~ubuntu-branches/ubuntu/utopic/ceph/utopic-proposed

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
// -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*- 
// vim: ts=8 sw=2 smarttab
/*
 * Ceph - scalable distributed file system
 *
 * Copyright (C) 2004-2006 Sage Weil <sage@newdream.net>
 *
 * This is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License version 2.1, as published by the Free Software 
 * Foundation.  See file COPYING.
 * 
 */

#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>

#include <iostream>
#include <string>
using namespace std;

#include "common/config.h"
#include "include/ceph_features.h"

#include "mon/MonMap.h"
#include "mon/Monitor.h"
#include "mon/MonitorDBStore.h"
#include "mon/MonClient.h"

#include "msg/Messenger.h"

#include "include/CompatSet.h"

#include "common/ceph_argparse.h"
#include "common/pick_address.h"
#include "common/Timer.h"
#include "common/errno.h"
#include "common/Preforker.h"

#include "global/global_init.h"
#include "global/signal_handler.h"

#include "include/assert.h"

#include "erasure-code/ErasureCodePlugin.h"

#define dout_subsys ceph_subsys_mon

Monitor *mon = NULL;

void handle_mon_signal(int signum)
{
  if (mon)
    mon->handle_signal(signum);
}


int obtain_monmap(MonitorDBStore &store, bufferlist &bl)
{
  dout(10) << __func__ << dendl;
  /*
   * the monmap may be in one of three places:
   *  'monmap:<latest_version_no>' - the monmap we'd really like to have
   *  'mon_sync:latest_monmap'     - last monmap backed up for the last sync
   *  'mkfs:monmap'                - a monmap resulting from mkfs
   */

  if (store.exists("monmap", "last_committed")) {
    version_t latest_ver = store.get("monmap", "last_committed");
    if (store.exists("monmap", latest_ver)) {
      int err = store.get("monmap", latest_ver, bl);
      assert(err == 0);
      assert(bl.length() > 0);
      dout(10) << __func__ << " read last committed monmap ver "
               << latest_ver << dendl;
      return 0;
    }
  }

  if (store.exists("mon_sync", "in_sync")
      || store.exists("mon_sync", "force_sync")) {
    dout(10) << __func__ << " detected aborted sync" << dendl;
    if (store.exists("mon_sync", "latest_monmap")) {
      int err = store.get("mon_sync", "latest_monmap", bl);
      assert(err == 0);
      assert(bl.length() > 0);
      dout(10) << __func__ << " read backup monmap" << dendl;
      return 0;
    }
  }

  if (store.exists("mkfs", "monmap")) {
    dout(10) << __func__ << " found mkfs monmap" << dendl;
    int err = store.get("mkfs", "monmap", bl);
    assert(err == 0);
    assert(bl.length() > 0);
    return 0;
  }

  derr << __func__ << " unable to find a monmap" << dendl;
  return -ENOENT;
}

int mon_data_exists(bool *r)
{
  string mon_data = g_conf->mon_data;
  struct stat buf;
  if (::stat(mon_data.c_str(), &buf)) {
    if (errno == ENOENT) {
      *r = false;
    } else {
      cerr << "stat(" << mon_data << ") " << cpp_strerror(errno) << std::endl;
      return -errno;
    }
  } else {
    *r = true;
  }
  return 0;
}

int mon_data_empty(bool *r)
{
  string mon_data = g_conf->mon_data;

  DIR *dir = ::opendir(mon_data.c_str());
  if (!dir) {
    cerr << "opendir(" << mon_data << ") " << cpp_strerror(errno) << std::endl;
    return -errno;
  }
  char buf[offsetof(struct dirent, d_name) + PATH_MAX + 1];

  *r = false;
  int code = 0;
  struct dirent *de;
  errno = 0;
  while (!::readdir_r(dir, reinterpret_cast<struct dirent*>(buf), &de)) {
    if (!de) {
      if (errno) {
	cerr << "readdir(" << mon_data << ") " << cpp_strerror(errno) << std::endl;
	code = -errno;
      }
      break;
    }
    if (string(".") != de->d_name &&
	string("..") != de->d_name) {
      *r = true;
      break;
    }
  }

