]> git.proxmox.com Git - ceph.git/blob - ceph/src/ceph_mon.cc
update sources to v12.1.1
[ceph.git] / ceph / src / ceph_mon.cc
1 // -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*-
2 // vim: ts=8 sw=2 smarttab
3 /*
4 * Ceph - scalable distributed file system
5 *
6 * Copyright (C) 2004-2006 Sage Weil <sage@newdream.net>
7 *
8 * This is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License version 2.1, as published by the Free Software
11 * Foundation. See file COPYING.
12 *
13 */
14
15 #include <sys/types.h>
16 #include <sys/stat.h>
17 #include <fcntl.h>
18
19 #include <iostream>
20 #include <string>
21 using namespace std;
22
23 #include "common/config.h"
24 #include "include/ceph_features.h"
25
26 #include "mon/MonMap.h"
27 #include "mon/Monitor.h"
28 #include "mon/MonitorDBStore.h"
29 #include "mon/MonClient.h"
30
31 #include "msg/Messenger.h"
32
33 #include "include/CompatSet.h"
34
35 #include "common/ceph_argparse.h"
36 #include "common/pick_address.h"
37 #include "common/Timer.h"
38 #include "common/errno.h"
39 #include "common/Preforker.h"
40
41 #include "global/global_init.h"
42 #include "global/signal_handler.h"
43
44 #include "perfglue/heap_profiler.h"
45
46 #include "include/assert.h"
47
48 #define dout_subsys ceph_subsys_mon
49
50 Monitor *mon = NULL;
51
52 void handle_mon_signal(int signum)
53 {
54 if (mon)
55 mon->handle_signal(signum);
56 }
57
58
59 int obtain_monmap(MonitorDBStore &store, bufferlist &bl)
60 {
61 dout(10) << __func__ << dendl;
62 /*
63 * the monmap may be in one of three places:
64 * 'monmap:<latest_version_no>' - the monmap we'd really like to have
65 * 'mon_sync:latest_monmap' - last monmap backed up for the last sync
66 * 'mkfs:monmap' - a monmap resulting from mkfs
67 */
68
69 if (store.exists("monmap", "last_committed")) {
70 version_t latest_ver = store.get("monmap", "last_committed");
71 if (store.exists("monmap", latest_ver)) {
72 int err = store.get("monmap", latest_ver, bl);
73 assert(err == 0);
74 assert(bl.length() > 0);
75 dout(10) << __func__ << " read last committed monmap ver "
76 << latest_ver << dendl;
77 return 0;
78 }
79 }
80
81 if (store.exists("mon_sync", "in_sync")
82 || store.exists("mon_sync", "force_sync")) {
83 dout(10) << __func__ << " detected aborted sync" << dendl;
84 if (store.exists("mon_sync", "latest_monmap")) {
85 int err = store.get("mon_sync", "latest_monmap", bl);
86 assert(err == 0);
87 assert(bl.length() > 0);
88 dout(10) << __func__ << " read backup monmap" << dendl;
89 return 0;
90 }
91 }
92
93 if (store.exists("mkfs", "monmap")) {
94 dout(10) << __func__ << " found mkfs monmap" << dendl;
95 int err = store.get("mkfs", "monmap", bl);
96 assert(err == 0);
97 assert(bl.length() > 0);
98 return 0;
99 }
100
101 derr << __func__ << " unable to find a monmap" << dendl;
102 return -ENOENT;
103 }
104
105 int check_mon_data_exists()
106 {
107 string mon_data = g_conf->mon_data;
108 struct stat buf;
109 if (::stat(mon_data.c_str(), &buf)) {
110 if (errno != ENOENT) {
111 derr << "stat(" << mon_data << ") " << cpp_strerror(errno) << dendl;
112 }
113 return -errno;
114 }
115 return 0;
116 }
117
118 /** Check whether **mon data** is empty.
119 *
120 * Being empty means mkfs has not been run and there's no monitor setup
121 * at **g_conf->mon_data**.
