-
Notifications
You must be signed in to change notification settings - Fork 253
/
repmgrd-physical.c
5628 lines (4578 loc) · 160 KB
/
repmgrd-physical.c
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
/*
* repmgrd-physical.c - physical (streaming) replication functionality for repmgrd
*
* Copyright (c) EnterpriseDB Corporation, 2010-2021
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <signal.h>
#include "repmgr.h"
#include "repmgrd.h"
#include "repmgrd-physical.h"
typedef enum
{
FAILOVER_STATE_UNKNOWN = -1,
FAILOVER_STATE_NONE,
FAILOVER_STATE_PROMOTED,
FAILOVER_STATE_PROMOTION_FAILED,
FAILOVER_STATE_PRIMARY_REAPPEARED,
FAILOVER_STATE_LOCAL_NODE_FAILURE,
FAILOVER_STATE_WAITING_NEW_PRIMARY,
FAILOVER_STATE_FOLLOW_NEW_PRIMARY,
FAILOVER_STATE_REQUIRES_MANUAL_FAILOVER,
FAILOVER_STATE_FOLLOWED_NEW_PRIMARY,
FAILOVER_STATE_FOLLOWING_ORIGINAL_PRIMARY,
FAILOVER_STATE_NO_NEW_PRIMARY,
FAILOVER_STATE_FOLLOW_FAIL,
FAILOVER_STATE_NODE_NOTIFICATION_ERROR,
FAILOVER_STATE_ELECTION_RERUN
} FailoverState;
typedef enum
{
ELECTION_NOT_CANDIDATE = -1,
ELECTION_WON,
ELECTION_LOST,
ELECTION_CANCELLED,
ELECTION_RERUN
} ElectionResult;
typedef struct election_stats
{
int visible_nodes;
int shared_upstream_nodes;
int all_nodes;
} election_stats;
typedef struct t_child_node_info
{
int node_id;
char node_name[NAMEDATALEN];
t_server_type type;
NodeAttached attached;
instr_time detached_time;
struct t_child_node_info *next;
} t_child_node_info;
typedef struct t_child_node_info_list
{
t_child_node_info *head;
t_child_node_info *tail;
int node_count;
} t_child_node_info_list;
#define T_CHILD_NODE_INFO_LIST_INITIALIZER { \
NULL, \
NULL, \
0 \
}
static PGconn *upstream_conn = NULL;
static PGconn *primary_conn = NULL;
static FailoverState failover_state = FAILOVER_STATE_UNKNOWN;
static int primary_node_id = UNKNOWN_NODE_ID;
static t_node_info upstream_node_info = T_NODE_INFO_INITIALIZER;
static instr_time last_monitoring_update;
static bool child_nodes_disconnect_command_executed = false;
static ElectionResult do_election(NodeInfoList *sibling_nodes, int *new_primary_id);
static const char *_print_election_result(ElectionResult result);
static FailoverState promote_self(void);
static void notify_followers(NodeInfoList *standby_nodes, int follow_node_id);
static void check_connection(t_node_info *node_info, PGconn **conn);
static bool check_primary_status(int degraded_monitoring_elapsed);
static void check_primary_child_nodes(t_child_node_info_list *local_child_nodes);
static bool wait_primary_notification(int *new_primary_id);
static FailoverState follow_new_primary(int new_primary_id);
static FailoverState witness_follow_new_primary(int new_primary_id);
static void reset_node_voting_status(void);
static bool do_primary_failover(void);
static bool do_upstream_standby_failover(void);
static bool do_witness_failover(void);
static bool update_monitoring_history(void);
static void handle_sighup(PGconn **conn, t_server_type server_type);
static const char *format_failover_state(FailoverState failover_state);
static ElectionResult execute_failover_validation_command(t_node_info *node_info, election_stats *stats);
static void parse_failover_validation_command(const char *template, t_node_info *node_info, election_stats *stats, PQExpBufferData *out);
static bool check_node_can_follow(PGconn *local_conn, XLogRecPtr local_xlogpos, PGconn *follow_target_conn, t_node_info *follow_target_node_info);
static void check_witness_attached(t_node_info *node_info, bool startup);
static t_child_node_info *append_child_node_record(t_child_node_info_list *nodes, int node_id, const char *node_name, t_server_type type, NodeAttached attached);
static void remove_child_node_record(t_child_node_info_list *nodes, int node_id);
static void clear_child_node_info_list(t_child_node_info_list *nodes);
static void parse_child_nodes_disconnect_command(char *parsed_command, char *template, int reporting_node_id);
static void execute_child_nodes_disconnect_command(NodeInfoList *db_child_node_records, t_child_node_info_list *local_child_nodes);
static int try_primary_reconnect(PGconn **conn, PGconn *local_conn, t_node_info *node_info);
void
handle_sigint_physical(SIGNAL_ARGS)
