forked from alibaba/AliSQL
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathClusterMgr.cpp
More file actions
1441 lines (1254 loc) · 36.2 KB
/
ClusterMgr.cpp
File metadata and controls
1441 lines (1254 loc) · 36.2 KB
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
/*
Copyright (c) 2003, 2011, Oracle and/or its affiliates. All rights reserved.
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; version 2 of the License.
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, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include <ndb_global.h>
#include <ndb_limits.h>
#include <util/version.h>
#include "TransporterFacade.hpp"
#include <kernel/GlobalSignalNumbers.h>
#include "ClusterMgr.hpp"
#include <IPCConfig.hpp>
#include "NdbApiSignal.hpp"
#include <NdbSleep.h>
#include <NdbOut.hpp>
#include <NdbTick.h>
#include <signaldata/NodeFailRep.hpp>
#include <signaldata/NFCompleteRep.hpp>
#include <signaldata/ApiRegSignalData.hpp>
#include <signaldata/AlterTable.hpp>
#include <signaldata/SumaImpl.hpp>
#include <mgmapi.h>
#include <mgmapi_configuration.hpp>
#include <mgmapi_config_parameters.h>
int global_flag_skip_invalidate_cache = 0;
int global_flag_skip_waiting_for_clean_cache = 0;
//#define DEBUG_REG
// Just a C wrapper for threadMain
extern "C"
void*
runClusterMgr_C(void * me)
{
((ClusterMgr*) me)->threadMain();
return NULL;
}
ClusterMgr::ClusterMgr(TransporterFacade & _facade):
theStop(0),
theFacade(_facade),
theArbitMgr(NULL),
m_connect_count(0),
m_max_api_reg_req_interval(~0),
noOfAliveNodes(0),
noOfConnectedNodes(0),
minDbVersion(0),
theClusterMgrThread(NULL),
waitingForHB(false),
m_cluster_state(CS_waiting_for_clean_cache)
{
DBUG_ENTER("ClusterMgr::ClusterMgr");
clusterMgrThreadMutex = NdbMutex_Create();
waitForHBCond= NdbCondition_Create();
m_auto_reconnect = -1;
Uint32 ret = this->open(&theFacade, API_CLUSTERMGR);
if (unlikely(ret == 0))
{
ndbout_c("Failed to register ClusterMgr! ret: %d", ret);
abort();
}
DBUG_VOID_RETURN;
}
ClusterMgr::~ClusterMgr()
{
DBUG_ENTER("ClusterMgr::~ClusterMgr");
doStop();
if (theArbitMgr != 0)
{
delete theArbitMgr;
theArbitMgr = 0;
}
this->close(); // disconnect from TransporterFacade
NdbCondition_Destroy(waitForHBCond);
NdbMutex_Destroy(clusterMgrThreadMutex);
DBUG_VOID_RETURN;
}
void
ClusterMgr::configure(Uint32 nodeId,
const ndb_mgm_configuration* config)
{
ndb_mgm_configuration_iterator iter(* config, CFG_SECTION_NODE);
for(iter.first(); iter.valid(); iter.next()){
Uint32 nodeId = 0;
if(iter.get(CFG_NODE_ID, &nodeId))
continue;
// Check array bounds + don't allow node 0 to be touched
assert(nodeId > 0 && nodeId < MAX_NODES);
trp_node& theNode = theNodes[nodeId];
theNode.defined = true;
unsigned type;
if(iter.get(CFG_TYPE_OF_SECTION, &type))
continue;
switch(type){
case NODE_TYPE_DB:
theNode.m_info.m_type = NodeInfo::DB;
break;
case NODE_TYPE_API:
theNode.m_info.m_type = NodeInfo::API;
break;
case NODE_TYPE_MGM:
theNode.m_info.m_type = NodeInfo::MGM;
break;
default:
type = type;
break;
}
}
/* Mark all non existing nodes as not defined */
for(Uint32 i = 0; i<MAX_NODES; i++) {
if (iter.first())
continue;
if (iter.find(CFG_NODE_ID, i))
