forked from confluentinc/librdkafka
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathrdkafka.c
1936 lines (1532 loc) · 50.5 KB
/
rdkafka.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
/*
* librdkafka - Apache Kafka C library
*
* Copyright (c) 2012-2013, Magnus Edenhill
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright notice,
* this list of conditions and the following disclaimer in the documentation
* and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
#define _GNU_SOURCE
#include <errno.h>
#include <string.h>
#include <stdarg.h>
#include <syslog.h>
#include <pthread.h>
#include <signal.h>
#include <stdlib.h>
#include "rdkafka_int.h"
#include "rdkafka_msg.h"
#include "rdkafka_broker.h"
#include "rdkafka_topic.h"
#include "rdkafka_offset.h"
#include "rdtime.h"
static pthread_once_t rd_kafka_global_init_once = PTHREAD_ONCE_INIT;
/**
* Current number of threads created by rdkafka.
* This is used in regression tests.
*/
int rd_kafka_thread_cnt_curr = 0;
int rd_kafka_thread_cnt (void) {
return rd_kafka_thread_cnt_curr;
}
/**
* Current number of live rd_kafka_t handles.
* This is used by rd_kafka_wait_destroyed() to know when the library
* has fully cleaned up after itself.
*/
static int rd_kafka_handle_cnt_curr = 0; /* atomic */
/**
* Wait for all rd_kafka_t objects to be destroyed.
* Returns 0 if all kafka objects are now destroyed, or -1 if the
* timeout was reached.
*/
int rd_kafka_wait_destroyed (int timeout_ms) {
rd_ts_t timeout = rd_clock() + (timeout_ms * 1000);
while (rd_kafka_thread_cnt() > 0 ||
rd_kafka_handle_cnt_curr > 0) {
if (rd_clock() >= timeout) {
errno = ETIMEDOUT;
return -1;
}
usleep(25000); /* 25ms */
}
return 0;
}
/**
* Wrapper for pthread_cond_timedwait() that makes it simpler to use
* for delta timeouts.
* `timeout_ms' is the delta timeout in milliseconds.
*/
int pthread_cond_timedwait_ms (pthread_cond_t *cond,
pthread_mutex_t *mutex,
int timeout_ms) {
struct timeval tv;
struct timespec ts;
gettimeofday(&tv, NULL);
TIMEVAL_TO_TIMESPEC(&tv, &ts);
ts.tv_sec += timeout_ms / 1000;
ts.tv_nsec += (timeout_ms % 1000) * 1000000;
if (ts.tv_nsec >= 1000000000) {
ts.tv_sec++;
ts.tv_nsec -= 1000000000;
}
return pthread_cond_timedwait(cond, mutex, &ts);
}
void rd_kafka_log_buf (const rd_kafka_t *rk, int level,
const char *fac, const char *buf) {
if (!rk->rk_conf.log_cb || level > rk->rk_conf.log_level)
return;
rk->rk_conf.log_cb(rk, level, fac, buf);
}
void rd_kafka_log0 (const rd_kafka_t *rk, const char *extra, int level,
const char *fac, const char *fmt, ...) {
char buf[2048];
va_list ap;
int elen = 0;
if (!rk->rk_conf.log_cb || level > rk->rk_conf.log_level)
return;
if (extra) {
elen = snprintf(buf, sizeof(buf), "%s: ", extra);
if (unlikely(elen >= sizeof(buf)))
elen = sizeof(buf);
}
va_start(ap, fmt);
vsnprintf(buf+elen, sizeof(buf)-elen, fmt, ap);
va_end(ap);
rk->rk_conf.log_cb(rk, level, fac, buf);
}