  ::closedir(dir);

  return code;
}

int mon_exists(bool *r)
{
  int code = mon_data_exists(r);
  if (code || *r == false)
    return code;
  return mon_data_empty(r);
}

void usage()
{
  cerr << "usage: ceph-mon -i monid [flags]" << std::endl;
  cerr << "  --debug_mon n\n";
  cerr << "        debug monitor level (e.g. 10)\n";
  cerr << "  --mkfs\n";
  cerr << "        build fresh monitor fs\n";
  cerr << "  --force-sync\n";
  cerr << "        force a sync from another mon by wiping local data (BE CAREFUL)\n";
  cerr << "  --yes-i-really-mean-it\n";
  cerr << "        mandatory safeguard for --force-sync\n";
  cerr << "  --compact\n";
  cerr << "        compact the monitor store\n";
  cerr << "  --osdmap <filename>\n";
  cerr << "        only used when --mkfs is provided: load the osdmap from <filename>\n";
  cerr << "  --inject-monmap <filename>\n";
  cerr << "        write the <filename> monmap to the local monitor store and exit\n";
  cerr << "  --extract-monmap <filename>\n";
  cerr << "        extract the monmap from the local monitor store and exit\n";
  cerr << "  --mon-data <directory>\n";
  cerr << "        where the mon store and keyring are located\n";
  generic_server_usage();
}

int preload_erasure_code()
{
  string directory = g_conf->osd_pool_default_erasure_code_directory;
  string plugins = g_conf->osd_erasure_code_plugins;
  stringstream ss;
  int r = ErasureCodePluginRegistry::instance().preload(plugins,
							directory,
							ss);
  if (r)
    derr << ss.str() << dendl;
  else
    dout(10) << ss.str() << dendl;
  return r;
}

int main(int argc, const char **argv) 
{
  int err;

  bool mkfs = false;
  bool compact = false;
  bool force_sync = false;
  bool yes_really = false;
  std::string osdmapfn, inject_monmap, extract_monmap;

  vector<const char*> args;
  argv_to_vec(argc, argv, args);
  env_to_vec(args);

  int flags = 0;
  {
    vector<const char*> args_copy = args;
    std::string val;
    for (std::vector<const char*>::iterator i = args_copy.begin();
	 i != args_copy.end(); ) {
      if (ceph_argparse_double_dash(args_copy, i)) {
	break;
      } else if (ceph_argparse_flag(args_copy, i, "--mkfs", (char*)NULL)) {
	flags |= CINIT_FLAG_NO_DAEMON_ACTIONS;
      } else if (ceph_argparse_witharg(args_copy, i, &val, "--inject_monmap", (char*)NULL)) {
	flags |= CINIT_FLAG_NO_DAEMON_ACTIONS;
      } else if (ceph_argparse_witharg(args_copy, i, &val, "--extract-monmap", (char*)NULL)) {
	flags |= CINIT_FLAG_NO_DAEMON_ACTIONS;
      } else {
	++i;
      }
    }
  }

  global_init(NULL, args, CEPH_ENTITY_TYPE_MON, CODE_ENVIRONMENT_DAEMON, flags);

  uuid_d fsid;
  std::string val;
  for (std::vector<const char*>::iterator i = args.begin(); i != args.end(); ) {
    if (ceph_argparse_double_dash(args, i)) {
      break;
    } else if (ceph_argparse_flag(args, i, "-h", "--help", (char*)NULL)) {
      usage();
      exit(0);
    } else if (ceph_argparse_flag(args, i, "--mkfs", (char*)NULL)) {
      mkfs = true;
    } else if (ceph_argparse_flag(args, i, "--compact", (char*)NULL)) {
      compact = true;
    } else if (ceph_argparse_flag(args, i, "--force-sync", (char*)NULL)) {
      force_sync = true;
    } else if (ceph_argparse_flag(args, i, "--yes-i-really-mean-it", (char*)NULL)) {
      yes_really = true;
    } else if (ceph_argparse_witharg(args, i, &val, "--osdmap", (char*)NULL)) {
      osdmapfn = val;
    } else if (ceph_argparse_witharg(args, i, &val, "--inject_monmap", (char*)NULL)) {
      inject_monmap = val;
    } else if (ceph_argparse_witharg(args, i, &val, "--extract-monmap", (char*)NULL)) {
      extract_monmap = val;
    } else {
      ++i;
    }
  }
  if (!args.empty()) {
    cerr << "too many arguments: " << args << std::endl;
    usage();
  }

