-
Notifications
You must be signed in to change notification settings - Fork 103
/
mcserver.cc
1384 lines (1224 loc) · 48.6 KB
/
mcserver.cc
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: 4; c-basic-offset: 4; indent-tabs-mode: nil -*- */
/*
* Copyright 2011-2020 Couchbase, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <cstring>
#include "internal.h"
#include "collections.h"
#include "logging.h"
#include "settings.h"
#include "negotiate.h"
#include "bucketconfig/clconfig.h"
#include "mc/mcreq-flush-inl.h"
#include <lcbio/ssl.h>
#include <include/memcached/protocol_binary.h>
#include "ctx-log-inl.h"
#include "sllist.h"
#include "sllist-inl.h"
#include "mc/compress.h"
#define LOGARGS(c, lvl) (c)->settings, "server", LCB_LOG_##lvl, __FILE__, __LINE__
#define LOGARGS_T(lvl) LOGARGS(this, lvl)
#define LOGFMT CTX_LOGFMT_PRE ",SRV=%p,IX=%d) "
#define PKTFMT "OP=0x%x, RC=0x%x, SEQ=%u"
#define PKTARGS(pkt) (pkt).opcode(), (pkt).status(), (pkt).opaque()
#define LOGID(server) CTX_LOGID(server->connctx), (void *)server, server->index
#define LOGID_T() LOGID(this)
#define MCREQ_MAXIOV 32
#define LCBCONN_UNWANT(conn, flags) (conn)->want &= ~(flags)
using namespace lcb;
static void on_error(lcbio_CTX *ctx, lcb_STATUS err);
static void on_flush_ready(lcbio_CTX *ctx)
{
Server *server = Server::get(ctx);
nb_IOV iov[MCREQ_MAXIOV] = {};
int ready;
do {
int niov = 0;
unsigned nb;
nb = mcreq_flush_iov_fill(server, iov, MCREQ_MAXIOV, &niov);
if (!nb) {
return;
}
#ifdef LCB_DUMP_PACKETS
{
char *b64 = nullptr;
int nb64 = 0;
lcb_base64_encode_iov((lcb_IOV *)iov, niov, nb, &b64, &nb64);
lcb_log(LOGARGS(server, TRACE), LOGFMT "pkt,snd,fill: size=%d, %.*s", LOGID(server), nb64, nb64, b64);
free(b64);
}
#endif
ready = lcbio_ctx_put_ex(ctx, (lcb_IOV *)iov, niov, nb);
} while (ready);
lcbio_ctx_wwant(ctx);
}
static void on_flush_done(lcbio_CTX *ctx, unsigned expected, unsigned actual)
{
Server *server = Server::get(ctx);
lcb_U64 now = 0;
if (server->settings->readj_ts_wait) {
now = gethrtime();
}
#ifdef LCB_DUMP_PACKETS
lcb_log(LOGARGS(server, TRACE), LOGFMT "pkt,snd,flush: expected=%u, actual=%u", LOGID(server), expected, actual);
#endif
mcreq_flush_done_ex(server, actual, expected, now);
server->check_closed();
}
void Server::flush()
{
/** Call into the wwant stuff.. */
if (!connctx->rdwant) {
lcbio_ctx_rwant(connctx, 24);
}
lcbio_ctx_wwant(connctx);
lcbio_ctx_schedule(connctx);
if (!lcbio_timer_armed(io_timer)) {
/**
* XXX: Maybe use get_next_timeout(), although here we can assume
* that a command was just scheduled
*/
lcbio_timer_rearm(io_timer, default_timeout());
}
}
LIBCOUCHBASE_API
void lcb_sched_flush(lcb_INSTANCE *instance)
{
for (size_t ii = 0; ii < LCBT_NSERVERS(instance); ii++) {
Server *server = instance->get_server(ii);
if (!server->has_pending()) {
continue;
}
server->flush_start(server);
}
}
/**
* Invoked when get a NOT_MY_VBUCKET response. If the response contains a JSON
* payload then we refresh the configuration with it.
*
* This function returns 1 if the operation was successfully rescheduled;
* otherwise it returns 0. If it returns 0 then we give the error back to the
* user.
*/
bool Server::handle_nmv(MemcachedResponse &resinfo, mc_PACKET *oldpkt)
{
protocol_binary_request_header hdr;
lcb_STATUS err = LCB_ERR_GENERIC;
lcb_U16 vbid;
lcb::clconfig::Provider *cccp = instance->confmon->get_provider(lcb::clconfig::CLCONFIG_CCCP);
MC_INCR_METRIC(this, packets_nmv, 1);
mcreq_read_hdr(oldpkt, &hdr);
vbid = ntohs(hdr.request.vbucket);
lcb_log(LOGARGS_T(WARN), LOGFMT "NOT_MY_VBUCKET. Packet=%p (S=%u). VBID=%u, has_config=%s", LOGID_T(),
(void *)oldpkt, oldpkt->opaque, vbid, resinfo.vallen() ? "yes" : "no");
/* Notify of new map */
lcb_vbguess_remap(instance, vbid, index);
if (resinfo.vallen() && cccp->enabled) {
/*
* apply new configuration, so that when the command will be retried, it will be routed using new vbucket map
*/
err = lcb::clconfig::cccp_update(cccp, curhost->host, resinfo.inflated_value());
}
if (err != LCB_SUCCESS) {
int bs_options;
if (instance->cur_configinfo->get_origin() == lcb::clconfig::CLCONFIG_CCCP) {
/**
* XXX: Not enough to see if cccp was enabled, since cccp might
* be requested by a user, but would still not actually be active
* for clusters < 2.5 If our current config is from CCCP
* then we can be fairly certain that CCCP is indeed working.