void rd_kafka_log_print (const rd_kafka_t *rk, int level,
const char *fac, const char *buf) {
struct timeval tv;
gettimeofday(&tv, NULL);
fprintf(stderr, "%%%i|%u.%03u|%s|%s| %s\n",
level, (int)tv.tv_sec, (int)(tv.tv_usec / 1000),
fac, rk ? rk->rk_name : "", buf);
}
void rd_kafka_log_syslog (const rd_kafka_t *rk, int level,
const char *fac, const char *buf) {
static int initialized = 0;
if (!initialized)
openlog("rdkafka", LOG_PID|LOG_CONS, LOG_USER);
syslog(level, "%s: %s: %s", fac, rk ? rk->rk_name : "", buf);
}
void rd_kafka_set_logger (rd_kafka_t *rk,
void (*func) (const rd_kafka_t *rk, int level,
const char *fac, const char *buf)) {
rk->rk_conf.log_cb = func;
}
void rd_kafka_set_log_level (rd_kafka_t *rk, int level) {
rk->rk_conf.log_level = level;
}
rd_kafka_op_t *rd_kafka_op_new (rd_kafka_op_type_t type) {
rd_kafka_op_t *rko;
rko = calloc(1, sizeof(*rko));
rko->rko_type = type;
return rko;
}
void rd_kafka_op_destroy (rd_kafka_op_t *rko) {
/* Decrease refcount on rkbuf to eventually free the shared buffer */
if (rko->rko_rkbuf)
rd_kafka_buf_destroy(rko->rko_rkbuf);
else if (rko->rko_payload && rko->rko_flags & RD_KAFKA_OP_F_FREE)
free(rko->rko_payload);
if (rko->rko_rkt)
rd_kafka_topic_destroy0(rko->rko_rkt);
if (rko->rko_rktp)
rd_kafka_toppar_destroy(rko->rko_rktp);
if (rko->rko_metadata)
rd_kafka_metadata_destroy(rko->rko_metadata);
free(rko);
}
/**
* Destroy a queue. The queue must be empty.
*/
int rd_kafka_q_destroy (rd_kafka_q_t *rkq) {
if (rd_atomic_sub(&rkq->rkq_refcnt, 1) > 0)
return 0;
assert(!rkq->rkq_fwdq);
rd_kafka_q_purge(rkq);
pthread_mutex_destroy(&rkq->rkq_lock);
pthread_cond_destroy(&rkq->rkq_cond);
if (rkq->rkq_flags & RD_KAFKA_Q_F_ALLOCATED)
free(rkq);
return 1;
}
/**
* Initialize a queue.
*/
void rd_kafka_q_init (rd_kafka_q_t *rkq) {
TAILQ_INIT(&rkq->rkq_q);
rkq->rkq_fwdq = NULL;
rkq->rkq_qlen = 0;
rkq->rkq_qsize = 0;
rkq->rkq_refcnt = 1;
rkq->rkq_flags = 0;
pthread_mutex_init(&rkq->rkq_lock, NULL);
pthread_cond_init(&rkq->rkq_cond, NULL);
}
/**
* Allocate a new queue and initialize it.
*/
rd_kafka_q_t *rd_kafka_q_new (void) {
rd_kafka_q_t *rkq = malloc(sizeof(*rkq));
rd_kafka_q_init(rkq);
rkq->rkq_flags |= RD_KAFKA_Q_F_ALLOCATED;
return rkq;
}
/**
* Set/clear forward queue.
* Queue forwarding enables message routing inside rdkafka.
* Typical use is to re-route all fetched messages for all partitions
* to one single queue.
*
* All access to rkq_fwdq are protected by rkq_lock.
*/
void rd_kafka_q_fwd_set (rd_kafka_q_t *rkq, rd_kafka_q_t *fwdq) {
pthread_mutex_lock(&rkq->rkq_lock);
if (rkq->rkq_fwdq) {
rd_kafka_q_destroy(rkq->rkq_fwdq);
rkq->rkq_fwdq = NULL;
}
if (fwdq) {
rd_kafka_q_keep(fwdq);
rkq->rkq_fwdq = fwdq;
/* If rkq has ops in queue, append them to fwdq's queue.
* This is an irreversible operation. */
if (rkq->rkq_qlen > 0)
rd_kafka_q_concat(fwdq, rkq);
}
pthread_mutex_unlock(&rkq->rkq_lock);
}
/**
* Purge all entries from a queue.