theNodes[i]= Node();
}
#if 0
print_nodes("init");
#endif
// Configure arbitrator
Uint32 rank = 0;
iter.first();
iter.find(CFG_NODE_ID, nodeId); // let not found in config mean rank=0
iter.get(CFG_NODE_ARBIT_RANK, &rank);
if (rank > 0)
{
// The arbitrator should be active
if (!theArbitMgr)
theArbitMgr = new ArbitMgr(* this);
theArbitMgr->setRank(rank);
Uint32 delay = 0;
iter.get(CFG_NODE_ARBIT_DELAY, &delay);
theArbitMgr->setDelay(delay);
}
else if (theArbitMgr)
{
// No arbitrator should be started
theArbitMgr->doStop(NULL);
delete theArbitMgr;
theArbitMgr= NULL;
}
}
void
ClusterMgr::startThread() {
Guard g(clusterMgrThreadMutex);
theStop = -1;
theClusterMgrThread = NdbThread_Create(runClusterMgr_C,
(void**)this,
0, // default stack size
"ndb_clustermgr",
NDB_THREAD_PRIO_HIGH);
Uint32 cnt = 0;
while (theStop == -1 && cnt < 60)
{
NdbCondition_WaitTimeout(waitForHBCond, clusterMgrThreadMutex, 1000);
}
assert(theStop == 0);
}
void
ClusterMgr::doStop( ){
DBUG_ENTER("ClusterMgr::doStop");
{
Guard g(clusterMgrThreadMutex);
if(theStop == 1){
DBUG_VOID_RETURN;
}
}
void *status;
theStop = 1;
if (theClusterMgrThread) {
NdbThread_WaitFor(theClusterMgrThread, &status);
NdbThread_Destroy(&theClusterMgrThread);
}
if (theArbitMgr != NULL)
{
theArbitMgr->doStop(NULL);
}
DBUG_VOID_RETURN;
}
void
ClusterMgr::forceHB()
{
theFacade.lock_mutex();
if(waitingForHB)
{
NdbCondition_WaitTimeout(waitForHBCond, theFacade.theMutexPtr, 1000);
theFacade.unlock_mutex();
return;
}
waitingForHB= true;
NodeBitmask ndb_nodes;
ndb_nodes.clear();
waitForHBFromNodes.clear();
for(Uint32 i = 1; i < MAX_NDB_NODES; i++)
{
const trp_node &node= getNodeInfo(i);
if(!node.defined)
continue;
if(node.m_info.getType() == NodeInfo::DB)
{
ndb_nodes.set(i);
waitForHBFromNodes.bitOR(node.m_state.m_connected_nodes);
}
}
waitForHBFromNodes.bitAND(ndb_nodes);
theFacade.unlock_mutex();
#ifdef DEBUG_REG
char buf[128];
ndbout << "Waiting for HB from " << waitForHBFromNodes.getText(buf) << endl;
#endif
NdbApiSignal signal(numberToRef(API_CLUSTERMGR, theFacade.ownId()));
signal.theVerId_signalNumber = GSN_API_REGREQ;
signal.theReceiversBlockNumber = QMGR;
signal.theTrace = 0;
signal.theLength = ApiRegReq::SignalLength;
ApiRegReq * req = CAST_PTR(ApiRegReq, signal.getDataPtrSend());
req->ref = numberToRef(API_CLUSTERMGR, theFacade.ownId());
req->version = NDB_VERSION;
req->mysql_version = NDB_MYSQL_VERSION_D;
{
lock();
int nodeId= 0;
for(int i=0;
(int) NodeBitmask::NotFound != (nodeId= waitForHBFromNodes.find(i));
i= nodeId+1)
{
#ifdef DEBUG_REG
ndbout << "FORCE HB to " << nodeId << endl;
#endif
raw_sendSignal(&signal, nodeId);
}
unlock();
}
/* Wait for nodes to reply - if any heartbeats was sent */
theFacade.lock_mutex();
if (!waitForHBFromNodes.isclear())
NdbCondition_WaitTimeout(waitForHBCond, theFacade.theMutexPtr, 1000);
waitingForHB= false;
#ifdef DEBUG_REG
ndbout << "Still waiting for HB from " << waitForHBFromNodes.getText(buf) << endl;
#endif
theFacade.unlock_mutex();
}
void
ClusterMgr::startup()
{
assert(theStop == -1);
Uint32 nodeId = getOwnNodeId();
Node & cm_node = theNodes[nodeId];
trp_node & theNode = cm_node;
assert(theNode.defined);
lock();