  if (force_sync && !yes_really) {
    cerr << "are you SURE you want to force a sync?  this will erase local data and may\n"
	 << "break your mon cluster.  pass --yes-i-really-mean-it if you do." << std::endl;
    exit(1);
  }

  if (g_conf->mon_data.empty()) {
    cerr << "must specify '--mon-data=foo' data path" << std::endl;
    usage();
  }

  if (g_conf->name.get_id().empty()) {
    cerr << "must specify id (--id <id> or --name mon.<id>)" << std::endl;
    usage();
  }

  bool exists;
  if (mon_exists(&exists))
    exit(1);

  if (mkfs && exists) {
    cerr << g_conf->mon_data << " already exists" << std::endl;
    exit(0);
  }

  // -- mkfs --
  if (mkfs) {

    if (mon_data_exists(&exists))
      exit(1);

    if (!exists) {
      if (::mkdir(g_conf->mon_data.c_str(), 0755)) {
	cerr << "mkdir(" << g_conf->mon_data << ") : "
	     << cpp_strerror(errno) << std::endl;
	exit(1);
      }
    }

    // resolve public_network -> public_addr
    pick_addresses(g_ceph_context, CEPH_PICK_ADDRESS_PUBLIC);

    common_init_finish(g_ceph_context, flags);

    bufferlist monmapbl, osdmapbl;
    std::string error;
    MonMap monmap;

    // load or generate monmap
    if (g_conf->monmap.length()) {
      int err = monmapbl.read_file(g_conf->monmap.c_str(), &error);
      if (err < 0) {
	cerr << argv[0] << ": error reading " << g_conf->monmap << ": " << error << std::endl;
	exit(1);
      }
      try {
	monmap.decode(monmapbl);

	// always mark seed/mkfs monmap as epoch 0
	monmap.set_epoch(0);
      }
      catch (const buffer::error& e) {
	cerr << argv[0] << ": error decoding monmap " << g_conf->monmap << ": " << e.what() << std::endl;
	exit(1);
      }      
    } else {
      int err = monmap.build_initial(g_ceph_context, cerr);
      if (err < 0) {
	cerr << argv[0] << ": warning: no initial monitors; must use admin socket to feed hints" << std::endl;
      }

      // am i part of the initial quorum?
      if (monmap.contains(g_conf->name.get_id())) {
	// hmm, make sure the ip listed exists on the current host?
	// maybe later.
      } else if (!g_conf->public_addr.is_blank_ip()) {
	entity_addr_t a = g_conf->public_addr;
	if (a.get_port() == 0)
	  a.set_port(CEPH_MON_PORT);
	if (monmap.contains(a)) {
	  string name;
	  monmap.get_addr_name(a, name);
	  monmap.rename(name, g_conf->name.get_id());
	  cout << argv[0] << ": renaming mon." << name << " " << a
	       << " to mon." << g_conf->name.get_id() << std::endl;
	}
      } else {
	// is a local address listed without a name?  if so, name myself.
	list<entity_addr_t> ls;
	monmap.list_addrs(ls);
	entity_addr_t local;

	if (have_local_addr(g_ceph_context, ls, &local)) {
	  string name;
	  monmap.get_addr_name(local, name);

	  if (name.find("noname-") == 0) {
	    cout << argv[0] << ": mon." << name << " " << local
		 << " is local, renaming to mon." << g_conf->name.get_id() << std::endl;
	    monmap.rename(name, g_conf->name.get_id());
	  } else {
	    cout << argv[0] << ": mon." << name << " " << local
		 << " is local, but not 'noname-' + something; not assuming it's me" << std::endl;
	  }
	}
      }
    }

    if (!g_conf->fsid.is_zero()) {
      monmap.fsid = g_conf->fsid;
      cout << argv[0] << ": set fsid to " << g_conf->fsid << std::endl;
    }
    
    if (monmap.fsid.is_zero()) {
      cerr << argv[0] << ": generated monmap has no fsid; use '--fsid <uuid>'" << std::endl;
      exit(10);
    }

    //monmap.print(cout);

    // osdmap
    if (osdmapfn.length()) {
      err = osdmapbl.read_file(osdmapfn.c_str(), &error);
      if (err < 0) {
	cerr << argv[0] << ": error reading " << osdmapfn << ": "
	     << error << std::endl;
	exit(1);
      }
    }