122 *
123 * If the directory g_conf->mon_data is not empty we will return -ENOTEMPTY.
124 * Otherwise we will return 0. Any other negative returns will represent
125 * a failure to be handled by the caller.
126 *
127 * @return **0** on success, -ENOTEMPTY if not empty or **-errno** otherwise.
128 */
129 int check_mon_data_empty()
130 {
131 string mon_data = g_conf->mon_data;
132
133 DIR *dir = ::opendir(mon_data.c_str());
134 if (!dir) {
135 derr << "opendir(" << mon_data << ") " << cpp_strerror(errno) << dendl;
136 return -errno;
137 }
138 int code = 0;
139 struct dirent *de = nullptr;
140 errno = 0;
141 while ((de = ::readdir(dir))) {
142 if (string(".") != de->d_name &&
143 string("..") != de->d_name &&
144 string("kv_backend") != de->d_name) {
145 code = -ENOTEMPTY;
146 break;
147 }
148 }
149 if (!de && errno) {
150 derr << "readdir(" << mon_data << ") " << cpp_strerror(errno) << dendl;
151 code = -errno;
152 }
153
154 ::closedir(dir);
155
156 return code;
157 }
158
159 static void usage()
160 {
161 cout << "usage: ceph-mon -i <ID> [flags]\n"
162 << " --debug_mon n\n"
163 << " debug monitor level (e.g. 10)\n"
164 << " --mkfs\n"
165 << " build fresh monitor fs\n"
166 << " --force-sync\n"
167 << " force a sync from another mon by wiping local data (BE CAREFUL)\n"
168 << " --yes-i-really-mean-it\n"
169 << " mandatory safeguard for --force-sync\n"
170 << " --compact\n"
171 << " compact the monitor store\n"
172 << " --osdmap <filename>\n"
173 << " only used when --mkfs is provided: load the osdmap from <filename>\n"
174 << " --inject-monmap <filename>\n"
175 << " write the <filename> monmap to the local monitor store and exit\n"
176 << " --extract-monmap <filename>\n"
177 << " extract the monmap from the local monitor store and exit\n"
178 << " --mon-data <directory>\n"
179 << " where the mon store and keyring are located\n"
180 << std::endl;
181 generic_server_usage();
182 }
183
184 #ifdef BUILDING_FOR_EMBEDDED
185 void cephd_preload_embedded_plugins();
186 extern "C" int cephd_mon(int argc, const char **argv)
187 #else
188 int main(int argc, const char **argv)
189 #endif
190 {
191 int err;
192
193 bool mkfs = false;
194 bool compact = false;
195 bool force_sync = false;
196 bool yes_really = false;
197 std::string osdmapfn, inject_monmap, extract_monmap;
198
199 vector<const char*> args;
200 argv_to_vec(argc, argv, args);
201 env_to_vec(args);
202
203 // We need to specify some default values that may be overridden by the
204 // user, that are specific to the monitor. The options we are overriding
205 // are also used on the OSD (or in any other component that uses leveldb),
206 // so changing them directly in common/config_opts.h is not an option.
207 // This is not the prettiest way of doing this, especially since it has us
208 // having a different place than common/config_opts.h defining default
209 // values, but it's not horribly wrong enough to prevent us from doing it :)
210 //
211 // NOTE: user-defined options will take precedence over ours.
