int create_listen_socket(void) {
listen_backlog = (int) config_get_number("global", "http port listen backlog", LISTEN_BACKLOG);
listen_port = (int) config_get_number("global", "port", LISTEN_PORT);
if(listen_port < 1 || listen_port > 65535) {
error("Invalid listen port %d given. Defaulting to %d.", listen_port, LISTEN_PORT);
listen_port = LISTEN_PORT;
}
else debug(D_OPTIONS, "Listen port set to %d.", listen_port);
int ip = 0;
char *ipv = config_get("global", "ip version", "any");
if(!strcmp(ipv, "any") || !strcmp(ipv, "both") || !strcmp(ipv, "all")) ip = 0;
else if(!strcmp(ipv, "ipv4") || !strcmp(ipv, "IPV4") || !strcmp(ipv, "IPv4") || !strcmp(ipv, "4")) ip = 4;
else if(!strcmp(ipv, "ipv6") || !strcmp(ipv, "IPV6") || !strcmp(ipv, "IPv6") || !strcmp(ipv, "6")) ip = 6;
else error("Cannot understand ip version '%s'. Assuming 'any'.", ipv);
if(ip == 0 || ip == 6) listen_fd = create_listen_socket6(config_get("global", "bind socket to IP", "*"), listen_port, listen_backlog);
if(listen_fd < 0) {
listen_fd = create_listen_socket4(config_get("global", "bind socket to IP", "*"), listen_port, listen_backlog);
// if(listen_fd >= 0 && ip != 4) info("Managed to open an IPv4 socket on port %d.", listen_port);
}
return listen_fd;
}
void read_cgroup_plugin_configuration() {
cgroup_check_for_new_every = config_get_number("plugin:cgroups", "check for new cgroups every", cgroup_check_for_new_every);
cgroup_enable_cpuacct_stat = config_get_boolean_ondemand("plugin:cgroups", "enable cpuacct stat", cgroup_enable_cpuacct_stat);
cgroup_enable_cpuacct_usage = config_get_boolean_ondemand("plugin:cgroups", "enable cpuacct usage", cgroup_enable_cpuacct_usage);
cgroup_enable_memory = config_get_boolean_ondemand("plugin:cgroups", "enable memory", cgroup_enable_memory);
cgroup_enable_blkio = config_get_boolean_ondemand("plugin:cgroups", "enable blkio", cgroup_enable_blkio);
char filename[FILENAME_MAX + 1], *s;
struct mountinfo *mi, *root = mountinfo_read();
mi = mountinfo_find_by_filesystem_mount_source(root, "cgroup", "cpuacct");
if(!mi) mi = mountinfo_find_by_filesystem_super_option(root, "cgroup", "cpuacct");
if(!mi) {
error("Cannot find cgroup cpuacct mountinfo. Assuming default: /sys/fs/cgroup/cpuacct");
s = "/sys/fs/cgroup/cpuacct";
}
else s = mi->mount_point;
snprintfz(filename, FILENAME_MAX, "%s%s", global_host_prefix, s);
cgroup_cpuacct_base = config_get("plugin:cgroups", "path to /sys/fs/cgroup/cpuacct", filename);
mi = mountinfo_find_by_filesystem_mount_source(root, "cgroup", "blkio");
if(!mi) mi = mountinfo_find_by_filesystem_super_option(root, "cgroup", "blkio");
if(!mi) {
error("Cannot find cgroup blkio mountinfo. Assuming default: /sys/fs/cgroup/blkio");
s = "/sys/fs/cgroup/blkio";
}
else s = mi->mount_point;
snprintfz(filename, FILENAME_MAX, "%s%s", global_host_prefix, s);
cgroup_blkio_base = config_get("plugin:cgroups", "path to /sys/fs/cgroup/blkio", filename);
mi = mountinfo_find_by_filesystem_mount_source(root, "cgroup", "memory");
if(!mi) mi = mountinfo_find_by_filesystem_super_option(root, "cgroup", "memory");
if(!mi) {
error("Cannot find cgroup memory mountinfo. Assuming default: /sys/fs/cgroup/memory");
s = "/sys/fs/cgroup/memory";
}
else s = mi->mount_point;
snprintfz(filename, FILENAME_MAX, "%s%s", global_host_prefix, s);
cgroup_memory_base = config_get("plugin:cgroups", "path to /sys/fs/cgroup/memory", filename);
mi = mountinfo_find_by_filesystem_mount_source(root, "cgroup", "devices");
if(!mi) mi = mountinfo_find_by_filesystem_super_option(root, "cgroup", "devices");
if(!mi) {
error("Cannot find cgroup devices mountinfo. Assuming default: /sys/fs/cgroup/devices");
s = "/sys/fs/cgroup/devices";
}
else s = mi->mount_point;
snprintfz(filename, FILENAME_MAX, "%s%s", global_host_prefix, s);
cgroup_devices_base = config_get("plugin:cgroups", "path to /sys/fs/cgroup/devices", filename);
cgroup_root_max = config_get_number("plugin:cgroups", "max cgroups to allow", cgroup_root_max);
cgroup_max_depth = config_get_number("plugin:cgroups", "max cgroups depth to monitor", cgroup_max_depth);
cgroup_enable_new_cgroups_detected_at_runtime = config_get_boolean("plugin:cgroups", "enable new cgroups detected at run time", cgroup_enable_new_cgroups_detected_at_runtime);
mountinfo_free(root);
}
/*
* Sets up a socket on which we listen for clients. Also creates the listening
* thread.
*/
int client_interface_set_up() {
char ourname[HOSTNAME_BUF_SIZE+1];
struct sockaddr_in sa;
struct hostent *hp;
memset(&sa, 0, sizeof(struct sockaddr_in));
gethostname(ourname, HOSTNAME_BUF_SIZE);
hp = gethostbyname(ourname);
if(hp == NULL) { // we apparently don't exist - what is this, the matrix?
perror("Finding hostname");
return -1;
}
sa.sin_family = hp->h_addrtype;
sa.sin_port = htons(config_get_number("Client_Listen_Port"));
// create socket
if((client_socket = socket(AF_INET, SOCK_STREAM, 0)) < 0) {
perror("Creating listening socket");
return 255;
}
// bind socket
if(bind(client_socket, (const struct sockaddr *) &sa, sizeof(struct sockaddr_in)) < 0) {
close(client_socket);
perror("Binding listening socket");
return 512;
}
listen(client_socket, (int) config_get_number("Client_Max_Queue"));
// Set up the listening thread
client_should_run = 1;
int error = pthread_create(&client_service_thread, NULL, client_interface_listening_thread, NULL);
if(error) {
client_should_run = 0;
close(client_socket);
perror("Creating listening thread");
return error | 0x8000;
}
return 0;
}
/*
* This thread processes connections.
*/
void *client_interface_listening_thread(void *param) {
int t; // holds the connection we establish with clients
printf("Client listening thread active. Listening for clients on port %lli...\n"
, config_get_number("Client_Listen_Port"));
fflush(stdout);
while(client_should_run == 1) {
if((t = client_interface_get_connection()) < 0) {
if(errno == EINTR) { // The socket might sometimes get EINTR
continue;
}
perror("Accept client connection"); // die, we got an unknown error
} else {
pthread_t thread; // we don't really care about this thread later, it'll die eventually
pthread_create(&thread, NULL, client_interface_connection_handler, (void *) &t);
}
}
close(client_socket);
client_has_quit = 1;
pthread_exit(NULL);
}
static inline void tc_device_classes_cleanup(struct tc_device *d) {
static int cleanup_every = 999;
if(unlikely(cleanup_every > 0)) {
cleanup_every = (int) config_get_number("plugin:tc", "cleanup unused classes every", 120);
if(cleanup_every < 0) cleanup_every = -cleanup_every;
}
d->name_updated = 0;
d->family_updated = 0;
struct tc_class *c = d->classes;
while(c) {
if(unlikely(cleanup_every && c->unupdated >= cleanup_every)) {
struct tc_class *nc = c->next;
tc_class_free(d, c);
c = nc;
}
else {
c->updated = 0;
c->name_updated = 0;
c = c->next;
}
}
}
int sock_fd_init (void)
{
static int initialized = 0;
if (fds)
return (0);
#if defined (NUTTX_RTOS)
fds = malloc (sizeof (struct pollfd) * FD_INC_SIZE);
fcts = malloc (sizeof (RSDATAFCT) * FD_INC_SIZE);
ud = malloc (sizeof (void *) * FD_INC_SIZE);
names = malloc (sizeof (const char *) * FD_INC_SIZE);
#else
fds = xmalloc (sizeof (struct pollfd) * FD_INC_SIZE);
fcts = xmalloc (sizeof (RSDATAFCT) * FD_INC_SIZE);
ud = xmalloc (sizeof (void *) * FD_INC_SIZE);
names = xmalloc (sizeof (const char *) * FD_INC_SIZE);
#endif
if (!fds || !fcts || !ud || !names)
return (1);
atomic_set_w (num_fds, 0);
atomic_set_w (max_fds, FD_INC_SIZE);
if (!initialized) {
lock_init_nr (sock_lock, "lock");
fd_max_size = config_get_number (DC_IP_Sockets, FD_MAX_SIZE);
initialized = 1;
}
return (0);
}
void web_server_threading_selection(void) {
int threaded = config_get_boolean("global", "multi threaded web server", 1);
int i;
for(i = 0; static_threads[i].name ; i++) {
if(static_threads[i].start_routine == socket_listen_main_multi_threaded)
static_threads[i].enabled = threaded?1:0;
if(static_threads[i].start_routine == socket_listen_main_single_threaded)
static_threads[i].enabled = threaded?0:1;
}
web_client_timeout = (int) config_get_number("global", "disconnect idle web clients after seconds", DEFAULT_DISCONNECT_IDLE_WEB_CLIENTS_AFTER_SECONDS);
web_donotrack_comply = config_get_boolean("global", "respect web browser do not track policy", web_donotrack_comply);
#ifdef NETDATA_WITH_ZLIB
web_enable_gzip = config_get_boolean("global", "enable web responses gzip compression", web_enable_gzip);
char *s = config_get("global", "web compression strategy", "default");
if(!strcmp(s, "default"))
web_gzip_strategy = Z_DEFAULT_STRATEGY;
else if(!strcmp(s, "filtered"))
web_gzip_strategy = Z_FILTERED;
else if(!strcmp(s, "huffman only"))
web_gzip_strategy = Z_HUFFMAN_ONLY;
else if(!strcmp(s, "rle"))
web_gzip_strategy = Z_RLE;
else if(!strcmp(s, "fixed"))
web_gzip_strategy = Z_FIXED;
else {
error("Invalid compression strategy '%s'. Valid strategies are 'default', 'filtered', 'huffman only', 'rle' and 'fixed'. Proceeding with 'default'.", s);
web_gzip_strategy = Z_DEFAULT_STRATEGY;
}
web_gzip_level = (int)config_get_number("global", "web compression level", 3);
if(web_gzip_level < 1) {
error("Invalid compression level %d. Valid levels are 1 (fastest) to 9 (best ratio). Proceeding with level 1 (fastest compression).", web_gzip_level);
web_gzip_level = 1;
}
else if(web_gzip_level > 9) {
error("Invalid compression level %d. Valid levels are 1 (fastest) to 9 (best ratio). Proceeding with level 9 (best compression).", web_gzip_level);
web_gzip_level = 9;
}
#endif /* NETDATA_WITH_ZLIB */
}
void web_server_config_options(void) {
web_client_timeout = (int) config_get_number(CONFIG_SECTION_WEB, "disconnect idle clients after seconds", web_client_timeout);
web_client_first_request_timeout = (int) config_get_number(CONFIG_SECTION_WEB, "timeout for first request", web_client_first_request_timeout);
respect_web_browser_do_not_track_policy = config_get_boolean(CONFIG_SECTION_WEB, "respect do not track policy", respect_web_browser_do_not_track_policy);
web_x_frame_options = config_get(CONFIG_SECTION_WEB, "x-frame-options response header", "");
if(!*web_x_frame_options) web_x_frame_options = NULL;
web_allow_connections_from = simple_pattern_create(config_get(CONFIG_SECTION_WEB, "allow connections from", "localhost *"), NULL, SIMPLE_PATTERN_EXACT);
web_allow_dashboard_from = simple_pattern_create(config_get(CONFIG_SECTION_WEB, "allow dashboard from", "localhost *"), NULL, SIMPLE_PATTERN_EXACT);
web_allow_badges_from = simple_pattern_create(config_get(CONFIG_SECTION_WEB, "allow badges from", "*"), NULL, SIMPLE_PATTERN_EXACT);
web_allow_registry_from = simple_pattern_create(config_get(CONFIG_SECTION_REGISTRY, "allow from", "*"), NULL, SIMPLE_PATTERN_EXACT);
web_allow_streaming_from = simple_pattern_create(config_get(CONFIG_SECTION_WEB, "allow streaming from", "*"), NULL, SIMPLE_PATTERN_EXACT);
web_allow_netdataconf_from = simple_pattern_create(config_get(CONFIG_SECTION_WEB, "allow netdata.conf from", "localhost fd* 10.* 192.168.* 172.16.* 172.17.* 172.18.* 172.19.* 172.20.* 172.21.* 172.22.* 172.23.* 172.24.* 172.25.* 172.26.* 172.27.* 172.28.* 172.29.* 172.30.* 172.31.*"), NULL, SIMPLE_PATTERN_EXACT);
#ifdef NETDATA_WITH_ZLIB
web_enable_gzip = config_get_boolean(CONFIG_SECTION_WEB, "enable gzip compression", web_enable_gzip);
char *s = config_get(CONFIG_SECTION_WEB, "gzip compression strategy", "default");
if(!strcmp(s, "default"))
web_gzip_strategy = Z_DEFAULT_STRATEGY;
else if(!strcmp(s, "filtered"))
web_gzip_strategy = Z_FILTERED;
else if(!strcmp(s, "huffman only"))
web_gzip_strategy = Z_HUFFMAN_ONLY;
else if(!strcmp(s, "rle"))
web_gzip_strategy = Z_RLE;
else if(!strcmp(s, "fixed"))
web_gzip_strategy = Z_FIXED;
else {
error("Invalid compression strategy '%s'. Valid strategies are 'default', 'filtered', 'huffman only', 'rle' and 'fixed'. Proceeding with 'default'.", s);
web_gzip_strategy = Z_DEFAULT_STRATEGY;
}
web_gzip_level = (int)config_get_number(CONFIG_SECTION_WEB, "gzip compression level", 3);
if(web_gzip_level < 1) {
error("Invalid compression level %d. Valid levels are 1 (fastest) to 9 (best ratio). Proceeding with level 1 (fastest compression).", web_gzip_level);
web_gzip_level = 1;
}
else if(web_gzip_level > 9) {
error("Invalid compression level %d. Valid levels are 1 (fastest) to 9 (best ratio). Proceeding with level 9 (best compression).", web_gzip_level);
web_gzip_level = 9;
}
#endif /* NETDATA_WITH_ZLIB */
}
void log_init(void) {
char filename[FILENAME_MAX + 1];
snprintfz(filename, FILENAME_MAX, "%s/debug.log", netdata_configured_log_dir);
stdout_filename = config_get(CONFIG_SECTION_GLOBAL, "debug log", filename);
snprintfz(filename, FILENAME_MAX, "%s/error.log", netdata_configured_log_dir);
stderr_filename = config_get(CONFIG_SECTION_GLOBAL, "error log", filename);
snprintfz(filename, FILENAME_MAX, "%s/access.log", netdata_configured_log_dir);
stdaccess_filename = config_get(CONFIG_SECTION_GLOBAL, "access log", filename);
error_log_throttle_period_backup =
error_log_throttle_period = config_get_number(CONFIG_SECTION_GLOBAL, "errors flood protection period", error_log_throttle_period);
error_log_errors_per_period = (unsigned long)config_get_number(CONFIG_SECTION_GLOBAL, "errors to trigger flood protection", (long long int)error_log_errors_per_period);
setenv("NETDATA_ERRORS_THROTTLE_PERIOD", config_get(CONFIG_SECTION_GLOBAL, "errors flood protection period" , ""), 1);
setenv("NETDATA_ERRORS_PER_PERIOD", config_get(CONFIG_SECTION_GLOBAL, "errors to trigger flood protection", ""), 1);
}
void *cpuidlejitter_main(void *ptr) {
struct netdata_static_thread *static_thread = (struct netdata_static_thread *)ptr;
info("IDLEJITTER thread created with task id %d", gettid());
if(pthread_setcanceltype(PTHREAD_CANCEL_DEFERRED, NULL) != 0)
error("Cannot set pthread cancel type to DEFERRED.");
if(pthread_setcancelstate(PTHREAD_CANCEL_ENABLE, NULL) != 0)
error("Cannot set pthread cancel state to ENABLE.");
int sleep_ms = (int) config_get_number("plugin:idlejitter", "loop time in ms", CPU_IDLEJITTER_SLEEP_TIME_MS);
if(sleep_ms <= 0) {
config_set_number("plugin:idlejitter", "loop time in ms", CPU_IDLEJITTER_SLEEP_TIME_MS);
sleep_ms = CPU_IDLEJITTER_SLEEP_TIME_MS;
}
RRDSET *st = rrdset_find_localhost("system.idlejitter");
if(!st) {
st = rrdset_create_localhost("system", "idlejitter", NULL, "processes", NULL, "CPU Idle Jitter"
, "microseconds lost/s", 9999, localhost->rrd_update_every, RRDSET_TYPE_LINE);
rrddim_add(st, "jitter", NULL, 1, 1, RRD_ALGORITHM_ABSOLUTE);
}
struct timeval before, after;
unsigned long long counter;
for(counter = 0; 1 ;counter++) {
usec_t usec = 0, susec = 0;
if(netdata_exit) break;
while(susec < (localhost->rrd_update_every * USEC_PER_SEC)) {
now_monotonic_timeval(&before);
sleep_usec(sleep_ms * 1000);
now_monotonic_timeval(&after);
// calculate the time it took for a full loop
usec = dt_usec(&after, &before);
susec += usec;
}
usec -= (sleep_ms * 1000);
if(counter) rrdset_next(st);
rrddim_set(st, "jitter", usec);
rrdset_done(st);
}
info("IDLEJITTER thread exiting");
static_thread->enabled = 0;
pthread_exit(NULL);
return NULL;
}
int create_listen_sockets(void) {
listen_backlog = (int) config_get_number("global", "http port listen backlog", LISTEN_BACKLOG);
if(config_exists("global", "bind socket to IP") && !config_exists("global", "bind to"))
config_rename("global", "bind socket to IP", "bind to");
if(config_exists("global", "port") && !config_exists("global", "default port"))
config_rename("global", "port", "default port");
listen_port = (int) config_get_number("global", "default port", LISTEN_PORT);
if(listen_port < 1 || listen_port > 65535) {
error("Invalid listen port %d given. Defaulting to %d.", listen_port, LISTEN_PORT);
listen_port = (int) config_set_number("global", "default port", LISTEN_PORT);
}
debug(D_OPTIONS, "Default listen port set to %d.", listen_port);
char *s = config_get("global", "bind to", "*");
while(*s) {
char *e = s;
// skip separators, moving both s(tart) and e(nd)
while(isspace(*e) || *e == ',') s = ++e;
// move e(nd) to the first separator
while(*e && !isspace(*e) && *e != ',') e++;
// is there anything?