theFacade.doConnect(nodeId);
unlock();
for (Uint32 i = 0; i<3000; i++)
{
lock();
theFacade.theTransporterRegistry->update_connections();
unlock();
if (theNode.is_connected())
break;
NdbSleep_MilliSleep(20);
}
assert(theNode.is_connected());
Guard g(clusterMgrThreadMutex);
theStop = 0;
NdbCondition_Broadcast(waitForHBCond);
}
void
ClusterMgr::threadMain()
{
startup();
NdbApiSignal signal(numberToRef(API_CLUSTERMGR, theFacade.ownId()));
signal.theVerId_signalNumber = GSN_API_REGREQ;
signal.theTrace = 0;
signal.theLength = ApiRegReq::SignalLength;
ApiRegReq * req = CAST_PTR(ApiRegReq, signal.getDataPtrSend());
req->ref = numberToRef(API_CLUSTERMGR, theFacade.ownId());
req->version = NDB_VERSION;
req->mysql_version = NDB_MYSQL_VERSION_D;
NdbApiSignal nodeFail_signal(numberToRef(API_CLUSTERMGR, getOwnNodeId()));
nodeFail_signal.theVerId_signalNumber = GSN_NODE_FAILREP;
nodeFail_signal.theReceiversBlockNumber = API_CLUSTERMGR;
nodeFail_signal.theTrace = 0;
nodeFail_signal.theLength = NodeFailRep::SignalLengthLong;
NDB_TICKS timeSlept = 100;
NDB_TICKS now = NdbTick_CurrentMillisecond();
while(!theStop)
{
/* Sleep at 100ms between each heartbeat check */
NDB_TICKS before = now;
for (Uint32 i = 0; i<10; i++)
{
NdbSleep_MilliSleep(10);
{
Guard g(clusterMgrThreadMutex);
/**
* Protect from ArbitMgr sending signals while we poll
*/
start_poll();
do_poll(0);
complete_poll();
}
}
now = NdbTick_CurrentMillisecond();
timeSlept = (now - before);
if (m_cluster_state == CS_waiting_for_clean_cache &&
theFacade.m_globalDictCache)
{
if (!global_flag_skip_waiting_for_clean_cache)
{
theFacade.m_globalDictCache->lock();
unsigned sz= theFacade.m_globalDictCache->get_size();
theFacade.m_globalDictCache->unlock();
if (sz)
continue;
}
m_cluster_state = CS_waiting_for_first_connect;
}
NodeFailRep * nodeFailRep = CAST_PTR(NodeFailRep,
nodeFail_signal.getDataPtrSend());
nodeFailRep->noOfNodes = 0;
NodeBitmask::clear(nodeFailRep->theNodes);
trp_client::lock();
for (int i = 1; i < MAX_NODES; i++){
/**
* Send register request (heartbeat) to all available nodes
* at specified timing intervals
*/
const NodeId nodeId = i;
// Check array bounds + don't allow node 0 to be touched
assert(nodeId > 0 && nodeId < MAX_NODES);
Node & cm_node = theNodes[nodeId];
trp_node & theNode = cm_node;
if (!theNode.defined)
continue;
if (theNode.is_connected() == false){
theFacade.doConnect(nodeId);
continue;
}
if (!theNode.compatible){
continue;
}
if (nodeId == getOwnNodeId() && theNode.is_confirmed())
{
/**
* Don't send HB to self more than once
* (once needed to avoid weird special cases in e.g ConfigManager)
*/
continue;
}
cm_node.hbCounter += (Uint32)timeSlept;
if (cm_node.hbCounter >= m_max_api_reg_req_interval ||
cm_node.hbCounter >= cm_node.hbFrequency)
{
/**
* It is now time to send a new Heartbeat
*/
if (cm_node.hbCounter >= cm_node.hbFrequency)
{
cm_node.hbMissed++;
cm_node.hbCounter = 0;
}
if (theNode.m_info.m_type != NodeInfo::DB)
signal.theReceiversBlockNumber = API_CLUSTERMGR;
else
signal.theReceiversBlockNumber = QMGR;
#ifdef DEBUG_REG
ndbout_c("ClusterMgr: Sending API_REGREQ to node %d", (int)nodeId);
#endif
raw_sendSignal(&signal, nodeId);
