forked from ceph/ceph
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathOSDMap.h
1571 lines (1378 loc) · 49.3 KB
/
OSDMap.h
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
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
// -*- 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 <[email protected]>
* Copyright (C) 2013,2014 Cloudwatt <[email protected]>
*
* Author: Loic Dachary <[email protected]>
*
* 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.
*
*/
#ifndef CEPH_OSDMAP_H
#define CEPH_OSDMAP_H
/*
* describe properties of the OSD cluster.
* disks, disk groups, total # osds,
*
*/
#include <vector>
#include <list>
#include <set>
#include <map>
#include <memory>
#include <boost/smart_ptr/local_shared_ptr.hpp>
#include "include/btree_map.h"
#include "include/common_fwd.h"
#include "include/types.h"
#include "common/ceph_releases.h"
#include "osd_types.h"
//#include "include/ceph_features.h"
#include "crush/CrushWrapper.h"
// forward declaration
class CrushWrapper;
class health_check_map_t;
/*
* we track up to two intervals during which the osd was alive and
* healthy. the most recent is [up_from,up_thru), where up_thru is
* the last epoch the osd is known to have _started_. i.e., a lower
* bound on the actual osd death. down_at (if it is > up_from) is an
* upper bound on the actual osd death.
*
* the second is the last_clean interval [begin,end). in that case,
* the last interval is the last epoch known to have been either
* _finished_, or during which the osd cleanly shut down. when
* possible, we push this forward to the epoch the osd was eventually
* marked down.
*
* the lost_at is used to allow build_prior to proceed without waiting
* for an osd to recover. In certain cases, progress may be blocked
* because an osd is down that may contain updates (i.e., a pg may have
* gone rw during an interval). If the osd can't be brought online, we
* can force things to proceed knowing that we _might_ be losing some
* acked writes. If the osd comes back to life later, that's fine to,
* but those writes will still be lost (the divergent objects will be
* thrown out).
*/
struct osd_info_t {
epoch_t last_clean_begin; // last interval that ended with a clean osd shutdown
epoch_t last_clean_end;
epoch_t up_from; // epoch osd marked up
epoch_t up_thru; // lower bound on actual osd death (if > up_from)
epoch_t down_at; // upper bound on actual osd death (if > up_from)
epoch_t lost_at; // last epoch we decided data was "lost"
osd_info_t() : last_clean_begin(0), last_clean_end(0),
up_from(0), up_thru(0), down_at(0), lost_at(0) {}
void dump(ceph::Formatter *f) const;
void encode(ceph::buffer::list& bl) const;
void decode(ceph::buffer::list::const_iterator& bl);
static void generate_test_instances(std::list<osd_info_t*>& o);
};
WRITE_CLASS_ENCODER(osd_info_t)
std::ostream& operator<<(std::ostream& out, const osd_info_t& info);
struct osd_xinfo_t {
utime_t down_stamp; ///< timestamp when we were last marked down
float laggy_probability; ///< encoded as __u32: 0 = definitely not laggy, 0xffffffff definitely laggy
__u32 laggy_interval; ///< average interval between being marked laggy and recovering
uint64_t features; ///< features supported by this osd we should know about
__u32 old_weight; ///< weight prior to being auto marked out
utime_t last_purged_snaps_scrub; ///< last scrub of purged_snaps
epoch_t dead_epoch = 0; ///< last epoch we were confirmed dead (not just down)