if(!*s || s == e) break;
char buf[e - s + 1];
strncpyz(buf, s, e - s);
bind_to_one(buf, listen_port, listen_backlog);
s = e;
}
if(!listen_fds_count)
fatal("Cannot listen on any socket. Exiting...");
return listen_fds_count;
}
int rrdpush_init() {
default_rrdpush_enabled = appconfig_get_boolean(&stream_config, CONFIG_SECTION_STREAM, "enabled", default_rrdpush_enabled);
default_rrdpush_destination = appconfig_get(&stream_config, CONFIG_SECTION_STREAM, "destination", "");
default_rrdpush_api_key = appconfig_get(&stream_config, CONFIG_SECTION_STREAM, "api key", "");
rrdhost_free_orphan_time = config_get_number(CONFIG_SECTION_GLOBAL, "cleanup orphan hosts after seconds", rrdhost_free_orphan_time);
if(default_rrdpush_enabled && (!default_rrdpush_destination || !*default_rrdpush_destination || !default_rrdpush_api_key || !*default_rrdpush_api_key)) {
error("STREAM [send]: cannot enable sending thread - information is missing.");
default_rrdpush_enabled = 0;
}
return default_rrdpush_enabled;
}
void *cpuidlejitter_main(void *ptr)
{
if(ptr) { ; }
info("CPU Idle Jitter thread created with task id %d", gettid());
if(pthread_setcanceltype(PTHREAD_CANCEL_DEFERRED, NULL) != 0)
error("Cannot set pthread cancel type to DEFERRED.");
if(pthread_setcancelstate(PTHREAD_CANCEL_ENABLE, NULL) != 0)
error("Cannot set pthread cancel state to ENABLE.");
int sleep_ms = (int) config_get_number("plugin:idlejitter", "loop time in ms", CPU_IDLEJITTER_SLEEP_TIME_MS);
if(sleep_ms <= 0) {
config_set_number("plugin:idlejitter", "loop time in ms", CPU_IDLEJITTER_SLEEP_TIME_MS);
sleep_ms = CPU_IDLEJITTER_SLEEP_TIME_MS;
}
RRDSET *st = rrdset_find("system.idlejitter");
if(!st) {
st = rrdset_create("system", "idlejitter", NULL, "cpu", "CPU Idle Jitter", "microseconds lost/s", 9999, rrd_update_every, RRDSET_TYPE_LINE);
rrddim_add(st, "jitter", NULL, 1, 1, RRDDIM_ABSOLUTE);
}
struct timeval before, after;
unsigned long long counter;
for(counter = 0; 1 ;counter++) {
unsigned long long usec = 0, susec = 0;
while(susec < (rrd_update_every * 1000000ULL)) {
gettimeofday(&before, NULL);
usleep(sleep_ms * 1000);
gettimeofday(&after, NULL);
// calculate the time it took for a full loop
usec = usecdiff(&after, &before);
susec += usec;
}
usec -= (sleep_ms * 1000);
if(counter) rrdset_next(st);
rrddim_set(st, "jitter", usec);
rrdset_done(st);
}
return NULL;
}
static inline void tc_device_classes_cleanup(struct tc_device *d) {
static int cleanup_every = 999;
if(cleanup_every > 0) {
cleanup_every = (int) -config_get_number("plugin:tc", "cleanup unused classes every", 60);
if(cleanup_every > 0) cleanup_every = -cleanup_every;
if(cleanup_every == 0) cleanup_every = -1;
}
struct tc_class *c = d->classes;
while(c) {
if(c->seen < cleanup_every) {
struct tc_class *nc = c->next;
tc_class_free(d, c);
c = nc;
}
else c = c->next;
if(c) {
c->updated = 0;
c->seen--;
}
}
}
int main(int argc, char **argv)
{
int i;
int config_loaded = 0;
int dont_fork = 0;
// global initialization
get_HZ();
// set the name for logging
program_name = "netdata";
// parse the arguments
for(i = 1; i < argc ; i++) {
if(strcmp(argv[i], "-c") == 0 && (i+1) < argc) {
if(load_config(argv[i+1], 1) != 1) {
error("Cannot load configuration file %s.", argv[i+1]);
exit(1);
}
else {
debug(D_OPTIONS, "Configuration loaded from %s.", argv[i+1]);
config_loaded = 1;
}
i++;
}
else if(strcmp(argv[i], "-df") == 0 && (i+1) < argc) { config_set("global", "debug flags", argv[i+1]); debug_flags = strtoull(argv[i+1], NULL, 0); i++; }
else if(strcmp(argv[i], "-p") == 0 && (i+1) < argc) { config_set("global", "port", argv[i+1]); i++; }
else if(strcmp(argv[i], "-u") == 0 && (i+1) < argc) { config_set("global", "run as user", argv[i+1]); i++; }
else if(strcmp(argv[i], "-l") == 0 && (i+1) < argc) { config_set("global", "history", argv[i+1]); i++; }
else if(strcmp(argv[i], "-t") == 0 && (i+1) < argc) { config_set("global", "update every", argv[i+1]); i++; }
else if(strcmp(argv[i], "-ch") == 0 && (i+1) < argc) { config_set("global", "host access prefix", argv[i+1]); i++; }
else if(strcmp(argv[i], "-nodaemon") == 0 || strcmp(argv[i], "-nd") == 0) dont_fork = 1;
else if(strcmp(argv[i], "--unittest") == 0) {
rrd_update_every = 1;
if(run_all_mockup_tests()) exit(1);
if(unit_test_storage()) exit(1);
fprintf(stderr, "\n\nALL TESTS PASSED\n\n");
exit(0);
}
else {
fprintf(stderr, "Cannot understand option '%s'.\n", argv[i]);
fprintf(stderr, "\nUSAGE: %s [-d] [-l LINES_TO_SAVE] [-u UPDATE_TIMER] [-p LISTEN_PORT] [-df debug flags].\n\n", argv[0]);
fprintf(stderr, " -c CONFIG FILE the configuration file to load. Default: %s.\n", CONFIG_DIR "/" CONFIG_FILENAME);
fprintf(stderr, " -l LINES_TO_SAVE can be from 5 to %d lines in JSON data. Default: %d.\n", RRD_HISTORY_ENTRIES_MAX, RRD_DEFAULT_HISTORY_ENTRIES);
fprintf(stderr, " -t UPDATE_TIMER can be from 1 to %d seconds. Default: %d.\n", UPDATE_EVERY_MAX, UPDATE_EVERY);
fprintf(stderr, " -p LISTEN_PORT can be from 1 to %d. Default: %d.\n", 65535, LISTEN_PORT);
fprintf(stderr, " -u USERNAME can be any system username to run as. Default: none.\n");
fprintf(stderr, " -ch path to access host /proc and /sys when running in a container. Default: empty.\n");
fprintf(stderr, " -nd or -nodeamon to disable forking in the background. Default: unset.\n");
fprintf(stderr, " -df FLAGS debug options. Default: 0x%08llx.\n", debug_flags);
exit(1);
}
}
if(!config_loaded) load_config(NULL, 0);
// prepare configuration environment variables for the plugins
setenv("NETDATA_CONFIG_DIR" , config_get("global", "config directory" , CONFIG_DIR) , 1);
setenv("NETDATA_PLUGINS_DIR", config_get("global", "plugins directory" , PLUGINS_DIR), 1);
setenv("NETDATA_WEB_DIR" , config_get("global", "web files directory", WEB_DIR) , 1);
setenv("NETDATA_CACHE_DIR" , config_get("global", "cache directory" , CACHE_DIR) , 1);
setenv("NETDATA_LOG_DIR" , config_get("global", "log directory" , LOG_DIR) , 1);
setenv("NETDATA_HOST_PREFIX", config_get("global", "host access prefix" , "") , 1);
// avoid extended to stat(/etc/localtime)
// http://stackoverflow.com/questions/4554271/how-to-avoid-excessive-stat-etc-localtime-calls-in-strftime-on-linux
setenv("TZ", ":/etc/localtime", 0);
// cd to /tmp to avoid any plugins writing files at random places
if(chdir("/tmp")) error("netdata: ERROR: Cannot cd to /tmp");
char *input_log_file = NULL;
char *output_log_file = NULL;
char *error_log_file = NULL;
char *access_log_file = NULL;
char *user = NULL;
{
char buffer[1024];
// --------------------------------------------------------------------
sprintf(buffer, "0x%08llx", 0ULL);
char *flags = config_get("global", "debug flags", buffer);
setenv("NETDATA_DEBUG_FLAGS", flags, 1);
debug_flags = strtoull(flags, NULL, 0);
debug(D_OPTIONS, "Debug flags set to '0x%8llx'.", debug_flags);
if(debug_flags != 0) {
struct rlimit rl = { RLIM_INFINITY, RLIM_INFINITY };
if(setrlimit(RLIMIT_CORE, &rl) != 0)
info("Cannot request unlimited core dumps for debugging... Proceeding anyway...");
prctl(PR_SET_DUMPABLE, 1, 0, 0, 0);
}
// --------------------------------------------------------------------
#ifdef MADV_MERGEABLE
enable_ksm = config_get_boolean("global", "memory deduplication (ksm)", enable_ksm);
#else
#warning "Kernel memory deduplication (KSM) is not available"
#endif
// --------------------------------------------------------------------
global_host_prefix = config_get("global", "host access prefix", "");
setenv("NETDATA_HOST_PREFIX", global_host_prefix, 1);
// --------------------------------------------------------------------
output_log_file = config_get("global", "debug log", LOG_DIR "/debug.log");
if(strcmp(output_log_file, "syslog") == 0) {
output_log_syslog = 1;
output_log_file = NULL;
}
else if(strcmp(output_log_file, "none") == 0) {
output_log_syslog = 0;
output_log_file = NULL;
}
else output_log_syslog = 0;
// --------------------------------------------------------------------
error_log_file = config_get("global", "error log", LOG_DIR "/error.log");
if(strcmp(error_log_file, "syslog") == 0) {
error_log_syslog = 1;
error_log_file = NULL;
}
else if(strcmp(error_log_file, "none") == 0) {
error_log_syslog = 0;
error_log_file = NULL;
// optimization - do not even generate debug log entries
}
else error_log_syslog = 0;
// --------------------------------------------------------------------
access_log_file = config_get("global", "access log", LOG_DIR "/access.log");
if(strcmp(access_log_file, "syslog") == 0) {
access_log_syslog = 1;
access_log_file = NULL;
}
else if(strcmp(access_log_file, "none") == 0) {
access_log_syslog = 0;
access_log_file = NULL;
}
else access_log_syslog = 0;
// --------------------------------------------------------------------
rrd_memory_mode = rrd_memory_mode_id(config_get("global", "memory mode", rrd_memory_mode_name(rrd_memory_mode)));
// --------------------------------------------------------------------
if(gethostname(buffer, HOSTNAME_MAX) == -1)
error("WARNING: Cannot get machine hostname.");
hostname = config_get("global", "hostname", buffer);
debug(D_OPTIONS, "hostname set to '%s'", hostname);
// --------------------------------------------------------------------
rrd_default_history_entries = (int) config_get_number("global", "history", RRD_DEFAULT_HISTORY_ENTRIES);
if(rrd_default_history_entries < 5 || rrd_default_history_entries > RRD_HISTORY_ENTRIES_MAX) {
fprintf(stderr, "Invalid save lines %d given. Defaulting to %d.\n", rrd_default_history_entries, RRD_DEFAULT_HISTORY_ENTRIES);
rrd_default_history_entries = RRD_DEFAULT_HISTORY_ENTRIES;
}
else {
debug(D_OPTIONS, "save lines set to %d.", rrd_default_history_entries);
}
// --------------------------------------------------------------------
rrd_update_every = (int) config_get_number("global", "update every", UPDATE_EVERY);
if(rrd_update_every < 1 || rrd_update_every > 600) {
fprintf(stderr, "Invalid update timer %d given. Defaulting to %d.\n", rrd_update_every, UPDATE_EVERY_MAX);
rrd_update_every = UPDATE_EVERY;
}
else debug(D_OPTIONS, "update timer set to %d.", rrd_update_every);
// let the plugins know the min update_every
{
char buf[50];
snprintf(buf, 50, "%d", rrd_update_every);
setenv("NETDATA_UPDATE_EVERY", buf, 1);
}
// --------------------------------------------------------------------
for (i = 0; static_threads[i].name != NULL ; i++) {
struct netdata_static_thread *st = &static_threads[i];
if(st->config_name) st->enabled = config_get_boolean(st->config_section, st->config_name, st->enabled);
if(st->enabled && st->init_routine) st->init_routine();
}
// --------------------------------------------------------------------
prepare_rundir();
user = config_get("global", "run as user" , (getuid() == 0)?NETDATA_USER:"");
web_files_uid();
// --------------------------------------------------------------------
listen_backlog = (int) config_get_number("global", "http port listen backlog", LISTEN_BACKLOG);
listen_port = (int) config_get_number("global", "port", LISTEN_PORT);
if(listen_port < 1 || listen_port > 65535) {
fprintf(stderr, "Invalid listen port %d given. Defaulting to %d.\n", listen_port, LISTEN_PORT);
listen_port = LISTEN_PORT;
}
else debug(D_OPTIONS, "Listen port set to %d.", listen_port);
int ip = 0;
char *ipv = config_get("global", "ip version", "any");
if(!strcmp(ipv, "any") || !strcmp(ipv, "both") || !strcmp(ipv, "all")) ip = 0;
else if(!strcmp(ipv, "ipv4") || !strcmp(ipv, "IPV4") || !strcmp(ipv, "IPv4") || !strcmp(ipv, "4")) ip = 4;
else if(!strcmp(ipv, "ipv6") || !strcmp(ipv, "IPV6") || !strcmp(ipv, "IPv6") || !strcmp(ipv, "6")) ip = 6;
else fprintf(stderr, "Cannot understand ip version '%s'. Assumming 'any'.", ipv);
if(ip == 0 || ip == 6) listen_fd = create_listen_socket6(listen_port, listen_backlog);
if(listen_fd < 0) {
listen_fd = create_listen_socket4(listen_port, listen_backlog);
if(listen_fd >= 0 && ip != 4) fprintf(stderr, "Managed to open an IPv4 socket on port %d.", listen_port);
}
if(listen_fd < 0) fatal("Cannot listen socket.");
}
// never become a problem
if(nice(20) == -1) error("Cannot lower my CPU priority.");
if(become_daemon(dont_fork, 0, user, input_log_file, output_log_file, error_log_file, access_log_file, &access_fd, &stdaccess) == -1) {
fatal("Cannot demonize myself.");
exit(1);
}
if(debug_flags != 0) {
struct rlimit rl = { RLIM_INFINITY, RLIM_INFINITY };
if(setrlimit(RLIMIT_CORE, &rl) != 0)
info("Cannot request unlimited core dumps for debugging... Proceeding anyway...");
prctl(PR_SET_DUMPABLE, 1, 0, 0, 0);
}
if(output_log_syslog || error_log_syslog || access_log_syslog)
openlog("netdata", LOG_PID, LOG_DAEMON);
info("NetData started on pid %d", getpid());
// catch all signals
for (i = 1 ; i < 65 ;i++) {
switch(i) {
case SIGKILL: // not catchable
case SIGSTOP: // not catchable
break;
case SIGSEGV:
case SIGFPE:
case SIGCHLD:
signal(i, SIG_DFL);
break;
default:
signal(i, sig_handler);
break;
}
}
for (i = 0; static_threads[i].name != NULL ; i++) {
struct netdata_static_thread *st = &static_threads[i];
if(st->enabled) {
st->thread = malloc(sizeof(pthread_t));
if(!st->thread)
fatal("Cannot allocate pthread_t memory");
info("Starting thread %s.", st->name);
if(pthread_create(st->thread, NULL, st->start_routine, NULL))
error("failed to create new thread for %s.", st->name);
else if(pthread_detach(*st->thread))
error("Cannot request detach of newly created %s thread.", st->name);
}
else info("Not starting thread %s.", st->name);
}
// for future use - the main thread
while(1) {
if(netdata_exit != 0) {
netdata_exit++;
if(netdata_exit > 5) {
netdata_cleanup_and_exit(0);
exit(0);
}
}
sleep(2);
}
exit(0);
}
void *proc_diskspace_main(void *ptr) {
struct netdata_static_thread *static_thread = (struct netdata_static_thread *)ptr;
info("DISKSPACE thread created with task id %d", gettid());
if(pthread_setcanceltype(PTHREAD_CANCEL_DEFERRED, NULL) != 0)
error("DISKSPACE: Cannot set pthread cancel type to DEFERRED.");
if(pthread_setcancelstate(PTHREAD_CANCEL_ENABLE, NULL) != 0)
error("DISKSPACE: Cannot set pthread cancel state to ENABLE.");
int vdo_cpu_netdata = config_get_boolean("plugin:proc", "netdata server resources", 1);
int update_every = (int)config_get_number(CONFIG_SECTION_DISKSPACE, "update every", localhost->rrd_update_every);
if(update_every < localhost->rrd_update_every)
update_every = localhost->rrd_update_every;
check_for_new_mountpoints_every = (int)config_get_number(CONFIG_SECTION_DISKSPACE, "check for new mount points every", check_for_new_mountpoints_every);
if(check_for_new_mountpoints_every < update_every)
check_for_new_mountpoints_every = update_every;
struct rusage thread;
usec_t duration = 0;
usec_t step = update_every * USEC_PER_SEC;
heartbeat_t hb;
heartbeat_init(&hb);
for(;;) {
duration = heartbeat_dt_usec(&hb);
/* usec_t hb_dt = */ heartbeat_next(&hb, step);
if(unlikely(netdata_exit)) break;