}//if
if (cm_node.hbMissed == 4 && cm_node.hbFrequency > 0)
{
nodeFailRep->noOfNodes++;
NodeBitmask::set(nodeFailRep->theNodes, nodeId);
}
}
if (nodeFailRep->noOfNodes)
{
raw_sendSignal(&nodeFail_signal, getOwnNodeId());
}
trp_client::unlock();
}
}
void
ClusterMgr::trp_deliver_signal(const NdbApiSignal* sig,
const LinearSectionPtr ptr[3])
{
const Uint32 gsn = sig->theVerId_signalNumber;
const Uint32 * theData = sig->getDataPtr();
switch (gsn){
case GSN_API_REGREQ:
execAPI_REGREQ(theData);
break;
case GSN_API_REGCONF:
execAPI_REGCONF(sig, ptr);
break;
case GSN_API_REGREF:
execAPI_REGREF(theData);
break;
case GSN_NODE_FAILREP:
execNODE_FAILREP(sig, ptr);
break;
case GSN_NF_COMPLETEREP:
execNF_COMPLETEREP(sig, ptr);
break;
case GSN_ARBIT_STARTREQ:
if (theArbitMgr != NULL)
theArbitMgr->doStart(theData);
break;
case GSN_ARBIT_CHOOSEREQ:
if (theArbitMgr != NULL)
theArbitMgr->doChoose(theData);
break;
case GSN_ARBIT_STOPORD:
if(theArbitMgr != NULL)
theArbitMgr->doStop(theData);
break;
case GSN_ALTER_TABLE_REP:
{
if (theFacade.m_globalDictCache == NULL)
break;
const AlterTableRep* rep = (const AlterTableRep*)theData;
theFacade.m_globalDictCache->lock();
theFacade.m_globalDictCache->
alter_table_rep((const char*)ptr[0].p,
rep->tableId,
rep->tableVersion,
rep->changeType == AlterTableRep::CT_ALTERED);
theFacade.m_globalDictCache->unlock();
break;
}
case GSN_SUB_GCP_COMPLETE_REP:
{
/**
* Report
*/
theFacade.for_each(this, sig, ptr);
/**
* Reply
*/
{
BlockReference ownRef = numberToRef(API_CLUSTERMGR, theFacade.ownId());
NdbApiSignal tSignal(* sig);
Uint32* send= tSignal.getDataPtrSend();
memcpy(send, theData, tSignal.getLength() << 2);
CAST_PTR(SubGcpCompleteAck, send)->rep.senderRef = ownRef;
Uint32 ref= sig->theSendersBlockRef;
Uint32 aNodeId= refToNode(ref);
tSignal.theReceiversBlockNumber= refToBlock(ref);
tSignal.theVerId_signalNumber= GSN_SUB_GCP_COMPLETE_ACK;
tSignal.theSendersBlockRef = API_CLUSTERMGR;
safe_sendSignal(&tSignal, aNodeId);
}
break;
}
case GSN_TAKE_OVERTCCONF:
{
/**
* Report
*/
theFacade.for_each(this, sig, ptr);
return;
}
case GSN_CONNECT_REP:
{
execCONNECT_REP(sig, ptr);
return;
}
case GSN_DISCONNECT_REP:
{
execDISCONNECT_REP(sig, ptr);
return;
}
default:
break;
}
return;
}
ClusterMgr::Node::Node()
: hbFrequency(0), hbCounter(0)
{
}
/**
* recalcMinDbVersion
*
* This method is called whenever the 'minimum DB node
* version' data for the connected DB nodes changes
* It calculates the minimum version of all the connected
* DB nodes.
* This information is cached by Ndb object instances.
* This information is useful when implementing API compatibility
* with older DB nodes
*/
void
ClusterMgr::recalcMinDbVersion()
{
Uint32 newMinDbVersion = ~ (Uint32) 0;
for (Uint32 i = 0; i < MAX_NODES; i++)
{
trp_node& node = theNodes[i];
if (node.is_connected() &&
node.is_confirmed() &&
node.m_info.getType() == NodeInfo::DB)
{
/* Include this node in the set of nodes used to
* compute the lowest current DB node version
*/
assert(node.m_info.m_version);
if (node.minDbVersion < newMinDbVersion)
{
newMinDbVersion = node.minDbVersion;
}
}
}
/* Now update global min Db version if we have one.