*
* For this reason, we don't use if (cccp->enabled) {...}
*/
bs_options = BS_REFRESH_THROTTLE;
} else {
bs_options = BS_REFRESH_ALWAYS;
}
instance->bootstrap(bs_options);
}
lcb_RETRY_ACTION retry = lcb_kv_should_retry(settings, oldpkt, LCB_ERR_NOT_MY_VBUCKET);
if (!retry.should_retry) {
return false;
}
/** Reschedule the packet again .. */
mc_PACKET *newpkt = mcreq_renew_packet(oldpkt);
newpkt->flags &= ~MCREQ_STATE_FLAGS;
instance->retryq->nmvadd((mc_EXPACKET *)newpkt);
return true;
}
struct packet_wrapper {
lcb_KEYBUF key{};
const char *scope = nullptr;
size_t nscope = 0;
const char *collection = nullptr;
size_t ncollection = 0;
uint32_t cid = 0;
uint32_t timeout = 0;
mc_PACKET *pkt = nullptr;
lcb_INSTANCE *instance = nullptr;
std::string scope_;
std::string collection_;
packet_wrapper()
{
key.type = LCB_KV_COPY;
key.contig.bytes = nullptr;
key.contig.nbytes = 0;
key.vbid = 0;
}
packet_wrapper(const packet_wrapper &other)
{
key = other.key;
cid = other.cid;
timeout = other.timeout;
pkt = other.pkt;
instance = other.instance;
scope_ = other.scope_;
collection_ = other.collection_;
update_pointers();
}
void assign_name(const std::string &name)
{
size_t dot = name.find('.');
scope_ = name.substr(0, dot);
collection_ = name.substr(dot + 1);
update_pointers();
}
private:
void update_pointers()
{
scope = scope_.c_str();
nscope = scope_.size();
collection = collection_.c_str();
ncollection = collection_.size();
}
};
static lcb_STATUS reschedule_clone(const packet_wrapper *src, packet_wrapper **dst)
{
*dst = new packet_wrapper(*src);
return LCB_SUCCESS;
}
static lcb_STATUS reschedule_destroy(packet_wrapper *wrapper)
{
delete wrapper;
return LCB_SUCCESS;
}
bool Server::handle_unknown_collection(MemcachedResponse &resp, mc_PACKET *oldpkt)
{
auto orig_status = static_cast<protocol_binary_response_status>(resp.status());
lcb_STATUS orig_err = resp.status() == PROTOCOL_BINARY_RESPONSE_UNKNOWN_SCOPE ? LCB_ERR_SCOPE_NOT_FOUND
: LCB_ERR_COLLECTION_NOT_FOUND;
protocol_binary_request_header req;
memcpy(&req, SPAN_BUFFER(&oldpkt->kh_span), sizeof(req));
lcb_RETRY_ACTION retry = lcb_kv_should_retry(settings, oldpkt, orig_err);
if (!retry.should_retry) {
return false;
}
hrtime_t now = gethrtime();
if (now > MCREQ_PKT_RDATA(oldpkt)->deadline) {
return false;
}
if (req.request.opcode == PROTOCOL_BINARY_CMD_COLLECTIONS_GET_CID) {
mc_PACKET *newpkt = mcreq_renew_packet(oldpkt);
newpkt->flags &= ~MCREQ_STATE_FLAGS;
instance->retryq->ucadd((mc_EXPACKET *)newpkt, LCB_ERR_TIMEOUT, orig_status);
return true;
}
int cid_set = 0;
uint32_t cid = mcreq_get_cid(instance, oldpkt, &cid_set);
if ((collections == MCREQ_COLLECTIONS_UNSUPPORTTED && cid_set) /* we need to strip collection and retry */
|| (collections == MCREQ_COLLECTIONS_SUPPORTED && !cid_set) /* we need to prepend collection and retry */) {
lcb_log(LOGARGS_T(WARN),
LOGFMT
"UNKNOWN_COLLECTION. Packet=%p (M=0x%x, S=%u, OP=0x%x), CID=%u, collections=%d (set=%d). Retrying",
LOGID_T(), (void *)oldpkt, (int)req.request.magic, oldpkt->opaque, (int)req.request.opcode,
(unsigned)cid, (int)collections, cid_set);
mc_PACKET *newpkt = mcreq_renew_packet(oldpkt);
newpkt->flags &= ~MCREQ_STATE_FLAGS;
instance->retryq->ucadd((mc_EXPACKET *)newpkt, LCB_ERR_TIMEOUT, orig_status);
return true;
}
std::string name = instance->collcache->id_to_name(cid);
packet_wrapper wrapper;
mcreq_get_key(oldpkt, (const char **)&wrapper.key.contig.bytes, &wrapper.key.contig.nbytes);
lcb_log(LOGARGS_T(WARN), LOGFMT "UNKNOWN_COLLECTION. Packet=%p (M=0x%x, S=%u, OP=0x%x), CID=%u, CNAME=%s",
LOGID_T(), (void *)oldpkt, (int)req.request.magic, oldpkt->opaque, (int)req.request.opcode, (unsigned)cid,
name.c_str());
wrapper.assign_name(name);