{
PGconn *writeable_conn;
PQExpBufferData event_details;
initPQExpBuffer(&event_details);
appendPQExpBuffer(&event_details,
_("%s signal received"),
postgres_signal_arg == SIGTERM
? "TERM" : "INT");
log_notice("%s", event_details.data);
if (local_node_info.type == PRIMARY)
writeable_conn = local_conn;
else
writeable_conn = primary_conn;
if (PQstatus(writeable_conn) == CONNECTION_OK)
create_event_notification(writeable_conn,
&config_file_options,
config_file_options.node_id,
"repmgrd_shutdown",
true,
event_details.data);
termPQExpBuffer(&event_details);
terminate(SUCCESS);
}
/* perform some sanity checks on the node's configuration */
void
do_physical_node_check(PGconn *conn)
{
/*
* If node record is "inactive"; if not, attempt to set it to "active".
*
* Usually it will have become inactive due to e.g. a standby being shut down
* while repmgrd was running in an unpaused state. In this case it's
* perfectly reasonable to automatically mark the node as "active".
*/
if (local_node_info.active == false)
{
char *hint = "Check that \"repmgr (primary|standby) register\" was executed for this node";
RecoveryType recovery_type = get_recovery_type(conn);
/*
* If the local node's recovery status is incompatible with its registered
* status, e.g. registered as primary but running as a standby, refuse to start.
*
* This typically happens when a failed primary is recloned but the node was not
* re-registered, leaving the cluster in a potentially ambiguous state. In
* this case it would not be possible or desirable to attempt to set the
* node to active; the user should ensure the cluster is in the correct state.
*/
if (recovery_type != RECTYPE_UNKNOWN && local_node_info.type != UNKNOWN)
{
bool require_reregister = false;
PQExpBufferData event_details;
initPQExpBuffer(&event_details);
if (recovery_type == RECTYPE_STANDBY && local_node_info.type != STANDBY)
{
appendPQExpBuffer(&event_details,
_("node is registered as a %s but running as a standby"),
get_node_type_string(local_node_info.type));
require_reregister = true;
}
else if (recovery_type == RECTYPE_PRIMARY && local_node_info.type == STANDBY)
{
log_error(_("node is registered as a standby but running as a %s"), get_node_type_string(local_node_info.type));
require_reregister = true;
}
if (require_reregister == true)
{
log_error("%s", event_details.data);
log_hint(_("%s"), hint);
create_event_notification(NULL,
&config_file_options,
config_file_options.node_id,
"repmgrd_start",
false,
event_details.data);
termPQExpBuffer(&event_details);
terminate(ERR_BAD_CONFIG);
}
termPQExpBuffer(&event_details);
}
/*
* Attempt to set node record active (unless explicitly configured not to)
*/
log_notice(_("setting node record for node \"%s\" (ID: %i) to \"active\""),
local_node_info.node_name,
local_node_info.node_id);
if (config_file_options.repmgrd_exit_on_inactive_node == false)
{
PGconn *primary_conn = get_primary_connection(conn, NULL, NULL);
bool success = true;
if (PQstatus(primary_conn) != CONNECTION_OK)
{
log_error(_("unable to connect to the primary node to activate the node record"));
success = false;
}
else
{
success = update_node_record_set_active(primary_conn, local_node_info.node_id, true);
PQfinish(primary_conn);
}
if (success == true)
{
local_node_info.active = true;
}
}
/*
* Corner-case where it was not possible to set the node to "active"
*/
if (local_node_info.active == false)
{
switch (config_file_options.failover)
{
/* "failover" is an enum, all values should be covered here */
case FAILOVER_AUTOMATIC:
log_error(_("this node is marked as inactive and cannot be used as a failover target"));
log_hint(_("%s"), hint);
create_event_notification(NULL,
&config_file_options,
config_file_options.node_id,
"repmgrd_start",
false,
"node is inactive and cannot be used as a failover target");
terminate(ERR_BAD_CONFIG);
break;
case FAILOVER_MANUAL:
log_warning(_("this node is marked as inactive and will be passively monitored only"));
log_hint(_("%s"), hint);
break;
}
}
}
if (config_file_options.failover == FAILOVER_AUTOMATIC)
{
/*
* Check that "promote_command" and "follow_command" are defined, otherwise repmgrd
* won't be able to perform any useful action in a failover situation.