* Otherwise set it to 0
*/
newMinDbVersion = (newMinDbVersion == ~ (Uint32) 0) ?
0 :
newMinDbVersion;
//#ifdef DEBUG_MINVER
#ifdef DEBUG_MINVER
if (newMinDbVersion != minDbVersion)
{
ndbout << "Previous min Db node version was "
<< NdbVersion(minDbVersion)
<< " new min is "
<< NdbVersion(newMinDbVersion)
<< endl;
}
else
{
ndbout << "MinDbVersion recalculated, but is same : "
<< NdbVersion(minDbVersion)
<< endl;
}
#endif
minDbVersion = newMinDbVersion;
}
/******************************************************************************
* API_REGREQ and friends
******************************************************************************/
void
ClusterMgr::execAPI_REGREQ(const Uint32 * theData){
const ApiRegReq * const apiRegReq = (ApiRegReq *)&theData[0];
const NodeId nodeId = refToNode(apiRegReq->ref);
#ifdef DEBUG_REG
ndbout_c("ClusterMgr: Recd API_REGREQ from node %d", nodeId);
#endif
assert(nodeId > 0 && nodeId < MAX_NODES);
Node & cm_node = theNodes[nodeId];
trp_node & node = cm_node;
assert(node.defined == true);
assert(node.is_connected() == true);
if(node.m_info.m_version != apiRegReq->version){
node.m_info.m_version = apiRegReq->version;
node.m_info.m_mysql_version = apiRegReq->mysql_version;
if (node.m_info.m_version < NDBD_SPLIT_VERSION)
node.m_info.m_mysql_version = 0;
if (getMajor(node.m_info.m_version) < getMajor(NDB_VERSION) ||
getMinor(node.m_info.m_version) < getMinor(NDB_VERSION)) {
node.compatible = false;
} else {
node.compatible = true;
}
}
NdbApiSignal signal(numberToRef(API_CLUSTERMGR, theFacade.ownId()));
signal.theVerId_signalNumber = GSN_API_REGCONF;
signal.theReceiversBlockNumber = API_CLUSTERMGR;
signal.theTrace = 0;
signal.theLength = ApiRegConf::SignalLength;
ApiRegConf * const conf = CAST_PTR(ApiRegConf, signal.getDataPtrSend());
conf->qmgrRef = numberToRef(API_CLUSTERMGR, theFacade.ownId());
conf->version = NDB_VERSION;
conf->mysql_version = NDB_MYSQL_VERSION_D;
conf->apiHeartbeatFrequency = cm_node.hbFrequency;
conf->minDbVersion= 0;
conf->nodeState= node.m_state;
node.set_confirmed(true);
if (safe_sendSignal(&signal, nodeId) != 0)
node.set_confirmed(false);
}
void
ClusterMgr::execAPI_REGCONF(const NdbApiSignal * signal,
const LinearSectionPtr ptr[])
{
const ApiRegConf * apiRegConf = CAST_CONSTPTR(ApiRegConf,
signal->getDataPtr());
const NodeId nodeId = refToNode(apiRegConf->qmgrRef);
#ifdef DEBUG_REG
ndbout_c("ClusterMgr: Recd API_REGCONF from node %d", nodeId);
#endif
assert(nodeId > 0 && nodeId < MAX_NODES);
Node & cm_node = theNodes[nodeId];
trp_node & node = cm_node;
assert(node.defined == true);
assert(node.is_connected() == true);
if(node.m_info.m_version != apiRegConf->version){
node.m_info.m_version = apiRegConf->version;
node.m_info.m_mysql_version = apiRegConf->mysql_version;
if (node.m_info.m_version < NDBD_SPLIT_VERSION)
node.m_info.m_mysql_version = 0;
if(theNodes[theFacade.ownId()].m_info.m_type == NodeInfo::MGM)
node.compatible = ndbCompatible_mgmt_ndb(NDB_VERSION,
node.m_info.m_version);
else
node.compatible = ndbCompatible_api_ndb(NDB_VERSION,
node.m_info.m_version);
}
node.set_confirmed(true);
if (node.minDbVersion != apiRegConf->minDbVersion)
{
node.minDbVersion = apiRegConf->minDbVersion;
recalcMinDbVersion();
}
if (node.m_info.m_version >= NDBD_255_NODES_VERSION)
{
node.m_state = apiRegConf->nodeState;
}
else
{
/**
* from 2 to 8 words = 6 words diff, 6*4 = 24
*/
memcpy(&node.m_state, &apiRegConf->nodeState, sizeof(node.m_state) - 24);
}
if (node.m_info.m_type == NodeInfo::DB)
{
/**
* Only set DB nodes to "alive"
*/
if (node.compatible && (node.m_state.startLevel == NodeState::SL_STARTED ||
node.m_state.getSingleUserMode()))
{
set_node_alive(node, true);
}
else
{
set_node_alive(node, false);
}
}
cm_node.hbMissed = 0;
cm_node.hbCounter = 0;
cm_node.hbFrequency = (apiRegConf->apiHeartbeatFrequency * 10) - 50;
// Distribute signal to all threads/blocks
// TODO only if state changed...