// --------------------------------------------------------------------------
// this is smart enough not to reload it every time
mountinfo_reload(0);
// --------------------------------------------------------------------------
// disk space metrics
struct mountinfo *mi;
for(mi = disk_mountinfo_root; mi; mi = mi->next) {
if(unlikely(mi->flags & (MOUNTINFO_IS_DUMMY | MOUNTINFO_IS_BIND)))
continue;
do_disk_space_stats(mi, update_every);
if(unlikely(netdata_exit)) break;
}
if(unlikely(netdata_exit)) break;
dictionary_get_all(dict_mountpoints, mount_point_cleanup, NULL);
if(vdo_cpu_netdata) {
static RRDSET *stcpu_thread = NULL, *st_duration = NULL;
static RRDDIM *rd_user = NULL, *rd_system = NULL, *rd_duration = NULL;
// ----------------------------------------------------------------
getrusage(RUSAGE_THREAD, &thread);
if(!stcpu_thread) {
stcpu_thread = rrdset_find_localhost("netdata.plugin_diskspace");
if(!stcpu_thread)
stcpu_thread = rrdset_create_localhost(
"netdata"
, "plugin_diskspace"
, NULL
, "diskspace"
, NULL
, "NetData Disk Space Plugin CPU usage"
, "milliseconds/s"
, 132020
, update_every
, RRDSET_TYPE_STACKED
);
rd_user = rrddim_add(stcpu_thread, "user", NULL, 1, 1000, RRD_ALGORITHM_INCREMENTAL);
rd_system = rrddim_add(stcpu_thread, "system", NULL, 1, 1000, RRD_ALGORITHM_INCREMENTAL);
}
else
rrdset_next(stcpu_thread);
rrddim_set_by_pointer(stcpu_thread, rd_user, thread.ru_utime.tv_sec * 1000000ULL + thread.ru_utime.tv_usec);
rrddim_set_by_pointer(stcpu_thread, rd_system, thread.ru_stime.tv_sec * 1000000ULL + thread.ru_stime.tv_usec);
rrdset_done(stcpu_thread);
// ----------------------------------------------------------------
if(!st_duration) {
st_duration = rrdset_find_localhost("netdata.plugin_diskspace_dt");
if(!st_duration)
st_duration = rrdset_create_localhost(
"netdata"
, "plugin_diskspace_dt"
, NULL
, "diskspace"
, NULL
, "NetData Disk Space Plugin Duration"
, "milliseconds/run"
, 132021
, update_every
, RRDSET_TYPE_AREA
);
rd_duration = rrddim_add(st_duration, "duration", NULL, 1, 1000, RRD_ALGORITHM_ABSOLUTE);
}
else
rrdset_next(st_duration);
rrddim_set_by_pointer(st_duration, rd_duration, duration);
rrdset_done(st_duration);
// ----------------------------------------------------------------
if(unlikely(netdata_exit)) break;
}
}
info("DISKSPACE thread exiting");
static_thread->enabled = 0;
pthread_exit(NULL);
return NULL;
}
RRDDIM *rrddim_add(RRDSET *st, const char *id, const char *name, long multiplier, long divisor, int algorithm)
{
char filename[FILENAME_MAX + 1];
char fullfilename[FILENAME_MAX + 1];
char varname[CONFIG_MAX_NAME + 1];
RRDDIM *rd = NULL;
unsigned long size = sizeof(RRDDIM) + (st->entries * sizeof(storage_number));
debug(D_RRD_CALLS, "Adding dimension '%s/%s'.", st->id, id);
rrdset_strncpyz_name(filename, id, FILENAME_MAX);
snprintfz(fullfilename, FILENAME_MAX, "%s/%s.db", st->cache_dir, filename);
if(rrd_memory_mode != RRD_MEMORY_MODE_RAM) rd = (RRDDIM *)mymmap(fullfilename, size, ((rrd_memory_mode == RRD_MEMORY_MODE_MAP)?MAP_SHARED:MAP_PRIVATE), 1);
if(rd) {
struct timeval now;
gettimeofday(&now, NULL);
if(strcmp(rd->magic, RRDDIMENSION_MAGIC) != 0) {
errno = 0;
info("Initializing file %s.", fullfilename);
bzero(rd, size);
}
else if(rd->memsize != size) {
errno = 0;
error("File %s does not have the desired size. Clearing it.", fullfilename);
bzero(rd, size);
}
else if(rd->multiplier != multiplier) {
errno = 0;
error("File %s does not have the same multiplier. Clearing it.", fullfilename);
bzero(rd, size);
}
else if(rd->divisor != divisor) {
errno = 0;
error("File %s does not have the same divisor. Clearing it.", fullfilename);
bzero(rd, size);
}
else if(rd->algorithm != algorithm) {
errno = 0;
error("File %s does not have the same algorithm. Clearing it.", fullfilename);
bzero(rd, size);
}
else if(rd->update_every != st->update_every) {
errno = 0;
error("File %s does not have the same refresh frequency. Clearing it.", fullfilename);
bzero(rd, size);
}
else if(usecdiff(&now, &rd->last_collected_time) > (rd->entries * rd->update_every * 1000000ULL)) {
errno = 0;
error("File %s is too old. Clearing it.", fullfilename);
bzero(rd, size);
}
else if(strcmp(rd->id, id) != 0) {
errno = 0;
error("File %s contents are not for dimension %s. Clearing it.", fullfilename, id);
// munmap(rd, size);
// rd = NULL;
bzero(rd, size);
}
}
if(rd) {
// we have a file mapped for rd
rd->mapped = rrd_memory_mode;
rd->flags = 0x00000000;
rd->next = NULL;
rd->name = NULL;
}
else {
// if we didn't manage to get a mmap'd dimension, just create one
rd = calloc(1, size);
if(!rd) {
fatal("Cannot allocate RRD_DIMENSION %s/%s.", st->id, id);
return NULL;
}
rd->mapped = RRD_MEMORY_MODE_RAM;
}
rd->memsize = size;
strcpy(rd->magic, RRDDIMENSION_MAGIC);
strcpy(rd->cache_filename, fullfilename);
strncpyz(rd->id, id, RRD_ID_LENGTH_MAX);
rd->hash = simple_hash(rd->id);
snprintfz(varname, CONFIG_MAX_NAME, "dim %s name", rd->id);
rd->name = config_get(st->id, varname, (name && *name)?name:rd->id);
snprintfz(varname, CONFIG_MAX_NAME, "dim %s algorithm", rd->id);
rd->algorithm = rrddim_algorithm_id(config_get(st->id, varname, rrddim_algorithm_name(algorithm)));
snprintfz(varname, CONFIG_MAX_NAME, "dim %s multiplier", rd->id);
rd->multiplier = config_get_number(st->id, varname, multiplier);
snprintfz(varname, CONFIG_MAX_NAME, "dim %s divisor", rd->id);
rd->divisor = config_get_number(st->id, varname, divisor);
if(!rd->divisor) rd->divisor = 1;
rd->entries = st->entries;
rd->update_every = st->update_every;
// prevent incremental calculation spikes
rd->counter = 0;
rd->updated = 0;
rd->calculated_value = 0;
rd->last_calculated_value = 0;
rd->collected_value = 0;
rd->last_collected_value = 0;
rd->collected_volume = 0;
rd->stored_volume = 0;
rd->values[st->current_entry] = pack_storage_number(0, SN_NOT_EXISTS);
rd->last_collected_time.tv_sec = 0;
rd->last_collected_time.tv_usec = 0;
// append this dimension
pthread_rwlock_wrlock(&st->rwlock);
if(!st->dimensions)
st->dimensions = rd;
else {
RRDDIM *td = st->dimensions;
for(; td->next; td = td->next) ;
td->next = rd;
}
pthread_rwlock_unlock(&st->rwlock);
rrddim_index_add(st, rd);
return(rd);
}
/*
* This function is called in a new thread to process a client that connects. It
* handles the protocol and all that other fun stuff to translate messages the
* clients send into a push notification we can send out.
*/
void *client_interface_connection_handler(void *connection) {
if(connection == NULL) return NULL;
int sock = *((int *) connection);
printf("Client connected.\n");
char *msg_buf = calloc(config_get_number("Client_Max_Msg_Size") + 1, sizeof(char));
char *msg_buf_write_ptr = msg_buf;
long bytes_read_total = 0;
long bytes_read = 0;
// read one byte at a time
while((bytes_read = read(sock, msg_buf_write_ptr, 1)) > 0) {
bytes_read_total += bytes_read;
// check if we have \n
if(strcmp(msg_buf_write_ptr, "\n") == 0) {
// printf("Received \\n.\n");
break;
}
msg_buf_write_ptr += bytes_read;
// have we got MAX_CLIENT_MSG_SIZE bytes and no \n?
if(bytes_read_total == config_get_number("Client_Max_Msg_Size")) {
write(sock, "err", 3);
printf(ANSI_COLOR_RED
"Client tried to write too much request data - potential"
"buffer overflow exploit attempt!\n" ANSI_RESET);
write(sock, "overflow", 8);
// Close socket
free(msg_buf);
close(sock);
pthread_exit(NULL);
}
}
// printf("Read %li bytes from client.\n%s\n", bytes_read_total, msg_buf);
// try and parse JSON
json_error_t json_error;
json_t *parsed = json_loads(msg_buf, 0, &json_error);
free(msg_buf);
if(!parsed) {
write(sock, "err", 3);
printf("Error parsing JSON at line %i: %s\n", json_error.line, json_error.text);
// free memory, close socket
free(msg_buf);
close(sock);
pthread_exit(NULL);
} else {
if(!json_is_object(parsed)) {
write(sock, "err_type", 8);
printf("Expected JSON object, got %i", parsed->type);
// free memory, close socket
free(parsed);
close(sock);
pthread_exit(NULL);
}
}
// allocate memory for the message
push_msg *message = malloc(sizeof(push_msg));
memset(message, 0x00, sizeof(push_msg));
if(message == NULL) {
write(sock, "nomem", 5);
printf("Could not allocate push_msg object!\n");
// free memory, close socket
free(parsed);
close(sock);
pthread_exit(NULL);
}
message->text = copy_json_info(json_object_get(parsed, "text"));
message->sound = copy_json_info(json_object_get(parsed, "sound"));
message->badgeNumber = (int) json_integer_value(json_object_get(parsed, "badge"));
message->custPayload = copy_json_info(json_object_get(parsed, "custom"));
message->deviceID = copy_json_info(json_object_get(parsed, "key"));
message->buttonTitle = copy_json_info(json_object_get(parsed, "btnTitle"));
message->localized_template = copy_json_info(json_object_get(parsed, "localized_text"));
message->localized_arguments.argument0 = copy_json_info(json_object_get(parsed, "localized_arg0"));
message->localized_arguments.argument1 = copy_json_info(json_object_get(parsed, "localized_arg1"));
message->localized_arguments.argument2 = copy_json_info(json_object_get(parsed, "localized_arg2"));
message->localized_arguments.argument3 = copy_json_info(json_object_get(parsed, "localized_arg3"));
write(sock, "ok", 2);
// Free memory, close socket
free(parsed);
close(sock);
// printf("Text: %s\nDevice: %s\n", message->text, message->deviceID);
int error = msg_queue_insert(message); // shove message into queue
if(error) {
printf("Error adding to queue: %i\n\n", error);
}
fflush(stdout);
pthread_exit(NULL);
}
int main(int argc, char **argv) {
int i;
int config_loaded = 0;
int dont_fork = 0;
size_t default_stacksize;
// set the name for logging
program_name = "netdata";
// parse depercated options
// TODO: Remove this block with the next major release.
{
i = 1;
while(i < argc) {
if(strcmp(argv[i], "-pidfile") == 0 && (i+1) < argc) {
strncpyz(pidfile, argv[i+1], FILENAME_MAX);
fprintf(stderr, "%s: deprecated option -- %s -- please use -P instead.\n", argv[0], argv[i]);
remove_option(i, &argc, argv);
}
else if(strcmp(argv[i], "-nodaemon") == 0 || strcmp(argv[i], "-nd") == 0) {
dont_fork = 1;
fprintf(stderr, "%s: deprecated option -- %s -- please use -D instead.\n ", argv[0], argv[i]);
remove_option(i, &argc, argv);
}
else if(strcmp(argv[i], "-ch") == 0 && (i+1) < argc) {
config_set(CONFIG_SECTION_GLOBAL, "host access prefix", argv[i+1]);
fprintf(stderr, "%s: deprecated option -- %s -- please use -s instead.\n", argv[0], argv[i]);
remove_option(i, &argc, argv);
}
else if(strcmp(argv[i], "-l") == 0 && (i+1) < argc) {
config_set(CONFIG_SECTION_GLOBAL, "history", argv[i+1]);
fprintf(stderr, "%s: deprecated option -- %s -- This option will be removed with V2.*.\n", argv[0], argv[i]);
remove_option(i, &argc, argv);
}
else i++;
}
}
// parse options
{
int num_opts = sizeof(option_definitions) / sizeof(struct option_def);
char optstring[(num_opts * 2) + 1];
int string_i = 0;
for( i = 0; i < num_opts; i++ ) {
optstring[string_i] = option_definitions[i].val;
string_i++;
if(option_definitions[i].arg_name) {
optstring[string_i] = ':';
string_i++;
}
}
// terminate optstring
optstring[string_i] ='\0';
optstring[(num_opts *2)] ='\0';
int opt;
while( (opt = getopt(argc, argv, optstring)) != -1 ) {
switch(opt) {
case 'c':
if(config_load(optarg, 1) != 1) {
error("Cannot load configuration file %s.", optarg);
return 1;
}
else {
debug(D_OPTIONS, "Configuration loaded from %s.", optarg);
config_loaded = 1;
}
break;
case 'D':
dont_fork = 1;
break;
case 'h':
return help(0);
case 'i':
config_set(CONFIG_SECTION_WEB, "bind to", optarg);
break;
case 'P':
strncpy(pidfile, optarg, FILENAME_MAX);
pidfile[FILENAME_MAX] = '\0';
break;
case 'p':
config_set(CONFIG_SECTION_GLOBAL, "default port", optarg);
break;
case 's':
config_set(CONFIG_SECTION_GLOBAL, "host access prefix", optarg);
break;
case 't':
config_set(CONFIG_SECTION_GLOBAL, "update every", optarg);
break;
case 'u':
config_set(CONFIG_SECTION_GLOBAL, "run as user", optarg);
break;
case 'v':
case 'V':
printf("%s %s\n", program_name, program_version);
return 0;
case 'W':
{
char* stacksize_string = "stacksize=";
char* debug_flags_string = "debug_flags=";
if(strcmp(optarg, "unittest") == 0) {
if(unit_test_buffer()) return 1;
if(unit_test_str2ld()) return 1;
//default_rrd_update_every = 1;
//default_rrd_memory_mode = RRD_MEMORY_MODE_RAM;
//if(!config_loaded) config_load(NULL, 0);
get_netdata_configured_variables();
default_rrd_update_every = 1;
default_rrd_memory_mode = RRD_MEMORY_MODE_RAM;
default_health_enabled = 0;
rrd_init("unittest");
default_rrdpush_enabled = 0;
if(run_all_mockup_tests()) return 1;
if(unit_test_storage()) return 1;
fprintf(stderr, "\n\nALL TESTS PASSED\n\n");
return 0;
}
else if(strcmp(optarg, "simple-pattern") == 0) {
if(optind + 2 > argc) {
fprintf(stderr, "%s", "\nUSAGE: -W simple-pattern 'pattern' 'string'\n\n"
" Checks if 'pattern' matches the given 'string'.\n"
" - 'pattern' can be one or more space separated words.\n"
" - each 'word' can contain one or more asterisks.\n"
" - words starting with '!' give negative matches.\n"
" - words are processed left to right\n"
"\n"
"Examples:\n"
"\n"
" > match all veth interfaces, except veth0:\n"
"\n"
" -W simple-pattern '!veth0 veth*' 'veth12'\n"
"\n"
"\n"
" > match all *.ext files directly in /path/:\n"
" (this will not match *.ext files in a subdir of /path/)\n"
"\n"
" -W simple-pattern '!/path/*/*.ext /path/*.ext' '/path/test.ext'\n"
"\n"
);
return 1;
}
const char *heystack = argv[optind];
const char *needle = argv[optind + 1];
size_t len = strlen(needle) + 1;
char wildcarded[len];
SIMPLE_PATTERN *p = simple_pattern_create(heystack, NULL, SIMPLE_PATTERN_EXACT);
int ret = simple_pattern_matches_extract(p, needle, wildcarded, len);
simple_pattern_free(p);
if(ret) {
fprintf(stdout, "RESULT: MATCHED - pattern '%s' matches '%s', wildcarded '%s'\n", heystack, needle, wildcarded);
return 0;
}
else {
fprintf(stdout, "RESULT: NOT MATCHED - pattern '%s' does not match '%s', wildcarded '%s'\n", heystack, needle, wildcarded);
return 1;
}
}
else if(strncmp(optarg, stacksize_string, strlen(stacksize_string)) == 0) {
optarg += strlen(stacksize_string);
config_set(CONFIG_SECTION_GLOBAL, "pthread stack size", optarg);
}
else if(strncmp(optarg, debug_flags_string, strlen(debug_flags_string)) == 0) {
optarg += strlen(debug_flags_string);
config_set(CONFIG_SECTION_GLOBAL, "debug flags", optarg);
debug_flags = strtoull(optarg, NULL, 0);
}
else if(strcmp(optarg, "set") == 0) {
if(optind + 3 > argc) {
fprintf(stderr, "%s", "\nUSAGE: -W set 'section' 'key' 'value'\n\n"
" Overwrites settings of netdata.conf.\n"
"\n"
" These options interact with: -c netdata.conf\n"
" If -c netdata.conf is given on the command line,\n"
" before -W set... the user may overwrite command\n"
" line parameters at netdata.conf\n"
" If -c netdata.conf is given after (or missing)\n"
" -W set... the user cannot overwrite the command line\n"
" parameters."