wrapper.pkt = mcreq_renew_packet(oldpkt);
wrapper.instance = instance;
wrapper.timeout = LCB_NS2US(MCREQ_PKT_RDATA(wrapper.pkt)->deadline - now);
auto operation = [orig_status](const lcb_RESPGETCID *, packet_wrapper *wrp) {
if ((wrp->pkt->flags & MCREQ_F_NOCID) == 0) {
wrp->pkt = mcreq_set_cid(wrp->pkt, wrp->cid);
}
/** Reschedule the packet again .. */
wrp->pkt->flags &= ~MCREQ_STATE_FLAGS;
wrp->instance->retryq->ucadd((mc_EXPACKET *)wrp->pkt, LCB_ERR_TIMEOUT, orig_status);
return LCB_SUCCESS;
};
lcb_STATUS rc = collcache_resolve(instance, &wrapper, operation, reschedule_clone, reschedule_destroy);
return rc == LCB_SUCCESS;
}
/**
* Determine if this is an error code that we can pass to the user, or can
* otherwise handle "innately"
*/
static bool is_fastpath_error(uint16_t rc)
{
switch (rc) {
case PROTOCOL_BINARY_RESPONSE_SUCCESS:
case PROTOCOL_BINARY_RESPONSE_KEY_ENOENT:
case PROTOCOL_BINARY_RESPONSE_KEY_EEXISTS:
case PROTOCOL_BINARY_RESPONSE_E2BIG:
case PROTOCOL_BINARY_RESPONSE_NOT_STORED:
case PROTOCOL_BINARY_RESPONSE_DELTA_BADVAL:
case PROTOCOL_BINARY_RESPONSE_ERANGE:
case PROTOCOL_BINARY_RESPONSE_NOT_SUPPORTED:
case PROTOCOL_BINARY_RESPONSE_UNKNOWN_COMMAND:
case PROTOCOL_BINARY_RESPONSE_ETMPFAIL:
case PROTOCOL_BINARY_RESPONSE_ENOMEM:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_PATH_ENOENT:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_PATH_EEXISTS:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_PATH_MISMATCH:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_PATH_EINVAL:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_PATH_E2BIG:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_VALUE_CANTINSERT:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_VALUE_ETOODEEP:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_DOC_NOTJSON:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_NUM_ERANGE:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_DELTA_ERANGE:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_INVALID_COMBO:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_MULTI_PATH_FAILURE:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_SUCCESS_DELETED:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_XATTR_INVALID_FLAG_COMBO:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_XATTR_INVALID_KEY_COMBO:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_XATTR_UNKNOWN_MACRO:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_XATTR_UNKNOWN_VATTR:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_XATTR_CANT_MODIFY_VATTR:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_MULTI_PATH_FAILURE_DELETED:
case PROTOCOL_BINARY_RESPONSE_SUBDOC_INVALID_XATTR_ORDER:
case PROTOCOL_BINARY_RESPONSE_EACCESS:
case PROTOCOL_BINARY_RESPONSE_DURABILITY_INVALID_LEVEL:
case PROTOCOL_BINARY_RESPONSE_DURABILITY_IMPOSSIBLE:
case PROTOCOL_BINARY_RESPONSE_SYNC_WRITE_IN_PROGRESS:
case PROTOCOL_BINARY_RESPONSE_SYNC_WRITE_AMBIGUOUS:
case PROTOCOL_BINARY_RESPONSE_LOCKED:
case PROTOCOL_BINARY_RATE_LIMITED_NETWORK_INGRESS:
case PROTOCOL_BINARY_RATE_LIMITED_NETWORK_EGRESS:
case PROTOCOL_BINARY_RATE_LIMITED_MAX_CONNECTIONS:
case PROTOCOL_BINARY_RATE_LIMITED_MAX_COMMANDS:
case PROTOCOL_BINARY_SCOPE_SIZE_LIMIT_EXCEEDED:
return true;
default:
if (rc >= 0xc0 && rc <= 0xcc) {
// other subdoc?
return true;
} else {
return false;
}
break;
}
}
#define ERRMAP_HANDLE_CONTINUE 0
#define ERRMAP_HANDLE_DISCONN 1
#define ERRMAP_HANDLE_RETRY 2
/**
* Handle an unknown memcached error
*
* @param mcresp Response which contains the unknown error
* @param[out] newerr more user-friendly based on error map attributes
*
* @return true if this function handled the error specially (by disconnecting)
* or false if normal handling should continue.
