/
ChangeLog
4137 lines (2829 loc) · 135 KB
/
ChangeLog
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
2009-11-15 17:27 t-ishii
* pgpool.conf.sample, pool.h, pool_config.c, pool_config.l,
pool_process_query.c, pool_proto_modules.c, pool_relcache.c: Add
new directive log_per_node_statement. If true, print all statements
to the log. Similar to log_statement except that prints DB node id
and backend process id info.
2009-11-15 17:05 t-ishii
* pool_config.c, pool_config.l: Set change default value of
reset_query_list to "ABORT, DISCARD ALL"
2009-11-14 22:22 t-ishii
* child.c, pool.h, pool_process_query.c, pool_proto_modules.c,
pool_proto_modules.h: Fix connection_cache = false case. In this
mode, if frontend exits abnormaly, some global state variables are
not reset. For this purpose, new function reset_variables() added.
Remove unnecessary argument of end_load_balance().
2009-11-14 20:41 t-ishii
* pool_config.c: Update pool_config.l
2009-11-14 20:40 t-ishii
* pool_config.l: Allow to change weight with reloading pgpool.conf
2009-11-14 20:35 t-ishii
* main.c: Limit listen(2)'s backlog parameter to
num_init_children*2 at largest. This is almost same algorithm as
PostgreSQL. I hope pgpool-II does not consume so much system
resource any more...
2009-11-14 08:35 t-ishii
* pool_timestamp.c, test/timestamp/expected/insert.out,
test/timestamp/expected/misc.out,
test/timestamp/expected/update.out: Fix some query rewriting
cases.
2009-11-12 16:37 t-ishii
* pgpool.conf.sample: Change reset_query_list default to PostgreSQL
8.3.
2009-11-12 16:04 t-ishii
* pool_process_query.c: Prepare for bind_rewrite_timestamp()
returns NULL case. This could happen if SELECT now() is called
inside an aborted transaction.
Fix SimpleForwardToBackend free rewrite_msg too early.
2009-11-11 10:53 t-ishii
* pool_timestamp.c: Fix comment.
2009-11-10 19:03 t-ishii
* Makefile.am, Makefile.in, pool_process_query.c,
pool_proto_modules.c, pool_proto_modules.h, pool_timestamp.c,
pool_timestamp.h, test/timestamp/Makefile, test/timestamp/main.c,
test/timestamp/parse_schedule, test/timestamp/run-test,
test/timestamp/expected/insert.out,
test/timestamp/expected/misc.out,
test/timestamp/expected/update.out,
test/timestamp/input/insert.sql, test/timestamp/input/misc.sql,
test/timestamp/input/update.sql: Add support for
CURRENT_TIMESTAMP and friends. Patch provided by Akio Ishida.
2009-11-10 11:24 t-ishii
* main.c: Fix comment.
2009-11-05 15:26 t-ishii
* pool_relcache.c: Fix relcache replacement logic bug. Patches
provided by Akio Ishida.
2009-11-05 12:43 t-ishii
* ChangeLog, TODO, parser/outfuncs.c,
test/parser/expected/insert.out, test/parser/expected/select.out,
test/parser/input/insert.sql, test/parser/input/select.sql: Fix
nodeToString() bug. example test case: If nodeToString() is given
the parse tree of "insert into r1 values(1,2), (3,4)", it returns
"insert into r1 values(1,2), values(3,4)", which is apparently
wrong. Patch provided by Akio Ishida.
2009-11-04 22:59 t-ishii
* pool_process_query.c: Fix do_query() to return more than 1
characters. Patch provided by Akio Ishida.
2009-11-03 20:14 devrim
* pgpool.spec:
- Remove init script from all runlevels before uninstall. Per #RH
Bugzilla 532177
2009-11-03 11:00 t-ishii
* pool.h, pool_auth.c, pool_process_query.c: When authentication
fails, report to frontend proper error messages, rather than
unhelpfull "server closed the connection unexpectedly" message.
Patch provided by Glyn Astill with minor tweak by Tatsuo Ishii.
2009-11-03 10:50 t-ishii
* pool_stream.c: Fix pool_flush_it() so that it does not complain
when failed to flush to frontend. Also add info to the message if it
was backend or frontend to avoid confusion.
2009-10-30 14:11 t-ishii
* pool_proto_modules.c: Allow is_temp_table to work with pre 8.4
PostgreSQL.