212 //
213 // leveldb_write_buffer_size = 32*1024*1024 = 33554432 // 32MB
214 // leveldb_cache_size = 512*1024*1204 = 536870912 // 512MB
215 // leveldb_block_size = 64*1024 = 65536 // 64KB
216 // leveldb_compression = false
217 // leveldb_log = ""
218 vector<const char*> def_args;
219 def_args.push_back("--leveldb-write-buffer-size=33554432");
220 def_args.push_back("--leveldb-cache-size=536870912");
221 def_args.push_back("--leveldb-block-size=65536");
222 def_args.push_back("--leveldb-compression=false");
223 def_args.push_back("--leveldb-log=");
224
225 int flags = 0;
226 {
227 vector<const char*> args_copy = args;
228 std::string val;
229 for (std::vector<const char*>::iterator i = args_copy.begin();
230 i != args_copy.end(); ) {
231 if (ceph_argparse_double_dash(args_copy, i)) {
232 break;
233 } else if (ceph_argparse_flag(args_copy, i, "--mkfs", (char*)NULL)) {
234 flags |= CINIT_FLAG_NO_DAEMON_ACTIONS;
235 } else if (ceph_argparse_witharg(args_copy, i, &val, "--inject_monmap", (char*)NULL)) {
236 flags |= CINIT_FLAG_NO_DAEMON_ACTIONS;
237 } else if (ceph_argparse_witharg(args_copy, i, &val, "--extract-monmap", (char*)NULL)) {
238 flags |= CINIT_FLAG_NO_DAEMON_ACTIONS;
239 } else {
240 ++i;
241 }
242 }
243 }
244
245 auto cct = global_init(&def_args, args,
246 CEPH_ENTITY_TYPE_MON, CODE_ENVIRONMENT_DAEMON,
247 flags, "mon_data");
248 ceph_heap_profiler_init();
249
250 uuid_d fsid;
251 std::string val;
252 for (std::vector<const char*>::iterator i = args.begin(); i != args.end(); ) {
253 if (ceph_argparse_double_dash(args, i)) {
254 break;
255 } else if (ceph_argparse_flag(args, i, "-h", "--help", (char*)NULL)) {
256 usage();
257 } else if (ceph_argparse_flag(args, i, "--mkfs", (char*)NULL)) {
258 mkfs = true;
259 } else if (ceph_argparse_flag(args, i, "--compact", (char*)NULL)) {
260 compact = true;
261 } else if (ceph_argparse_flag(args, i, "--force-sync", (char*)NULL)) {
262 force_sync = true;
263 } else if (ceph_argparse_flag(args, i, "--yes-i-really-mean-it", (char*)NULL)) {
264 yes_really = true;
265 } else if (ceph_argparse_witharg(args, i, &val, "--osdmap", (char*)NULL)) {
266 osdmapfn = val;
267 } else if (ceph_argparse_witharg(args, i, &val, "--inject_monmap", (char*)NULL)) {
268 inject_monmap = val;
269 } else if (ceph_argparse_witharg(args, i, &val, "--extract-monmap", (char*)NULL)) {
270 extract_monmap = val;
271 } else {
272 ++i;
273 }
274 }
275 if (!args.empty()) {
276 derr << "too many arguments: " << args << dendl;
277 usage();
278 }
279
280 if (force_sync && !yes_really) {
281 derr << "are you SURE you want to force a sync? this will erase local data and may\n"
282 << "break your mon cluster. pass --yes-i-really-mean-it if you do." << dendl;
283 exit(1);
284 }
285
286 if (g_conf->mon_data.empty()) {
287 derr << "must specify '--mon-data=foo' data path" << dendl;
288 usage();
289 }
290
291 if (g_conf->name.get_id().empty()) {
292 derr << "must specify id (--id <id> or --name mon.<id>)" << dendl;
293 usage();
294 }
295
296 // -- mkfs --
297 if (mkfs) {
298
299 int err = check_mon_data_exists();
300 if (err == -ENOENT) {
301 if (::mkdir(g_conf->mon_data.c_str(), 0755)) {
302 derr << "mkdir(" << g_conf->mon_data << ") : "
303 << cpp_strerror(errno) << dendl;
304 exit(1);
305 }
306 } else if (err < 0) {
307 derr << "error opening '" << g_conf->mon_data << "': "
308 << cpp_strerror(-err) << dendl;
309 exit(-err);
310 }
311
312 err = check_mon_data_empty();
313 if (err == -ENOTEMPTY) {
314 // Mon may exist. Let the user know and exit gracefully.