*/
int Server::handle_unknown_error(const mc_PACKET *request, const MemcachedResponse &mcresp, lcb_STATUS &newerr)
{
if (!settings->use_errmap || !settings->errmap->isLoaded()) {
// If there's no error map, just return false
return ERRMAP_HANDLE_CONTINUE;
}
// Look up the error map definition for this error
const errmap::Error &err = settings->errmap->getError(mcresp.status());
if (!err.isValid() || err.hasAttribute(errmap::SPECIAL_HANDLING)) {
lcb_log(LOGARGS_T(ERR), LOGFMT "Received error not in error map or requires special handling! " PKTFMT,
LOGID_T(), PKTARGS(mcresp));
lcbio_ctx_senderr(connctx, LCB_ERR_PROTOCOL_ERROR);
return ERRMAP_HANDLE_DISCONN;
} else {
lcb_log(LOGARGS_T(WARN), LOGFMT "Received server error %s (0x%x) on packet: " PKTFMT, LOGID_T(),
err.shortname.c_str(), err.code, PKTARGS(mcresp));
}
if (err.hasAttribute(errmap::FETCH_CONFIG)) {
instance->bootstrap(BS_REFRESH_THROTTLE);
}
if (err.hasAttribute(errmap::TEMPORARY)) {
newerr = LCB_ERR_TEMPORARY_FAILURE;
}
if (err.hasAttribute(errmap::AUTH)) {
newerr = LCB_ERR_AUTHENTICATION_FAILURE;
}
/* TODO: remove masking LOCKED in 3.0 release */
if (err.hasAttribute(errmap::ITEM_LOCKED)) {
switch (mcresp.opcode()) {
case PROTOCOL_BINARY_CMD_SET:
case PROTOCOL_BINARY_CMD_REPLACE:
case PROTOCOL_BINARY_CMD_DELETE:
newerr = LCB_ERR_DOCUMENT_EXISTS;
break;
default:
newerr = LCB_ERR_TEMPORARY_FAILURE;
}
}
int rv = 0;
if (err.hasAttribute(errmap::AUTO_RETRY)) {
errmap::RetrySpec *spec = err.getRetrySpec();
mc_PACKET *newpkt = mcreq_renew_packet(request);
newpkt->flags &= ~MCREQ_STATE_FLAGS;
instance->retryq->add((mc_EXPACKET *)newpkt, newerr ? newerr : LCB_ERR_GENERIC,
static_cast<protocol_binary_response_status>(mcresp.status()), spec);
rv |= ERRMAP_HANDLE_RETRY;
}
if (err.hasAttribute(errmap::CONN_STATE_INVALIDATED)) {
if (newerr != LCB_SUCCESS) {
newerr = LCB_ERR_GENERIC;
}
lcbio_ctx_senderr(connctx, newerr);
rv |= ERRMAP_HANDLE_DISCONN;
}
return rv;
}
lcb_STATUS lcb_map_error(lcb_INSTANCE *instance, int in);
static bool is_warmup_issue(uint16_t status)
{
return status == PROTOCOL_BINARY_RESPONSE_NO_BUCKET || status == PROTOCOL_BINARY_RESPONSE_NOT_INITIALIZED;
}
/**
* The handler extracts epoch and revision from the notification message, and pass it to configuration provider to
* perform PROTOCOL_BINARY_CMD_GET_CLUSTER_CONFIG (0xb5) operation to fetch configuration body.
*/
void Server::handle_clustermap_notification(const MemcachedResponse &request)
{
std::int64_t epoch{0};
std::int64_t revision{0};
if (request.extlen() == 2 * sizeof(std::uint64_t)) {
const auto *ext = request.ext();
memcpy(&epoch, ext, sizeof(std::uint64_t));
epoch = lcb_ntohll(epoch);
ext += sizeof(std::uint64_t);
memcpy(&revision, ext, sizeof(std::uint64_t));
revision = lcb_ntohll(revision);
}
lcb_log(LOGARGS_T(TRACE),
LOGFMT "Received payload clustermap notification. (key=\"%.*s\", epoch=%" PRId64 ", revision=%" PRId64 ")",
LOGID_T(), (int)request.keylen(), request.key(), epoch, revision);
lcb::clconfig::Provider *cccp = instance->confmon->get_provider(lcb::clconfig::CLCONFIG_CCCP);
if (cccp == nullptr || !cccp->enabled) {
lcb_log(LOGARGS_T(ERR),
LOGFMT
"CCCP configuration provider is not enabled, ignoring notification. (key=\"%.*s\", epoch=%" PRId64
", revision=%" PRId64 ")",
LOGID_T(), (int)request.keylen(), request.key(), epoch, revision);
return;
}
/*
* Notify configuration provider that the connection seen notification about existence of newer configuration
* version.