*/
void rd_kafka_q_purge (rd_kafka_q_t *rkq) {
rd_kafka_op_t *rko, *next;
TAILQ_HEAD(, rd_kafka_op_s) tmpq = TAILQ_HEAD_INITIALIZER(tmpq);
pthread_mutex_lock(&rkq->rkq_lock);
if (rkq->rkq_fwdq) {
rd_kafka_q_purge(rkq->rkq_fwdq);
pthread_mutex_unlock(&rkq->rkq_lock);
return;
}
/* Move ops queue to tmpq to avoid lock-order issue
* by locks taken from rd_kafka_op_destroy(). */
TAILQ_MOVE(&tmpq, &rkq->rkq_q, rko_link);
/* Zero out queue */
TAILQ_INIT(&rkq->rkq_q);
rkq->rkq_qlen = 0;
rkq->rkq_qsize = 0;
pthread_mutex_unlock(&rkq->rkq_lock);
/* Destroy the ops */
next = TAILQ_FIRST(&tmpq);
while ((rko = next)) {
next = TAILQ_NEXT(next, rko_link);
rd_kafka_op_destroy(rko);
}
}
/**
* Move 'cnt' entries from 'srcq' to 'dstq'.
* Returns the number of entries moved.
*/
size_t rd_kafka_q_move_cnt (rd_kafka_q_t *dstq, rd_kafka_q_t *srcq,
size_t cnt) {
rd_kafka_op_t *rko;
size_t mcnt = 0;
pthread_mutex_lock(&srcq->rkq_lock);
pthread_mutex_lock(&dstq->rkq_lock);
if (!dstq->rkq_fwdq && !srcq->rkq_fwdq) {
/* Optimization, if 'cnt' is equal/larger than all
* items of 'srcq' we can move the entire queue. */
if (cnt >= srcq->rkq_qlen) {
TAILQ_CONCAT(&dstq->rkq_q, &srcq->rkq_q, rko_link);
TAILQ_INIT(&srcq->rkq_q);
mcnt = srcq->rkq_qlen;
(void)rd_atomic_add(&dstq->rkq_qlen, srcq->rkq_qlen);
(void)rd_atomic_add(&dstq->rkq_qsize, srcq->rkq_qsize);
srcq->rkq_qlen = 0;
srcq->rkq_qsize = 0;
} else {
while (mcnt < cnt &&
(rko = TAILQ_FIRST(&srcq->rkq_q))) {
TAILQ_REMOVE(&srcq->rkq_q, rko, rko_link);
TAILQ_INSERT_TAIL(&dstq->rkq_q, rko, rko_link);
(void)rd_atomic_sub(&srcq->rkq_qlen, 1);
(void)rd_atomic_add(&dstq->rkq_qlen, 1);
(void)rd_atomic_sub(&srcq->rkq_qsize,
rko->rko_len);
(void)rd_atomic_add(&dstq->rkq_qsize,
rko->rko_len);
mcnt++;
}
}
} else
mcnt = rd_kafka_q_move_cnt(dstq->rkq_fwdq ? : dstq,
srcq->rkq_fwdq ? : srcq,
cnt);
pthread_mutex_unlock(&dstq->rkq_lock);
pthread_mutex_unlock(&srcq->rkq_lock);
return mcnt;
}
/**
* Pop an op from a queue.
*
* Locality: any thread.
*/
rd_kafka_op_t *rd_kafka_q_pop (rd_kafka_q_t *rkq, int timeout_ms) {
rd_kafka_op_t *rko;
pthread_mutex_lock(&rkq->rkq_lock);
if (!rkq->rkq_fwdq) {
while (!(rko = TAILQ_FIRST(&rkq->rkq_q)) &&
timeout_ms != RD_POLL_NOWAIT) {
if (timeout_ms != RD_POLL_INFINITE) {
if (pthread_cond_timedwait_ms(&rkq->rkq_cond,
&rkq->rkq_lock,
timeout_ms) ==
ETIMEDOUT) {
pthread_mutex_unlock(&rkq->rkq_lock);
return NULL;
}
timeout_ms = 0;
} else
pthread_cond_wait(&rkq->rkq_cond,
&rkq->rkq_lock);
}
if (rko) {
TAILQ_REMOVE(&rkq->rkq_q, rko, rko_link);
(void)rd_atomic_sub(&rkq->rkq_qlen, 1);
(void)rd_atomic_sub(&rkq->rkq_qsize, rko->rko_len);
}
} else
rko = rd_kafka_q_pop(rkq->rkq_fwdq, timeout_ms);
pthread_mutex_unlock(&rkq->rkq_lock);
return rko;
}
/**
* Pop all available ops from a queue and call the provided
* callback for each op.