theFacade.for_each(this, signal, ptr);
check_wait_for_hb(nodeId);
}
void
ClusterMgr::check_wait_for_hb(NodeId nodeId)
{
if(waitingForHB)
{
waitForHBFromNodes.clear(nodeId);
if(waitForHBFromNodes.isclear())
{
waitingForHB= false;
NdbCondition_Broadcast(waitForHBCond);
}
}
return;
}
void
ClusterMgr::execAPI_REGREF(const Uint32 * theData){
ApiRegRef * ref = (ApiRegRef*)theData;
const NodeId nodeId = refToNode(ref->ref);
assert(nodeId > 0 && nodeId < MAX_NODES);
Node & cm_node = theNodes[nodeId];
trp_node & node = cm_node;
assert(node.is_connected() == true);
assert(node.defined == true);
/* Only DB nodes will send API_REGREF */
assert(node.m_info.getType() == NodeInfo::DB);
node.compatible = false;
set_node_alive(node, false);
node.m_state = NodeState::SL_NOTHING;
node.m_info.m_version = ref->version;
switch(ref->errorCode){
case ApiRegRef::WrongType:
ndbout_c("Node %d reports that this node should be a NDB node", nodeId);
abort();
case ApiRegRef::UnsupportedVersion:
default:
break;
}
check_wait_for_hb(nodeId);
}
void
ClusterMgr::execNF_COMPLETEREP(const NdbApiSignal* signal,
const LinearSectionPtr ptr[3])
{
const NFCompleteRep * nfComp = CAST_CONSTPTR(NFCompleteRep,
signal->getDataPtr());
const NodeId nodeId = nfComp->failedNodeId;
assert(nodeId > 0 && nodeId < MAX_NODES);
trp_node & node = theNodes[nodeId];
if (node.nfCompleteRep == false)
{
node.nfCompleteRep = true;
theFacade.for_each(this, signal, ptr);
}
}
void
ClusterMgr::reportConnected(NodeId nodeId)
{
DBUG_ENTER("ClusterMgr::reportConnected");
DBUG_PRINT("info", ("nodeId: %u", nodeId));
/**
* Ensure that we are sending heartbeat every 100 ms
* until we have got the first reply from NDB providing
* us with the real time-out period to use.
*/
assert(nodeId > 0 && nodeId < MAX_NODES);
if (nodeId == getOwnNodeId())
{
noOfConnectedNodes--; // Don't count self...
}
noOfConnectedNodes++;
Node & cm_node = theNodes[nodeId];
trp_node & theNode = cm_node;
cm_node.hbMissed = 0;
cm_node.hbCounter = 0;
cm_node.hbFrequency = 0;
assert(theNode.is_connected() == false);
/**
* make sure the node itself is marked connected even
* if first API_REGCONF has not arrived
*/
theNode.set_connected(true);
theNode.m_state.m_connected_nodes.set(nodeId);
theNode.m_info.m_version = 0;
theNode.compatible = true;
theNode.nfCompleteRep = true;
theNode.m_node_fail_rep = false;
theNode.m_state.startLevel = NodeState::SL_NOTHING;
theNode.minDbVersion = 0;
/**
* We know that we have clusterMgrThreadMutex and trp_client::mutex
* but we don't know if we are polling...and for_each can
* only be used by a poller...