"\n"
);
return 1;
}
const char *section = argv[optind];
const char *key = argv[optind + 1];
const char *value = argv[optind + 2];
optind += 3;
// set this one as the default
// only if it is not already set in the config file
// so the caller can use -c netdata.conf before or
// after this parameter to prevent or allow overwriting
// variables at netdata.conf
config_set_default(section, key, value);
// fprintf(stderr, "SET section '%s', key '%s', value '%s'\n", section, key, value);
}
else if(strcmp(optarg, "get") == 0) {
if(optind + 3 > argc) {
fprintf(stderr, "%s", "\nUSAGE: -W get 'section' 'key' 'value'\n\n"
" Prints settings of netdata.conf.\n"
"\n"
" These options interact with: -c netdata.conf\n"
" -c netdata.conf has to be given before -W get.\n"
"\n"
);
return 1;
}
if(!config_loaded) {
fprintf(stderr, "warning: no configuration file has been loaded. Use -c CONFIG_FILE, before -W get. Using default config.\n");
config_load(NULL, 0);
}
backwards_compatible_config();
get_netdata_configured_variables();
const char *section = argv[optind];
const char *key = argv[optind + 1];
const char *def = argv[optind + 2];
const char *value = config_get(section, key, def);
printf("%s\n", value);
return 0;
}
else {
fprintf(stderr, "Unknown -W parameter '%s'\n", optarg);
return help(1);
}
}
break;
default: /* ? */
fprintf(stderr, "Unknown parameter '%c'\n", opt);
return help(1);
}
}
}
#ifdef _SC_OPEN_MAX
// close all open file descriptors, except the standard ones
// the caller may have left open files (lxc-attach has this issue)
{
int fd;
for(fd = (int) (sysconf(_SC_OPEN_MAX) - 1); fd > 2; fd--)
if(fd_is_valid(fd)) close(fd);
}
#endif
if(!config_loaded)
config_load(NULL, 0);
// ------------------------------------------------------------------------
// initialize netdata
{
char *pmax = config_get(CONFIG_SECTION_GLOBAL, "glibc malloc arena max for plugins", "1");
if(pmax && *pmax)
setenv("MALLOC_ARENA_MAX", pmax, 1);
#if defined(HAVE_C_MALLOPT)
i = (int)config_get_number(CONFIG_SECTION_GLOBAL, "glibc malloc arena max for netdata", 1);
if(i > 0)
mallopt(M_ARENA_MAX, 1);
#endif
// prepare configuration environment variables for the plugins
get_netdata_configured_variables();
set_global_environment();
// work while we are cd into config_dir
// to allow the plugins refer to their config
// files using relative filenames
if(chdir(netdata_configured_config_dir) == -1)
fatal("Cannot cd to '%s'", netdata_configured_config_dir);
}
char *user = NULL;
{
// --------------------------------------------------------------------
// get the debugging flags from the configuration file
char *flags = config_get(CONFIG_SECTION_GLOBAL, "debug flags", "0x0000000000000000");
setenv("NETDATA_DEBUG_FLAGS", flags, 1);
debug_flags = strtoull(flags, NULL, 0);
debug(D_OPTIONS, "Debug flags set to '0x%" PRIX64 "'.", debug_flags);
if(debug_flags != 0) {
struct rlimit rl = { RLIM_INFINITY, RLIM_INFINITY };
if(setrlimit(RLIMIT_CORE, &rl) != 0)
error("Cannot request unlimited core dumps for debugging... Proceeding anyway...");
#ifdef HAVE_SYS_PRCTL_H
prctl(PR_SET_DUMPABLE, 1, 0, 0, 0);
#endif
}
// --------------------------------------------------------------------
// get log filenames and settings
log_init();
error_log_limit_unlimited();
// --------------------------------------------------------------------
// load stream.conf
{
char filename[FILENAME_MAX + 1];
snprintfz(filename, FILENAME_MAX, "%s/stream.conf", netdata_configured_config_dir);
appconfig_load(&stream_config, filename, 0);
}
// --------------------------------------------------------------------
// setup process signals
// block signals while initializing threads.
// this causes the threads to block signals.
signals_block();
// setup the signals we want to use
signals_init();
// setup threads configs
default_stacksize = netdata_threads_init();
// --------------------------------------------------------------------
// check which threads are enabled and initialize them
for (i = 0; static_threads[i].name != NULL ; i++) {
struct netdata_static_thread *st = &static_threads[i];
if(st->config_name)
st->enabled = config_get_boolean(st->config_section, st->config_name, st->enabled);
if(st->enabled && st->init_routine)
st->init_routine();
}
// --------------------------------------------------------------------
// get the user we should run
// IMPORTANT: this is required before web_files_uid()
if(getuid() == 0) {
user = config_get(CONFIG_SECTION_GLOBAL, "run as user", NETDATA_USER);
}
else {
struct passwd *passwd = getpwuid(getuid());
user = config_get(CONFIG_SECTION_GLOBAL, "run as user", (passwd && passwd->pw_name)?passwd->pw_name:"");
}
// --------------------------------------------------------------------
// create the listening sockets
web_client_api_v1_init();
web_server_threading_selection();
if(web_server_mode != WEB_SERVER_MODE_NONE)
api_listen_sockets_setup();
}
// initialize the log files
open_all_log_files();
#ifdef NETDATA_INTERNAL_CHECKS
if(debug_flags != 0) {
struct rlimit rl = { RLIM_INFINITY, RLIM_INFINITY };
if(setrlimit(RLIMIT_CORE, &rl) != 0)
error("Cannot request unlimited core dumps for debugging... Proceeding anyway...");
#ifdef HAVE_SYS_PRCTL_H
prctl(PR_SET_DUMPABLE, 1, 0, 0, 0);
#endif
}
#endif /* NETDATA_INTERNAL_CHECKS */
// get the max file limit
if(getrlimit(RLIMIT_NOFILE, &rlimit_nofile) != 0)
error("getrlimit(RLIMIT_NOFILE) failed");
else
info("resources control: allowed file descriptors: soft = %zu, max = %zu", rlimit_nofile.rlim_cur, rlimit_nofile.rlim_max);
// fork, switch user, create pid file, set process priority
if(become_daemon(dont_fork, user) == -1)
fatal("Cannot daemonize myself.");
info("netdata started on pid %d.", getpid());
// IMPORTANT: these have to run once, while single threaded
// but after we have switched user
web_files_uid();
web_files_gid();
netdata_threads_init_after_fork((size_t)config_get_number(CONFIG_SECTION_GLOBAL, "pthread stack size", (long)default_stacksize));
// ------------------------------------------------------------------------
// initialize rrd, registry, health, rrdpush, etc.
rrd_init(netdata_configured_hostname);
// ------------------------------------------------------------------------
// enable log flood protection
error_log_limit_reset();
// ------------------------------------------------------------------------
// spawn the threads
web_server_config_options();
for (i = 0; static_threads[i].name != NULL ; i++) {
struct netdata_static_thread *st = &static_threads[i];
if(st->enabled) {
st->thread = mallocz(sizeof(netdata_thread_t));
debug(D_SYSTEM, "Starting thread %s.", st->name);
netdata_thread_create(st->thread, st->name, NETDATA_THREAD_OPTION_DEFAULT, st->start_routine, st);
}
else debug(D_SYSTEM, "Not starting thread %s.", st->name);
}
info("netdata initialization completed. Enjoy real-time performance monitoring!");
// ------------------------------------------------------------------------
// unblock signals
signals_unblock();
// ------------------------------------------------------------------------
// Handle signals
signals_handle();
// should never reach this point
// but we need it for rpmlint #2752
return 1;
}
void *pluginsd_main(void *ptr)
{
if(ptr) { ; }
info("PLUGINS.D thread created with task id %d", gettid());
if(pthread_setcanceltype(PTHREAD_CANCEL_DEFERRED, NULL) != 0)
error("Cannot set pthread cancel type to DEFERRED.");
if(pthread_setcancelstate(PTHREAD_CANCEL_ENABLE, NULL) != 0)
error("Cannot set pthread cancel state to ENABLE.");
char *dir_name = config_get("plugins", "plugins directory", PLUGINS_DIR);
int automatic_run = config_get_boolean("plugins", "enable running new plugins", 1);
int scan_frequency = (int) config_get_number("plugins", "check for new plugins every", 60);
DIR *dir = NULL;
struct dirent *file = NULL;
struct plugind *cd;
// enable the apps plugin by default
// config_get_boolean("plugins", "apps", 1);
if(scan_frequency < 1) scan_frequency = 1;
while(likely(1)) {
if(unlikely(netdata_exit)) break;
dir = opendir(dir_name);
if(unlikely(!dir)) {
error("Cannot open directory '%s'.", dir_name);
return NULL;
}
while(likely((file = readdir(dir)))) {
if(unlikely(netdata_exit)) break;
debug(D_PLUGINSD, "PLUGINSD: Examining file '%s'", file->d_name);
if(unlikely(strcmp(file->d_name, ".") == 0 || strcmp(file->d_name, "..") == 0)) continue;
int len = (int) strlen(file->d_name);
if(unlikely(len <= (int)PLUGINSD_FILE_SUFFIX_LEN)) continue;
if(unlikely(strcmp(PLUGINSD_FILE_SUFFIX, &file->d_name[len - (int)PLUGINSD_FILE_SUFFIX_LEN]) != 0)) {
debug(D_PLUGINSD, "PLUGINSD: File '%s' does not end in '%s'.", file->d_name, PLUGINSD_FILE_SUFFIX);
continue;
}
char pluginname[CONFIG_MAX_NAME + 1];
snprintf(pluginname, CONFIG_MAX_NAME, "%.*s", (int)(len - PLUGINSD_FILE_SUFFIX_LEN), file->d_name);
int enabled = config_get_boolean("plugins", pluginname, automatic_run);
if(unlikely(!enabled)) {
debug(D_PLUGINSD, "PLUGINSD: plugin '%s' is not enabled", file->d_name);
continue;
}
// check if it runs already
for(cd = pluginsd_root ; likely(cd) ; cd = cd->next) {
if(unlikely(strcmp(cd->filename, file->d_name) == 0)) break;
}
if(likely(cd && !cd->obsolete)) {
debug(D_PLUGINSD, "PLUGINSD: plugin '%s' is already running", cd->filename);
continue;
}
// it is not running
// allocate a new one, or use the obsolete one
if(unlikely(!cd)) {
cd = calloc(sizeof(struct plugind), 1);
if(unlikely(!cd)) fatal("Cannot allocate memory for plugin.");
snprintf(cd->id, CONFIG_MAX_NAME, "plugin:%s", pluginname);
strncpy(cd->filename, file->d_name, FILENAME_MAX);
snprintf(cd->fullfilename, FILENAME_MAX, "%s/%s", dir_name, cd->filename);
cd->enabled = enabled;
cd->update_every = (int) config_get_number(cd->id, "update every", rrd_update_every);
cd->started_t = time(NULL);
char *def = "";
snprintf(cd->cmd, PLUGINSD_CMD_MAX, "exec %s %d %s", cd->fullfilename, cd->update_every, config_get(cd->id, "command options", def));
// link it
if(likely(pluginsd_root)) cd->next = pluginsd_root;
pluginsd_root = cd;
}
cd->obsolete = 0;
if(unlikely(!cd->enabled)) continue;
// spawn a new thread for it
if(unlikely(pthread_create(&cd->thread, NULL, pluginsd_worker_thread, cd) != 0)) {
error("PLUGINSD: failed to create new thread for plugin '%s'.", cd->filename);
cd->obsolete = 1;
}
else if(unlikely(pthread_detach(cd->thread) != 0))
error("PLUGINSD: Cannot request detach of newly created thread for plugin '%s'.", cd->filename);
}
closedir(dir);
sleep((unsigned int) scan_frequency);
}
return NULL;
}
void *proc_diskspace_main(void *ptr) {
netdata_thread_cleanup_push(diskspace_main_cleanup, ptr);
int vdo_cpu_netdata = config_get_boolean("plugin:proc", "netdata server resources", 1);
cleanup_mount_points = config_get_boolean(CONFIG_SECTION_DISKSPACE, "remove charts of unmounted disks" , cleanup_mount_points);
int update_every = (int)config_get_number(CONFIG_SECTION_DISKSPACE, "update every", localhost->rrd_update_every);
if(update_every < localhost->rrd_update_every)
update_every = localhost->rrd_update_every;
check_for_new_mountpoints_every = (int)config_get_number(CONFIG_SECTION_DISKSPACE, "check for new mount points every", check_for_new_mountpoints_every);
if(check_for_new_mountpoints_every < update_every)
check_for_new_mountpoints_every = update_every;
struct rusage thread;
usec_t duration = 0;
usec_t step = update_every * USEC_PER_SEC;
heartbeat_t hb;
heartbeat_init(&hb);
while(!netdata_exit) {
duration = heartbeat_dt_usec(&hb);
/* usec_t hb_dt = */ heartbeat_next(&hb, step);
if(unlikely(netdata_exit)) break;
// --------------------------------------------------------------------------
// this is smart enough not to reload it every time
mountinfo_reload(0);
// --------------------------------------------------------------------------
// disk space metrics
struct mountinfo *mi;
for(mi = disk_mountinfo_root; mi; mi = mi->next) {
if(unlikely(mi->flags & (MOUNTINFO_IS_DUMMY | MOUNTINFO_IS_BIND)))
continue;
do_disk_space_stats(mi, update_every);
if(unlikely(netdata_exit)) break;
}
if(unlikely(netdata_exit)) break;
if(dict_mountpoints)
dictionary_get_all(dict_mountpoints, mount_point_cleanup, NULL);
if(vdo_cpu_netdata) {
static RRDSET *stcpu_thread = NULL, *st_duration = NULL;
static RRDDIM *rd_user = NULL, *rd_system = NULL, *rd_duration = NULL;
// ----------------------------------------------------------------
getrusage(RUSAGE_THREAD, &thread);
if(unlikely(!stcpu_thread)) {
stcpu_thread = rrdset_create_localhost(
"netdata"
, "plugin_diskspace"
, NULL
, "diskspace"
, NULL
, "NetData Disk Space Plugin CPU usage"
, "milliseconds/s"
, "diskspace"
, NULL
, 132020
, update_every
, RRDSET_TYPE_STACKED
);
rd_user = rrddim_add(stcpu_thread, "user", NULL, 1, 1000, RRD_ALGORITHM_INCREMENTAL);
rd_system = rrddim_add(stcpu_thread, "system", NULL, 1, 1000, RRD_ALGORITHM_INCREMENTAL);
}
else
rrdset_next(stcpu_thread);
rrddim_set_by_pointer(stcpu_thread, rd_user, thread.ru_utime.tv_sec * 1000000ULL + thread.ru_utime.tv_usec);
rrddim_set_by_pointer(stcpu_thread, rd_system, thread.ru_stime.tv_sec * 1000000ULL + thread.ru_stime.tv_usec);
rrdset_done(stcpu_thread);
// ----------------------------------------------------------------
if(unlikely(!st_duration)) {
st_duration = rrdset_create_localhost(
"netdata"
, "plugin_diskspace_dt"
, NULL
, "diskspace"
, NULL
, "NetData Disk Space Plugin Duration"
, "milliseconds/run"
, "diskspace"
, NULL
, 132021
, update_every
, RRDSET_TYPE_AREA
);
rd_duration = rrddim_add(st_duration, "duration", NULL, 1, 1000, RRD_ALGORITHM_ABSOLUTE);
}
else
rrdset_next(st_duration);
rrddim_set_by_pointer(st_duration, rd_duration, duration);
rrdset_done(st_duration);
// ----------------------------------------------------------------
if(unlikely(netdata_exit)) break;
}
}
netdata_thread_cleanup_pop(1);
return NULL;
}
RRDHOST *rrdhost_create(const char *hostname,
const char *guid,
const char *os,
int update_every,
long entries,
RRD_MEMORY_MODE memory_mode,
int health_enabled,
int rrdpush_enabled,
char *rrdpush_destination,
char *rrdpush_api_key,
int is_localhost
) {
debug(D_RRDHOST, "Host '%s': adding with guid '%s'", hostname, guid);
rrd_check_wrlock();
RRDHOST *host = callocz(1, sizeof(RRDHOST));
host->rrd_update_every = (update_every > 0)?update_every:1;
host->rrd_history_entries = align_entries_to_pagesize(memory_mode, entries);
host->rrd_memory_mode = memory_mode;
host->health_enabled = (memory_mode == RRD_MEMORY_MODE_NONE)? 0 : health_enabled;
host->rrdpush_enabled = (rrdpush_enabled && rrdpush_destination && *rrdpush_destination && rrdpush_api_key && *rrdpush_api_key);
host->rrdpush_destination = (host->rrdpush_enabled)?strdupz(rrdpush_destination):NULL;
host->rrdpush_api_key = (host->rrdpush_enabled)?strdupz(rrdpush_api_key):NULL;
host->rrdpush_pipe[0] = -1;
host->rrdpush_pipe[1] = -1;
host->rrdpush_socket = -1;
netdata_mutex_init(&host->rrdpush_mutex);
netdata_rwlock_init(&host->rrdhost_rwlock);
rrdhost_init_hostname(host, hostname);
rrdhost_init_machine_guid(host, guid);
rrdhost_init_os(host, os);
avl_init_lock(&(host->rrdset_root_index), rrdset_compare);
avl_init_lock(&(host->rrdset_root_index_name), rrdset_compare_name);
avl_init_lock(&(host->rrdfamily_root_index), rrdfamily_compare);
avl_init_lock(&(host->variables_root_index), rrdvar_compare);
if(config_get_boolean(CONFIG_SECTION_GLOBAL, "delete obsolete charts files", 1))
rrdhost_flag_set(host, RRDHOST_DELETE_OBSOLETE_FILES);
if(config_get_boolean(CONFIG_SECTION_GLOBAL, "delete orphan hosts files", 1) && !is_localhost)
rrdhost_flag_set(host, RRDHOST_DELETE_ORPHAN_FILES);
// ------------------------------------------------------------------------
// initialize health variables
host->health_log.next_log_id = 1;
host->health_log.next_alarm_id = 1;
host->health_log.max = 1000;
host->health_log.next_log_id =
host->health_log.next_alarm_id = (uint32_t)now_realtime_sec();
long n = config_get_number(CONFIG_SECTION_HEALTH, "in memory max health log entries", host->health_log.max);
if(n < 10) {
error("Host '%s': health configuration has invalid max log entries %ld. Using default %u", host->hostname, n, host->health_log.max);
config_set_number(CONFIG_SECTION_HEALTH, "in memory max health log entries", (long)host->health_log.max);
}
else
host->health_log.max = (unsigned int)n;
netdata_rwlock_init(&host->health_log.alarm_log_rwlock);
char filename[FILENAME_MAX + 1];
if(is_localhost) {
host->cache_dir = strdupz(netdata_configured_cache_dir);
host->varlib_dir = strdupz(netdata_configured_varlib_dir);
}
else {
// this is not localhost - append our GUID to localhost path
snprintfz(filename, FILENAME_MAX, "%s/%s", netdata_configured_cache_dir, host->machine_guid);
host->cache_dir = strdupz(filename);
if(host->rrd_memory_mode == RRD_MEMORY_MODE_MAP || host->rrd_memory_mode == RRD_MEMORY_MODE_SAVE) {
int r = mkdir(host->cache_dir, 0775);
if(r != 0 && errno != EEXIST)
error("Host '%s': cannot create directory '%s'", host->hostname, host->cache_dir);
}
snprintfz(filename, FILENAME_MAX, "%s/%s", netdata_configured_varlib_dir, host->machine_guid);
host->varlib_dir = strdupz(filename);
if(host->health_enabled) {
int r = mkdir(host->varlib_dir, 0775);
if(r != 0 && errno != EEXIST)
error("Host '%s': cannot create directory '%s'", host->hostname, host->varlib_dir);
}
}
if(host->health_enabled) {
snprintfz(filename, FILENAME_MAX, "%s/health", host->varlib_dir);
int r = mkdir(filename, 0775);
if(r != 0 && errno != EEXIST)
error("Host '%s': cannot create directory '%s'", host->hostname, filename);
}
snprintfz(filename, FILENAME_MAX, "%s/health/health-log.db", host->varlib_dir);
host->health_log_filename = strdupz(filename);
snprintfz(filename, FILENAME_MAX, "%s/alarm-notify.sh", netdata_configured_plugins_dir);
host->health_default_exec = strdupz(config_get(CONFIG_SECTION_HEALTH, "script to execute on alarm", filename));