315 derr << "'" << g_conf->mon_data << "' already exists and is not empty"
316 << ": monitor may already exist" << dendl;
317 exit(0);
318 } else if (err < 0) {
319 derr << "error checking if '" << g_conf->mon_data << "' is empty: "
320 << cpp_strerror(-err) << dendl;
321 exit(-err);
322 }
323
324 // resolve public_network -> public_addr
325 pick_addresses(g_ceph_context, CEPH_PICK_ADDRESS_PUBLIC);
326
327 common_init_finish(g_ceph_context);
328
329 bufferlist monmapbl, osdmapbl;
330 std::string error;
331 MonMap monmap;
332
333 // load or generate monmap
334 if (g_conf->monmap.length()) {
335 int err = monmapbl.read_file(g_conf->monmap.c_str(), &error);
336 if (err < 0) {
337 derr << argv[0] << ": error reading " << g_conf->monmap << ": " << error << dendl;
338 exit(1);
339 }
340 try {
341 monmap.decode(monmapbl);
342
343 // always mark seed/mkfs monmap as epoch 0
344 monmap.set_epoch(0);
345 }
346 catch (const buffer::error& e) {
347 derr << argv[0] << ": error decoding monmap " << g_conf->monmap << ": " << e.what() << dendl;
348 exit(1);
349 }
350 } else {
351 ostringstream oss;
352 int err = monmap.build_initial(g_ceph_context, oss);
353 if (oss.tellp())
354 derr << oss.str() << dendl;
355 if (err < 0) {
356 derr << argv[0] << ": warning: no initial monitors; must use admin socket to feed hints" << dendl;
357 }
358
359 // am i part of the initial quorum?
360 if (monmap.contains(g_conf->name.get_id())) {
361 // hmm, make sure the ip listed exists on the current host?
362 // maybe later.
363 } else if (!g_conf->public_addr.is_blank_ip()) {
364 entity_addr_t a = g_conf->public_addr;
365 if (a.get_port() == 0)
366 a.set_port(CEPH_MON_PORT);
367 if (monmap.contains(a)) {
368 string name;
369 monmap.get_addr_name(a, name);
370 monmap.rename(name, g_conf->name.get_id());
371 dout(0) << argv[0] << ": renaming mon." << name << " " << a
372 << " to mon." << g_conf->name.get_id() << dendl;
373 }
374 } else {
375 // is a local address listed without a name? if so, name myself.
376 list<entity_addr_t> ls;
377 monmap.list_addrs(ls);
378 entity_addr_t local;
379
380 if (have_local_addr(g_ceph_context, ls, &local)) {
381 string name;
382 monmap.get_addr_name(local, name);
383
384 if (name.compare(0, 7, "noname-") == 0) {
385 dout(0) << argv[0] << ": mon." << name << " " << local
386 << " is local, renaming to mon." << g_conf->name.get_id() << dendl;
387 monmap.rename(name, g_conf->name.get_id());
388 } else {
389 dout(0) << argv[0] << ": mon." << name << " " << local
390 << " is local, but not 'noname-' + something; not assuming it's me" << dendl;
391 }
392 }
393 }
394 }
395
396 if (!g_conf->fsid.is_zero()) {
397 monmap.fsid = g_conf->fsid;
398 dout(0) << argv[0] << ": set fsid to " << g_conf->fsid << dendl;
399 }
400
401 if (monmap.fsid.is_zero()) {
402 derr << argv[0] << ": generated monmap has no fsid; use '--fsid <uuid>'" << dendl;
403 exit(10);
404 }
405
406 //monmap.print(cout);
407
408 // osdmap
409 if (osdmapfn.length()) {
410 err = osdmapbl.read_file(osdmapfn.c_str(), &error);
411 if (err < 0) {
412 derr << argv[0] << ": error reading " << osdmapfn << ": "
413 << error << dendl;
414 exit(1);
415 }
416 }
417
418 // go
419 MonitorDBStore store(g_conf->mon_data);
420 ostringstream oss;
421 int r = store.create_and_open(oss);
422 if (oss.tellp())
423 derr << oss.str() << dendl;
424 if (r < 0) {
425 derr << argv[0] << ": error opening mon data directory at '"
426 << g_conf->mon_data << "': " << cpp_strerror(r) << dendl;
427 exit(1);
428 }
429 assert(r == 0);
430
431 Monitor mon(g_ceph_context, g_conf->name.get_id(), &store, 0, 0, &monmap);
432 r = mon.mkfs(osdmapbl);