    // go
    MonitorDBStore store(g_conf->mon_data);
    int r = store.create_and_open(cerr);
    if (r < 0) {
      cerr << argv[0] << ": error opening mon data directory at '"
           << g_conf->mon_data << "': " << cpp_strerror(r) << std::endl;
      exit(1);
    }
    assert(r == 0);

    Monitor mon(g_ceph_context, g_conf->name.get_id(), &store, 0, &monmap);
    r = mon.mkfs(osdmapbl);
    if (r < 0) {
      cerr << argv[0] << ": error creating monfs: " << cpp_strerror(r) << std::endl;
      exit(1);
    }
    cout << argv[0] << ": created monfs at " << g_conf->mon_data 
	 << " for " << g_conf->name << std::endl;
    return 0;
  }

  {
    // check fs stats. don't start if it's critically close to full.
    ceph_data_stats_t stats;
    int err = get_fs_stats(stats, g_conf->mon_data.c_str());
    if (err < 0) {
      cerr << "error checking monitor data's fs stats: " << cpp_strerror(err)
           << std::endl;
      exit(-err);
    }
    if (stats.avail_percent <= g_conf->mon_data_avail_crit) {
      cerr << "error: monitor data filesystem reached concerning levels of"
           << " available storage space (available: "
           << stats.avail_percent << "% " << prettybyte_t(stats.byte_avail)
           << ")\nyou may adjust 'mon data avail crit' to a lower value"
           << " to make this go away (default: " << g_conf->mon_data_avail_crit
           << "%)\n" << std::endl;
      exit(ENOSPC);
    }
  }

  // we fork early to prevent leveldb's environment static state from
  // screwing us over
  Preforker prefork;
  if (!(flags & CINIT_FLAG_NO_DAEMON_ACTIONS)) {
    if (global_init_prefork(g_ceph_context, 0) >= 0) {
      prefork.prefork();
      if (prefork.is_parent()) {
	return prefork.parent_wait();
      }
      global_init_postfork_start(g_ceph_context);
    }
    common_init_finish(g_ceph_context);
    global_init_chdir(g_ceph_context);
    if (preload_erasure_code() < -1)
      prefork.exit(1);
  }

  MonitorDBStore *store = new MonitorDBStore(g_conf->mon_data);

  Monitor::StoreConverter converter(g_conf->mon_data, store);
  if (store->open(std::cerr) < 0) {
    int ret = store->create_and_open(std::cerr);
    if (ret < 0) {
      derr << "failed to create new leveldb store" << dendl;
      prefork.exit(1);
    }

    ret = converter.needs_conversion();
    if (ret < 0) {
      derr << "found errors while validating legacy unconverted monitor store: "
           << cpp_strerror(ret) << dendl;
      prefork.exit(1);
    }
    if (ret > 0) {
      dout(0) << "converting monitor store, please do not interrupt..." << dendl;
      int r = converter.convert();
      if (r) {
	derr << "failed to convert monitor store: " << cpp_strerror(r) << dendl;
	prefork.exit(1);
      }
    }
  } else if (converter.is_converting()) {
    derr << "there is an on-going (maybe aborted?) conversion." << dendl;
    derr << "you should check what happened" << dendl;
    derr << "remove store.db to restart conversion" << dendl;
    prefork.exit(1);
  }

  bufferlist magicbl;
  err = store->get(Monitor::MONITOR_NAME, "magic", magicbl);
  if (!magicbl.length()) {
    derr << "unable to read magic from mon data.. did you run mkcephfs?" << dendl;
    prefork.exit(1);
  }
  string magic(magicbl.c_str(), magicbl.length()-1);  // ignore trailing \n
  if (strcmp(magic.c_str(), CEPH_MON_ONDISK_MAGIC)) {
    derr << "mon fs magic '" << magic << "' != current '" << CEPH_MON_ONDISK_MAGIC << "'" << dendl;
    prefork.exit(1);
  }

  err = Monitor::check_features(store);
  if (err < 0) {
    derr << "error checking features: " << cpp_strerror(err) << dendl;
    prefork.exit(1);
  }

  // inject new monmap?
  if (!inject_monmap.empty()) {
    bufferlist bl;
    std::string error;
    int r = bl.read_file(inject_monmap.c_str(), &error);
    if (r) {
      derr << "unable to read monmap from " << inject_monmap << ": "
	   << error << dendl;
      prefork.exit(1);
    }