*/
bool required_param_missing = false;
if (config_file_options.promote_command[0] == '\0')
{
log_error(_("\"promote_command\" must be defined in the configuration file"));
if (config_file_options.service_promote_command[0] != '\0')
{
/*
* "service_promote_command" is *not* a substitute for "promote_command";
* it is intended for use in those systems (e.g. Debian) where there's a service
* level promote command (e.g. pg_ctlcluster).
*
* "promote_command" should either execute "repmgr standby promote" directly, or
* a script which executes "repmgr standby promote". This is essential, as the
* repmgr metadata is updated by "repmgr standby promote".
*
* "service_promote_command", if set, will be executed by "repmgr standby promote",
* but never by repmgrd.
*
*/
log_hint(_("\"service_promote_command\" is set, but can only be executed by \"repmgr standby promote\""));
}
required_param_missing = true;
}
if (config_file_options.follow_command[0] == '\0')
{
log_error(_("\"follow_command\" must be defined in the configuration file"));
required_param_missing = true;
}
if (required_param_missing == true)
{
log_hint(_("add the missing configuration parameter(s) and start repmgrd again"));
terminate(ERR_BAD_CONFIG);
}
}
}
/*
* repmgrd running on the primary server
*/
void
monitor_streaming_primary(void)
{
instr_time log_status_interval_start;
instr_time child_nodes_check_interval_start;
t_child_node_info_list local_child_nodes = T_CHILD_NODE_INFO_LIST_INITIALIZER;
reset_node_voting_status();
repmgrd_set_upstream_node_id(local_conn, NO_UPSTREAM_NODE);
{
PQExpBufferData event_details;
initPQExpBuffer(&event_details);
appendPQExpBuffer(&event_details,
_("monitoring cluster primary \"%s\" (ID: %i)"),
local_node_info.node_name,
local_node_info.node_id);
/* Log startup event */
if (startup_event_logged == false)
{
create_event_notification(local_conn,
&config_file_options,
config_file_options.node_id,
"repmgrd_start",
true,
event_details.data);
startup_event_logged = true;
}
else
{
create_event_notification(local_conn,
&config_file_options,
config_file_options.node_id,
"repmgrd_reload",
true,
event_details.data);
}
log_notice("%s", event_details.data);
termPQExpBuffer(&event_details);
}
INSTR_TIME_SET_CURRENT(log_status_interval_start);
INSTR_TIME_SET_CURRENT(child_nodes_check_interval_start);
local_node_info.node_status = NODE_STATUS_UP;
/*
* get list of expected and attached nodes
*/
{
NodeInfoList db_child_node_records = T_NODE_INFO_LIST_INITIALIZER;
bool success = get_child_nodes(local_conn, config_file_options.node_id, &db_child_node_records);
if (!success)
{
log_error(_("unable to retrieve list of child nodes"));
}
else
{
NodeInfoListCell *cell;
for (cell = db_child_node_records.head; cell; cell = cell->next)
{
/*
* At startup, if a node for which a repmgr record exists, is not found
* in pg_stat_replication, we can't know whether it has become detached, or
* (e.g. during a provisioning operation) is a new node which has not yet
* attached. We set the status to "NODE_ATTACHED_UNKNOWN" to stop repmgrd
* emitting bogus "node has become detached" alerts.