2009-10-29 22:44 t-ishii
* Makefile.am, Makefile.in, child.c, configure, configure.in,
pool.h, pool_process_query.c, pool_proto_modules.c: Add
pool_relcache.c to Makefile.am. Refactor need_insert_lock to use
pool_search_relcache and friends. Add LocalSessionId to child.c
which manage per session counter of relcache. Allow to use temp
tables with master/slave mode. Parse() failed. since Parse() was
always sent to master and slave and slave does not have the temp
table. To fix this, is_temp_table() added to pool_proto_modules.c
which inquire particular table is a temporary one or not when
INSERT/UPDATE/DELETE. If it's a temporary table, parse message is
sent to only master. Note that /*NO LOAD BALANCE*/ directive can be
used to control this. SELECT still needs /*NO LOAD BALANCE*/. It's
hard to find all tables in a SELECT statement.
2009-10-29 21:46 t-ishii
* pool_relcache.c: Per process relation cache module
2009-10-25 16:19 t-ishii
* test/jdbc/: CreateTempTableTest.java, expected/CreateTempTable,
RunTest.java: Add temp table test
2009-10-22 20:52 devrim
* redhat/: pgpool.init, pgpool.sysconfig: Fix logging, per report
from Toshihiro Kitagawa.
2009-10-21 23:39 t-ishii
* pool_process_query.c: Do not execute CREATE/DROP TABLE SPACE
inside a transaction block.
Teach do_command() that S (parameter status) should be ignored until
receiving command complete.
2009-10-21 23:05 t-ishii
* main.c: Let reaper() report the case when child died by segfault.
2009-10-12 11:21 t-ishii
* pgpool.8.in: Fix typo in man page. Patch provided by Rodolphe.
2009-10-12 10:54 t-ishii
* pool_process_query.c: REINDEX DATABASE or SYSTEM should not be
executed in a transaction block. See [pgpool-general-jp: 623] for
more details.
2009-10-06 00:12 devrim
* pgpool.spec:
Add 2 new docs, per Tatsuo.
2009-10-05 19:38 devrim
* pgpool.spec:
- Update to 2.2.5, for various fixes described at
http://lists.pgfoundry.org/pipermail/pgpool-general/2009-October/002188.html
- Re-apply a fix for Red Hat Bugzilla #442372
- Sync with Fedora spec.
2009-10-04 08:52 t-ishii
* doc/: tutorial-en.html, tutorial-ja.html: Add sample to use
rotatelogs
2009-10-03 21:39 t-ishii
* sample/pgpool_recovery_pitr: Enhance online recovery 2nd stage
script to sync sequence value
2009-10-03 20:55 t-ishii
* pool_proto_modules.h: Remove unnecessary declaration
"allow_close_transaction"
2009-10-02 17:00 t-ishii
* pool_proto_modules.c: Prevoius commit was incomplete
2009-10-02 16:56 t-ishii
* pool_proto_modules.c: Fix Parse() to use
wait_for_query_response() instead of synchronize().
2009-10-02 16:53 t-ishii
* pool_process_query.c: Fix comment
2009-09-26 18:14 t-ishii
* pool_proto_modules.c: Allow to print more accurate query when
Parse() detects kind mismatch errors.
2009-09-26 16:51 t-ishii
* pool_process_query.c: Fix reset_backend to deal with DEALLOCATE
error. In this case prepared objects are not removed and
reset_backend is infinitely called.
Fix do_command to guard against the case that after execution of an
SQL command, the backend does not return either N, E or C. This
should not happen but there was a report that it actually happend in
certain cases(maybe caused by asynchronized backend data?)
See:
Subject: Re: [Pgpool-general] pgpool 2.2.4: DEALLOCATED children
From: Xavier Noguer Date: Fri, 25 Sep 2009 03:35:57 -0400
For more details.
Patches by Xavier Noguer with minor tweak by Tatsuo Ishii.
2009-09-26 09:27 t-ishii
* pool_process_query.c: Remove error checks for
pool_write(frontend), pool_flush(frontend) in
SimpleForwardToFrontend.