osd_xinfo_t() : laggy_probability(0), laggy_interval(0),
features(0), old_weight(0) {}
void dump(ceph::Formatter *f) const;
void encode(ceph::buffer::list& bl, uint64_t features) const;
void decode(ceph::buffer::list::const_iterator& bl);
static void generate_test_instances(std::list<osd_xinfo_t*>& o);
};
WRITE_CLASS_ENCODER_FEATURES(osd_xinfo_t)
std::ostream& operator<<(std::ostream& out, const osd_xinfo_t& xi);
struct PGTempMap {
#if 1
ceph::buffer::list data;
typedef btree::btree_map<pg_t,ceph_le32*> map_t;
map_t map;
void encode(ceph::buffer::list& bl) const {
using ceph::encode;
uint32_t n = map.size();
encode(n, bl);
for (auto &p : map) {
encode(p.first, bl);
bl.append((char*)p.second, (*p.second + 1) * sizeof(ceph_le32));
}
}
void decode(ceph::buffer::list::const_iterator& p) {
using ceph::decode;
data.clear();
map.clear();
uint32_t n;
decode(n, p);
if (!n)
return;
auto pstart = p;
size_t start_off = pstart.get_off();
std::vector<std::pair<pg_t,size_t>> offsets;
offsets.resize(n);
for (unsigned i=0; i<n; ++i) {
pg_t pgid;
decode(pgid, p);
offsets[i].first = pgid;
offsets[i].second = p.get_off() - start_off;
uint32_t vn;
decode(vn, p);
p += vn * sizeof(int32_t);
}
size_t len = p.get_off() - start_off;
pstart.copy(len, data);
if (data.get_num_buffers() > 1) {
data.rebuild();
}
//map.reserve(n);
char *start = data.c_str();
for (auto i : offsets) {
map.insert(map.end(), std::make_pair(i.first, (ceph_le32*)(start + i.second)));
}
}
void rebuild() {
ceph::buffer::list bl;
encode(bl);
auto p = std::cbegin(bl);
decode(p);
}
friend bool operator==(const PGTempMap& l, const PGTempMap& r) {
return
l.map.size() == r.map.size() &&
l.data.contents_equal(r.data);
}
class iterator {
map_t::const_iterator it;
map_t::const_iterator end;
std::pair<pg_t,std::vector<int32_t>> current;
void init_current() {
if (it != end) {
current.first = it->first;
ceph_assert(it->second);
current.second.resize(*it->second);
ceph_le32 *p = it->second + 1;
for (uint32_t n = 0; n < *it->second; ++n, ++p) {
current.second[n] = *p;
}
}
}
public:
iterator(map_t::const_iterator p,
map_t::const_iterator e)
: it(p), end(e) {
init_current();
}
const std::pair<pg_t,std::vector<int32_t>>& operator*() const {
return current;
}
const std::pair<pg_t,std::vector<int32_t>>* operator->() const {
return ¤t;
}
friend bool operator==(const iterator& l, const iterator& r) {
return l.it == r.it;
}
friend bool operator!=(const iterator& l, const iterator& r) {
return l.it != r.it;
}
iterator& operator++() {
++it;
if (it != end)
init_current();
return *this;
}
iterator operator++(int) {
iterator r = *this;
++it;
if (it != end)
init_current();
return r;
}
};
iterator begin() const {
return iterator(map.begin(), map.end());
}
iterator end() const {
return iterator(map.end(), map.end());
}
iterator find(pg_t pgid) const {
return iterator(map.find(pgid), map.end());
}
size_t size() const {
return map.size();
}
size_t count(pg_t pgid) const {
return map.count(pgid);
}
void erase(pg_t pgid) {
map.erase(pgid);
}
void clear() {
map.clear();
data.clear();
}
void set(pg_t pgid, const mempool::osdmap::vector<int32_t>& v) {
using ceph::encode;
size_t need = sizeof(ceph_le32) * (1 + v.size());
if (need < data.get_append_buffer_unused_tail_length()) {
ceph::buffer::ptr z(data.get_append_buffer_unused_tail_length());
z.zero();
data.append(z.c_str(), z.length());
}
encode(v, data);
map[pgid] = (ceph_le32*)(data.back().end_c_str()) - (1 + v.size());
}