*
* Returns the number of ops served.
*
* Locality: any thread.
*/
int rd_kafka_q_serve (rd_kafka_q_t *rkq, int timeout_ms,
void (*callback) (rd_kafka_op_t *rko,
void *opaque),
void *opaque) {
rd_kafka_op_t *rko, *tmp;
rd_kafka_q_t localq;
TAILQ_INIT(&localq.rkq_q);
pthread_mutex_lock(&rkq->rkq_lock);
if (rkq->rkq_fwdq) {
int ret = rd_kafka_q_serve(rkq->rkq_fwdq, timeout_ms,
callback, opaque);
pthread_mutex_unlock(&rkq->rkq_lock);
return ret;
}
/* Wait for op */
while (!(rko = TAILQ_FIRST(&rkq->rkq_q)) && timeout_ms != 0) {
if (timeout_ms != RD_POLL_INFINITE) {
if (pthread_cond_timedwait_ms(&rkq->rkq_cond,
&rkq->rkq_lock,
timeout_ms) ==
ETIMEDOUT)
break;
timeout_ms = 0;
} else
pthread_cond_wait(&rkq->rkq_cond,
&rkq->rkq_lock);
}
if (!rko) {
pthread_mutex_unlock(&rkq->rkq_lock);
return 0;
}
/* Move all ops to local queue */
TAILQ_CONCAT(&localq.rkq_q, &rkq->rkq_q, rko_link);
localq.rkq_qlen = rkq->rkq_qlen;
localq.rkq_qsize = rkq->rkq_qsize;
/* Reset real queue */
TAILQ_INIT(&rkq->rkq_q);
rkq->rkq_qlen = 0;
rkq->rkq_qsize = 0;
pthread_mutex_unlock(&rkq->rkq_lock);
/* Call callback for each op */
TAILQ_FOREACH_SAFE(rko, &localq.rkq_q, rko_link, tmp) {
callback(rko, opaque);
rd_kafka_op_destroy(rko);
}
return localq.rkq_qlen;
}
void rd_kafka_message_destroy (rd_kafka_message_t *rkmessage) {
rd_kafka_op_t *rko;
if (likely((rko = (rd_kafka_op_t *)rkmessage->_private) != NULL))
rd_kafka_op_destroy(rko);
else
free(rkmessage);
}
static rd_kafka_message_t *rd_kafka_message_get (rd_kafka_op_t *rko) {
rd_kafka_message_t *rkmessage;
if (rko) {
rkmessage = &rko->rko_rkmessage;
rkmessage->_private = rko;
} else
rkmessage = calloc(1, sizeof(*rkmessage));
return rkmessage;
}
/**
* Populate 'rkmessages' array with messages from 'rkq'.
* If 'auto_commit' is set, each message's offset will be commited
* to the offset store for that toppar.
*
* Returns the number of messages added.