*/
(void) append_child_node_record(&local_child_nodes,
cell->node_info->node_id,
cell->node_info->node_name,
cell->node_info->type,
cell->node_info->attached == NODE_ATTACHED ? NODE_ATTACHED : NODE_ATTACHED_UNKNOWN);
/*
* witness will not be "attached" in the normal way
*/
if (cell->node_info->type == WITNESS)
{
check_witness_attached(cell->node_info, true);
}
if (cell->node_info->attached == NODE_ATTACHED)
{
log_info(_("child node \"%s\" (ID: %i) is attached"),
cell->node_info->node_name,
cell->node_info->node_id);
}
else
{
log_info(_("child node \"%s\" (ID: %i) is not yet attached"),
cell->node_info->node_name,
cell->node_info->node_id);
}
}
}
}
while (true)
{
/*
* TODO: cache node list here, refresh at `node_list_refresh_interval`
* also return reason for inavailability so we can log it
*/
(void) connection_ping(local_conn);
check_connection(&local_node_info, &local_conn);
if (PQstatus(local_conn) != CONNECTION_OK)
{
/* local node is down, we were expecting it to be up */
if (local_node_info.node_status == NODE_STATUS_UP)
{
instr_time local_node_unreachable_start;
INSTR_TIME_SET_CURRENT(local_node_unreachable_start);
{
PQExpBufferData event_details;
initPQExpBuffer(&event_details);
appendPQExpBufferStr(&event_details,
_("unable to connect to local node"));
log_warning("%s", event_details.data);
/*
* as we're monitoring the primary, no point in trying to
* write the event to the database
*
* TODO: possibly add pre-action event here
*/
create_event_notification(NULL,
&config_file_options,
config_file_options.node_id,
"repmgrd_local_disconnect",
true,
event_details.data);
termPQExpBuffer(&event_details);
}
local_node_info.node_status = NODE_STATUS_UNKNOWN;
try_reconnect(&local_conn, &local_node_info);
if (local_node_info.node_status == NODE_STATUS_UP)
{
int local_node_unreachable_elapsed = calculate_elapsed(local_node_unreachable_start);
int stored_local_node_id = UNKNOWN_NODE_ID;
PQExpBufferData event_details;
initPQExpBuffer(&event_details);
appendPQExpBuffer(&event_details,
_("reconnected to local node after %i seconds"),
local_node_unreachable_elapsed);
log_notice("%s", event_details.data);
create_event_notification(local_conn,
&config_file_options,
config_file_options.node_id,
"repmgrd_local_reconnect",
true,
event_details.data);
termPQExpBuffer(&event_details);
/*
* If the local node was restarted, we'll need to reinitialise values
* stored in shared memory.
*/
stored_local_node_id = repmgrd_get_local_node_id(local_conn);
if (stored_local_node_id == UNKNOWN_NODE_ID)
{
repmgrd_set_local_node_id(local_conn, config_file_options.node_id);
repmgrd_set_pid(local_conn, getpid(), pid_file);
}
/*
* check that the local node is still primary, otherwise switch
* to standby monitoring
*/
if (check_primary_status(NO_DEGRADED_MONITORING_ELAPSED) == false)
return;
goto loop;
}
monitoring_state = MS_DEGRADED;
INSTR_TIME_SET_CURRENT(degraded_monitoring_start);
log_notice(_("unable to connect to local node, falling back to degraded monitoring"));
}
}
if (monitoring_state == MS_DEGRADED)
{
int degraded_monitoring_elapsed = calculate_elapsed(degraded_monitoring_start);