mempool::osdmap::vector<int32_t> get(pg_t pgid) {
mempool::osdmap::vector<int32_t> v;
ceph_le32 *p = map[pgid];
size_t n = *p++;
v.resize(n);
for (size_t i = 0; i < n; ++i, ++p) {
v[i] = *p;
}
return v;
}
#else
// trivial implementation
mempool::osdmap::map<pg_t,mempool::osdmap::vector<int32_t> > pg_temp;
void encode(ceph::buffer::list& bl) const {
encode(pg_temp, bl);
}
void decode(ceph::buffer::list::const_iterator& p) {
decode(pg_temp, p);
}
friend bool operator==(const PGTempMap& l, const PGTempMap& r) {
return
l.pg_temp.size() == r.pg_temp.size() &&
l.pg_temp == r.pg_temp;
}
class iterator {
mempool::osdmap::map<pg_t,mempool::osdmap::vector<int32_t> >::const_iterator it;
public:
iterator(mempool::osdmap::map<pg_t,
mempool::osdmap::vector<int32_t> >::const_iterator p)
: it(p) {}
std::pair<pg_t,const mempool::osdmap::vector<int32_t>&> operator*() const {
return *it;
}
const std::pair<const pg_t,mempool::osdmap::vector<int32_t>>* operator->() const {
return &*it;
}
friend bool operator==(const iterator& l, const iterator& r) {
return l.it == r.it;
}
friend bool operator!=(const iterator& l, const iterator& r) {
return l.it != r.it;
}
iterator& operator++() {
++it;
return *this;
}
iterator operator++(int) {
iterator r = *this;
++it;
return r;
}
};
iterator begin() const {
return iterator(pg_temp.cbegin());
}
iterator end() const {
return iterator(pg_temp.cend());
}
iterator find(pg_t pgid) const {
return iterator(pg_temp.find(pgid));
}
size_t size() const {
return pg_temp.size();
}
size_t count(pg_t pgid) const {
return pg_temp.count(pgid);
}
void erase(pg_t pgid) {
pg_temp.erase(pgid);
}
void clear() {
pg_temp.clear();
}
void set(pg_t pgid, const mempool::osdmap::vector<int32_t>& v) {
pg_temp[pgid] = v;
}
const mempool::osdmap::vector<int32_t>& get(pg_t pgid) {
return pg_temp.at(pgid);
}
#endif
void dump(ceph::Formatter *f) const {
for (const auto &pg : *this) {
f->open_object_section("osds");
f->dump_stream("pgid") << pg.first;
f->open_array_section("osds");
for (const auto osd : pg.second)
f->dump_int("osd", osd);
f->close_section();
f->close_section();
}
}
};
WRITE_CLASS_ENCODER(PGTempMap)
/** OSDMap
*/
class OSDMap {
public:
MEMPOOL_CLASS_HELPERS();
class Incremental {
public:
MEMPOOL_CLASS_HELPERS();
/// feature bits we were encoded with. the subsequent OSDMap
/// encoding should match.
uint64_t encode_features;
uuid_d fsid;
epoch_t epoch; // new epoch; we are a diff from epoch-1 to epoch
utime_t modified;
int64_t new_pool_max; //incremented by the OSDMonitor on each pool create
int32_t new_flags;
ceph_release_t new_require_osd_release{0xff};
uint32_t new_stretch_bucket_count{0};
uint32_t new_degraded_stretch_mode{0};
uint32_t new_recovering_stretch_mode{0};
int32_t new_stretch_mode_bucket{0};
bool stretch_mode_enabled{false};
bool change_stretch_mode{false};
// full (rare)
ceph::buffer::list fullmap; // in lieu of below.
ceph::buffer::list crush;
// incremental
int32_t new_max_osd;
mempool::osdmap::map<int64_t,pg_pool_t> new_pools;
mempool::osdmap::map<int64_t,std::string> new_pool_names;
mempool::osdmap::set<int64_t> old_pools;
mempool::osdmap::map<std::string,std::map<std::string,std::string> > new_erasure_code_profiles;
mempool::osdmap::vector<std::string> old_erasure_code_profiles;
mempool::osdmap::map<int32_t,entity_addrvec_t> new_up_client;
mempool::osdmap::map<int32_t,entity_addrvec_t> new_up_cluster;
mempool::osdmap::map<int32_t,uint32_t> new_state; // XORed onto previous state.