433 if (r < 0) {
434 derr << argv[0] << ": error creating monfs: " << cpp_strerror(r) << dendl;
435 exit(1);
436 }
437 store.close();
438 dout(0) << argv[0] << ": created monfs at " << g_conf->mon_data
439 << " for " << g_conf->name << dendl;
440 return 0;
441 }
442
443 err = check_mon_data_exists();
444 if (err < 0 && err == -ENOENT) {
445 derr << "monitor data directory at '" << g_conf->mon_data << "'"
446 << " does not exist: have you run 'mkfs'?" << dendl;
447 exit(1);
448 } else if (err < 0) {
449 derr << "error accessing monitor data directory at '"
450 << g_conf->mon_data << "': " << cpp_strerror(-err) << dendl;
451 exit(1);
452 }
453
454 err = check_mon_data_empty();
455 if (err == 0) {
456 derr << "monitor data directory at '" << g_conf->mon_data
457 << "' is empty: have you run 'mkfs'?" << dendl;
458 exit(1);
459 } else if (err < 0 && err != -ENOTEMPTY) {
460 // we don't want an empty data dir by now
461 derr << "error accessing '" << g_conf->mon_data << "': "
462 << cpp_strerror(-err) << dendl;
463 exit(1);
464 }
465
466 {
467 // check fs stats. don't start if it's critically close to full.
468 ceph_data_stats_t stats;
469 int err = get_fs_stats(stats, g_conf->mon_data.c_str());
470 if (err < 0) {
471 derr << "error checking monitor data's fs stats: " << cpp_strerror(err)
472 << dendl;
473 exit(-err);
474 }
475 if (stats.avail_percent <= g_conf->mon_data_avail_crit) {
476 derr << "error: monitor data filesystem reached concerning levels of"
477 << " available storage space (available: "
478 << stats.avail_percent << "% " << prettybyte_t(stats.byte_avail)
479 << ")\nyou may adjust 'mon data avail crit' to a lower value"
480 << " to make this go away (default: " << g_conf->mon_data_avail_crit
481 << "%)\n" << dendl;
482 exit(ENOSPC);
483 }
484 }
485
486 // we fork early to prevent leveldb's environment static state from
487 // screwing us over
488 Preforker prefork;
489 if (!(flags & CINIT_FLAG_NO_DAEMON_ACTIONS)) {
490 if (global_init_prefork(g_ceph_context) >= 0) {
491 string err_msg;
492 err = prefork.prefork(err_msg);
493 if (err < 0) {
494 derr << err_msg << dendl;
495 prefork.exit(err);
496 }
497 if (prefork.is_parent()) {
498 err = prefork.parent_wait(err_msg);
499 if (err < 0)
500 derr << err_msg << dendl;
501 prefork.exit(err);
502 }
503 setsid();
504 global_init_postfork_start(g_ceph_context);
505 }
506 common_init_finish(g_ceph_context);
507 global_init_chdir(g_ceph_context);
508 #ifndef BUILDING_FOR_EMBEDDED
509 if (global_init_preload_erasure_code(g_ceph_context) < 0)
510 prefork.exit(1);
511 #else
512 cephd_preload_embedded_plugins();
513 #endif
514 }
515
516 MonitorDBStore *store = new MonitorDBStore(g_conf->mon_data);
517 {
518 ostringstream oss;
519 err = store->open(oss);
520 if (oss.tellp())
521 derr << oss.str() << dendl;
522 if (err < 0) {
523 derr << "error opening mon data directory at '"
524 << g_conf->mon_data << "': " << cpp_strerror(err) << dendl;
525 prefork.exit(1);
526 }
527 }
528
529 bufferlist magicbl;
530 err = store->get(Monitor::MONITOR_NAME, "magic", magicbl);
531 if (err || !magicbl.length()) {
532 derr << "unable to read magic from mon data" << dendl;
533 prefork.exit(1);
534 }
535 string magic(magicbl.c_str(), magicbl.length()-1); // ignore trailing \n
536 if (strcmp(magic.c_str(), CEPH_MON_ONDISK_MAGIC)) {
537 derr << "mon fs magic '" << magic << "' != current '" << CEPH_MON_ONDISK_MAGIC << "'" << dendl;
538 prefork.exit(1);
539 }
540
541 err = Monitor::check_features(store);
542 if (err < 0) {
543 derr << "error checking features: " << cpp_strerror(err) << dendl;
544 prefork.exit(1);
545 }
546
547 // inject new monmap?