*/
lcb::clconfig::schedule_get_config(cccp, has_valid_host() ? &get_host() : nullptr,
lcb::clconfig::config_version{epoch, revision});
}
void Server::handle_config_only(const mc_PACKET *oldpkt)
{
const auto *address = has_valid_host() ? &get_host() : nullptr;
lcb_log(LOGARGS_T(DEBUG),
LOGFMT "The bucket is configured in config-only mode on " LCB_HOST_FMT
", refresh configuration and retry operation",
LOGID_T(), LCB_HOST_ARG(this->settings, address));
auto *cccp = instance->confmon->get_provider(lcb::clconfig::CLCONFIG_CCCP);
if (cccp != nullptr && cccp->enabled) {
lcb::clconfig::schedule_get_config(cccp);
} else {
lcb_log(LOGARGS_T(DEBUG), LOGFMT "CCCP configuration provider is not enabled, using next available provider",
LOGID_T());
instance->confmon->do_next_provider();
}
mc_PACKET *newpkt = mcreq_renew_packet(oldpkt);
newpkt->flags &= ~MCREQ_STATE_FLAGS;
instance->retryq->config_only_add((mc_EXPACKET *)newpkt);
}
/**
* Handles requests that are initiated by the KV engine. It only happens when PROTOCOL_BINARY_FEATURE_DUPLEX (0x0c) is
* enabled for the connection.
*/
void Server::handle_server_request(const MemcachedResponse &request)
{
switch (request.opcode()) {
case PROTOCOL_BINARY_CMD_CLUSTERMAP_CHANGE_NOTIFICATION:
handle_clustermap_notification(request);
break;
default:
lcb_log(LOGARGS_T(DEBUG),
LOGFMT "Server sent us unknown notification packet, ignoring it. (OP=0x%x, RC=0x%x, SEQ=%u)",
LOGID_T(), request.opcode(), request.status(), request.opaque());
break;
}
}
/* This function is called within a loop to process a single packet.
*
* If a full packet is available, it will process the packet and return
* PKT_READ_COMPLETE, resulting in the `on_read()` function calling this
* function in a loop.
*
* When a complete packet is not available, PKT_READ_PARTIAL will be returned
* and the `on_read()` loop will exit, scheduling any required pending I/O.
*/
Server::ReadState Server::try_read(lcbio_CTX *ctx, rdb_IOROPE *ior)
{
MemcachedResponse mcresp;
mc_PACKET *request;
unsigned pktsize = 24, is_last = 1;
#define RETURN_NEED_MORE(n) \
if (has_pending()) { \
lcbio_ctx_rwant(ctx, n); \
} \
return PKT_READ_PARTIAL
#define DO_ASSIGN_PAYLOAD() \
rdb_consumed(ior, mcresp.hdrsize()); \
if (mcresp.bodylen()) { \
mcresp.payload = rdb_get_consolidated(ior, mcresp.bodylen()); \
} \
{
#define DO_SWALLOW_PAYLOAD() \
} \
if (mcresp.bodylen()) { \
rdb_consumed(ior, mcresp.bodylen()); \
}
if (rdb_get_nused(ior) < pktsize) {
RETURN_NEED_MORE(pktsize);
}
MC_INCR_METRIC(this, packets_read, 1);
/* copy bytes into the info structure */
rdb_copyread(ior, mcresp.hdrbytes(), mcresp.hdrsize());
pktsize += mcresp.bodylen();
if (rdb_get_nused(ior) < pktsize) {
RETURN_NEED_MORE(pktsize);
}
if (mcresp.res.response.magic == PROTOCOL_BINARY_SREQ) {
/*
* Server-initiated request, we don't need to look for SDK-request object here.
* Handle and return from the function.
*/
DO_ASSIGN_PAYLOAD()
handle_server_request(mcresp);
DO_SWALLOW_PAYLOAD()
return PKT_READ_COMPLETE;
}
/* Find the packet */
if (mcresp.opcode() == PROTOCOL_BINARY_CMD_STAT && mcresp.keylen() != 0) {
is_last = 0;
request = mcreq_pipeline_find(this, mcresp.opaque());
} else {
is_last = 1;
request = mcreq_pipeline_remove(this, mcresp.opaque());
}
if (!request) {
if (mcresp.opcode() == PROTOCOL_BINARY_CMD_SELECT_BUCKET) {
rdb_consumed(ior, pktsize);
switch (mcresp.status()) {
case PROTOCOL_BINARY_RESPONSE_SUCCESS:
lcb_log(LOGARGS_T(TRACE), R"(<%s:%s> (SRV=%p) Selected bucket "%.*s" (SEQ=0x%02x))", curhost->host,
curhost->port, (void *)this, (int)bucket.size(), bucket.c_str(), mcresp.opaque());