mempool::osdmap::map<int32_t,uint32_t> new_weight;
mempool::osdmap::map<pg_t,mempool::osdmap::vector<int32_t> > new_pg_temp; // [] to remove
mempool::osdmap::map<pg_t, int32_t> new_primary_temp; // [-1] to remove
mempool::osdmap::map<int32_t,uint32_t> new_primary_affinity;
mempool::osdmap::map<int32_t,epoch_t> new_up_thru;
mempool::osdmap::map<int32_t,std::pair<epoch_t,epoch_t> > new_last_clean_interval;
mempool::osdmap::map<int32_t,epoch_t> new_lost;
mempool::osdmap::map<int32_t,uuid_d> new_uuid;
mempool::osdmap::map<int32_t,osd_xinfo_t> new_xinfo;
mempool::osdmap::map<entity_addr_t,utime_t> new_blocklist;
mempool::osdmap::vector<entity_addr_t> old_blocklist;
mempool::osdmap::map<int32_t, entity_addrvec_t> new_hb_back_up;
mempool::osdmap::map<int32_t, entity_addrvec_t> new_hb_front_up;
mempool::osdmap::map<pg_t,mempool::osdmap::vector<int32_t>> new_pg_upmap;
mempool::osdmap::map<pg_t,mempool::osdmap::vector<std::pair<int32_t,int32_t>>> new_pg_upmap_items;
mempool::osdmap::set<pg_t> old_pg_upmap, old_pg_upmap_items;
mempool::osdmap::map<int64_t, snap_interval_set_t> new_removed_snaps;
mempool::osdmap::map<int64_t, snap_interval_set_t> new_purged_snaps;
mempool::osdmap::map<int32_t,uint32_t> new_crush_node_flags;
mempool::osdmap::map<int32_t,uint32_t> new_device_class_flags;
std::string cluster_snapshot;
float new_nearfull_ratio = -1;
float new_backfillfull_ratio = -1;
float new_full_ratio = -1;
ceph_release_t new_require_min_compat_client{0xff};
utime_t new_last_up_change, new_last_in_change;
mutable bool have_crc; ///< crc values are defined
uint32_t full_crc; ///< crc of the resulting OSDMap
mutable uint32_t inc_crc; ///< crc of this incremental
int get_net_marked_out(const OSDMap *previous) const;
int get_net_marked_down(const OSDMap *previous) const;
int identify_osd(uuid_d u) const;
void encode_client_old(ceph::buffer::list& bl) const;
void encode_classic(ceph::buffer::list& bl, uint64_t features) const;
void encode(ceph::buffer::list& bl, uint64_t features=CEPH_FEATURES_ALL) const;
void decode_classic(ceph::buffer::list::const_iterator &p);
void decode(ceph::buffer::list::const_iterator &bl);
void dump(ceph::Formatter *f) const;
static void generate_test_instances(std::list<Incremental*>& o);
explicit Incremental(epoch_t e=0) :
encode_features(0),
epoch(e), new_pool_max(-1), new_flags(-1), new_max_osd(-1),
have_crc(false), full_crc(0), inc_crc(0) {
}
explicit Incremental(ceph::buffer::list &bl) {
auto p = std::cbegin(bl);
decode(p);
}
explicit Incremental(ceph::buffer::list::const_iterator &p) {
decode(p);
}
pg_pool_t *get_new_pool(int64_t pool, const pg_pool_t *orig) {
if (new_pools.count(pool) == 0)
new_pools[pool] = *orig;
return &new_pools[pool];
}
bool has_erasure_code_profile(const std::string &name) const {
auto i = new_erasure_code_profiles.find(name);
return i != new_erasure_code_profiles.end();
}
void set_erasure_code_profile(const std::string &name,
const std::map<std::string,std::string>& profile) {
new_erasure_code_profiles[name] = profile;
}
mempool::osdmap::map<std::string,std::map<std::string,std::string>> get_erasure_code_profiles() const {
return new_erasure_code_profiles;
}
/// propagate update pools' snap metadata to any of their tiers
int propagate_snaps_to_tiers(CephContext *cct, const OSDMap &base);
/// filter out osds with any pending state changing
size_t get_pending_state_osds(std::vector<int> *osds) {
ceph_assert(osds);
osds->clear();
for (auto &p : new_state) {
osds->push_back(p.first);
}
return osds->size();
}
bool pending_osd_has_state(int osd, unsigned state) {
return new_state.count(osd) && (new_state[osd] & state) != 0;
}
bool pending_osd_state_set(int osd, unsigned state) {
if (pending_osd_has_state(osd, state))
return false;
new_state[osd] |= state;
return true;
}
// cancel the specified pending osd state if there is any
// return ture on success, false otherwise.
bool pending_osd_state_clear(int osd, unsigned state) {
if (!pending_osd_has_state(osd, state)) {
// never has been set or already has been cancelled.
return false;
}
new_state[osd] &= ~state;
if (!new_state[osd]) {
// all flags cleared
new_state.erase(osd);
}
return true;
}
bool in_new_removed_snaps(int64_t pool, snapid_t snap) const {
auto p = new_removed_snaps.find(pool);
if (p == new_removed_snaps.end()) {
return false;
}
return p->second.contains(snap);
}
};
private:
uuid_d fsid;
epoch_t epoch; // what epoch of the osd cluster descriptor is this
utime_t created, modified; // epoch start time
int32_t pool_max; // the largest pool num, ever
uint32_t flags;
int num_osd; // not saved; see calc_num_osds
int num_up_osd; // not saved; see calc_num_osds
int num_in_osd; // not saved; see calc_num_osds
int32_t max_osd;
std::vector<uint32_t> osd_state;
mempool::osdmap::map<int32_t,uint32_t> crush_node_flags; // crush node -> CEPH_OSD_* flags
mempool::osdmap::map<int32_t,uint32_t> device_class_flags; // device class -> CEPH_OSD_* flags
utime_t last_up_change, last_in_change;
// These features affect OSDMap[::Incremental] encoding, or the
// encoding of some type embedded therein (CrushWrapper, something
// from osd_types, etc.).