if (config_file_options.degraded_monitoring_timeout > 0
&& degraded_monitoring_elapsed > config_file_options.degraded_monitoring_timeout)
{
PQExpBufferData event_details;
initPQExpBuffer(&event_details);
appendPQExpBuffer(&event_details,
_("degraded monitoring timeout (%i seconds) exceeded, terminating"),
degraded_monitoring_elapsed);
log_notice("%s", event_details.data);
create_event_notification(NULL,
&config_file_options,
config_file_options.node_id,
"repmgrd_shutdown",
true,
event_details.data);
termPQExpBuffer(&event_details);
terminate(ERR_MONITORING_TIMEOUT);
}
log_debug("monitoring node in degraded state for %i seconds", degraded_monitoring_elapsed);
if (is_server_available(local_node_info.conninfo) == true)
{
close_connection(&local_conn);
local_conn = establish_db_connection(local_node_info.conninfo, false);
if (PQstatus(local_conn) != CONNECTION_OK)
{
log_warning(_("node appears to be up but no connection could be made"));
close_connection(&local_conn);
}
else
{
local_node_info.node_status = NODE_STATUS_UP;
if (check_primary_status(degraded_monitoring_elapsed) == false)
return;
goto loop;
}
}
/*
* possibly attempt to find another node from cached list check if
* there's a new primary - if so add hook for fencing? loop, if
* starts up check status, switch monitoring mode
*/
}
else
{
if (config_file_options.child_nodes_check_interval > 0)
{
int child_nodes_check_interval_elapsed = calculate_elapsed(child_nodes_check_interval_start);
if (child_nodes_check_interval_elapsed >= config_file_options.child_nodes_check_interval)
{
INSTR_TIME_SET_CURRENT(child_nodes_check_interval_start);
check_primary_child_nodes(&local_child_nodes);
}
}
}
loop:
/* check node is still primary, if not restart monitoring */
if (check_primary_status(NO_DEGRADED_MONITORING_ELAPSED) == false)
return;
/* emit "still alive" log message at regular intervals, if requested */
if (config_file_options.log_status_interval > 0)
{
int log_status_interval_elapsed = calculate_elapsed(log_status_interval_start);
if (log_status_interval_elapsed >= config_file_options.log_status_interval)
{
log_info(_("monitoring primary node \"%s\" (ID: %i) in %s state"),
local_node_info.node_name,
local_node_info.node_id,
print_monitoring_state(monitoring_state));
if (monitoring_state == MS_DEGRADED)
{
log_detail(_("waiting for the node to become available"));
}
INSTR_TIME_SET_CURRENT(log_status_interval_start);
}
}
if (got_SIGHUP)
{
handle_sighup(&local_conn, PRIMARY);
}
log_verbose(LOG_DEBUG, "sleeping %i seconds (parameter \"monitor_interval_secs\")",
config_file_options.monitor_interval_secs);
sleep(config_file_options.monitor_interval_secs);
}
}
/*
* If monitoring a primary, it's possible that after an outage of the local node
* (due to e.g. a switchover), the node has come back as a standby. We therefore
* need to verify its status and if everything looks OK, restart monitoring in
* standby mode.
*
* Returns "true" to indicate repmgrd should continue monitoring the node as
* a primary; "false" indicates repmgrd should start monitoring the node as
* a standby.