return PKT_READ_COMPLETE;
case PROTOCOL_BINARY_RESPONSE_EACCESS:
case PROTOCOL_BINARY_RESPONSE_KEY_ENOENT:
lcb_log(LOGARGS_T(WARN),
LOGFMT "unable to select bucket with SELECT_BUCKET (OP=0x%x, RC=0x%x, SEQ=%u)", LOGID_T(),
mcresp.opcode(), mcresp.status(), mcresp.opaque());
return PKT_READ_ABORT;
default:
lcb_log(LOGARGS_T(DEBUG),
LOGFMT "unexpected status received for SELECT_BUCKET (OP=0x%x, RC=0x%x, SEQ=%u)", LOGID_T(),
mcresp.opcode(), mcresp.status(), mcresp.opaque());
return PKT_READ_COMPLETE;
}
}
MC_INCR_METRIC(this, packets_ownerless, 1);
lcb_log(LOGARGS_T(DEBUG), LOGFMT "Server sent us reply for a timed-out command. (OP=0x%x, RC=0x%x, SEQ=%u)",
LOGID_T(), mcresp.opcode(), mcresp.status(), mcresp.opaque());
rdb_consumed(ior, pktsize);
return PKT_READ_COMPLETE;
}
lcb_STATUS err_override = LCB_SUCCESS;
ReadState rdstate = PKT_READ_COMPLETE;
int unknown_err_rv;
auto status = static_cast<protocol_binary_response_status>(mcresp.status());
if (is_warmup_issue(status)) {
DO_ASSIGN_PAYLOAD()
mc_PACKET *newpkt = mcreq_renew_packet(request);
newpkt->flags &= ~MCREQ_STATE_FLAGS;
instance->retryq->add((mc_EXPACKET *)newpkt, lcb_map_error(instance, status), status, nullptr);
DO_SWALLOW_PAYLOAD()
rdstate = PKT_READ_ABORT;
goto GT_DONE;
} else if (is_fastpath_error(status)) {
/* Check if the status code is one which must be handled carefully by the client */
lcb_STATUS err = lcb_map_error(instance, status);
if (err != LCB_SUCCESS && maybe_retry_packet(request, err, status)) {
DO_ASSIGN_PAYLOAD()
DO_SWALLOW_PAYLOAD()
goto GT_DONE;
}
} else if (status == PROTOCOL_BINARY_RESPONSE_NOT_MY_VBUCKET) {
/* consume the header */
DO_ASSIGN_PAYLOAD()
if (!handle_nmv(mcresp, request)) {
mcreq_dispatch_response(this, request, &mcresp, LCB_ERR_NOT_MY_VBUCKET);
}
DO_SWALLOW_PAYLOAD()
goto GT_DONE;
} else if (status == PROTOCOL_BINARY_RESPONSE_CONFIG_ONLY) {
/* consume the header */
DO_ASSIGN_PAYLOAD()
handle_config_only(request);
DO_SWALLOW_PAYLOAD()
goto GT_DONE;
} else if (status == PROTOCOL_BINARY_RESPONSE_UNKNOWN_COLLECTION ||
status == PROTOCOL_BINARY_RESPONSE_UNKNOWN_SCOPE) {
/* consume the header */
DO_ASSIGN_PAYLOAD()
if (!handle_unknown_collection(mcresp, request)) {
mcreq_dispatch_response(this, request, &mcresp, LCB_ERR_TIMEOUT);
}
DO_SWALLOW_PAYLOAD()
goto GT_DONE;
} else if ((unknown_err_rv = handle_unknown_error(request, mcresp, err_override)) != ERRMAP_HANDLE_CONTINUE) {
DO_ASSIGN_PAYLOAD()
if (!(unknown_err_rv & ERRMAP_HANDLE_RETRY)) {
mcreq_dispatch_response(this, request, &mcresp, err_override);
}
DO_SWALLOW_PAYLOAD()
if (unknown_err_rv & ERRMAP_HANDLE_DISCONN) {
rdstate = PKT_READ_ABORT;
}
goto GT_DONE;
}
/* Figure out if the request is 'ufwd' or not */
if (!(request->flags & MCREQ_F_UFWD)) {
DO_ASSIGN_PAYLOAD()
mcresp.bufh = rdb_get_first_segment(ior);
mcreq_dispatch_response(this, request, &mcresp, err_override);
DO_SWALLOW_PAYLOAD()
} else {
/* figure out how many buffers we want to use as an upper limit for the
* IOV arrays. Currently we'll keep it simple and ensure the entire
* response is contiguous. */
lcb_PKTFWDRESP resp = {0}; /* TODO: next ABI version should include is_last flag */
rdb_ROPESEG *segs;
nb_IOV iov;
rdb_consolidate(ior, pktsize);
rdb_refread_ex(ior, &iov, &segs, 1, pktsize);
resp.bufs = &segs;
resp.iovs = (lcb_IOV *)&iov;
resp.nitems = 1;
resp.header = mcresp.hdrbytes();
instance->callbacks.pktfwd(instance, MCREQ_PKT_COOKIE(request), LCB_SUCCESS, &resp);
rdb_consumed(ior, pktsize);
}
GT_DONE:
if (is_last) {
mcreq_packet_handled(this, request);
}
return rdstate;
}