static constexpr uint64_t SIGNIFICANT_FEATURES =
CEPH_FEATUREMASK_PGID64 |
CEPH_FEATUREMASK_PGPOOL3 |
CEPH_FEATUREMASK_OSDENC |
CEPH_FEATUREMASK_OSDMAP_ENC |
CEPH_FEATUREMASK_OSD_POOLRESEND |
CEPH_FEATUREMASK_NEW_OSDOP_ENCODING |
CEPH_FEATUREMASK_MSG_ADDR2 |
CEPH_FEATUREMASK_CRUSH_TUNABLES5 |
CEPH_FEATUREMASK_CRUSH_CHOOSE_ARGS |
CEPH_FEATUREMASK_SERVER_LUMINOUS |
CEPH_FEATUREMASK_SERVER_MIMIC |
CEPH_FEATUREMASK_SERVER_NAUTILUS |
CEPH_FEATUREMASK_SERVER_OCTOPUS;
struct addrs_s {
mempool::osdmap::vector<std::shared_ptr<entity_addrvec_t> > client_addrs;
mempool::osdmap::vector<std::shared_ptr<entity_addrvec_t> > cluster_addrs;
mempool::osdmap::vector<std::shared_ptr<entity_addrvec_t> > hb_back_addrs;
mempool::osdmap::vector<std::shared_ptr<entity_addrvec_t> > hb_front_addrs;
};
std::shared_ptr<addrs_s> osd_addrs;
entity_addrvec_t _blank_addrvec;
mempool::osdmap::vector<__u32> osd_weight; // 16.16 fixed point, 0x10000 = "in", 0 = "out"
mempool::osdmap::vector<osd_info_t> osd_info;
std::shared_ptr<PGTempMap> pg_temp; // temp pg mapping (e.g. while we rebuild)
std::shared_ptr< mempool::osdmap::map<pg_t,int32_t > > primary_temp; // temp primary mapping (e.g. while we rebuild)
std::shared_ptr< mempool::osdmap::vector<__u32> > osd_primary_affinity; ///< 16.16 fixed point, 0x10000 = baseline
// remap (post-CRUSH, pre-up)
mempool::osdmap::map<pg_t,mempool::osdmap::vector<int32_t>> pg_upmap; ///< remap pg
mempool::osdmap::map<pg_t,mempool::osdmap::vector<std::pair<int32_t,int32_t>>> pg_upmap_items; ///< remap osds in up set
mempool::osdmap::map<int64_t,pg_pool_t> pools;
mempool::osdmap::map<int64_t,std::string> pool_name;
mempool::osdmap::map<std::string, std::map<std::string,std::string>> erasure_code_profiles;
mempool::osdmap::map<std::string,int64_t, std::less<>> name_pool;
std::shared_ptr< mempool::osdmap::vector<uuid_d> > osd_uuid;
mempool::osdmap::vector<osd_xinfo_t> osd_xinfo;
mempool::osdmap::unordered_map<entity_addr_t,utime_t> blocklist;
/// queue of snaps to remove
mempool::osdmap::map<int64_t, snap_interval_set_t> removed_snaps_queue;
/// removed_snaps additions this epoch
mempool::osdmap::map<int64_t, snap_interval_set_t> new_removed_snaps;
/// removed_snaps removals this epoch
mempool::osdmap::map<int64_t, snap_interval_set_t> new_purged_snaps;
epoch_t cluster_snapshot_epoch;
std::string cluster_snapshot;
bool new_blocklist_entries;
float full_ratio = 0, backfillfull_ratio = 0, nearfull_ratio = 0;
/// min compat client we want to support
ceph_release_t require_min_compat_client{ceph_release_t::unknown};
public:
/// require osds to run at least this release
ceph_release_t require_osd_release{ceph_release_t::unknown};
private:
mutable uint64_t cached_up_osd_features;
mutable bool crc_defined;
mutable uint32_t crc;
void _calc_up_osd_features();
public:
bool have_crc() const { return crc_defined; }
uint32_t get_crc() const { return crc; }
std::shared_ptr<CrushWrapper> crush; // hierarchical map
bool stretch_mode_enabled; // we are in stretch mode, requiring multiple sites
uint32_t stretch_bucket_count; // number of sites we expect to be in
uint32_t degraded_stretch_mode; // 0 if not degraded; else count of up sites
uint32_t recovering_stretch_mode; // 0 if not recovering; else 1
int32_t stretch_mode_bucket; // the bucket type we're stretched across
private:
uint32_t crush_version = 1;
friend class OSDMonitor;
public:
OSDMap() : epoch(0),
pool_max(0),
flags(0),
num_osd(0), num_up_osd(0), num_in_osd(0),
max_osd(0),
