST_NOTIFICATION | NSLOG_SERVICE_NOTIFICATION | NSLOG_EVENT_HANDLER | NSLOG_EXTERNAL_COMMAND | NSLOG_PASSIVE_CHECK | NSLOG_HOST_UP | NSLOG_HOST_DOWN | NSLOG_HOST_UNREACHABLE | NSLOG_SERVICE_OK | NSLOG_SERVICE_WARNING | NSLOG_SERVICE_UNKNOWN | NSLOG_SERVICE_CRITICAL | NSLOG_INFO_MESSAGE;
syslog_options = NSLOG_RUNTIME_ERROR | NSLOG_RUNTIME_WARNING | NSLOG_VERIFICATION_ERROR | NSLOG_VERIFICATION_WARNING | NSLOG_CONFIG_ERROR | NSLOG_CONFIG_WARNING | NSLOG_PROCESS_INFO | NSLOG_HOST_NOTIFICATION | NSLOG_SERVICE_NOTIFICATION | NSLOG_EVENT_HANDLER | NSLOG_EXTERNAL_COMMAND | NSLOG_PASSIVE_CHECK | NSLOG_HOST_UP | NSLOG_HOST_DOWN | NSLOG_HOST_UNREACHABLE | NSLOG_SERVICE_OK | NSLOG_SERVICE_WARNING | NSLOG_SERVICE_UNKNOWN | NSLOG_SERVICE_CRITICAL | NSLOG_INFO_MESSAGE;
modified_host_process_attributes = MODATTR_NONE;
modified_service_process_attributes = MODATTR_NONE;
last_log_rotation = 0L;
currently_running_service_checks = 0;
next_comment_id = 0L; /* comment and downtime id get initialized to nonzero elsewhere */
next_downtime_id = 0L;
next_event_id = 1;
next_notification_id = 1;
free_child_process_memory = FALSE;
child_processes_fork_twice = FALSE;
/* initialize macros */
init_macros();
global_host_event_handler_ptr = NULL;
global_service_event_handler_ptr = NULL;
ocsp_command_ptr = NULL;
ochp_command_ptr = NULL;
/* reset umask */
umask(S_IWGRP | S_IWOTH);
return OK;
}
/* try and detect any problems with sys limits
we're specifically interested in NPROC
but could easily add NOFILE here if necessary */
#ifdef DETECT_RLIMIT_PROBLEM
void rlimit_problem_detection(int desired_workers) {
log_debug_info(DEBUGL_PROCESS, 2, "rlimit_problem_detection()\n");
struct rlimit rlim;
int ilim;
host * temp_host = NULL;
service * temp_service = NULL;
/* how many times a worker forks/execs to run a plugin */
int forks_per_worker_per_check = 2;
/* time period to calculate over (in minutes) */
int time_period_calc = 5;
/* how many processes do we just want to account for? */
int arbitrary_user_processes = 300;
double checks_per_time_period = 0.0;
double this_interval = 0.0;
int total_num_procs = 0;
/* first, we grab the NPROC limit, then we check if it isn't unlimited
if it isn't, but the max is, then we set the soft limit to the max
if that doesn't work, or the max isn't unlimited, then we try and
calculate what the current usage is really, and how many workers
we expect to have, and calculate what our usage (generally) is
going to be */
ilim = getrlimit(RLIMIT_NPROC, &rlim);
if (ilim != 0) {
/* nothing we can do here, so just let it keep moving along */
logit(NSLOG_PROCESS_INFO, TRUE, "WARNING: getrlimit(RLIMIT_NPROC) failed with errno: %s\n", strerror(errno));
return;
}
if (rlim.rlim_cur == RLIM_INFINITY) {
/* we won't have any problems due to fork constraints */
log_debug_info(DEBUGL_PROCESS, 0, " * RLIMIT_NPROC is unlimited, no need to continue checking.\n");
return;
}
desired_workers = get_desired_workers(desired_workers);
/* calculate the amount of checks
and the worst-case (estimation) frequency in which they repeat */
for (temp_host = host_list; temp_host != NULL; temp_host = temp_host->next) {
this_interval = 0;
if (!(temp_host->checks_enabled)) {
continue;
}
/* get the smallest possible */
if (temp_host->check_interval > temp_host->retry_interval
&& temp_host->retry_interval != 0) {
this_interval = temp_host->retry_interval;
} else {
this_interval = temp_host->check_interval;
}
/* get them on an average scale (5 min) */
if (this_interval > 0) {
this_interval = ceil((double) time_period_calc / this_interval);
}
checks_per_time_period += this_interval;
}
for (temp_service = service_list; temp_service != NULL; temp_service = temp_service->next) {
this_interval = 0;
if (!(temp_service->checks_enabled)) {
continue;
}
/* get the smallest possible */
if (temp_service->check_interval > temp_service->retry_interval
&& temp_service->retry_interval != 0) {
this_interval = temp_service->retry_interval;
} else {
this_interval = temp_service->check_interval;
}
/* get them on an average scale (5 min) */
if (this_interval > 0) {
this_interval = ceil(time_period_calc / this_interval);
}
checks_per_time_period += this_interval;
}
total_num_procs = checks_per_time_period * forks_per_worker_per_check;
total_num_procs += desired_workers;
total_num_procs += arbitrary_user_processes;
log_debug_info(DEBUGL_PROCESS, 0, " * total_num_procs is: %d\n", total_num_procs);
log_debug_info(DEBUGL_PROCESS, 0, " * using forks_per_worker_per_check: %d\n", forks_per_worker_per_check);
log_debug_info(DEBUGL_PROCESS, 0, " * using desired_workers: %d\n", desired_workers);
log_debug_info(DEBUGL_PROCESS, 0, " * using arbitrary_user_processes: %d\n", arbitrary_user_processes);
if (rlim.rlim_cur > total_num_procs) {
log_debug_info(DEBUGL_PROCESS, 0, " * RLIMIT_NPROC is %d, total max estimated processes is %d, everything looks okay!\n",
(int) rlim.rlim_cur, total_num_procs);
} else {
/* just warn the user - no need to bail out */
logit(NSLOG_RUNTIME_WARNING, TRUE, "WARNING: RLIMIT_NPROC is %d, total max estimated processes is %d! You should increase your limits (ulimit -u, or limits.conf)\n",
(int) rlim.rlim_cur, total_num_procs);
}
}
#endif