Check error for send_deallocate() in reset_backend. Otherwise
send_deallocate will be called infinitely. Per bug report from
Agustin Almonte Ferrada. See:
Subject: Re: [Pgpool-general] pgpool 2.2.4: DEALLOCATED children
From: Agustin Almonte Ferrada <aalmonte@antica.cl> To: Tatsuo Ishii
<ishii@sraoss.co.jp> Cc: pgpool-general@pgfoundry.org Date: Wed, 23
Sep 2009 01:55:54 -0400
2009-09-26 09:05 t-ishii
* pool_process_query.c: FIX parallel query
1. Fix JOIN Expr
Subquery was used on larg or rarg.
Subquery was used on ON-Cluase.
Full OUTER JOIN was not optimized.
2. Fail_on_error=false was passed to the 3rd arguments of
dblink.
This patch was applied to only 2.2-STABLE tree by y-mori on
2009/09/08.
2009-09-25 16:35 t-ishii
* main.c: Fix write_pid_file. It forgot to write out NULL
termination of pid. Pgpool occasionally won't stop because of this
bug.
2009-09-23 11:03 t-ishii
* pool_auth.c: Fix pool_read_kind and pool_read_int so that the
error message for kind is shown in hexadecimal. Other part of
pgpool-II already does this.
2009-09-17 20:32 t-ishii
* pool_process_query.c, pool_proto_modules.c, pool_proto_modules.h:
Teach do_command() to use wait_for_query_response(). insert_lock()
calls do_command() to issue LOCK TABLE to lock tables having SERIAL
column when insert_lock is true. Problem is, insert_lock() waits
forever in do_command() if the target table is locked by someone
else. This patch should give a way insert_lock() to espcape from the
situation.
2009-09-06 12:56 t-ishii
* child.c, pool.h: Fix connection_count_down() so that it does
decrement too much in some corner cases. If failed to read a start
up packet, or receive cancel request etc., connection_count_down()
is called and goes back to the onnection accept loop. Problem is, at
the very beginning of the connection accept loop, if we have
received a signal, we call child_exit() which calls
connection_count_down() again. Problem report by Ryoichi TANABE.
See [pgpool-general-jp: 616] for more details.
Also unneccesary public declarations for connection_count_up() and
connection_count_down() are removed.
2009-09-06 12:54 t-ishii
* pool_process_query.c: Undef #define SEND_NOTICE_ON_PROTO2, which
was defined by mistake.
2009-08-22 13:04 t-ishii
* child.c, main.c, md5.c, pcp_child.c, pg_md5.c, pool_auth.c,
pool_config.c, pool_connection_pool.c, pool_error.c, pool_hba.c,
pool_ip.c, pool_params.c, pool_path.c, pool_process_query.c,
pool_proto_modules.c, pool_query_cache.c,
pool_rewrite_outfuncs.c, pool_rewrite_query.c, pool_sema.c,
pool_shmem.c, pool_signal.c, pool_stream.c, pool_system.c,
ps_status.c, recovery.c, strlcpy.c, pcp/md5.c: Remove unnecessary
spaces and tabs at end of line. Patch conributed by Jun Kuriyama.
2009-08-20 15:26 t-ishii
* pool_process_query.c: Fix possible bug introduced in pgpool-II
2.2.2. Load balance control variables may remain not be restored
and subsequent DML/DDL call might sent to only master node.
2009-08-18 17:41 t-ishii
* pool_process_query.c, pool_proto_modules.c, pool_proto_modules.h:
Change wait_for_query_response() so that it sends NOTICE message to
frontend periodically. It appears that pgpool-II 2.2.3
unconditionaly sends param packet to client even it uses version 2
protocol, which is apparentlt wrong. Also tweak checking period from
1 second to 30 seconds since 1 second seems too aggressive. Note
that you need to edit pool_process_query.c to define
SEND_NOTICE_ON_PROTO2 around line 879 to enable this feature in
version 2 protocol.
2009-08-18 17:30 t-ishii
* child.c: Remove unneccessary call to send_frontend_exit() in
do_child().
2009-08-13 22:40 t-ishii
* child.c: Remove unnessary call of send_frontend_exits() in die().
2009-08-13 22:39 t-ishii
* main.c: Block signals before forking children rather after.
Otherwise parent will be killed by failover signal if it receives
before establish signal handler. See [pgpool-general-jp: 586] for
more details.