*/
static int rd_kafka_q_serve_rkmessages (rd_kafka_q_t *rkq, int timeout_ms,
rd_kafka_message_t **rkmessages,
size_t rkmessages_size) {
int cnt = 0;
struct timeval tv;
struct timespec ts;
pthread_mutex_lock(&rkq->rkq_lock);
if (rkq->rkq_fwdq) {
cnt = rd_kafka_q_serve_rkmessages(rkq->rkq_fwdq, timeout_ms,
rkmessages, rkmessages_size);
pthread_mutex_unlock(&rkq->rkq_lock);
return cnt;
}
gettimeofday(&tv, NULL);
TIMEVAL_TO_TIMESPEC(&tv, &ts);
ts.tv_sec += timeout_ms / 1000;
ts.tv_nsec += (timeout_ms % 1000) * 1000000;
if (ts.tv_nsec >= 1000000000) {
ts.tv_sec++;
ts.tv_nsec -= 1000000000;
}
while (cnt < rkmessages_size) {
rd_kafka_op_t *rko;
while (!(rko = TAILQ_FIRST(&rkq->rkq_q))) {
if (pthread_cond_timedwait(&rkq->rkq_cond,
&rkq->rkq_lock,
&ts) == ETIMEDOUT)
break;
}
if (!rko)
break; /* Timed out */
TAILQ_REMOVE(&rkq->rkq_q, rko, rko_link);
(void)rd_atomic_sub(&rkq->rkq_qlen, 1);
(void)rd_atomic_sub(&rkq->rkq_qsize, rko->rko_len);
/* Auto-commit offset, if enabled. */
if (!rko->rko_err &&
rko->rko_rkmessage.rkt->rkt_conf.auto_commit)
rd_kafka_offset_store0(rko->rko_rktp,
rko->rko_offset, 1/*lock*/);
/* Get rkmessage from rko and append to array. */
rkmessages[cnt++] = rd_kafka_message_get(rko);
}
pthread_mutex_unlock(&rkq->rkq_lock);
return cnt;
}
void rd_kafka_queue_destroy (rd_kafka_queue_t *rkqu) {
if (!rd_kafka_q_destroy(&rkqu->rkqu_q))
return; /* Still references */
free(rkqu);
}
rd_kafka_queue_t *rd_kafka_queue_new (rd_kafka_t *rk) {
rd_kafka_queue_t *rkqu;
rkqu = calloc(1, sizeof(*rkqu));
rd_kafka_q_init(&rkqu->rkqu_q);
return rkqu;
}
/**
* Send an op back to the application.
*
* Locality: Kafka threads
*/
void rd_kafka_op_reply0 (rd_kafka_t *rk, rd_kafka_op_t *rko,
rd_kafka_op_type_t type,
rd_kafka_resp_err_t err,
void *payload, int len) {
rko->rko_type = type;
rko->rko_flags |= RD_KAFKA_OP_F_FREE;
rko->rko_payload = payload;
rko->rko_len = len;
rko->rko_err = err;
}
/**
* Send an op back to the application.
*
* Locality: Kafka thread
*/
void rd_kafka_op_reply (rd_kafka_t *rk,
rd_kafka_op_type_t type,
rd_kafka_resp_err_t err,
void *payload, int len) {
rd_kafka_op_t *rko;
rko = calloc(1, sizeof(*rko));
if (err && !payload) {
/* Provide human readable error string if not provided. */
/* Provide more info for some errors. */
if (err == RD_KAFKA_RESP_ERR_OFFSET_OUT_OF_RANGE) {
char tmp[512];
snprintf(tmp, sizeof(tmp), "%s (%"PRIu64")",
rd_kafka_err2str(err),
rk->rk_consumer.offset);
payload = strdup(tmp);
} else
payload = strdup(rd_kafka_err2str(err));
len = strlen(payload);
}
rd_kafka_op_reply0(rk, rko, type, err, payload, len);
rd_kafka_q_enq(&rk->rk_rep, rko);
}
void rd_kafka_op_reply2 (rd_kafka_t *rk, rd_kafka_op_t *rko) {
rd_kafka_q_enq(&rk->rk_rep, rko);
}
/**
* Propogate an error event to the application.
* If no error_cb has been set by the application the error will
* be logged instead.
*/
void rd_kafka_op_err (rd_kafka_t *rk, rd_kafka_resp_err_t err,
const char *fmt, ...) {
va_list ap;
char buf[2048];
va_start(ap, fmt);
vsnprintf(buf, sizeof(buf), fmt, ap);
va_end(ap);
if (rk->rk_conf.error_cb)
rd_kafka_op_reply(rk, RD_KAFKA_OP_ERR, err,
strdup(buf), strlen(buf));
else
rd_kafka_log_buf(rk, LOG_ERR, "ERROR", buf);
}
/**
* Send request to queue, wait for response.