host->health_default_recipient = strdup("root");
// ------------------------------------------------------------------------
// load health configuration
if(host->health_enabled) {
health_alarm_log_load(host);
health_alarm_log_open(host);
rrdhost_wrlock(host);
health_readdir(host, health_config_dir());
rrdhost_unlock(host);
}
// ------------------------------------------------------------------------
// link it and add it to the index
if(is_localhost) {
host->next = localhost;
localhost = host;
}
else {
if(localhost) {
host->next = localhost->next;
localhost->next = host;
}
else localhost = host;
}
RRDHOST *t = rrdhost_index_add(host);
if(t != host) {
error("Host '%s': cannot add host with machine guid '%s' to index. It already exists as host '%s' with machine guid '%s'.", host->hostname, host->machine_guid, t->hostname, t->machine_guid);
rrdhost_free(host);
host = NULL;
}
else {
info("Host '%s' with guid '%s' initialized"
", os %s"
", update every %d"
", memory mode %s"
", history entries %ld"
", streaming %s"
" (to '%s' with api key '%s')"
", health %s"
", cache_dir '%s'"
", varlib_dir '%s'"
", health_log '%s'"
", alarms default handler '%s'"
", alarms default recipient '%s'"
, host->hostname
, host->machine_guid
, host->os
, host->rrd_update_every
, rrd_memory_mode_name(host->rrd_memory_mode)
, host->rrd_history_entries
, host->rrdpush_enabled?"enabled":"disabled"
, host->rrdpush_destination?host->rrdpush_destination:""
, host->rrdpush_api_key?host->rrdpush_api_key:""
, host->health_enabled?"enabled":"disabled"
, host->cache_dir
, host->varlib_dir
, host->health_log_filename
, host->health_default_exec
, host->health_default_recipient
);
}
rrd_hosts_available++;
return host;
}
static void get_netdata_configured_variables() {
backwards_compatible_config();
// ------------------------------------------------------------------------
// get the hostname
char buf[HOSTNAME_MAX + 1];
if(gethostname(buf, HOSTNAME_MAX) == -1)
error("Cannot get machine hostname.");
netdata_configured_hostname = config_get(CONFIG_SECTION_GLOBAL, "hostname", buf);
debug(D_OPTIONS, "hostname set to '%s'", netdata_configured_hostname);
// ------------------------------------------------------------------------
// get default database size
default_rrd_history_entries = (int) config_get_number(CONFIG_SECTION_GLOBAL, "history", align_entries_to_pagesize(default_rrd_memory_mode, RRD_DEFAULT_HISTORY_ENTRIES));
long h = align_entries_to_pagesize(default_rrd_memory_mode, default_rrd_history_entries);
if(h != default_rrd_history_entries) {
config_set_number(CONFIG_SECTION_GLOBAL, "history", h);
default_rrd_history_entries = (int)h;
}
if(default_rrd_history_entries < 5 || default_rrd_history_entries > RRD_HISTORY_ENTRIES_MAX) {
error("Invalid history entries %d given. Defaulting to %d.", default_rrd_history_entries, RRD_DEFAULT_HISTORY_ENTRIES);
default_rrd_history_entries = RRD_DEFAULT_HISTORY_ENTRIES;
}
// ------------------------------------------------------------------------
// get default database update frequency
default_rrd_update_every = (int) config_get_number(CONFIG_SECTION_GLOBAL, "update every", UPDATE_EVERY);
if(default_rrd_update_every < 1 || default_rrd_update_every > 600) {
error("Invalid data collection frequency (update every) %d given. Defaulting to %d.", default_rrd_update_every, UPDATE_EVERY_MAX);
default_rrd_update_every = UPDATE_EVERY;
}
// ------------------------------------------------------------------------
// get system paths
netdata_configured_config_dir = config_get(CONFIG_SECTION_GLOBAL, "config directory", CONFIG_DIR);
netdata_configured_log_dir = config_get(CONFIG_SECTION_GLOBAL, "log directory", LOG_DIR);
netdata_configured_web_dir = config_get(CONFIG_SECTION_GLOBAL, "web files directory", WEB_DIR);
netdata_configured_cache_dir = config_get(CONFIG_SECTION_GLOBAL, "cache directory", CACHE_DIR);
netdata_configured_varlib_dir = config_get(CONFIG_SECTION_GLOBAL, "lib directory", VARLIB_DIR);
netdata_configured_home_dir = config_get(CONFIG_SECTION_GLOBAL, "home directory", CACHE_DIR);
{
char plugins_dirs[(FILENAME_MAX * 2) + 1];
snprintfz(plugins_dirs, FILENAME_MAX * 2, "\"%s\" \"%s/custom-plugins.d\"", PLUGINS_DIR, CONFIG_DIR);
netdata_configured_plugins_dir_base = strdupz(config_get(CONFIG_SECTION_GLOBAL, "plugins directory", plugins_dirs));
quoted_strings_splitter(netdata_configured_plugins_dir_base, plugin_directories, PLUGINSD_MAX_DIRECTORIES, config_isspace);
netdata_configured_plugins_dir = plugin_directories[0];
}
// ------------------------------------------------------------------------
// get default memory mode for the database
default_rrd_memory_mode = rrd_memory_mode_id(config_get(CONFIG_SECTION_GLOBAL, "memory mode", rrd_memory_mode_name(default_rrd_memory_mode)));
// ------------------------------------------------------------------------
netdata_configured_host_prefix = config_get(CONFIG_SECTION_GLOBAL, "host access prefix", "");
verify_netdata_host_prefix();
// --------------------------------------------------------------------
// get KSM settings
#ifdef MADV_MERGEABLE
enable_ksm = config_get_boolean(CONFIG_SECTION_GLOBAL, "memory deduplication (ksm)", enable_ksm);
#endif
// --------------------------------------------------------------------
// get various system parameters
get_system_HZ();
get_system_cpus();
get_system_pid_max();
}
int registry_init(void) {
char filename[FILENAME_MAX + 1];
// registry enabled?
if(web_server_mode != WEB_SERVER_MODE_NONE) {
registry.enabled = config_get_boolean(CONFIG_SECTION_REGISTRY, "enabled", 0);
}
else {
info("Registry is disabled - use the central netdata");
config_set_boolean(CONFIG_SECTION_REGISTRY, "enabled", 0);
registry.enabled = 0;
}
// pathnames
snprintfz(filename, FILENAME_MAX, "%s/registry", netdata_configured_varlib_dir);
registry.pathname = config_get(CONFIG_SECTION_REGISTRY, "registry db directory", filename);
if(mkdir(registry.pathname, 0770) == -1 && errno != EEXIST)
fatal("Cannot create directory '%s'.", registry.pathname);
// filenames
snprintfz(filename, FILENAME_MAX, "%s/netdata.public.unique.id", registry.pathname);
registry.machine_guid_filename = config_get(CONFIG_SECTION_REGISTRY, "netdata unique id file", filename);
snprintfz(filename, FILENAME_MAX, "%s/registry.db", registry.pathname);
registry.db_filename = config_get(CONFIG_SECTION_REGISTRY, "registry db file", filename);
snprintfz(filename, FILENAME_MAX, "%s/registry-log.db", registry.pathname);
registry.log_filename = config_get(CONFIG_SECTION_REGISTRY, "registry log file", filename);
// configuration options
registry.save_registry_every_entries = (unsigned long long)config_get_number(CONFIG_SECTION_REGISTRY, "registry save db every new entries", 1000000);
registry.persons_expiration = config_get_number(CONFIG_SECTION_REGISTRY, "registry expire idle persons days", 365) * 86400;
registry.registry_domain = config_get(CONFIG_SECTION_REGISTRY, "registry domain", "");
registry.registry_to_announce = config_get(CONFIG_SECTION_REGISTRY, "registry to announce", "https://registry.my-netdata.io");
registry.hostname = config_get(CONFIG_SECTION_REGISTRY, "registry hostname", netdata_configured_hostname);
registry.verify_cookies_redirects = config_get_boolean(CONFIG_SECTION_REGISTRY, "verify browser cookies support", 1);
setenv("NETDATA_REGISTRY_HOSTNAME", registry.hostname, 1);
setenv("NETDATA_REGISTRY_URL", registry.registry_to_announce, 1);
registry.max_url_length = (size_t)config_get_number(CONFIG_SECTION_REGISTRY, "max URL length", 1024);
if(registry.max_url_length < 10) {
registry.max_url_length = 10;
config_set_number(CONFIG_SECTION_REGISTRY, "max URL length", (long long)registry.max_url_length);
}
registry.max_name_length = (size_t)config_get_number(CONFIG_SECTION_REGISTRY, "max URL name length", 50);
if(registry.max_name_length < 10) {
registry.max_name_length = 10;
config_set_number(CONFIG_SECTION_REGISTRY, "max URL name length", (long long)registry.max_name_length);
}
// initialize entries counters
registry.persons_count = 0;
registry.machines_count = 0;
registry.usages_count = 0;
registry.urls_count = 0;
registry.persons_urls_count = 0;
registry.machines_urls_count = 0;
// initialize memory counters
registry.persons_memory = 0;
registry.machines_memory = 0;
registry.urls_memory = 0;
registry.persons_urls_memory = 0;
registry.machines_urls_memory = 0;
// initialize locks
netdata_mutex_init(®istry.lock);
// create dictionaries
registry.persons = dictionary_create(DICTIONARY_FLAGS);
registry.machines = dictionary_create(DICTIONARY_FLAGS);
avl_init(®istry.registry_urls_root_index, registry_url_compare);
// load the registry database
if(registry.enabled) {
registry_log_open();
registry_db_load();
registry_log_load();
if(unlikely(registry_db_should_be_saved()))
registry_db_save();
}
return 0;
}
int main(int argc, char **argv)
{
int i;
int config_loaded = 0;
int dont_fork = 0;
size_t wanted_stacksize = 0, stacksize = 0;
pthread_attr_t attr;
// global initialization
get_HZ();
// set the name for logging
program_name = "netdata";
// parse command line.
// parse depercated options
// TODO: Remove this block with the next major release.
{
i = 1;
while(i < argc) {
if(strcmp(argv[i], "-pidfile") == 0 && (i+1) < argc) {
strncpyz(pidfile, argv[i+1], FILENAME_MAX);
fprintf(stderr, "%s: deprecated option -- %s -- please use -P instead.\n", argv[0], argv[i]);
remove_option(i, &argc, argv);
}
else if(strcmp(argv[i], "-nodaemon") == 0 || strcmp(argv[i], "-nd") == 0) {
dont_fork = 1;
fprintf(stderr, "%s: deprecated option -- %s -- please use -D instead.\n ", argv[0], argv[i]);
remove_option(i, &argc, argv);
}
else if(strcmp(argv[i], "-ch") == 0 && (i+1) < argc) {
config_set("global", "host access prefix", argv[i+1]);
fprintf(stderr, "%s: deprecated option -- %s -- please use -s instead.\n", argv[0], argv[i]);
remove_option(i, &argc, argv);
}
else if(strcmp(argv[i], "-l") == 0 && (i+1) < argc) {
config_set("global", "history", argv[i+1]);
fprintf(stderr, "%s: deprecated option -- %s -- This option will be removed with V2.*.\n", argv[0], argv[i]);
remove_option(i, &argc, argv);
}
else i++;
}
}
// parse options
{
int num_opts = sizeof(options) / sizeof(struct option_def);
char optstring[(num_opts * 2) + 1];
int string_i = 0;
for( i = 0; i < num_opts; i++ ) {
optstring[string_i] = options[i].val;
string_i++;
if(options[i].arg_name) {
optstring[string_i] = ':';
string_i++;
}
}
int opt;
while( (opt = getopt(argc, argv, optstring)) != -1 ) {
switch(opt) {
case 'c':
if(load_config(optarg, 1) != 1) {
error("Cannot load configuration file %s.", optarg);
exit(1);
}
else {
debug(D_OPTIONS, "Configuration loaded from %s.", optarg);
config_loaded = 1;
}
break;
case 'D':
dont_fork = 1;
break;
case 'h':
help(0);
break;
case 'i':
config_set("global", "bind to", optarg);
break;
case 'P':
strncpy(pidfile, optarg, FILENAME_MAX);
pidfile[FILENAME_MAX] = '\0';
break;
case 'p':
config_set("global", "default port", optarg);
break;
case 's':
config_set("global", "host access prefix", optarg);
break;
case 't':
config_set("global", "update every", optarg);
break;
case 'u':
config_set("global", "run as user", optarg);
break;
case 'v':
// TODO: Outsource version to makefile which can compute version from git.
printf("netdata 1.2.1_master\n");
return 0;
break;
case 'W':
{
char* stacksize = "stacksize=";
char* debug_flags_string = "debug_flags=";
if(strcmp(optarg, "unittest") == 0) {
rrd_update_every = 1;
if(run_all_mockup_tests()) exit(1);
if(unit_test_storage()) exit(1);
fprintf(stderr, "\n\nALL TESTS PASSED\n\n");
exit(0);
} else if(strncmp(optarg, stacksize, strlen(stacksize)) == 0) {
optarg += strlen(stacksize);
config_set("global", "pthread stack size", optarg);
} else if(strncmp(optarg, debug_flags_string, strlen(debug_flags_string)) == 0) {
optarg += strlen(debug_flags_string);
config_set("global", "debug flags", optarg);
debug_flags = strtoull(optarg, NULL, 0);
}
}
break;
default: /* ? */
help(1);
break;
}
}
}
if(!config_loaded) load_config(NULL, 0);
// prepare configuration environment variables for the plugins
setenv("NETDATA_CONFIG_DIR" , config_get("global", "config directory" , CONFIG_DIR) , 1);
setenv("NETDATA_PLUGINS_DIR", config_get("global", "plugins directory" , PLUGINS_DIR), 1);
setenv("NETDATA_WEB_DIR" , config_get("global", "web files directory", WEB_DIR) , 1);
setenv("NETDATA_CACHE_DIR" , config_get("global", "cache directory" , CACHE_DIR) , 1);
setenv("NETDATA_LIB_DIR" , config_get("global", "lib directory" , VARLIB_DIR) , 1);
setenv("NETDATA_LOG_DIR" , config_get("global", "log directory" , LOG_DIR) , 1);
setenv("NETDATA_HOST_PREFIX", config_get("global", "host access prefix" , "") , 1);
setenv("HOME" , config_get("global", "home directory" , CACHE_DIR) , 1);
// disable buffering for python plugins
setenv("PYTHONUNBUFFERED", "1", 1);
// avoid flood calls to stat(/etc/localtime)
// http://stackoverflow.com/questions/4554271/how-to-avoid-excessive-stat-etc-localtime-calls-in-strftime-on-linux
setenv("TZ", ":/etc/localtime", 0);
{
char path[1024 + 1], *p = getenv("PATH");
if(!p) p = "/bin:/usr/bin";
snprintfz(path, 1024, "%s:%s", p, "/sbin:/usr/sbin:/usr/local/bin:/usr/local/sbin");
setenv("PATH", config_get("plugins", "PATH environment variable", path), 1);
}
// cd to /tmp to avoid any plugins writing files at random places
if(chdir("/tmp")) error("netdata: ERROR: Cannot cd to /tmp");
char *input_log_file = NULL;
char *output_log_file = NULL;
char *error_log_file = NULL;
char *access_log_file = NULL;
char *user = NULL;
{
char *flags = config_get("global", "debug flags", "0x00000000");
setenv("NETDATA_DEBUG_FLAGS", flags, 1);
debug_flags = strtoull(flags, NULL, 0);
debug(D_OPTIONS, "Debug flags set to '0x%8llx'.", debug_flags);
if(debug_flags != 0) {
struct rlimit rl = { RLIM_INFINITY, RLIM_INFINITY };
if(setrlimit(RLIMIT_CORE, &rl) != 0)
info("Cannot request unlimited core dumps for debugging... Proceeding anyway...");
prctl(PR_SET_DUMPABLE, 1, 0, 0, 0);
}
// --------------------------------------------------------------------
#ifdef MADV_MERGEABLE
enable_ksm = config_get_boolean("global", "memory deduplication (ksm)", enable_ksm);
#else
#warning "Kernel memory deduplication (KSM) is not available"
#endif
// --------------------------------------------------------------------
global_host_prefix = config_get("global", "host access prefix", "");
setenv("NETDATA_HOST_PREFIX", global_host_prefix, 1);
// --------------------------------------------------------------------
output_log_file = config_get("global", "debug log", LOG_DIR "/debug.log");
if(strcmp(output_log_file, "syslog") == 0) {
output_log_syslog = 1;
output_log_file = NULL;
}
else if(strcmp(output_log_file, "none") == 0) {
output_log_syslog = 0;
output_log_file = NULL;
}
else output_log_syslog = 0;
// --------------------------------------------------------------------
error_log_file = config_get("global", "error log", LOG_DIR "/error.log");
if(strcmp(error_log_file, "syslog") == 0) {
error_log_syslog = 1;
error_log_file = NULL;
}
else if(strcmp(error_log_file, "none") == 0) {
error_log_syslog = 0;
error_log_file = NULL;
// optimization - do not even generate debug log entries
}
else error_log_syslog = 0;
error_log_throttle_period = config_get_number("global", "errors flood protection period", error_log_throttle_period);
setenv("NETDATA_ERRORS_THROTTLE_PERIOD", config_get("global", "errors flood protection period" , ""), 1);
error_log_errors_per_period = (unsigned long)config_get_number("global", "errors to trigger flood protection", error_log_errors_per_period);
setenv("NETDATA_ERRORS_PER_PERIOD" , config_get("global", "errors to trigger flood protection", ""), 1);
// --------------------------------------------------------------------
access_log_file = config_get("global", "access log", LOG_DIR "/access.log");
if(strcmp(access_log_file, "syslog") == 0) {
access_log_syslog = 1;
access_log_file = NULL;
}
else if(strcmp(access_log_file, "none") == 0) {
access_log_syslog = 0;
access_log_file = NULL;
}
else access_log_syslog = 0;
// --------------------------------------------------------------------
rrd_memory_mode = rrd_memory_mode_id(config_get("global", "memory mode", rrd_memory_mode_name(rrd_memory_mode)));
// --------------------------------------------------------------------
{
char hostnamebuf[HOSTNAME_MAX + 1];
if(gethostname(hostnamebuf, HOSTNAME_MAX) == -1)
error("WARNING: Cannot get machine hostname.");
hostname = config_get("global", "hostname", hostnamebuf);
debug(D_OPTIONS, "hostname set to '%s'", hostname);
}
// --------------------------------------------------------------------
rrd_default_history_entries = (int) config_get_number("global", "history", RRD_DEFAULT_HISTORY_ENTRIES);
if(rrd_default_history_entries < 5 || rrd_default_history_entries > RRD_HISTORY_ENTRIES_MAX) {
info("Invalid save lines %d given. Defaulting to %d.", rrd_default_history_entries, RRD_DEFAULT_HISTORY_ENTRIES);
rrd_default_history_entries = RRD_DEFAULT_HISTORY_ENTRIES;
}
else {
debug(D_OPTIONS, "save lines set to %d.", rrd_default_history_entries);
}
// --------------------------------------------------------------------
rrd_update_every = (int) config_get_number("global", "update every", UPDATE_EVERY);
if(rrd_update_every < 1 || rrd_update_every > 600) {
info("Invalid update timer %d given. Defaulting to %d.", rrd_update_every, UPDATE_EVERY_MAX);
rrd_update_every = UPDATE_EVERY;
}
else debug(D_OPTIONS, "update timer set to %d.", rrd_update_every);
// let the plugins know the min update_every
{
char buf[16];
snprintfz(buf, 15, "%d", rrd_update_every);
setenv("NETDATA_UPDATE_EVERY", buf, 1);
}
// --------------------------------------------------------------------
// block signals while initializing threads.