    // get next version
    version_t v = store->get("monmap", "last_committed");
    dout(0) << "last committed monmap epoch is " << v << ", injected map will be " << (v+1)
            << dendl;
    v++;

    // set the version
    MonMap tmp;
    tmp.decode(bl);
    if (tmp.get_epoch() != v) {
      dout(0) << "changing monmap epoch from " << tmp.get_epoch()
           << " to " << v << dendl;
      tmp.set_epoch(v);
    }
    bufferlist mapbl;
    tmp.encode(mapbl, CEPH_FEATURES_ALL);
    bufferlist final;
    ::encode(v, final);
    ::encode(mapbl, final);

    MonitorDBStore::Transaction t;
    // save it
    t.put("monmap", v, mapbl);
    t.put("monmap", "latest", final);
    t.put("monmap", "last_committed", v);
    store->apply_transaction(t);

    dout(0) << "done." << dendl;
    prefork.exit(0);
  }

  // monmap?
  MonMap monmap;
  {
    // note that even if we don't find a viable monmap, we should go ahead
    // and try to build it up in the next if-else block.
    bufferlist mapbl;
    int err = obtain_monmap(*store, mapbl);
    if (err >= 0) {
      try {
        monmap.decode(mapbl);
      } catch (const buffer::error& e) {
        cerr << "can't decode monmap: " << e.what() << std::endl;
      }
    } else {
      derr << "unable to obtain a monmap: " << cpp_strerror(err) << dendl;
    }
    if (!extract_monmap.empty()) {
      int r = mapbl.write_file(extract_monmap.c_str());
      if (r < 0) {
	r = -errno;
	derr << "error writing monmap to " << extract_monmap << ": " << cpp_strerror(r) << dendl;
	prefork.exit(1);
      }
      derr << "wrote monmap to " << extract_monmap << dendl;
      prefork.exit(0);
    }
  }

  // this is what i will bind to
  entity_addr_t ipaddr;

  if (monmap.contains(g_conf->name.get_id())) {
    ipaddr = monmap.get_addr(g_conf->name.get_id());

    // print helpful warning if the conf file doesn't match
    entity_addr_t conf_addr;
    std::vector <std::string> my_sections;
    g_conf->get_my_sections(my_sections);
    std::string mon_addr_str;
    if (g_conf->get_val_from_conf_file(my_sections, "mon addr",
				       mon_addr_str, true) == 0) {
      if (conf_addr.parse(mon_addr_str.c_str()) && (ipaddr != conf_addr)) {
	derr << "WARNING: 'mon addr' config option " << conf_addr
	     << " does not match monmap file" << std::endl
	     << "         continuing with monmap configuration" << dendl;
      }
    }
  } else {
    dout(0) << g_conf->name << " does not exist in monmap, will attempt to join an existing cluster" << dendl;

    pick_addresses(g_ceph_context, CEPH_PICK_ADDRESS_PUBLIC);
    if (!g_conf->public_addr.is_blank_ip()) {
      ipaddr = g_conf->public_addr;
      if (ipaddr.get_port() == 0)
	ipaddr.set_port(CEPH_MON_PORT);
      dout(0) << "using public_addr " << g_conf->public_addr << " -> "
	      << ipaddr << dendl;
    } else {
      MonMap tmpmap;
      int err = tmpmap.build_initial(g_ceph_context, cerr);
      if (err < 0) {
	derr << argv[0] << ": error generating initial monmap: "
             << cpp_strerror(err) << dendl;
	usage();
	prefork.exit(1);
      }
      if (tmpmap.contains(g_conf->name.get_id())) {
	ipaddr = tmpmap.get_addr(g_conf->name.get_id());
      } else {
	derr << "no public_addr or public_network specified, and " << g_conf->name
	     << " not present in monmap or ceph.conf" << dendl;
	prefork.exit(1);
      }
    }
  }