*/
bool
check_primary_status(int degraded_monitoring_elapsed)
{
PGconn *new_primary_conn;
RecordStatus record_status;
bool resume_monitoring = true;
RecoveryType recovery_type = get_recovery_type(local_conn);
if (recovery_type == RECTYPE_UNKNOWN)
{
log_warning(_("unable to determine node recovery status"));
/* "true" to indicate repmgrd should continue monitoring in degraded state */
return true;
}
/* node is still primary - resume monitoring */
if (recovery_type == RECTYPE_PRIMARY)
{
if (degraded_monitoring_elapsed != NO_DEGRADED_MONITORING_ELAPSED)
{
PQExpBufferData event_details;
monitoring_state = MS_NORMAL;
initPQExpBuffer(&event_details);
appendPQExpBuffer(&event_details,
_("reconnected to primary node after %i seconds, resuming monitoring"),
degraded_monitoring_elapsed);
create_event_notification(local_conn,
&config_file_options,
config_file_options.node_id,
"repmgrd_local_reconnect",
true,
event_details.data);
log_notice("%s", event_details.data);
termPQExpBuffer(&event_details);
}
return true;
}
/* the node is now a standby */
{
PQExpBufferData event_details;
initPQExpBuffer(&event_details);
if (degraded_monitoring_elapsed != NO_DEGRADED_MONITORING_ELAPSED)
{
appendPQExpBuffer(&event_details,
_("reconnected to node after %i seconds, node is now a standby, switching to standby monitoring"),
degraded_monitoring_elapsed);
}
else
{
appendPQExpBufferStr(&event_details,
_("node is now a standby, switching to standby monitoring"));
}
log_notice("%s", event_details.data);
termPQExpBuffer(&event_details);
}
primary_node_id = UNKNOWN_NODE_ID;
new_primary_conn = get_primary_connection_quiet(local_conn, &primary_node_id, NULL);
if (PQstatus(new_primary_conn) != CONNECTION_OK)
{
if (primary_node_id == UNKNOWN_NODE_ID)
{
log_warning(_("unable to determine a new primary node"));
}
else
{
log_warning(_("unable to connect to new primary node %i"), primary_node_id);
log_detail("\n%s", PQerrorMessage(new_primary_conn));
}
close_connection(&new_primary_conn);
/* "true" to indicate repmgrd should continue monitoring in degraded state */
return true;
}
log_debug("primary node ID is now %i", primary_node_id);
record_status = get_node_record(new_primary_conn, config_file_options.node_id, &local_node_info);
/*
* If, for whatever reason, the new primary has no record of this node,
* we won't be able to perform proper monitoring. In that case
* terminate and let the user sort out the situation.
*/
if (record_status == RECORD_NOT_FOUND)
{
PQExpBufferData event_details;
initPQExpBuffer(&event_details);
appendPQExpBuffer(&event_details,
_("no metadata record found for this node on current primary %i"),
primary_node_id);
log_error("%s", event_details.data);
log_hint(_("check that 'repmgr (primary|standby) register' was executed for this node"));
close_connection(&new_primary_conn);
create_event_notification(NULL,
&config_file_options,
config_file_options.node_id,
"repmgrd_shutdown",
false,
event_details.data);
termPQExpBuffer(&event_details);
terminate(ERR_BAD_CONFIG);
}
log_debug("node %i is registered with type = %s",
config_file_options.node_id,
get_node_type_string(local_node_info.type));
/*
* node has recovered but metadata not updated - we can do that ourselves,
*/
if (local_node_info.type == PRIMARY)
{
log_notice(_("node \"%s\" (ID: %i) still registered as primary, setting to standby"),
config_file_options.node_name,
config_file_options.node_id);
if (update_node_record_set_active_standby(new_primary_conn, config_file_options.node_id) == false)
{
resume_monitoring = false;
}
else
{
/* refresh our copy of the node record from the primary */
record_status = get_node_record(new_primary_conn, config_file_options.node_id, &local_node_info);
/* this is unlikely to happen */
if (record_status != RECORD_FOUND)
{
log_warning(_("unable to retrieve local node record from primary node %i"), primary_node_id);
resume_monitoring = false;
}
}
}
if (resume_monitoring == true)
{
PQExpBufferData event_details;
initPQExpBuffer(&event_details);
if (degraded_monitoring_elapsed != NO_DEGRADED_MONITORING_ELAPSED)
{
monitoring_state = MS_NORMAL;
log_notice(_("former primary has been restored as standby after %i seconds, updating node record and resuming monitoring"),
degraded_monitoring_elapsed);
appendPQExpBuffer(&event_details,
_("node restored as standby after %i seconds, monitoring connection to upstream node %i"),
degraded_monitoring_elapsed,
local_node_info.upstream_node_id);
}
else
{
if (local_node_info.upstream_node_id == UNKNOWN_NODE_ID)
{
/*
* If upstream_node_id is not set, it's possible that following a switchover
* of some kind (possibly forced in some way), the updated node record has
* not yet propagated to the local node. In this case however we can safely
* assume we're monitoring the primary.