548 if (!inject_monmap.empty()) {
549 bufferlist bl;
550 std::string error;
551 int r = bl.read_file(inject_monmap.c_str(), &error);
552 if (r) {
553 derr << "unable to read monmap from " << inject_monmap << ": "
554 << error << dendl;
555 prefork.exit(1);
556 }
557
558 // get next version
559 version_t v = store->get("monmap", "last_committed");
560 dout(0) << "last committed monmap epoch is " << v << ", injected map will be " << (v+1)
561 << dendl;
562 v++;
563
564 // set the version
565 MonMap tmp;
566 tmp.decode(bl);
567 if (tmp.get_epoch() != v) {
568 dout(0) << "changing monmap epoch from " << tmp.get_epoch()
569 << " to " << v << dendl;
570 tmp.set_epoch(v);
571 }
572 bufferlist mapbl;
573 tmp.encode(mapbl, CEPH_FEATURES_ALL);
574 bufferlist final;
575 ::encode(v, final);
576 ::encode(mapbl, final);
577
578 auto t(std::make_shared<MonitorDBStore::Transaction>());
579 // save it
580 t->put("monmap", v, mapbl);
581 t->put("monmap", "latest", final);
582 t->put("monmap", "last_committed", v);
583 store->apply_transaction(t);
584
585 dout(0) << "done." << dendl;
586 prefork.exit(0);
587 }
588
589 // monmap?
590 MonMap monmap;
591 {
592 // note that even if we don't find a viable monmap, we should go ahead
593 // and try to build it up in the next if-else block.
594 bufferlist mapbl;
595 int err = obtain_monmap(*store, mapbl);
596 if (err >= 0) {
597 try {
598 monmap.decode(mapbl);
599 } catch (const buffer::error& e) {
600 derr << "can't decode monmap: " << e.what() << dendl;
601 }
602 } else {
603 derr << "unable to obtain a monmap: " << cpp_strerror(err) << dendl;
604 }
605 if (!extract_monmap.empty()) {
606 int r = mapbl.write_file(extract_monmap.c_str());
607 if (r < 0) {
608 r = -errno;
609 derr << "error writing monmap to " << extract_monmap << ": " << cpp_strerror(r) << dendl;
610 prefork.exit(1);
611 }
612 derr << "wrote monmap to " << extract_monmap << dendl;
613 prefork.exit(0);
614 }
615 }
616
617 // this is what i will bind to
618 entity_addr_t ipaddr;
619
620 if (monmap.contains(g_conf->name.get_id())) {
621 ipaddr = monmap.get_addr(g_conf->name.get_id());
622
623 // print helpful warning if the conf file doesn't match
624 entity_addr_t conf_addr;
625 std::vector <std::string> my_sections;
626 g_conf->get_my_sections(my_sections);
627 std::string mon_addr_str;
628 if (g_conf->get_val_from_conf_file(my_sections, "mon addr",
629 mon_addr_str, true) == 0) {
630 if (conf_addr.parse(mon_addr_str.c_str()) && (ipaddr != conf_addr)) {
631 derr << "WARNING: 'mon addr' config option " << conf_addr
632 << " does not match monmap file" << std::endl