// this causes the threads to block signals.
sigset_t sigset;
sigfillset(&sigset);
if(pthread_sigmask(SIG_BLOCK, &sigset, NULL) == -1) {
error("Could not block signals for threads");
}
// Catch signals which we want to use to quit savely
struct sigaction sa;
sigemptyset(&sa.sa_mask);
sigaddset(&sa.sa_mask, SIGHUP);
sigaddset(&sa.sa_mask, SIGINT);
sigaddset(&sa.sa_mask, SIGTERM);
sa.sa_handler = sig_handler_exit;
sa.sa_flags = 0;
if(sigaction(SIGHUP, &sa, NULL) == -1) {
error("Failed to change signal handler for SIGHUP");
}
if(sigaction(SIGINT, &sa, NULL) == -1) {
error("Failed to change signal handler for SIGINT");
}
if(sigaction(SIGTERM, &sa, NULL) == -1) {
error("Failed to change signal handler for SIGTERM");
}
// save database on SIGUSR1
sa.sa_handler = sig_handler_save;
if(sigaction(SIGUSR1, &sa, NULL) == -1) {
error("Failed to change signal handler for SIGUSR1");
}
// Ignore SIGPIPE completely.
// INFO: If we add signals here we have to unblock them
// at popen.c when running a external plugin.
sa.sa_handler = SIG_IGN;
if(sigaction(SIGPIPE, &sa, NULL) == -1) {
error("Failed to change signal handler for SIGPIPE");
}
// --------------------------------------------------------------------
i = pthread_attr_init(&attr);
if(i != 0)
fatal("pthread_attr_init() failed with code %d.", i);
i = pthread_attr_getstacksize(&attr, &stacksize);
if(i != 0)
fatal("pthread_attr_getstacksize() failed with code %d.", i);
else
debug(D_OPTIONS, "initial pthread stack size is %zu bytes", stacksize);
wanted_stacksize = config_get_number("global", "pthread stack size", stacksize);
// --------------------------------------------------------------------
for (i = 0; static_threads[i].name != NULL ; i++) {
struct netdata_static_thread *st = &static_threads[i];
if(st->config_name) st->enabled = config_get_boolean(st->config_section, st->config_name, st->enabled);
if(st->enabled && st->init_routine) st->init_routine();
}
// --------------------------------------------------------------------
// get the user we should run
// IMPORTANT: this is required before web_files_uid()
user = config_get("global", "run as user" , (getuid() == 0)?NETDATA_USER:"");
// IMPORTANT: these have to run once, while single threaded
web_files_uid(); // IMPORTANT: web_files_uid() before web_files_gid()
web_files_gid();
// --------------------------------------------------------------------
create_listen_sockets();
}
// never become a problem
if(nice(20) == -1) error("Cannot lower my CPU priority.");
if(become_daemon(dont_fork, 0, user, input_log_file, output_log_file, error_log_file, access_log_file, &access_fd, &stdaccess) == -1)
fatal("Cannot demonize myself.");
#ifdef NETDATA_INTERNAL_CHECKS
if(debug_flags != 0) {
struct rlimit rl = { RLIM_INFINITY, RLIM_INFINITY };
if(setrlimit(RLIMIT_CORE, &rl) != 0)
info("Cannot request unlimited core dumps for debugging... Proceeding anyway...");
prctl(PR_SET_DUMPABLE, 1, 0, 0, 0);
}
#endif /* NETDATA_INTERNAL_CHECKS */
if(output_log_syslog || error_log_syslog || access_log_syslog)
openlog("netdata", LOG_PID, LOG_DAEMON);
info("NetData started on pid %d", getpid());
// ------------------------------------------------------------------------
// get default pthread stack size
if(stacksize < wanted_stacksize) {
i = pthread_attr_setstacksize(&attr, wanted_stacksize);
if(i != 0)
fatal("pthread_attr_setstacksize() to %zu bytes, failed with code %d.", wanted_stacksize, i);
else
info("Successfully set pthread stacksize to %zu bytes", wanted_stacksize);
}
// --------------------------------------------------------------------
// initialize the registry
registry_init();
// ------------------------------------------------------------------------
// spawn the threads
web_server_threading_selection();
for (i = 0; static_threads[i].name != NULL ; i++) {
struct netdata_static_thread *st = &static_threads[i];
if(st->enabled) {
st->thread = malloc(sizeof(pthread_t));
if(!st->thread)
fatal("Cannot allocate pthread_t memory");
info("Starting thread %s.", st->name);
if(pthread_create(st->thread, &attr, st->start_routine, NULL))
error("failed to create new thread for %s.", st->name);
else if(pthread_detach(*st->thread))
error("Cannot request detach of newly created %s thread.", st->name);
}
else info("Not starting thread %s.", st->name);
}
// ------------------------------------------------------------------------
// block signals while initializing threads.
sigset_t sigset;
sigfillset(&sigset);
if(pthread_sigmask(SIG_UNBLOCK, &sigset, NULL) == -1) {
error("Could not unblock signals for threads");
}
// Handle flags set in the signal handler.
while(1) {
pause();
if(netdata_exit) {
info("Exit main loop of netdata.");
netdata_cleanup_and_exit(0);
exit(0);
}
}
}
void *backends_main(void *ptr) {
netdata_thread_cleanup_push(backends_main_cleanup, ptr);
int default_port = 0;
int sock = -1;
BUFFER *b = buffer_create(1), *response = buffer_create(1);
int (*backend_request_formatter)(BUFFER *, const char *, RRDHOST *, const char *, RRDSET *, RRDDIM *, time_t, time_t, uint32_t) = NULL;
int (*backend_response_checker)(BUFFER *) = NULL;
// ------------------------------------------------------------------------
// collect configuration options
struct timeval timeout = {
.tv_sec = 0,
.tv_usec = 0
};
int enabled = config_get_boolean(CONFIG_SECTION_BACKEND, "enabled", 0);
const char *source = config_get(CONFIG_SECTION_BACKEND, "data source", "average");
const char *type = config_get(CONFIG_SECTION_BACKEND, "type", "graphite");
const char *destination = config_get(CONFIG_SECTION_BACKEND, "destination", "localhost");
backend_prefix = config_get(CONFIG_SECTION_BACKEND, "prefix", "netdata");
const char *hostname = config_get(CONFIG_SECTION_BACKEND, "hostname", localhost->hostname);
backend_update_every = (int)config_get_number(CONFIG_SECTION_BACKEND, "update every", backend_update_every);
int buffer_on_failures = (int)config_get_number(CONFIG_SECTION_BACKEND, "buffer on failures", 10);
long timeoutms = config_get_number(CONFIG_SECTION_BACKEND, "timeout ms", backend_update_every * 2 * 1000);
backend_send_names = config_get_boolean(CONFIG_SECTION_BACKEND, "send names instead of ids", backend_send_names);
charts_pattern = simple_pattern_create(config_get(CONFIG_SECTION_BACKEND, "send charts matching", "*"), NULL, SIMPLE_PATTERN_EXACT);
hosts_pattern = simple_pattern_create(config_get(CONFIG_SECTION_BACKEND, "send hosts matching", "localhost *"), NULL, SIMPLE_PATTERN_EXACT);
// ------------------------------------------------------------------------
// validate configuration options
// and prepare for sending data to our backend
backend_options = backend_parse_data_source(source, backend_options);
if(timeoutms < 1) {
error("BACKEND: invalid timeout %ld ms given. Assuming %d ms.", timeoutms, backend_update_every * 2 * 1000);
timeoutms = backend_update_every * 2 * 1000;
}
timeout.tv_sec = (timeoutms * 1000) / 1000000;
timeout.tv_usec = (timeoutms * 1000) % 1000000;
if(!enabled || backend_update_every < 1)
goto cleanup;
// ------------------------------------------------------------------------
// select the backend type
if(!strcmp(type, "graphite") || !strcmp(type, "graphite:plaintext")) {
default_port = 2003;
backend_response_checker = process_graphite_response;
if((backend_options & BACKEND_SOURCE_BITS) == BACKEND_SOURCE_DATA_AS_COLLECTED)
backend_request_formatter = format_dimension_collected_graphite_plaintext;
else
backend_request_formatter = format_dimension_stored_graphite_plaintext;
}
else if(!strcmp(type, "opentsdb") || !strcmp(type, "opentsdb:telnet")) {
default_port = 4242;
backend_response_checker = process_opentsdb_response;
if((backend_options & BACKEND_SOURCE_BITS) == BACKEND_SOURCE_DATA_AS_COLLECTED)
backend_request_formatter = format_dimension_collected_opentsdb_telnet;
else
backend_request_formatter = format_dimension_stored_opentsdb_telnet;
}
else if (!strcmp(type, "json") || !strcmp(type, "json:plaintext")) {
default_port = 5448;
backend_response_checker = process_json_response;
if ((backend_options & BACKEND_SOURCE_BITS) == BACKEND_SOURCE_DATA_AS_COLLECTED)
backend_request_formatter = format_dimension_collected_json_plaintext;
else
backend_request_formatter = format_dimension_stored_json_plaintext;
}
else {
error("BACKEND: Unknown backend type '%s'", type);
goto cleanup;
}
if(backend_request_formatter == NULL || backend_response_checker == NULL) {
error("BACKEND: backend is misconfigured - disabling it.");
goto cleanup;
}
// ------------------------------------------------------------------------
// prepare the charts for monitoring the backend operation
struct rusage thread;
collected_number
chart_buffered_metrics = 0,
chart_lost_metrics = 0,
chart_sent_metrics = 0,
chart_buffered_bytes = 0,
chart_received_bytes = 0,
chart_sent_bytes = 0,
chart_receptions = 0,
chart_transmission_successes = 0,
chart_transmission_failures = 0,
chart_data_lost_events = 0,
chart_lost_bytes = 0,
chart_backend_reconnects = 0,
chart_backend_latency = 0;
RRDSET *chart_metrics = rrdset_create_localhost("netdata", "backend_metrics", NULL, "backend", NULL, "Netdata Buffered Metrics", "metrics", "backends", NULL, 130600, backend_update_every, RRDSET_TYPE_LINE);
rrddim_add(chart_metrics, "buffered", NULL, 1, 1, RRD_ALGORITHM_ABSOLUTE);
rrddim_add(chart_metrics, "lost", NULL, 1, 1, RRD_ALGORITHM_ABSOLUTE);
rrddim_add(chart_metrics, "sent", NULL, 1, 1, RRD_ALGORITHM_ABSOLUTE);
RRDSET *chart_bytes = rrdset_create_localhost("netdata", "backend_bytes", NULL, "backend", NULL, "Netdata Backend Data Size", "KB", "backends", NULL, 130610, backend_update_every, RRDSET_TYPE_AREA);
rrddim_add(chart_bytes, "buffered", NULL, 1, 1024, RRD_ALGORITHM_ABSOLUTE);
rrddim_add(chart_bytes, "lost", NULL, 1, 1024, RRD_ALGORITHM_ABSOLUTE);
rrddim_add(chart_bytes, "sent", NULL, 1, 1024, RRD_ALGORITHM_ABSOLUTE);
rrddim_add(chart_bytes, "received", NULL, 1, 1024, RRD_ALGORITHM_ABSOLUTE);
RRDSET *chart_ops = rrdset_create_localhost("netdata", "backend_ops", NULL, "backend", NULL, "Netdata Backend Operations", "operations", "backends", NULL, 130630, backend_update_every, RRDSET_TYPE_LINE);
rrddim_add(chart_ops, "write", NULL, 1, 1, RRD_ALGORITHM_ABSOLUTE);
rrddim_add(chart_ops, "discard", NULL, 1, 1, RRD_ALGORITHM_ABSOLUTE);
rrddim_add(chart_ops, "reconnect", NULL, 1, 1, RRD_ALGORITHM_ABSOLUTE);
rrddim_add(chart_ops, "failure", NULL, 1, 1, RRD_ALGORITHM_ABSOLUTE);
rrddim_add(chart_ops, "read", NULL, 1, 1, RRD_ALGORITHM_ABSOLUTE);
/*
* this is misleading - we can only measure the time we need to send data
* this time is not related to the time required for the data to travel to
* the backend database and the time that server needed to process them
*
* issue #1432 and https://www.softlab.ntua.gr/facilities/documentation/unix/unix-socket-faq/unix-socket-faq-2.html
*
RRDSET *chart_latency = rrdset_create_localhost("netdata", "backend_latency", NULL, "backend", NULL, "Netdata Backend Latency", "ms", "backends", NULL, 130620, backend_update_every, RRDSET_TYPE_AREA);
rrddim_add(chart_latency, "latency", NULL, 1, 1000, RRD_ALGORITHM_ABSOLUTE);
*/
RRDSET *chart_rusage = rrdset_create_localhost("netdata", "backend_thread_cpu", NULL, "backend", NULL, "NetData Backend Thread CPU usage", "milliseconds/s", "backends", NULL, 130630, backend_update_every, RRDSET_TYPE_STACKED);
rrddim_add(chart_rusage, "user", NULL, 1, 1000, RRD_ALGORITHM_INCREMENTAL);
rrddim_add(chart_rusage, "system", NULL, 1, 1000, RRD_ALGORITHM_INCREMENTAL);
// ------------------------------------------------------------------------
// prepare the backend main loop
info("BACKEND: configured ('%s' on '%s' sending '%s' data, every %d seconds, as host '%s', with prefix '%s')", type, destination, source, backend_update_every, hostname, backend_prefix);
usec_t step_ut = backend_update_every * USEC_PER_SEC;
time_t after = now_realtime_sec();
int failures = 0;
heartbeat_t hb;
heartbeat_init(&hb);
while(!netdata_exit) {
// ------------------------------------------------------------------------
// Wait for the next iteration point.