osd_addrs(std::make_shared<addrs_s>()),
pg_temp(std::make_shared<PGTempMap>()),
primary_temp(std::make_shared<mempool::osdmap::map<pg_t,int32_t>>()),
osd_uuid(std::make_shared<mempool::osdmap::vector<uuid_d>>()),
cluster_snapshot_epoch(0),
new_blocklist_entries(false),
cached_up_osd_features(0),
crc_defined(false), crc(0),
crush(std::make_shared<CrushWrapper>()),
stretch_mode_enabled(false), stretch_bucket_count(0),
degraded_stretch_mode(0), recovering_stretch_mode(0), stretch_mode_bucket(0) {
}
private:
OSDMap(const OSDMap& other) = default;
OSDMap& operator=(const OSDMap& other) = default;
public:
/// return feature mask subset that is relevant to OSDMap encoding
static uint64_t get_significant_features(uint64_t features) {
return SIGNIFICANT_FEATURES & features;
}
uint64_t get_encoding_features() const;
void deepish_copy_from(const OSDMap& o) {
*this = o;
primary_temp.reset(new mempool::osdmap::map<pg_t,int32_t>(*o.primary_temp));
pg_temp.reset(new PGTempMap(*o.pg_temp));
osd_uuid.reset(new mempool::osdmap::vector<uuid_d>(*o.osd_uuid));
if (o.osd_primary_affinity)
osd_primary_affinity.reset(new mempool::osdmap::vector<__u32>(*o.osd_primary_affinity));
// NOTE: this still references shared entity_addrvec_t's.
osd_addrs.reset(new addrs_s(*o.osd_addrs));
// NOTE: we do not copy crush. note that apply_incremental will
// allocate a new CrushWrapper, though.
}
// map info
const uuid_d& get_fsid() const { return fsid; }
void set_fsid(uuid_d& f) { fsid = f; }
epoch_t get_epoch() const { return epoch; }
void inc_epoch() { epoch++; }
void set_epoch(epoch_t e);
uint32_t get_crush_version() const {
return crush_version;
}
/* stamps etc */
const utime_t& get_created() const { return created; }
const utime_t& get_modified() const { return modified; }
bool is_blocklisted(const entity_addr_t& a) const;
bool is_blocklisted(const entity_addrvec_t& a) const;
void get_blocklist(std::list<std::pair<entity_addr_t,utime_t > > *bl) const;
void get_blocklist(std::set<entity_addr_t> *bl) const;
std::string get_cluster_snapshot() const {
if (cluster_snapshot_epoch == epoch)
return cluster_snapshot;
return std::string();
}
float get_full_ratio() const {
return full_ratio;
}
float get_backfillfull_ratio() const {
return backfillfull_ratio;
}
float get_nearfull_ratio() const {
return nearfull_ratio;
}
void get_full_pools(CephContext *cct,
std::set<int64_t> *full,
std::set<int64_t> *backfillfull,
std::set<int64_t> *nearfull) const;
void get_full_osd_counts(std::set<int> *full, std::set<int> *backfill,
std::set<int> *nearfull) const;
/***** cluster state *****/
/* osds */
int get_max_osd() const { return max_osd; }
void set_max_osd(int m);
unsigned get_num_osds() const {
return num_osd;
}
unsigned get_num_up_osds() const {
return num_up_osd;
}
unsigned get_num_in_osds() const {
return num_in_osd;
}
/// recalculate cached values for get_num{,_up,_in}_osds
int calc_num_osds();
void get_all_osds(std::set<int32_t>& ls) const;
void get_up_osds(std::set<int32_t>& ls) const;
void get_out_existing_osds(std::set<int32_t>& ls) const;
unsigned get_num_pg_temp() const {
return pg_temp->size();
}
int get_flags() const { return flags; }
bool test_flag(int f) const { return flags & f; }
void set_flag(int f) { flags |= f; }
void clear_flag(int f) { flags &= ~f; }
void get_flag_set(std::set<std::string> *flagset) const;
static void calc_state_set(int state, std::set<std::string>& st);
int get_state(int o) const {
ceph_assert(o < max_osd);
return osd_state[o];