*/
rd_kafka_op_t *rd_kafka_op_req0 (rd_kafka_q_t *destq,
rd_kafka_q_t *recvq,
rd_kafka_op_t *rko,
int timeout_ms) {
rd_kafka_op_t *reply;
/* Bump refcount for destination, destination will decrease refcount
* after posting reply. */
rd_kafka_q_keep(recvq);
/* Indicate to destination where to send reply. */
rko->rko_replyq = recvq;
/* Wait for reply */
reply = rd_kafka_q_pop(recvq, timeout_ms);
/* May be NULL for timeout */
return reply;
}
/**
* Send request to queue, wait for response.
* Creates a temporary reply queue.
*/
rd_kafka_op_t *rd_kafka_op_req (rd_kafka_q_t *destq,
rd_kafka_op_t *rko,
int timeout_ms) {
rd_kafka_q_t *recvq;
rd_kafka_op_t *reply;
recvq = rd_kafka_q_new();
reply = rd_kafka_op_req0(destq, recvq, rko, timeout_ms);
rd_kafka_q_destroy(recvq);
return reply;
}
static const char *rd_kafka_type2str (rd_kafka_type_t type) {
static const char *types[] = {
[RD_KAFKA_PRODUCER] = "producer",
[RD_KAFKA_CONSUMER] = "consumer",
};
return types[type];
}
const char *rd_kafka_err2str (rd_kafka_resp_err_t err) {
static __thread char ret[32];
switch (err)
{
case RD_KAFKA_RESP_ERR__BAD_MSG:
return "Local: Bad message format";
case RD_KAFKA_RESP_ERR__BAD_COMPRESSION:
return "Local: Invalid compressed data";
case RD_KAFKA_RESP_ERR__DESTROY:
return "Local: Broker handle destroyed";
case RD_KAFKA_RESP_ERR__FAIL:
return "Local: Communication failure with broker";
case RD_KAFKA_RESP_ERR__TRANSPORT:
return "Local: Broker transport failure";
case RD_KAFKA_RESP_ERR__CRIT_SYS_RESOURCE:
return "Local: Critical system resource failure";
case RD_KAFKA_RESP_ERR__RESOLVE:
return "Local: Host resolution failure";
case RD_KAFKA_RESP_ERR__MSG_TIMED_OUT:
return "Local: Message timed out";
case RD_KAFKA_RESP_ERR__PARTITION_EOF:
return "Broker: No more messages";
case RD_KAFKA_RESP_ERR__UNKNOWN_PARTITION:
return "Local: Unknown partition";
case RD_KAFKA_RESP_ERR__FS:
return "Local: File or filesystem error";
case RD_KAFKA_RESP_ERR__UNKNOWN_TOPIC:
return "Local: Unknown topic";
case RD_KAFKA_RESP_ERR__ALL_BROKERS_DOWN:
return "Local: All broker connections are down";
case RD_KAFKA_RESP_ERR__INVALID_ARG:
return "Local: Invalid argument or configuration";
case RD_KAFKA_RESP_ERR__TIMED_OUT:
return "Local: Timed out";
case RD_KAFKA_RESP_ERR__QUEUE_FULL:
return "Local: Queue full";
case RD_KAFKA_RESP_ERR__ISR_INSUFF:
return "Local: ISR count insufficient";
case RD_KAFKA_RESP_ERR_UNKNOWN:
return "Unknown error";
case RD_KAFKA_RESP_ERR_NO_ERROR:
return "Success";
case RD_KAFKA_RESP_ERR_OFFSET_OUT_OF_RANGE:
return "Broker: Offset out of range";
case RD_KAFKA_RESP_ERR_INVALID_MSG:
return "Broker: Invalid message";
case RD_KAFKA_RESP_ERR_UNKNOWN_TOPIC_OR_PART:
return "Broker: Unknown topic or partition";