heartbeat_next(&hb, step_ut);
time_t before = now_realtime_sec();
debug(D_BACKEND, "BACKEND: preparing buffer for timeframe %lu to %lu", (unsigned long)after, (unsigned long)before);
// ------------------------------------------------------------------------
// add to the buffer the data we need to send to the backend
netdata_thread_disable_cancelability();
size_t count_hosts = 0;
size_t count_charts_total = 0;
size_t count_dims_total = 0;
rrd_rdlock();
RRDHOST *host;
rrdhost_foreach_read(host) {
if(unlikely(!rrdhost_flag_check(host, RRDHOST_FLAG_BACKEND_SEND|RRDHOST_FLAG_BACKEND_DONT_SEND))) {
char *name = (host == localhost)?"localhost":host->hostname;
if (!hosts_pattern || simple_pattern_matches(hosts_pattern, name)) {
rrdhost_flag_set(host, RRDHOST_FLAG_BACKEND_SEND);
info("enabled backend for host '%s'", name);
}
else {
rrdhost_flag_set(host, RRDHOST_FLAG_BACKEND_DONT_SEND);
info("disabled backend for host '%s'", name);
}
}
if(unlikely(!rrdhost_flag_check(host, RRDHOST_FLAG_BACKEND_SEND)))
continue;
rrdhost_rdlock(host);
count_hosts++;
size_t count_charts = 0;
size_t count_dims = 0;
size_t count_dims_skipped = 0;
const char *__hostname = (host == localhost)?hostname:host->hostname;
RRDSET *st;
rrdset_foreach_read(st, host) {
if(likely(backends_can_send_rrdset(backend_options, st))) {
rrdset_rdlock(st);
count_charts++;
RRDDIM *rd;
rrddim_foreach_read(rd, st) {
if (likely(rd->last_collected_time.tv_sec >= after)) {
chart_buffered_metrics += backend_request_formatter(b, backend_prefix, host, __hostname, st, rd, after, before, backend_options);
count_dims++;
}
else {
debug(D_BACKEND, "BACKEND: not sending dimension '%s' of chart '%s' from host '%s', its last data collection (%lu) is not within our timeframe (%lu to %lu)", rd->id, st->id, __hostname, (unsigned long)rd->last_collected_time.tv_sec, (unsigned long)after, (unsigned long)before);
count_dims_skipped++;
}
}
rrdset_unlock(st);
}
}
debug(D_BACKEND, "BACKEND: sending host '%s', metrics of %zu dimensions, of %zu charts. Skipped %zu dimensions.", __hostname, count_dims, count_charts, count_dims_skipped);
count_charts_total += count_charts;
count_dims_total += count_dims;
rrdhost_unlock(host);
}
rrd_unlock();
netdata_thread_enable_cancelability();
debug(D_BACKEND, "BACKEND: buffer has %zu bytes, added metrics for %zu dimensions, of %zu charts, from %zu hosts", buffer_strlen(b), count_dims_total, count_charts_total, count_hosts);
// ------------------------------------------------------------------------
chart_buffered_bytes = (collected_number)buffer_strlen(b);
// reset the monitoring chart counters
chart_received_bytes =
chart_sent_bytes =
chart_sent_metrics =
chart_lost_metrics =
chart_transmission_successes =
chart_transmission_failures =
chart_data_lost_events =
chart_lost_bytes =
chart_backend_reconnects =
chart_backend_latency = 0;
if(unlikely(netdata_exit)) break;
//fprintf(stderr, "\nBACKEND BEGIN:\n%s\nBACKEND END\n", buffer_tostring(b)); // FIXME
//fprintf(stderr, "after = %lu, before = %lu\n", after, before);
// prepare for the next iteration
// to add incrementally data to buffer
after = before;
// ------------------------------------------------------------------------
// if we are connected, receive a response, without blocking
if(likely(sock != -1)) {
errno = 0;
// loop through to collect all data
while(sock != -1 && errno != EWOULDBLOCK) {
buffer_need_bytes(response, 4096);
ssize_t r = recv(sock, &response->buffer[response->len], response->size - response->len, MSG_DONTWAIT);
if(likely(r > 0)) {
// we received some data
response->len += r;
chart_received_bytes += r;
chart_receptions++;
}
else if(r == 0) {
error("BACKEND: '%s' closed the socket", destination);
close(sock);
sock = -1;
}
else {
// failed to receive data
if(errno != EAGAIN && errno != EWOULDBLOCK) {
error("BACKEND: cannot receive data from backend '%s'.", destination);
}
}
}
// if we received data, process them
if(buffer_strlen(response))
backend_response_checker(response);
}
// ------------------------------------------------------------------------
// if we are not connected, connect to a backend server
if(unlikely(sock == -1)) {
usec_t start_ut = now_monotonic_usec();
size_t reconnects = 0;
sock = connect_to_one_of(destination, default_port, &timeout, &reconnects, NULL, 0);
chart_backend_reconnects += reconnects;
chart_backend_latency += now_monotonic_usec() - start_ut;
}
if(unlikely(netdata_exit)) break;
// ------------------------------------------------------------------------
// if we are connected, send our buffer to the backend server
if(likely(sock != -1)) {
size_t len = buffer_strlen(b);
usec_t start_ut = now_monotonic_usec();
int flags = 0;
#ifdef MSG_NOSIGNAL
flags += MSG_NOSIGNAL;
#endif
ssize_t written = send(sock, buffer_tostring(b), len, flags);
chart_backend_latency += now_monotonic_usec() - start_ut;
if(written != -1 && (size_t)written == len) {
// we sent the data successfully
chart_transmission_successes++;
chart_sent_bytes += written;
chart_sent_metrics = chart_buffered_metrics;
// reset the failures count
failures = 0;
// empty the buffer
buffer_flush(b);
}
else {
// oops! we couldn't send (all or some of the) data
error("BACKEND: failed to write data to database backend '%s'. Willing to write %zu bytes, wrote %zd bytes. Will re-connect.", destination, len, written);
chart_transmission_failures++;
if(written != -1)
chart_sent_bytes += written;
// increment the counter we check for data loss
failures++;
// close the socket - we will re-open it next time
close(sock);
sock = -1;
}
}
else {
error("BACKEND: failed to update database backend '%s'", destination);
chart_transmission_failures++;
// increment the counter we check for data loss
failures++;
}
if(failures > buffer_on_failures) {
// too bad! we are going to lose data
chart_lost_bytes += buffer_strlen(b);
error("BACKEND: reached %d backend failures. Flushing buffers to protect this host - this results in data loss on back-end server '%s'", failures, destination);
buffer_flush(b);
failures = 0;
chart_data_lost_events++;
chart_lost_metrics = chart_buffered_metrics;
}
if(unlikely(netdata_exit)) break;
// ------------------------------------------------------------------------
// update the monitoring charts
if(likely(chart_ops->counter_done)) rrdset_next(chart_ops);
rrddim_set(chart_ops, "read", chart_receptions);
rrddim_set(chart_ops, "write", chart_transmission_successes);
rrddim_set(chart_ops, "discard", chart_data_lost_events);
rrddim_set(chart_ops, "failure", chart_transmission_failures);
rrddim_set(chart_ops, "reconnect", chart_backend_reconnects);
rrdset_done(chart_ops);
if(likely(chart_metrics->counter_done)) rrdset_next(chart_metrics);
rrddim_set(chart_metrics, "buffered", chart_buffered_metrics);
rrddim_set(chart_metrics, "lost", chart_lost_metrics);
rrddim_set(chart_metrics, "sent", chart_sent_metrics);
rrdset_done(chart_metrics);
if(likely(chart_bytes->counter_done)) rrdset_next(chart_bytes);
rrddim_set(chart_bytes, "buffered", chart_buffered_bytes);
rrddim_set(chart_bytes, "lost", chart_lost_bytes);
rrddim_set(chart_bytes, "sent", chart_sent_bytes);
rrddim_set(chart_bytes, "received", chart_received_bytes);
rrdset_done(chart_bytes);
/*
if(likely(chart_latency->counter_done)) rrdset_next(chart_latency);
rrddim_set(chart_latency, "latency", chart_backend_latency);
rrdset_done(chart_latency);
*/
getrusage(RUSAGE_THREAD, &thread);
if(likely(chart_rusage->counter_done)) rrdset_next(chart_rusage);
rrddim_set(chart_rusage, "user", thread.ru_utime.tv_sec * 1000000ULL + thread.ru_utime.tv_usec);
rrddim_set(chart_rusage, "system", thread.ru_stime.tv_sec * 1000000ULL + thread.ru_stime.tv_usec);
rrdset_done(chart_rusage);
if(likely(buffer_strlen(b) == 0))
chart_buffered_metrics = 0;
if(unlikely(netdata_exit)) break;
}
cleanup:
if(sock != -1)
close(sock);
buffer_free(b);
buffer_free(response);
netdata_thread_cleanup_pop(1);
return NULL;
}
int main(int argc, char **argv)
{
int i;
int config_loaded = 0;
int dont_fork = 0;
size_t wanted_stacksize = 0, stacksize = 0;
pthread_attr_t attr;
// global initialization
get_HZ();
// set the name for logging
program_name = "netdata";
// parse the arguments
for(i = 1; i < argc ; i++) {
if(strcmp(argv[i], "-c") == 0 && (i+1) < argc) {
if(load_config(argv[i+1], 1) != 1) {
error("Cannot load configuration file %s.", argv[i+1]);
exit(1);
}
else {
debug(D_OPTIONS, "Configuration loaded from %s.", argv[i+1]);
config_loaded = 1;
}
i++;
}
else if(strcmp(argv[i], "-df") == 0 && (i+1) < argc) { config_set("global", "debug flags", argv[i+1]); debug_flags = strtoull(argv[i+1], NULL, 0); i++; }
else if(strcmp(argv[i], "-p") == 0 && (i+1) < argc) { config_set("global", "port", argv[i+1]); i++; }
else if(strcmp(argv[i], "-u") == 0 && (i+1) < argc) { config_set("global", "run as user", argv[i+1]); i++; }
else if(strcmp(argv[i], "-l") == 0 && (i+1) < argc) { config_set("global", "history", argv[i+1]); i++; }
else if(strcmp(argv[i], "-t") == 0 && (i+1) < argc) { config_set("global", "update every", argv[i+1]); i++; }
else if(strcmp(argv[i], "-ch") == 0 && (i+1) < argc) { config_set("global", "host access prefix", argv[i+1]); i++; }
else if(strcmp(argv[i], "-stacksize") == 0 && (i+1) < argc) { config_set("global", "pthread stack size", argv[i+1]); i++; }
else if(strcmp(argv[i], "-nodaemon") == 0 || strcmp(argv[i], "-nd") == 0) dont_fork = 1;
else if(strcmp(argv[i], "-pidfile") == 0 && (i+1) < argc) {
i++;
strncpy(pidfile, argv[i], FILENAME_MAX);
pidfile[FILENAME_MAX] = '\0';
}
else if(strcmp(argv[i], "--unittest") == 0) {
rrd_update_every = 1;
if(run_all_mockup_tests()) exit(1);
if(unit_test_storage()) exit(1);
fprintf(stderr, "\n\nALL TESTS PASSED\n\n");
exit(0);
}
else {
fprintf(stderr, "Cannot understand option '%s'.\n", argv[i]);
fprintf(stderr, "\nUSAGE: %s [-d] [-l LINES_TO_SAVE] [-u UPDATE_TIMER] [-p LISTEN_PORT] [-df debug flags].\n\n", argv[0]);
fprintf(stderr, " -c CONFIG FILE the configuration file to load. Default: %s.\n", CONFIG_DIR "/" CONFIG_FILENAME);
fprintf(stderr, " -l LINES_TO_SAVE can be from 5 to %d lines in JSON data. Default: %d.\n", RRD_HISTORY_ENTRIES_MAX, RRD_DEFAULT_HISTORY_ENTRIES);
fprintf(stderr, " -t UPDATE_TIMER can be from 1 to %d seconds. Default: %d.\n", UPDATE_EVERY_MAX, UPDATE_EVERY);
fprintf(stderr, " -p LISTEN_PORT can be from 1 to %d. Default: %d.\n", 65535, LISTEN_PORT);
fprintf(stderr, " -u USERNAME can be any system username to run as. Default: none.\n");
fprintf(stderr, " -ch path to access host /proc and /sys when running in a container. Default: empty.\n");
fprintf(stderr, " -nd or -nodeamon to disable forking in the background. Default: unset.\n");
fprintf(stderr, " -df FLAGS debug options. Default: 0x%08llx.\n", debug_flags);
fprintf(stderr, " -stacksize BYTES to overwrite the pthread stack size.\n");
fprintf(stderr, " -pidfile FILENAME to save a pid while running.\n");
exit(1);
}
}
if(!config_loaded) load_config(NULL, 0);
// prepare configuration environment variables for the plugins
setenv("NETDATA_CONFIG_DIR" , config_get("global", "config directory" , CONFIG_DIR) , 1);
setenv("NETDATA_PLUGINS_DIR", config_get("global", "plugins directory" , PLUGINS_DIR), 1);
setenv("NETDATA_WEB_DIR" , config_get("global", "web files directory", WEB_DIR) , 1);
setenv("NETDATA_CACHE_DIR" , config_get("global", "cache directory" , CACHE_DIR) , 1);
setenv("NETDATA_LOG_DIR" , config_get("global", "log directory" , LOG_DIR) , 1);
setenv("NETDATA_HOST_PREFIX", config_get("global", "host access prefix" , "") , 1);
setenv("HOME" , config_get("global", "home directory" , CACHE_DIR) , 1);
// avoid extended to stat(/etc/localtime)
// http://stackoverflow.com/questions/4554271/how-to-avoid-excessive-stat-etc-localtime-calls-in-strftime-on-linux
setenv("TZ", ":/etc/localtime", 0);
// cd to /tmp to avoid any plugins writing files at random places
if(chdir("/tmp")) error("netdata: ERROR: Cannot cd to /tmp");
char *input_log_file = NULL;
char *output_log_file = NULL;
char *error_log_file = NULL;
char *access_log_file = NULL;
char *user = NULL;
{
char buffer[1024];
// --------------------------------------------------------------------
sprintf(buffer, "0x%08llx", 0ULL);
char *flags = config_get("global", "debug flags", buffer);
setenv("NETDATA_DEBUG_FLAGS", flags, 1);
debug_flags = strtoull(flags, NULL, 0);
debug(D_OPTIONS, "Debug flags set to '0x%8llx'.", debug_flags);
if(debug_flags != 0) {
struct rlimit rl = { RLIM_INFINITY, RLIM_INFINITY };
if(setrlimit(RLIMIT_CORE, &rl) != 0)
info("Cannot request unlimited core dumps for debugging... Proceeding anyway...");
prctl(PR_SET_DUMPABLE, 1, 0, 0, 0);
}
// --------------------------------------------------------------------
#ifdef MADV_MERGEABLE
enable_ksm = config_get_boolean("global", "memory deduplication (ksm)", enable_ksm);
#else
#warning "Kernel memory deduplication (KSM) is not available"
#endif
// --------------------------------------------------------------------
global_host_prefix = config_get("global", "host access prefix", "");
setenv("NETDATA_HOST_PREFIX", global_host_prefix, 1);
// --------------------------------------------------------------------
output_log_file = config_get("global", "debug log", LOG_DIR "/debug.log");
if(strcmp(output_log_file, "syslog") == 0) {
output_log_syslog = 1;
output_log_file = NULL;
}
else if(strcmp(output_log_file, "none") == 0) {
output_log_syslog = 0;
output_log_file = NULL;
}
else output_log_syslog = 0;
// --------------------------------------------------------------------
error_log_file = config_get("global", "error log", LOG_DIR "/error.log");
if(strcmp(error_log_file, "syslog") == 0) {
error_log_syslog = 1;
error_log_file = NULL;
}
else if(strcmp(error_log_file, "none") == 0) {
error_log_syslog = 0;
error_log_file = NULL;
// optimization - do not even generate debug log entries
}
else error_log_syslog = 0;
error_log_throttle_period = config_get_number("global", "errors flood protection period", error_log_throttle_period);
setenv("NETDATA_ERRORS_THROTTLE_PERIOD", config_get("global", "errors flood protection period" , ""), 1);
error_log_errors_per_period = config_get_number("global", "errors to trigger flood protection", error_log_errors_per_period);
setenv("NETDATA_ERRORS_PER_PERIOD" , config_get("global", "errors to trigger flood protection", ""), 1);
// --------------------------------------------------------------------
access_log_file = config_get("global", "access log", LOG_DIR "/access.log");
if(strcmp(access_log_file, "syslog") == 0) {
access_log_syslog = 1;
access_log_file = NULL;
}
else if(strcmp(access_log_file, "none") == 0) {
access_log_syslog = 0;
access_log_file = NULL;
}
else access_log_syslog = 0;
// --------------------------------------------------------------------
rrd_memory_mode = rrd_memory_mode_id(config_get("global", "memory mode", rrd_memory_mode_name(rrd_memory_mode)));
// --------------------------------------------------------------------
if(gethostname(buffer, HOSTNAME_MAX) == -1)
error("WARNING: Cannot get machine hostname.");
hostname = config_get("global", "hostname", buffer);
debug(D_OPTIONS, "hostname set to '%s'", hostname);
// --------------------------------------------------------------------
rrd_default_history_entries = (int) config_get_number("global", "history", RRD_DEFAULT_HISTORY_ENTRIES);
if(rrd_default_history_entries < 5 || rrd_default_history_entries > RRD_HISTORY_ENTRIES_MAX) {
info("Invalid save lines %d given. Defaulting to %d.", rrd_default_history_entries, RRD_DEFAULT_HISTORY_ENTRIES);
rrd_default_history_entries = RRD_DEFAULT_HISTORY_ENTRIES;
}
else {
debug(D_OPTIONS, "save lines set to %d.", rrd_default_history_entries);
}
// --------------------------------------------------------------------
rrd_update_every = (int) config_get_number("global", "update every", UPDATE_EVERY);
if(rrd_update_every < 1 || rrd_update_every > 600) {
info("Invalid update timer %d given. Defaulting to %d.", rrd_update_every, UPDATE_EVERY_MAX);
rrd_update_every = UPDATE_EVERY;
}
else debug(D_OPTIONS, "update timer set to %d.", rrd_update_every);
// let the plugins know the min update_every
{
char buf[50];
snprintf(buf, 50, "%d", rrd_update_every);
setenv("NETDATA_UPDATE_EVERY", buf, 1);
}
// --------------------------------------------------------------------
// block signals while initializing threads.
// this causes the threads to block signals.
sigset_t sigset;
sigfillset(&sigset);
if(pthread_sigmask(SIG_BLOCK, &sigset, NULL) == -1) {
error("Could not block signals for threads");
}
// Catch signals which we want to use to quit savely
struct sigaction sa;
sigemptyset(&sa.sa_mask);
sigaddset(&sa.sa_mask, SIGHUP);
sigaddset(&sa.sa_mask, SIGINT);
sigaddset(&sa.sa_mask, SIGTERM);
sa.sa_handler = sig_handler;
sa.sa_flags = 0;
if(sigaction(SIGHUP, &sa, NULL) == -1) {
error("Failed to change signal handler for SIGHUP");
}
if(sigaction(SIGINT, &sa, NULL) == -1) {
error("Failed to change signal handler for SIGINT");
}
if(sigaction(SIGTERM, &sa, NULL) == -1) {
error("Failed to change signal handler for SIGTERM");
}
// Ignore SIGPIPE completely.