  // bind
  int rank = monmap.get_rank(g_conf->name.get_id());
  Messenger *messenger = Messenger::create(g_ceph_context,
					   entity_name_t::MON(rank),
					   "mon",
					   0);
  messenger->set_cluster_protocol(CEPH_MON_PROTOCOL);
  messenger->set_default_send_priority(CEPH_MSG_PRIO_HIGH);

  uint64_t supported =
    CEPH_FEATURE_UID |
    CEPH_FEATURE_NOSRCADDR |
    CEPH_FEATURE_MONCLOCKCHECK |
    CEPH_FEATURE_PGID64 |
    CEPH_FEATURE_MSG_AUTH;
  messenger->set_default_policy(Messenger::Policy::stateless_server(supported, 0));
  messenger->set_policy(entity_name_t::TYPE_MON,
                        Messenger::Policy::lossless_peer_reuse(supported,
							       CEPH_FEATURE_UID |
							       CEPH_FEATURE_PGID64 |
							       CEPH_FEATURE_MON_SINGLE_PAXOS));
  messenger->set_policy(entity_name_t::TYPE_OSD,
                        Messenger::Policy::stateless_server(supported,
                                                            CEPH_FEATURE_PGID64 |
                                                            CEPH_FEATURE_OSDENC));
  messenger->set_policy(entity_name_t::TYPE_CLIENT,
			Messenger::Policy::stateless_server(supported, 0));
  messenger->set_policy(entity_name_t::TYPE_MDS,
			Messenger::Policy::stateless_server(supported, 0));


  // throttle client traffic
  Throttle *client_throttler = new Throttle(g_ceph_context, "mon_client_bytes",
					    g_conf->mon_client_bytes);
  messenger->set_policy_throttlers(entity_name_t::TYPE_CLIENT, client_throttler, NULL);

  // throttle daemon traffic
  // NOTE: actual usage on the leader may multiply by the number of
  // monitors if they forward large update messages from daemons.
  Throttle *daemon_throttler = new Throttle(g_ceph_context, "mon_daemon_bytes",
					    g_conf->mon_daemon_bytes);
  messenger->set_policy_throttlers(entity_name_t::TYPE_OSD, daemon_throttler, NULL);
  messenger->set_policy_throttlers(entity_name_t::TYPE_MDS, daemon_throttler, NULL);

  dout(0) << "starting " << g_conf->name << " rank " << rank
       << " at " << ipaddr
       << " mon_data " << g_conf->mon_data
       << " fsid " << monmap.get_fsid()
       << dendl;

  err = messenger->bind(ipaddr);
  if (err < 0) {
    derr << "unable to bind monitor to " << ipaddr << dendl;
    prefork.exit(1);
  }

  // start monitor
  mon = new Monitor(g_ceph_context, g_conf->name.get_id(), store, 
		    messenger, &monmap);

  if (force_sync) {
    derr << "flagging a forced sync ..." << dendl;
    mon->sync_force(NULL, cerr);
  }

  err = mon->preinit();
  if (err < 0) {
    derr << "failed to initialize" << dendl;
    prefork.exit(1);
  }

  if (compact || g_conf->mon_compact_on_start) {
    derr << "compacting monitor store ..." << dendl;
    mon->store->compact();
    derr << "done compacting" << dendl;
  }

  if (g_conf->daemonize) {
    global_init_postfork_finish(g_ceph_context, 0);
    prefork.daemonize();
  }

  messenger->start();

  mon->init();

  // set up signal handlers, now that we've daemonized/forked.
  init_async_signal_handler();
  register_async_signal_handler(SIGHUP, sighup_handler);
  register_async_signal_handler_oneshot(SIGINT, handle_mon_signal);
  register_async_signal_handler_oneshot(SIGTERM, handle_mon_signal);

  if (g_conf->inject_early_sigterm)
    kill(getpid(), SIGTERM);

  messenger->wait();

  unregister_async_signal_handler(SIGHUP, sighup_handler);
  unregister_async_signal_handler(SIGINT, handle_mon_signal);
  unregister_async_signal_handler(SIGTERM, handle_mon_signal);
  shutdown_async_signal_handler();

  delete mon;
  delete store;
  delete messenger;
  delete client_throttler;
  delete daemon_throttler;
  g_ceph_context->put();

  // cd on exit, so that gmon.out (if any) goes into a separate directory for each node.
  char s[20];
  snprintf(s, sizeof(s), "gmon/%d", getpid());
  if ((mkdir(s, 0755) == 0) && (chdir(s) == 0)) {
    dout(0) << "ceph-mon: gmon.out should be in " << s << dendl;
  }

  prefork.signal_exit(0);
  return 0;
}