case RD_KAFKA_RESP_ERR_INVALID_MSG_SIZE:
return "Broker: Invalid message size";
case RD_KAFKA_RESP_ERR_LEADER_NOT_AVAILABLE:
return "Broker: Leader not available";
case RD_KAFKA_RESP_ERR_NOT_LEADER_FOR_PARTITION:
return "Broker: Not leader for partition";
case RD_KAFKA_RESP_ERR_REQUEST_TIMED_OUT:
return "Broker: Request timed out";
case RD_KAFKA_RESP_ERR_BROKER_NOT_AVAILABLE:
return "Broker: Broker not available";
case RD_KAFKA_RESP_ERR_REPLICA_NOT_AVAILABLE:
return "Broker: Replica not available";
case RD_KAFKA_RESP_ERR_MSG_SIZE_TOO_LARGE:
return "Broker: Message size too large";
case RD_KAFKA_RESP_ERR_STALE_CTRL_EPOCH:
return "Broker: StaleControllerEpochCode";
case RD_KAFKA_RESP_ERR_OFFSET_METADATA_TOO_LARGE:
return "Broker: Offset metadata string too large";
default:
snprintf(ret, sizeof(ret), "Err-%i?", err);
return ret;
}
}
rd_kafka_resp_err_t rd_kafka_errno2err (int errnox) {
switch (errnox)
{
case EINVAL:
return RD_KAFKA_RESP_ERR__INVALID_ARG;
case ENOENT:
return RD_KAFKA_RESP_ERR__UNKNOWN_TOPIC;
case ESRCH:
return RD_KAFKA_RESP_ERR__UNKNOWN_PARTITION;
case ETIMEDOUT:
return RD_KAFKA_RESP_ERR__TIMED_OUT;
case EMSGSIZE:
return RD_KAFKA_RESP_ERR_MSG_SIZE_TOO_LARGE;
case ENOBUFS:
return RD_KAFKA_RESP_ERR__QUEUE_FULL;
default:
return RD_KAFKA_RESP_ERR__FAIL;
}
}
void rd_kafka_destroy0 (rd_kafka_t *rk) {
if (rd_atomic_sub(&rk->rk_refcnt, 1) > 0)
return;
/* Purge op-queue */
rd_kafka_q_purge(&rk->rk_rep);
rd_kafkap_str_destroy(rk->rk_clientid);
rd_kafkap_str_destroy(rk->rk_conf.group_id);
rd_kafka_anyconf_destroy(_RK_GLOBAL, &rk->rk_conf);
pthread_rwlock_destroy(&rk->rk_lock);
free(rk);
(void)rd_atomic_sub(&rd_kafka_handle_cnt_curr, 1);
}
/* NOTE: Must only be called by application.
* librdkafka itself must use rd_kafka_destroy0(). */
void rd_kafka_destroy (rd_kafka_t *rk) {
rd_kafka_topic_t *rkt, *rkt_tmp;
rd_kafka_broker_t *rkb;
rd_kafka_dbg(rk, GENERIC, "DESTROY", "Terminating instance");
(void)rd_atomic_add(&rk->rk_terminate, 1);
/* Decommission all topics */
rd_kafka_rdlock(rk);
TAILQ_FOREACH_SAFE(rkt, &rk->rk_topics, rkt_link, rkt_tmp) {
rd_kafka_unlock(rk);
rd_kafka_topic_partitions_remove(rkt);
rd_kafka_rdlock(rk);
}
rd_kafka_timers_interrupt(rk);
/* Interrupt all IO threads to speed up termination. */
if (rk->rk_conf.term_sig) {
TAILQ_FOREACH(rkb, &rk->rk_brokers, rkb_link)
pthread_kill(rkb->rkb_thread, rk->rk_conf.term_sig);
}
rd_kafka_unlock(rk);
/* Brokers pick up on rk_terminate automatically. */
rd_kafka_destroy0(rk);
}
/* Stats buffer printf */
#define _st_printf(fmt...) do { \
ssize_t r; \
ssize_t rem = size-of; \
r = snprintf(buf+of, rem, fmt); \
if (r > rem) { \
size *= 2; \
buf = realloc(buf, size); \
r = snprintf(buf+of, size-of, fmt); \
} \