*/
appendPQExpBuffer(&event_details,
_("node has become a standby, monitoring connection to primary node %i"),
primary_node_id);
}
else
{
appendPQExpBuffer(&event_details,
_("node has become a standby, monitoring connection to upstream node %i"),
local_node_info.upstream_node_id);
}
}
create_event_notification(new_primary_conn,
&config_file_options,
config_file_options.node_id,
"repmgrd_standby_reconnect",
true,
event_details.data);
termPQExpBuffer(&event_details);
close_connection(&new_primary_conn);
/* restart monitoring as standby */
return false;
}
/* continue monitoring as before */
return true;
}
static void
check_primary_child_nodes(t_child_node_info_list *local_child_nodes)
{
NodeInfoList db_child_node_records = T_NODE_INFO_LIST_INITIALIZER;
NodeInfoListCell *cell;
/* lists for newly attached and missing nodes */
t_child_node_info_list disconnected_child_nodes = T_CHILD_NODE_INFO_LIST_INITIALIZER;
t_child_node_info_list reconnected_child_nodes = T_CHILD_NODE_INFO_LIST_INITIALIZER;
t_child_node_info_list new_child_nodes = T_CHILD_NODE_INFO_LIST_INITIALIZER;
bool success = get_child_nodes(local_conn, config_file_options.node_id, &db_child_node_records);
if (!success)
{
/* unlikely this will happen, but if it does, we'll try again next time round */
log_error(_("unable to retrieve list of child nodes"));
return;
}
if (db_child_node_records.node_count == 0)
{
/* no registered child nodes - nothing to do */
return;
}
/*
* compare DB records with our internal list;
* this will tell us about:
* - previously known nodes and their current status
* - newly registered nodes we didn't know about
*
* We'll need to compare the opposite way to check for nodes
* which are in the internal list, but which have now vanished
*/
for (cell = db_child_node_records.head; cell; cell = cell->next)
{
t_child_node_info *local_child_node_rec;
bool local_child_node_rec_found = false;
/*
* witness will not be "attached" in the normal way
*/
if (cell->node_info->type == WITNESS)
{
check_witness_attached(cell->node_info, false);
}
log_debug("child node: %i; attached: %s",
cell->node_info->node_id,
cell->node_info->attached == NODE_ATTACHED ? "yes" : "no");
for (local_child_node_rec = local_child_nodes->head; local_child_node_rec; local_child_node_rec = local_child_node_rec->next)
{
if (local_child_node_rec->node_id == cell->node_info->node_id)
{
local_child_node_rec_found = true;
break;
}
}
if (local_child_node_rec_found == true)
{
/* our node record shows node attached, DB record indicates detached */
if (local_child_node_rec->attached == NODE_ATTACHED && cell->node_info->attached == NODE_DETACHED)
{
t_child_node_info *detached_child_node;
local_child_node_rec->attached = NODE_DETACHED;
INSTR_TIME_SET_CURRENT(local_child_node_rec->detached_time);
detached_child_node = append_child_node_record(&disconnected_child_nodes,
local_child_node_rec->node_id,
local_child_node_rec->node_name,
local_child_node_rec->type,
NODE_DETACHED);
detached_child_node->detached_time = local_child_node_rec->detached_time;
}
/* our node record shows node detached, DB record indicates attached */
else if (local_child_node_rec->attached == NODE_DETACHED && cell->node_info->attached == NODE_ATTACHED)
{
t_child_node_info *attached_child_node;
local_child_node_rec->attached = NODE_ATTACHED;
attached_child_node = append_child_node_record(&reconnected_child_nodes,
local_child_node_rec->node_id,
local_child_node_rec->node_name,
local_child_node_rec->type,
NODE_ATTACHED);
attached_child_node->detached_time = local_child_node_rec->detached_time;
INSTR_TIME_SET_ZERO(local_child_node_rec->detached_time);
}
else if (local_child_node_rec->attached == NODE_ATTACHED_UNKNOWN && cell->node_info->attached == NODE_ATTACHED)
{
local_child_node_rec->attached = NODE_ATTACHED;
append_child_node_record(&new_child_nodes,
local_child_node_rec->node_id,
local_child_node_rec->node_name,
local_child_node_rec->type,
NODE_ATTACHED);
}
}
else
{
/* node we didn't know about before */
NodeAttached attached = cell->node_info->attached;