*
* Send signal to self, so that we can do this when receiving a signal
*/
NdbApiSignal signal(numberToRef(API_CLUSTERMGR, getOwnNodeId()));
signal.theVerId_signalNumber = GSN_CONNECT_REP;
signal.theReceiversBlockNumber = API_CLUSTERMGR;
signal.theTrace = 0;
signal.theLength = 1;
signal.getDataPtrSend()[0] = nodeId;
raw_sendSignal(&signal, getOwnNodeId());
DBUG_VOID_RETURN;
}
void
ClusterMgr::execCONNECT_REP(const NdbApiSignal* sig,
const LinearSectionPtr ptr[])
{
theFacade.for_each(this, sig, 0);
}
void
ClusterMgr::set_node_dead(trp_node& theNode)
{
set_node_alive(theNode, false);
theNode.set_confirmed(false);
theNode.m_state.m_connected_nodes.clear();
theNode.m_state.startLevel = NodeState::SL_NOTHING;
theNode.m_info.m_connectCount ++;
theNode.nfCompleteRep = false;
}
void
ClusterMgr::reportDisconnected(NodeId nodeId)
{
assert(nodeId > 0 && nodeId < MAX_NODES);
assert(noOfConnectedNodes > 0);
/**
* We know that we have clusterMgrThreadMutex and trp_client::mutex
* but we don't know if we are polling...and for_each can
* only be used by a poller...
*
* Send signal to self, so that we can do this when receiving a signal
*/
NdbApiSignal signal(numberToRef(API_CLUSTERMGR, getOwnNodeId()));
signal.theVerId_signalNumber = GSN_DISCONNECT_REP;
signal.theReceiversBlockNumber = API_CLUSTERMGR;
signal.theTrace = 0;
signal.theLength = DisconnectRep::SignalLength;
DisconnectRep * rep = CAST_PTR(DisconnectRep, signal.getDataPtrSend());
rep->nodeId = nodeId;
rep->err = 0;
raw_sendSignal(&signal, getOwnNodeId());
}
void
ClusterMgr::execDISCONNECT_REP(const NdbApiSignal* sig,
const LinearSectionPtr ptr[])
{
const DisconnectRep * rep = CAST_CONSTPTR(DisconnectRep, sig->getDataPtr());
Uint32 nodeId = rep->nodeId;
assert(nodeId > 0 && nodeId < MAX_NODES);
Node & cm_node = theNodes[nodeId];
trp_node & theNode = cm_node;
bool node_failrep = theNode.m_node_fail_rep;
set_node_dead(theNode);
theNode.set_connected(false);
noOfConnectedNodes--;
if (noOfConnectedNodes == 0)
{
if (!global_flag_skip_invalidate_cache &&
theFacade.m_globalDictCache)
{
theFacade.m_globalDictCache->lock();
theFacade.m_globalDictCache->invalidate_all();
theFacade.m_globalDictCache->unlock();
m_connect_count ++;
m_cluster_state = CS_waiting_for_clean_cache;
}
if (m_auto_reconnect == 0)
{
theStop = 2;
}
}
if (node_failrep == false)
{
/**
* Inform API
*/
NdbApiSignal signal(numberToRef(API_CLUSTERMGR, getOwnNodeId()));
signal.theVerId_signalNumber = GSN_NODE_FAILREP;
signal.theReceiversBlockNumber = API_CLUSTERMGR;
signal.theTrace = 0;
signal.theLength = NodeFailRep::SignalLengthLong;
NodeFailRep * rep = CAST_PTR(NodeFailRep, signal.getDataPtrSend());
rep->failNo = 0;
rep->masterNodeId = 0;
rep->noOfNodes = 1;
NodeBitmask::clear(rep->theNodes);
NodeBitmask::set(rep->theNodes, nodeId);
execNODE_FAILREP(&signal, 0);
}
}
void
ClusterMgr::execNODE_FAILREP(const NdbApiSignal* sig,
const LinearSectionPtr ptr[])
{
const NodeFailRep * rep = CAST_CONSTPTR(NodeFailRep, sig->getDataPtr());
NdbApiSignal signal(sig->theSendersBlockRef);
signal.theVerId_signalNumber = GSN_NODE_FAILREP;
signal.theReceiversBlockNumber = API_CLUSTERMGR;
signal.theTrace = 0;
signal.theLength = NodeFailRep::SignalLengthLong;