633 << " continuing with monmap configuration" << dendl;
634 }
635 }
636 } else {
637 dout(0) << g_conf->name << " does not exist in monmap, will attempt to join an existing cluster" << dendl;
638
639 pick_addresses(g_ceph_context, CEPH_PICK_ADDRESS_PUBLIC);
640 if (!g_conf->public_addr.is_blank_ip()) {
641 ipaddr = g_conf->public_addr;
642 if (ipaddr.get_port() == 0)
643 ipaddr.set_port(CEPH_MON_PORT);
644 dout(0) << "using public_addr " << g_conf->public_addr << " -> "
645 << ipaddr << dendl;
646 } else {
647 MonMap tmpmap;
648 ostringstream oss;
649 int err = tmpmap.build_initial(g_ceph_context, oss);
650 if (oss.tellp())
651 derr << oss.str() << dendl;
652 if (err < 0) {
653 derr << argv[0] << ": error generating initial monmap: "
654 << cpp_strerror(err) << dendl;
655 usage();
656 prefork.exit(1);
657 }
658 if (tmpmap.contains(g_conf->name.get_id())) {
659 ipaddr = tmpmap.get_addr(g_conf->name.get_id());
660 } else {
661 derr << "no public_addr or public_network specified, and " << g_conf->name
662 << " not present in monmap or ceph.conf" << dendl;
663 prefork.exit(1);
664 }
665 }
666 }
667
668 // bind
669 int rank = monmap.get_rank(g_conf->name.get_id());
670 std::string public_msgr_type = g_conf->ms_public_type.empty() ? g_conf->get_val<std::string>("ms_type") : g_conf->ms_public_type;
671 Messenger *msgr = Messenger::create(g_ceph_context, public_msgr_type,
672 entity_name_t::MON(rank), "mon",
673 0, Messenger::HAS_MANY_CONNECTIONS);
674 if (!msgr)
675 exit(1);
676 msgr->set_cluster_protocol(CEPH_MON_PROTOCOL);
677 msgr->set_default_send_priority(CEPH_MSG_PRIO_HIGH);
678
679 msgr->set_default_policy(Messenger::Policy::stateless_server(0));
680 msgr->set_policy(entity_name_t::TYPE_MON,
681 Messenger::Policy::lossless_peer_reuse(
682 CEPH_FEATURE_UID |
683 CEPH_FEATURE_PGID64 |
684 CEPH_FEATURE_MON_SINGLE_PAXOS));
685 msgr->set_policy(entity_name_t::TYPE_OSD,
686 Messenger::Policy::stateless_server(
687 CEPH_FEATURE_PGID64 |
688 CEPH_FEATURE_OSDENC));
689 msgr->set_policy(entity_name_t::TYPE_CLIENT,
690 Messenger::Policy::stateless_server(0));
691 msgr->set_policy(entity_name_t::TYPE_MDS,
692 Messenger::Policy::stateless_server(0));
693
694 // throttle client traffic
695 Throttle *client_throttler = new Throttle(g_ceph_context, "mon_client_bytes",
696 g_conf->mon_client_bytes);
697 msgr->set_policy_throttlers(entity_name_t::TYPE_CLIENT,
698 client_throttler, NULL);
699
700 // throttle daemon traffic
701 // NOTE: actual usage on the leader may multiply by the number of
702 // monitors if they forward large update messages from daemons.