// INFO: If we add signals here we have to unblock them
// at popen.c when running a external plugin.
sa.sa_handler = SIG_IGN;
if(sigaction(SIGPIPE, &sa, NULL) == -1) {
error("Failed to change signal handler for SIGTERM");
}
// --------------------------------------------------------------------
i = pthread_attr_init(&attr);
if(i != 0)
fatal("pthread_attr_init() failed with code %d.", i);
i = pthread_attr_getstacksize(&attr, &stacksize);
if(i != 0)
fatal("pthread_attr_getstacksize() failed with code %d.", i);
else
debug(D_OPTIONS, "initial pthread stack size is %zu bytes", stacksize);
wanted_stacksize = config_get_number("global", "pthread stack size", stacksize);
// --------------------------------------------------------------------
for (i = 0; static_threads[i].name != NULL ; i++) {
struct netdata_static_thread *st = &static_threads[i];
if(st->config_name) st->enabled = config_get_boolean(st->config_section, st->config_name, st->enabled);
if(st->enabled && st->init_routine) st->init_routine();
}
// --------------------------------------------------------------------
// get the user we should run
// IMPORTANT: this is required before web_files_uid()
user = config_get("global", "run as user" , (getuid() == 0)?NETDATA_USER:"");
// IMPORTANT: these have to run once, while single threaded
web_files_uid(); // IMPORTANT: web_files_uid() before web_files_gid()
web_files_gid();
// --------------------------------------------------------------------
listen_backlog = (int) config_get_number("global", "http port listen backlog", LISTEN_BACKLOG);
listen_port = (int) config_get_number("global", "port", LISTEN_PORT);
if(listen_port < 1 || listen_port > 65535) {
info("Invalid listen port %d given. Defaulting to %d.", listen_port, LISTEN_PORT);
listen_port = LISTEN_PORT;
}
else debug(D_OPTIONS, "Listen port set to %d.", listen_port);
int ip = 0;
char *ipv = config_get("global", "ip version", "any");
if(!strcmp(ipv, "any") || !strcmp(ipv, "both") || !strcmp(ipv, "all")) ip = 0;
else if(!strcmp(ipv, "ipv4") || !strcmp(ipv, "IPV4") || !strcmp(ipv, "IPv4") || !strcmp(ipv, "4")) ip = 4;
else if(!strcmp(ipv, "ipv6") || !strcmp(ipv, "IPV6") || !strcmp(ipv, "IPv6") || !strcmp(ipv, "6")) ip = 6;
else info("Cannot understand ip version '%s'. Assuming 'any'.", ipv);
if(ip == 0 || ip == 6) listen_fd = create_listen_socket6(config_get("global", "bind socket to IP", "*"), listen_port, listen_backlog);
if(listen_fd < 0) {
listen_fd = create_listen_socket4(config_get("global", "bind socket to IP", "*"), listen_port, listen_backlog);
if(listen_fd >= 0 && ip != 4) info("Managed to open an IPv4 socket on port %d.", listen_port);
}
if(listen_fd < 0) fatal("Cannot listen socket.");
}
// never become a problem
if(nice(20) == -1) error("Cannot lower my CPU priority.");
if(become_daemon(dont_fork, 0, user, input_log_file, output_log_file, error_log_file, access_log_file, &access_fd, &stdaccess) == -1) {
fatal("Cannot demonize myself.");
exit(1);
}
if(debug_flags != 0) {
struct rlimit rl = { RLIM_INFINITY, RLIM_INFINITY };
if(setrlimit(RLIMIT_CORE, &rl) != 0)
info("Cannot request unlimited core dumps for debugging... Proceeding anyway...");
prctl(PR_SET_DUMPABLE, 1, 0, 0, 0);
}
if(output_log_syslog || error_log_syslog || access_log_syslog)
openlog("netdata", LOG_PID, LOG_DAEMON);
info("NetData started on pid %d", getpid());
// ------------------------------------------------------------------------
// get default pthread stack size
if(stacksize < wanted_stacksize) {
i = pthread_attr_setstacksize(&attr, wanted_stacksize);
if(i != 0)
fatal("pthread_attr_setstacksize() to %zu bytes, failed with code %d.", wanted_stacksize, i);
else
info("Successfully set pthread stacksize to %zu bytes", wanted_stacksize);
}
// ------------------------------------------------------------------------
// spawn the threads
for (i = 0; static_threads[i].name != NULL ; i++) {
struct netdata_static_thread *st = &static_threads[i];
if(st->enabled) {
st->thread = malloc(sizeof(pthread_t));
if(!st->thread)
fatal("Cannot allocate pthread_t memory");
info("Starting thread %s.", st->name);
if(pthread_create(st->thread, &attr, st->start_routine, NULL))
error("failed to create new thread for %s.", st->name);
else if(pthread_detach(*st->thread))
error("Cannot request detach of newly created %s thread.", st->name);
}
else info("Not starting thread %s.", st->name);
}
// ------------------------------------------------------------------------
// block signals while initializing threads.
sigset_t sigset;
sigfillset(&sigset);
if(pthread_sigmask(SIG_UNBLOCK, &sigset, NULL) == -1) {
error("Could not unblock signals for threads");
}
// Handle flags set in the signal handler.
while(1) {
pause();
if(netdata_exit) {
info("Exit main loop of netdata.");
netdata_cleanup_and_exit(0);
exit(0);
}
}
}
RRDSET *rrdset_create(const char *type, const char *id, const char *name, const char *family, const char *context, const char *title, const char *units, long priority, int update_every, int chart_type)
{
if(!type || !type[0]) {
fatal("Cannot create rrd stats without a type.");
return NULL;
}
if(!id || !id[0]) {
fatal("Cannot create rrd stats without an id.");
return NULL;
}
char fullid[RRD_ID_LENGTH_MAX + 1];
char fullfilename[FILENAME_MAX + 1];
RRDSET *st = NULL;
snprintfz(fullid, RRD_ID_LENGTH_MAX, "%s.%s", type, id);
st = rrdset_find(fullid);
if(st) {
error("Cannot create rrd stats for '%s', it already exists.", fullid);
return st;
}
long entries = config_get_number(fullid, "history", rrd_default_history_entries);
if(entries < 5) entries = config_set_number(fullid, "history", 5);
if(entries > RRD_HISTORY_ENTRIES_MAX) entries = config_set_number(fullid, "history", RRD_HISTORY_ENTRIES_MAX);
int enabled = config_get_boolean(fullid, "enabled", 1);
if(!enabled) entries = 5;
unsigned long size = sizeof(RRDSET);
char *cache_dir = rrdset_cache_dir(fullid);
debug(D_RRD_CALLS, "Creating RRD_STATS for '%s.%s'.", type, id);
snprintfz(fullfilename, FILENAME_MAX, "%s/main.db", cache_dir);
if(rrd_memory_mode != RRD_MEMORY_MODE_RAM) st = (RRDSET *)mymmap(fullfilename, size, ((rrd_memory_mode == RRD_MEMORY_MODE_MAP)?MAP_SHARED:MAP_PRIVATE), 0);
if(st) {
if(strcmp(st->magic, RRDSET_MAGIC) != 0) {
errno = 0;
info("Initializing file %s.", fullfilename);
bzero(st, size);
}
else if(strcmp(st->id, fullid) != 0) {
errno = 0;
error("File %s contents are not for chart %s. Clearing it.", fullfilename, fullid);
// munmap(st, size);
// st = NULL;
bzero(st, size);
}
else if(st->memsize != size || st->entries != entries) {
errno = 0;
error("File %s does not have the desired size. Clearing it.", fullfilename);
bzero(st, size);
}
else if(st->update_every != update_every) {
errno = 0;
error("File %s does not have the desired update frequency. Clearing it.", fullfilename);
bzero(st, size);
}
else if((time(NULL) - st->last_updated.tv_sec) > update_every * entries) {
errno = 0;
error("File %s is too old. Clearing it.", fullfilename);
bzero(st, size);
}
}
if(st) {
st->name = NULL;
st->type = NULL;
st->family = NULL;
st->context = NULL;
st->title = NULL;
st->units = NULL;
st->dimensions = NULL;
st->next = NULL;
st->mapped = rrd_memory_mode;
}
else {
st = calloc(1, size);
if(!st) {
fatal("Cannot allocate memory for RRD_STATS %s.%s", type, id);
return NULL;
}
st->mapped = RRD_MEMORY_MODE_RAM;
}
st->memsize = size;
st->entries = entries;
st->update_every = update_every;
if(st->current_entry >= st->entries) st->current_entry = 0;
strcpy(st->cache_filename, fullfilename);
strcpy(st->magic, RRDSET_MAGIC);
strcpy(st->id, fullid);
st->hash = simple_hash(st->id);
st->cache_dir = cache_dir;
st->chart_type = rrdset_type_id(config_get(st->id, "chart type", rrdset_type_name(chart_type)));
st->type = config_get(st->id, "type", type);
st->family = config_get(st->id, "family", family?family:st->type);
st->context = config_get(st->id, "context", context?context:st->id);
st->units = config_get(st->id, "units", units?units:"");
st->priority = config_get_number(st->id, "priority", priority);
st->enabled = enabled;
st->isdetail = 0;
st->debug = 0;
st->last_collected_time.tv_sec = 0;
st->last_collected_time.tv_usec = 0;
st->counter_done = 0;
st->gap_when_lost_iterations_above = (int) (
config_get_number(st->id, "gap when lost iterations above", RRD_DEFAULT_GAP_INTERPOLATIONS) + 2);
avl_init_lock(&st->dimensions_index, rrddim_compare);
pthread_rwlock_init(&st->rwlock, NULL);
pthread_rwlock_wrlock(&rrdset_root_rwlock);
if(name && *name) rrdset_set_name(st, name);
else rrdset_set_name(st, id);
{
char varvalue[CONFIG_MAX_VALUE + 1];
snprintfz(varvalue, CONFIG_MAX_VALUE, "%s (%s)", title?title:"", st->name);
st->title = config_get(st->id, "title", varvalue);
}
st->next = rrdset_root;
rrdset_root = st;
rrdset_index_add(st);
pthread_rwlock_unlock(&rrdset_root_rwlock);
return(st);
}
int become_user(const char *username, int pid_fd) {
int am_i_root = (getuid() == 0)?1:0;
struct passwd *pw = getpwnam(username);
if(!pw) {
error("User %s is not present.", username);
return -1;
}
uid_t uid = pw->pw_uid;
gid_t gid = pw->pw_gid;
create_needed_dir(netdata_configured_cache_dir, uid, gid);
create_needed_dir(netdata_configured_varlib_dir, uid, gid);
if(pidfile[0]) {
if(chown(pidfile, uid, gid) == -1)
error("Cannot chown '%s' to %u:%u", pidfile, (unsigned int)uid, (unsigned int)gid);
}
int ngroups = (int)sysconf(_SC_NGROUPS_MAX);
gid_t *supplementary_groups = NULL;
if(ngroups > 0) {
supplementary_groups = mallocz(sizeof(gid_t) * ngroups);
if(getgrouplist(username, gid, supplementary_groups, &ngroups) == -1) {
if(am_i_root)
error("Cannot get supplementary groups of user '%s'.", username);
ngroups = 0;
}
}
chown_open_file(STDOUT_FILENO, uid, gid);
chown_open_file(STDERR_FILENO, uid, gid);
chown_open_file(stdaccess_fd, uid, gid);
chown_open_file(pid_fd, uid, gid);
if(supplementary_groups && ngroups > 0) {
if(setgroups((size_t)ngroups, supplementary_groups) == -1) {
if(am_i_root)
error("Cannot set supplementary groups for user '%s'", username);
}
ngroups = 0;
}
if(supplementary_groups)
freez(supplementary_groups);
#ifdef __APPLE__
if(setregid(gid, gid) != 0) {
#else
if(setresgid(gid, gid, gid) != 0) {
#endif /* __APPLE__ */
error("Cannot switch to user's %s group (gid: %u).", username, gid);
return -1;
}
#ifdef __APPLE__
if(setreuid(uid, uid) != 0) {
#else
if(setresuid(uid, uid, uid) != 0) {
#endif /* __APPLE__ */
error("Cannot switch to user %s (uid: %u).", username, uid);
return -1;
}
if(setgid(gid) != 0) {
error("Cannot switch to user's %s group (gid: %u).", username, gid);
return -1;
}
if(setegid(gid) != 0) {
error("Cannot effectively switch to user's %s group (gid: %u).", username, gid);
return -1;
}
if(setuid(uid) != 0) {
error("Cannot switch to user %s (uid: %u).", username, uid);
return -1;
}
if(seteuid(uid) != 0) {
error("Cannot effectively switch to user %s (uid: %u).", username, uid);
return -1;
}
return(0);
}
#ifndef OOM_SCORE_ADJ_MAX
#define OOM_SCORE_ADJ_MAX 1000
#endif
#ifndef OOM_SCORE_ADJ_MIN
#define OOM_SCORE_ADJ_MIN -1000
#endif
static void oom_score_adj(void) {
char buf[30 + 1];
long long int old_score, wanted_score = OOM_SCORE_ADJ_MAX, final_score = 0;
// read the existing score
if(read_single_signed_number_file("/proc/self/oom_score_adj", &old_score)) {
error("Out-Of-Memory (OOM) score setting is not supported on this system.");
return;
}
if(old_score != 0)
wanted_score = old_score;
// check the environment
char *s = getenv("OOMScoreAdjust");
if(!s || !*s) {
snprintfz(buf, 30, "%d", (int)wanted_score);
s = buf;
}
// check netdata.conf configuration
s = config_get(CONFIG_SECTION_GLOBAL, "OOM score", s);
if(s && *s && (isdigit(*s) || *s == '-' || *s == '+'))
wanted_score = atoll(s);
else {
info("Out-Of-Memory (OOM) score not changed due to non-numeric setting: '%s' (running with %d)", s, (int)old_score);
return;
}
if(wanted_score < OOM_SCORE_ADJ_MIN) {
error("Wanted Out-Of-Memory (OOM) score %d is too small. Using %d", (int)wanted_score, (int)OOM_SCORE_ADJ_MIN);
wanted_score = OOM_SCORE_ADJ_MIN;
}
if(wanted_score > OOM_SCORE_ADJ_MAX) {
error("Wanted Out-Of-Memory (OOM) score %d is too big. Using %d", (int)wanted_score, (int)OOM_SCORE_ADJ_MAX);
wanted_score = OOM_SCORE_ADJ_MAX;
}
if(old_score == wanted_score) {
info("Out-Of-Memory (OOM) score is already set to the wanted value %d", (int)old_score);
return;
}
int written = 0;
int fd = open("/proc/self/oom_score_adj", O_WRONLY);
if(fd != -1) {
snprintfz(buf, 30, "%d", (int)wanted_score);
ssize_t len = strlen(buf);
if(len > 0 && write(fd, buf, (size_t)len) == len) written = 1;
close(fd);
if(written) {
if(read_single_signed_number_file("/proc/self/oom_score_adj", &final_score))
error("Adjusted my Out-Of-Memory (OOM) score to %d, but cannot verify it.", (int)wanted_score);
else if(final_score == wanted_score)
info("Adjusted my Out-Of-Memory (OOM) score from %d to %d.", (int)old_score, (int)final_score);
else
error("Adjusted my Out-Of-Memory (OOM) score from %d to %d, but it has been set to %d.", (int)old_score, (int)wanted_score, (int)final_score);
}
else
error("Failed to adjust my Out-Of-Memory (OOM) score to %d. Running with %d. (systemd systems may change it via netdata.service)", (int)wanted_score, (int)old_score);
}
else
error("Failed to adjust my Out-Of-Memory (OOM) score. Cannot open /proc/self/oom_score_adj for writing.");
}
static void process_nice_level(void) {
#ifdef HAVE_NICE
int nice_level = (int)config_get_number(CONFIG_SECTION_GLOBAL, "process nice level", 19);
if(nice(nice_level) == -1) error("Cannot set netdata CPU nice level to %d.", nice_level);
else debug(D_SYSTEM, "Set netdata nice level to %d.", nice_level);
#endif // HAVE_NICE
};
void *socket_listen_main(void *ptr)
{
if(ptr) { ; }
info("WEB SERVER thread created with task id %d", gettid());
struct web_client *w;
struct timeval tv;
int retval;
if(ptr) { ; }
if(pthread_setcanceltype(PTHREAD_CANCEL_DEFERRED, NULL) != 0)
error("Cannot set pthread cancel type to DEFERRED.");
if(pthread_setcancelstate(PTHREAD_CANCEL_ENABLE, NULL) != 0)
error("Cannot set pthread cancel state to ENABLE.");
web_client_timeout = (int) config_get_number("global", "disconnect idle web clients after seconds", DEFAULT_DISCONNECT_IDLE_WEB_CLIENTS_AFTER_SECONDS);
web_enable_gzip = config_get_boolean("global", "enable web responses gzip compression", web_enable_gzip);
if(listen_fd < 0) fatal("LISTENER: Listen socket is not ready.");
fd_set ifds, ofds, efds;
int fdmax = listen_fd;
FD_ZERO (&ifds);
FD_ZERO (&ofds);
FD_ZERO (&efds);
for(;;) {
tv.tv_sec = 0;
tv.tv_usec = 200000;
if(listen_fd >= 0) {
FD_SET(listen_fd, &ifds);
FD_SET(listen_fd, &efds);
}
// debug(D_WEB_CLIENT, "LISTENER: Waiting...");
retval = select(fdmax+1, &ifds, &ofds, &efds, &tv);
if(retval == -1) {
error("LISTENER: select() failed.");
continue;
}
else if(retval) {
// check for new incoming connections
if(FD_ISSET(listen_fd, &ifds)) {
w = web_client_create(listen_fd);
if(unlikely(!w)) {
// no need for error log - web_client_create already logged the error
continue;
}
if(pthread_create(&w->thread, NULL, web_client_main, w) != 0) {
error("%llu: failed to create new thread for web client.");
w->obsolete = 1;
}
else if(pthread_detach(w->thread) != 0) {
error("%llu: Cannot request detach of newly created web client thread.", w->id);
w->obsolete = 1;
}
}
else debug(D_WEB_CLIENT, "LISTENER: select() didn't do anything.");
}
//else {
// debug(D_WEB_CLIENT, "LISTENER: select() timeout.");
//}
// cleanup unused clients
for(w = web_clients; w ; w = w?w->next:NULL) {
if(w->obsolete) {
debug(D_WEB_CLIENT, "%llu: Removing client.", w->id);
// pthread_join(w->thread, NULL);
w = web_client_free(w);
log_allocations();
}
}
}
error("LISTENER: exit!");
if(listen_fd >= 0) close(listen_fd);
exit(2);
return NULL;
}