}
int get_state(int o, std::set<std::string>& st) const {
ceph_assert(o < max_osd);
unsigned t = osd_state[o];
calc_state_set(t, st);
return osd_state[o];
}
void set_state(int o, unsigned s) {
ceph_assert(o < max_osd);
osd_state[o] = s;
}
void set_weight(int o, unsigned w) {
ceph_assert(o < max_osd);
osd_weight[o] = w;
if (w)
osd_state[o] |= CEPH_OSD_EXISTS;
}
unsigned get_weight(int o) const {
ceph_assert(o < max_osd);
return osd_weight[o];
}
float get_weightf(int o) const {
return (float)get_weight(o) / (float)CEPH_OSD_IN;
}
void adjust_osd_weights(const std::map<int,double>& weights, Incremental& inc) const;
void set_primary_affinity(int o, int w) {
ceph_assert(o < max_osd);
if (!osd_primary_affinity)
osd_primary_affinity.reset(
new mempool::osdmap::vector<__u32>(
max_osd, CEPH_OSD_DEFAULT_PRIMARY_AFFINITY));
(*osd_primary_affinity)[o] = w;
}
unsigned get_primary_affinity(int o) const {
ceph_assert(o < max_osd);
if (!osd_primary_affinity)
return CEPH_OSD_DEFAULT_PRIMARY_AFFINITY;
return (*osd_primary_affinity)[o];
}
float get_primary_affinityf(int o) const {
return (float)get_primary_affinity(o) / (float)CEPH_OSD_MAX_PRIMARY_AFFINITY;
}
bool has_erasure_code_profile(const std::string &name) const {
auto i = erasure_code_profiles.find(name);
return i != erasure_code_profiles.end();
}
int get_erasure_code_profile_default(CephContext *cct,
std::map<std::string,std::string> &profile_map,
std::ostream *ss);
void set_erasure_code_profile(const std::string &name,
const std::map<std::string,std::string>& profile) {
erasure_code_profiles[name] = profile;
}
const std::map<std::string,std::string> &get_erasure_code_profile(
const std::string &name) const {
static std::map<std::string,std::string> empty;
auto i = erasure_code_profiles.find(name);
if (i == erasure_code_profiles.end())
return empty;
else
return i->second;
}
const mempool::osdmap::map<std::string,std::map<std::string,std::string>> &get_erasure_code_profiles() const {
return erasure_code_profiles;
}
bool exists(int osd) const {
//assert(osd >= 0);
return osd >= 0 && osd < max_osd && (osd_state[osd] & CEPH_OSD_EXISTS);
}
bool is_destroyed(int osd) const {
return exists(osd) && (osd_state[osd] & CEPH_OSD_DESTROYED);
}
bool is_up(int osd) const {
return exists(osd) && (osd_state[osd] & CEPH_OSD_UP);
}
bool has_been_up_since(int osd, epoch_t epoch) const {
return is_up(osd) && get_up_from(osd) <= epoch;
}
bool is_down(int osd) const {
return !is_up(osd);
}
bool is_stop(int osd) const {
return exists(osd) && is_down(osd) &&
(osd_state[osd] & CEPH_OSD_STOP);
}
bool is_out(int osd) const {
return !exists(osd) || get_weight(osd) == CEPH_OSD_OUT;
}
bool is_in(int osd) const {
return !is_out(osd);
}
bool is_dead(int osd) const {
if (!exists(osd)) {
return false; // unclear if they know they are removed from map
}
return get_xinfo(osd).dead_epoch > get_info(osd).up_from;
}
unsigned get_osd_crush_node_flags(int osd) const;
unsigned get_crush_node_flags(int id) const;
unsigned get_device_class_flags(int id) const;
bool is_noup_by_osd(int osd) const {
return exists(osd) && (osd_state[osd] & CEPH_OSD_NOUP);
}
bool is_nodown_by_osd(int osd) const {
return exists(osd) && (osd_state[osd] & CEPH_OSD_NODOWN);
}
bool is_noin_by_osd(int osd) const {
return exists(osd) && (osd_state[osd] & CEPH_OSD_NOIN);
}
bool is_noout_by_osd(int osd) const {
return exists(osd) && (osd_state[osd] & CEPH_OSD_NOOUT);
}
bool is_noup(int osd) const {
if (test_flag(CEPH_OSDMAP_NOUP)) // global?