of += r; \
} while (0)
/**
* Emit stats for toppar
*/
static inline void rd_kafka_stats_emit_toppar (char **bufp, size_t *sizep,
int *ofp,
rd_kafka_toppar_t *rktp,
int first) {
char *buf = *bufp;
size_t size = *sizep;
int of = *ofp;
int64_t consumer_lag = -1;
if (rktp->rktp_hi_offset != -1 && rktp->rktp_next_offset > 0) {
if (rktp->rktp_next_offset > rktp->rktp_hi_offset)
consumer_lag = 0;
else
consumer_lag = rktp->rktp_hi_offset -
rktp->rktp_next_offset;
}
_st_printf("%s\"%"PRId32"\": { "
"\"partition\":%"PRId32", "
"\"leader\":%"PRId32", "
"\"desired\":%s, "
"\"unknown\":%s, "
"\"msgq_cnt\":%i, "
"\"msgq_bytes\":%"PRIu64", "
"\"xmit_msgq_cnt\":%i, "
"\"xmit_msgq_bytes\":%"PRIu64", "
"\"fetchq_cnt\":%i, "
"\"fetchq_size\":%"PRIu64", "
"\"fetch_state\":\"%s\", "
"\"query_offset\":%"PRId64", "
"\"next_offset\":%"PRId64", "
"\"app_offset\":%"PRId64", "
"\"stored_offset\":%"PRId64", "
"\"commited_offset\":%"PRId64", " /*FIXME: issue #80 */
"\"committed_offset\":%"PRId64", "
"\"eof_offset\":%"PRId64", "
"\"lo_offset\":%"PRId64", "
"\"hi_offset\":%"PRId64", "
"\"consumer_lag\":%"PRId64", "
"\"txmsgs\":%"PRIu64", "
"\"txbytes\":%"PRIu64", "
"\"msgs\": %"PRIu64" "
"} ",
first ? "" : ", ",
rktp->rktp_partition,
rktp->rktp_partition,
rktp->rktp_leader ? rktp->rktp_leader->rkb_nodeid : -1,
(rktp->rktp_flags&RD_KAFKA_TOPPAR_F_DESIRED)?"true":"false",
(rktp->rktp_flags&RD_KAFKA_TOPPAR_F_UNKNOWN)?"true":"false",
rktp->rktp_msgq.rkmq_msg_cnt,
rktp->rktp_msgq.rkmq_msg_bytes,
rktp->rktp_xmit_msgq.rkmq_msg_cnt,
rktp->rktp_xmit_msgq.rkmq_msg_bytes,
rd_kafka_q_len(&rktp->rktp_fetchq),
rd_kafka_q_size(&rktp->rktp_fetchq),
rd_kafka_fetch_states[rktp->rktp_fetch_state],
rktp->rktp_query_offset,
rktp->rktp_next_offset,
rktp->rktp_app_offset,
rktp->rktp_stored_offset,
rktp->rktp_commited_offset, /* FIXME: issue #80 */
rktp->rktp_commited_offset,
rktp->rktp_eof_offset,
rktp->rktp_lo_offset,
rktp->rktp_hi_offset,
consumer_lag,
rktp->rktp_c.tx_msgs,
rktp->rktp_c.tx_bytes,
rktp->rktp_c.msgs);
*bufp = buf;
*sizep = size;
*ofp = of;
}
/**
* Emit all statistics
*/
static void rd_kafka_stats_emit_all (rd_kafka_t *rk) {
char *buf;
size_t size = 1024*rk->rk_refcnt;
int of = 0;
rd_kafka_broker_t *rkb;
rd_kafka_topic_t *rkt;
rd_kafka_toppar_t *rktp;
rd_ts_t now;
buf = malloc(size);
rd_kafka_rdlock(rk);
now = rd_clock();
_st_printf("{ "
"\"ts\":%"PRIu64", "
"\"time\":%lli, "
"\"replyq\":%i, "
"\"brokers\":{ "/*open brokers*/,
now,
(signed long long)time(NULL),
rd_kafka_q_len(&rk->rk_rep));
TAILQ_FOREACH(rkb, &rk->rk_brokers, rkb_link) {
rd_kafka_avg_t rtt;
rd_kafka_broker_lock(rkb);
rd_kafka_avg_rollover(&rtt, &rkb->rkb_avg_rtt);
_st_printf("%s\"%s\": { "/*open broker*/
"\"name\":\"%s\", "
"\"nodeid\":%"PRId32", "