703 Throttle *daemon_throttler = new Throttle(g_ceph_context, "mon_daemon_bytes",
704 g_conf->mon_daemon_bytes);
705 msgr->set_policy_throttlers(entity_name_t::TYPE_OSD, daemon_throttler,
706 NULL);
707 msgr->set_policy_throttlers(entity_name_t::TYPE_MDS, daemon_throttler,
708 NULL);
709
710 entity_addr_t bind_addr = ipaddr;
711 entity_addr_t public_addr = ipaddr;
712
713 // check if the public_bind_addr option is set
714 if (!g_conf->public_bind_addr.is_blank_ip()) {
715 bind_addr = g_conf->public_bind_addr;
716
717 // set the default port if not already set
718 if (bind_addr.get_port() == 0) {
719 bind_addr.set_port(CEPH_MON_PORT);
720 }
721 }
722
723 dout(0) << "starting " << g_conf->name << " rank " << rank
724 << " at public addr " << public_addr
725 << " at bind addr " << bind_addr
726 << " mon_data " << g_conf->mon_data
727 << " fsid " << monmap.get_fsid()
728 << dendl;
729
730 err = msgr->bind(bind_addr);
731 if (err < 0) {
732 derr << "unable to bind monitor to " << bind_addr << dendl;
733 prefork.exit(1);
734 }
735
736 // if the public and bind addr are different set the msgr addr
737 // to the public one, now that the bind is complete.
738 if (public_addr != bind_addr) {
739 msgr->set_addr(public_addr);
740 }
741
742 Messenger *mgr_msgr = Messenger::create(g_ceph_context, public_msgr_type,
743 entity_name_t::MON(rank), "mon-mgrc",
744 getpid(), 0);
745 if (!mgr_msgr) {
746 derr << "unable to create mgr_msgr" << dendl;
747 prefork.exit(1);
748 }
749
750 dout(0) << "starting " << g_conf->name << " rank " << rank
751 << " at " << ipaddr
752 << " mon_data " << g_conf->mon_data
753 << " fsid " << monmap.get_fsid()
754 << dendl;
755
756 // start monitor
757 mon = new Monitor(g_ceph_context, g_conf->name.get_id(), store,
758 msgr, mgr_msgr, &monmap);
759
760 if (force_sync) {
761 derr << "flagging a forced sync ..." << dendl;
762 ostringstream oss;
763 mon->sync_force(NULL, oss);
764 if (oss.tellp())
765 derr << oss.str() << dendl;
766 }
767
768 err = mon->preinit();
769 if (err < 0) {
770 derr << "failed to initialize" << dendl;
771 prefork.exit(1);
772 }
773
774 if (compact || g_conf->mon_compact_on_start) {
775 derr << "compacting monitor store ..." << dendl;
776 mon->store->compact();
777 derr << "done compacting" << dendl;
778 }
779
780 if (g_conf->daemonize) {
781 global_init_postfork_finish(g_ceph_context);
782 prefork.daemonize();
783 }
784
785 msgr->start();
786 mgr_msgr->start();
787
788 mon->init();
789
790 // set up signal handlers, now that we've daemonized/forked.
791 init_async_signal_handler();
792 register_async_signal_handler(SIGHUP, sighup_handler);
793 register_async_signal_handler_oneshot(SIGINT, handle_mon_signal);
794 register_async_signal_handler_oneshot(SIGTERM, handle_mon_signal);
795
796 if (g_conf->inject_early_sigterm)
797 kill(getpid(), SIGTERM);
798
799 msgr->wait();
800 mgr_msgr->wait();
801
802 store->close();
803
804 unregister_async_signal_handler(SIGHUP, sighup_handler);
805 unregister_async_signal_handler(SIGINT, handle_mon_signal);
806 unregister_async_signal_handler(SIGTERM, handle_mon_signal);
807 shutdown_async_signal_handler();
808
809 delete mon;
810 delete store;
811 delete msgr;
812 delete mgr_msgr;
813 delete client_throttler;
814 delete daemon_throttler;
815
816 // cd on exit, so that gmon.out (if any) goes into a separate directory for each node.
817 char s[20];
818 snprintf(s, sizeof(s), "gmon/%d", getpid());
819 if ((mkdir(s, 0755) == 0) && (chdir(s) == 0)) {
820 dout(0) << "ceph-mon: gmon.out should be in " << s << dendl;
821 }
822
823 prefork.signal_exit(0);
824 return 0;
825 }