static void on_read(lcbio_CTX *ctx, unsigned)
{
Server *server = Server::get(ctx);
rdb_IOROPE *ior = &ctx->ior;
if (server->check_closed()) {
return;
}
while (server->try_read(ctx, ior) == Server::PKT_READ_COMPLETE)
;
lcbio_ctx_schedule(ctx);
lcb_maybe_breakout(server->instance);
}
static void flush_noop(mc_PIPELINE *pipeline)
{
(void)pipeline;
}
static void server_connect(Server *server)
{
server->connect();
}
bool Server::maybe_retry_packet(mc_PACKET *pkt, lcb_STATUS err, protocol_binary_response_status status)
{
lcbvb_DISTMODE dist_t = lcbvb_get_distmode(parent->config);
if (dist_t != LCBVB_DIST_VBUCKET) {
/** memcached bucket */
return false;
}
lcb_RETRY_ACTION retry = lcb_kv_should_retry(settings, pkt, err);
if (!retry.should_retry) {
return false;
}
mc_PACKET *newpkt = mcreq_renew_packet(pkt);
newpkt->flags &= ~MCREQ_STATE_FLAGS;
// TODO: Load the 4th argument from the error map
instance->retryq->add((mc_EXPACKET *)newpkt, err, status, nullptr);
return true;
}
static void fail_callback(mc_PIPELINE *pipeline, mc_PACKET *pkt, lcb_STATUS err, void *)
{
static_cast<Server *>(pipeline)->purge_single(pkt, err);
}
static const char *opcode_name(uint8_t code)
{
switch (code) {
case PROTOCOL_BINARY_CMD_GET:
return "get";
case PROTOCOL_BINARY_CMD_SET:
return "set";
case PROTOCOL_BINARY_CMD_ADD:
return "add";
case PROTOCOL_BINARY_CMD_REPLACE:
return "replace";
case PROTOCOL_BINARY_CMD_DELETE:
return "delete";
case PROTOCOL_BINARY_CMD_INCREMENT:
return "incr";
case PROTOCOL_BINARY_CMD_DECREMENT:
return "decr";
case PROTOCOL_BINARY_CMD_FLUSH:
return "flush";
case PROTOCOL_BINARY_CMD_GETQ:
return "getq";
case PROTOCOL_BINARY_CMD_NOOP:
return "noop";
case PROTOCOL_BINARY_CMD_VERSION:
return "version";
case PROTOCOL_BINARY_CMD_APPEND:
return "append";
case PROTOCOL_BINARY_CMD_PREPEND:
return "prepend";
case PROTOCOL_BINARY_CMD_STAT:
return "stat";
case PROTOCOL_BINARY_CMD_VERBOSITY:
return "verbosity";
case PROTOCOL_BINARY_CMD_TOUCH:
return "touch";
case PROTOCOL_BINARY_CMD_GAT:
return "gat";
case PROTOCOL_BINARY_CMD_HELLO:
return "hello";
case PROTOCOL_BINARY_CMD_SASL_LIST_MECHS:
return "sasl_list_mechs";
case PROTOCOL_BINARY_CMD_SASL_AUTH:
return "sasl_auth";
case PROTOCOL_BINARY_CMD_SASL_STEP:
return "sasl_step";
case PROTOCOL_BINARY_CMD_GET_REPLICA:
return "get_replica";
case PROTOCOL_BINARY_CMD_SELECT_BUCKET:
return "select_bucket";
case PROTOCOL_BINARY_CMD_OBSERVE_SEQNO:
return "observe_seqno";
case PROTOCOL_BINARY_CMD_OBSERVE:
return "observe";
case PROTOCOL_BINARY_CMD_GET_LOCKED:
return "get_locked";
case PROTOCOL_BINARY_CMD_UNLOCK_KEY:
return "unlock_key";
case PROTOCOL_BINARY_CMD_GET_CLUSTER_CONFIG:
return "get_cluster_config";
case PROTOCOL_BINARY_CMD_SUBDOC_GET:
return "subdoc_get";
case PROTOCOL_BINARY_CMD_SUBDOC_EXISTS:
return "subdoc_exists";
case PROTOCOL_BINARY_CMD_SUBDOC_DICT_ADD:
return "subdoc_dict_add";
case PROTOCOL_BINARY_CMD_SUBDOC_DICT_UPSERT:
return "subdoc_dict_upsert";
case PROTOCOL_BINARY_CMD_SUBDOC_DELETE:
return "subdoc_delete";
case PROTOCOL_BINARY_CMD_SUBDOC_REPLACE:
return "subdoc_replace";
case PROTOCOL_BINARY_CMD_SUBDOC_ARRAY_PUSH_LAST:
return "subdoc_array_push_last";
case PROTOCOL_BINARY_CMD_SUBDOC_ARRAY_PUSH_FIRST:
return "subdoc_array_push_first";
case PROTOCOL_BINARY_CMD_SUBDOC_ARRAY_INSERT:
return "subdoc_array_insert";
case PROTOCOL_BINARY_CMD_SUBDOC_ARRAY_ADD_UNIQUE:
return "subdoc_array_add_unique";
case PROTOCOL_BINARY_CMD_SUBDOC_COUNTER:
return "subdoc_counter";
case PROTOCOL_BINARY_CMD_SUBDOC_MULTI_LOOKUP:
return "subdoc_multi_lookup";
case PROTOCOL_BINARY_CMD_SUBDOC_MULTI_MUTATION:
return "subdoc_multi_mutation";
case PROTOCOL_BINARY_CMD_SUBDOC_GET_COUNT:
return "subdoc_get_count";
case PROTOCOL_BINARY_CMD_GET_ERROR_MAP:
return "get_error_map";
case PROTOCOL_BINARY_CMD_GET_META:
return "exists";
default:
return "unknown";
}
}
void Server::purge_single(mc_PACKET *pkt, lcb_STATUS err)
{
if (err != LCB_ERR_REQUEST_CANCELED) {
if (maybe_retry_packet(pkt, err, PROTOCOL_BINARY_RESPONSE_UNSPECIFIED)) {
return;
}
}
if (err == LCB_ERR_AUTHENTICATION_FAILURE) {
/* In-situ auth errors are actually dead servers. Let's provide this
* as the actual error code. */
err = LCB_ERR_MAP_CHANGED;
}
if (err == LCB_ERR_TIMEOUT) {
lcb_STATUS tmperr = lcb::RetryQueue::error_for(pkt);
if (tmperr != LCB_SUCCESS) {
err = tmperr;
}
}
protocol_binary_request_header hdr;
memcpy(hdr.bytes, SPAN_BUFFER(&pkt->kh_span), sizeof(hdr.bytes));
MemcachedResponse resp(protocol_binary_command(hdr.request.opcode), hdr.request.opaque,
PROTOCOL_BINARY_RESPONSE_EINVAL);
lcbtrace_span_set_orphaned(MCREQ_PKT_RDATA(pkt)->span, true);
if (err == LCB_ERR_TIMEOUT && settings->use_tracing) {
Json::Value info;
char opid[30] = {};
snprintf(opid, sizeof(opid), "kv:%s", opcode_name(hdr.request.opcode));
info["s"] = opid;
if (settings->bucket) {
info["b"] = settings->bucket;
}
info["t"] = (Json::UInt64)LCB_NS2US(MCREQ_PKT_RDATA(pkt)->deadline - MCREQ_PKT_RDATA(pkt)->start);
if (this->has_valid_host()) {
const lcb_host_t &remote = get_host();
std::string rhost;
if (remote.ipv6) {
rhost.append("[").append(remote.host).append("]:").append(remote.port);
} else {
rhost.append(remote.host).append(":").append(remote.port);
}
info["r"] = rhost.c_str();
}
if (connctx) {
char local_id[54] = {};
snprintf(local_id, sizeof(local_id), "%016" PRIx64 "/%016" PRIx64 "/%x", settings->iid, connctx->sock->id,
(int)pkt->opaque);
info["i"] = local_id;
info["l"] = connctx->sock->info->ep_local_host_and_port;
}
std::string msg(Json::FastWriter().write(info));
if (msg.size() > 1) {
lcb_log(LOGARGS(instance, WARN), "Failing command with error %s: %.*s", lcb_strerror_short(err),
(int)msg.size(), msg.c_str());
}
} else {
lcb_log(LOGARGS_T(WARN), LOGFMT "Failing command (pkt=%p, opaque=%lu, opcode=0x%x) with error %s", LOGID_T(),
(void *)pkt, (unsigned long)pkt->opaque, hdr.request.opcode, lcb_strerror_short(err));
}
int rv = mcreq_dispatch_response(this, pkt, &resp, err);
lcb_assert(rv == 0);
}
int Server::purge(lcb_STATUS error, hrtime_t now, RefreshPolicy policy)
{
int affected;
if (now) {
affected = (int)mcreq_pipeline_timeout(this, error, fail_callback, nullptr, now);
} else {
mcreq_pipeline_fail(this, error, fail_callback, nullptr);
affected = -1;
}
MC_INCR_METRIC(this, packets_errored, affected);
if (policy == REFRESH_NEVER) {
return affected;
}
if (affected || policy == REFRESH_ALWAYS) {
instance->bootstrap(BS_REFRESH_THROTTLE | BS_REFRESH_INCRERR);
}
return affected;
}
static void flush_errdrain(mc_PIPELINE *pipeline)
{
/* Called when we are draining errors. */
auto *server = (Server *)pipeline;
if (!lcbio_timer_armed(server->io_timer)) {
lcbio_timer_rearm(server->io_timer, server->default_timeout());
}
}
uint32_t Server::next_timeout() const
{
hrtime_t now, expiry, diff, min = 0;
mc_PACKET *pkt = nullptr;
sllist_iterator iter;
SLLIST_ITERFOR(const_cast<sllist_root *>(&requests), &iter)
{
mc_PACKET *p = SLLIST_ITEM(iter.cur, mc_PACKET, slnode);
hrtime_t deadline = MCREQ_PKT_RDATA(p)->deadline;
if (pkt == nullptr || deadline < min) {
min = deadline;
pkt = p;
}
}
if (!pkt) {
return default_timeout();
}
now = gethrtime();
expiry = MCREQ_PKT_RDATA(pkt)->deadline;
if (expiry <= now) {