return true;
if (is_noup_by_osd(osd)) // by osd?
return true;
if (get_osd_crush_node_flags(osd) & CEPH_OSD_NOUP) // by crush-node?
return true;
if (auto class_id = crush->get_item_class_id(osd); class_id >= 0 &&
get_device_class_flags(class_id) & CEPH_OSD_NOUP) // by device-class?
return true;
return false;
}
bool is_nodown(int osd) const {
if (test_flag(CEPH_OSDMAP_NODOWN))
return true;
if (is_nodown_by_osd(osd))
return true;
if (get_osd_crush_node_flags(osd) & CEPH_OSD_NODOWN)
return true;
if (auto class_id = crush->get_item_class_id(osd); class_id >= 0 &&
get_device_class_flags(class_id) & CEPH_OSD_NODOWN)
return true;
return false;
}
bool is_noin(int osd) const {
if (test_flag(CEPH_OSDMAP_NOIN))
return true;
if (is_noin_by_osd(osd))
return true;
if (get_osd_crush_node_flags(osd) & CEPH_OSD_NOIN)
return true;
if (auto class_id = crush->get_item_class_id(osd); class_id >= 0 &&
get_device_class_flags(class_id) & CEPH_OSD_NOIN)
return true;
return false;
}
bool is_noout(int osd) const {
if (test_flag(CEPH_OSDMAP_NOOUT))
return true;
if (is_noout_by_osd(osd))
return true;
if (get_osd_crush_node_flags(osd) & CEPH_OSD_NOOUT)
return true;
if (auto class_id = crush->get_item_class_id(osd); class_id >= 0 &&
get_device_class_flags(class_id) & CEPH_OSD_NOOUT)
return true;
return false;
}
/**
* check if an entire crush subtree is down
*/
bool subtree_is_down(int id, std::set<int> *down_cache) const;
bool containing_subtree_is_down(CephContext *cct, int osd, int subtree_type, std::set<int> *down_cache) const;
bool subtree_type_is_down(CephContext *cct, int id, int subtree_type, std::set<int> *down_in_osds, std::set<int> *up_in_osds,
std::set<int> *subtree_up, std::unordered_map<int, std::set<int> > *subtree_type_down) const;
int identify_osd(const entity_addr_t& addr) const;
int identify_osd(const uuid_d& u) const;
int identify_osd_on_all_channels(const entity_addr_t& addr) const;
bool have_addr(const entity_addr_t& addr) const {
return identify_osd(addr) >= 0;
}
int find_osd_on_ip(const entity_addr_t& ip) const;
const entity_addrvec_t& get_addrs(int osd) const {
ceph_assert(exists(osd));
return osd_addrs->client_addrs[osd] ?
*osd_addrs->client_addrs[osd] : _blank_addrvec;
}
const entity_addrvec_t& get_most_recent_addrs(int osd) const {
return get_addrs(osd);
}
const entity_addrvec_t &get_cluster_addrs(int osd) const {
ceph_assert(exists(osd));
return osd_addrs->cluster_addrs[osd] ?
*osd_addrs->cluster_addrs[osd] : _blank_addrvec;
}
const entity_addrvec_t &get_hb_back_addrs(int osd) const {
ceph_assert(exists(osd));
return osd_addrs->hb_back_addrs[osd] ?
*osd_addrs->hb_back_addrs[osd] : _blank_addrvec;
}
const entity_addrvec_t &get_hb_front_addrs(int osd) const {
ceph_assert(exists(osd));
return osd_addrs->hb_front_addrs[osd] ?
*osd_addrs->hb_front_addrs[osd] : _blank_addrvec;
}
const uuid_d& get_uuid(int osd) const {
ceph_assert(exists(osd));
return (*osd_uuid)[osd];
}
const epoch_t& get_up_from(int osd) const {
ceph_assert(exists(osd));
return osd_info[osd].up_from;
}
const epoch_t& get_up_thru(int osd) const {
ceph_assert(exists(osd));
return osd_info[osd].up_thru;
}
const epoch_t& get_down_at(int osd) const {
ceph_assert(exists(osd));
return osd_info[osd].down_at;
}
const osd_info_t& get_info(int osd) const {