2009-08-11 22:23 t-ishii
* pool_connection_pool.c: Fix new_connection() so that it does
leave garbage in POOL_CONNECTION_POOL->slot[i]. This caused segfault
in child_exit(). Bug report at [pgpool-general-jp: 586].
2009-08-06 20:38 t-ishii
* Makefile.in: Update Makefile.am
2009-08-06 18:11 t-ishii
* Makefile.am: Add parser/test files. Patch contributed by Akio
Ishida.
2009-08-06 16:55 t-ishii
* pool_proto_modules.c: Print last query for kind mismatch error
message even extended protocol is used. Patch contributed by Akio
Ishida.
2009-08-06 16:46 t-ishii
* pool_process_query.c: Fix buffer overrun error which mangled kind
mismatch error message in pool_send_error_message(). Bug rpeort by
Takahiro Itagaki at [pgsql-jp: 39950]. Fix suggested by Akio Ishida
at [pgsql-jp: 39952].
2009-08-03 19:31 t-ishii
* parser/: kwlist.h, kwlookup.c: Missing files for PostgreSQL 8.4
paraser adaoptation. Patch contributed by Akio Ishida.
2009-08-01 20:41 t-ishii
* child.c: Check MASTER_CONNECTION(p)->sp) is non 0 before further
referring to sp. Current implementation may interrupt in the window,
after setting p and before setting sp.
2009-08-01 20:34 t-ishii
* pool_connection_pool.c: Fix pool_debug's mismatch between data
type and conversion specification for time_t. Check
MASTER_CONNECTION(p)->sp) is non 0 before further referring to sp.
Current implementation may interrupt in the window, after setting p
and before setting sp.
2009-08-01 12:20 t-ishii
* test/parser/: parse_schedule, input/v84.sql: parser test for
PostgreSQL 8.4 or later. Patch contributed by Akio Ishida.
2009-07-30 23:44 t-ishii
* pool_process_query.c, pool_rewrite_outfuncs.c, pool_type.h,
parser/Makefile.am, parser/Makefile.in, parser/copyfuncs.c,
parser/gram.c, parser/gram.h, parser/gram.y, parser/keywords.c,
parser/keywords.h, parser/makefuncs.c, parser/makefuncs.h,
parser/nodes.c, parser/nodes.h, parser/outfuncs.c,
parser/parsenodes.h, parser/parser.c, parser/pool_memory.h,
parser/pool_parser.h, parser/primnodes.h, test/parser/Makefile,
test/parser/expected/copy.out, test/parser/expected/create.out,
test/parser/expected/cursor.out, test/parser/expected/delete.out,
test/parser/expected/insert.out, test/parser/expected/misc.out,
test/parser/expected/prepare.out,
test/parser/expected/privileges.out,
test/parser/expected/select.out, test/parser/expected/update.out:
Adopt PostgreSQL 8.4 parser. Patch contributed by Akio Ishida.
2009-07-28 09:34 t-ishii
* pool_process_query.c: Fix typo in comment
2009-07-26 18:14 t-ishii
* pool_rewrite_outfuncs.c: Fix breakage by previous commit. Patch
by Guillaume Lelarge.
2009-07-26 12:09 t-ishii
* pool_rewrite_outfuncs.c: fix: Join statment in parallel query
sample)
Accouts table is defined by dist_def,
the below query is executed as loadbalance.
SELECT a.aid
FROM
accounts as a
JOIN
accounts as b
ON a.aid=b.aid
WHERE a.aid =1 or a.aid=2;
2009-07-26 12:01 t-ishii
* pool_process_query.c: Fix comment about load balancing.
2009-07-26 11:57 t-ishii
* pool_sema.c: Fix various pool_error/log format errors. Patch
provided by Akio Ishida.
2009-07-22 17:51 t-ishii
* child.c, main.c, pool.h, pool_process_query.c,
pool_proto_modules.c, pool_rewrite_outfuncs.c, pool_shmem.c: Fix
various pool_error/log format errors. Patch provided by Akio
Ishida.
2009-07-22 17:26 t-ishii
* pool_process_query.c, pool_proto_modules.c, pool_proto_modules.h:
Query cancel by statement timeout does not generate "kind mismatch"
error any more. How to reproduce the test case:
Directory connect to PostgreSQL. T1: BEGIN; T1: LOCK TABLE t1;
T2: SET statement_timeout TO 5000; T2: UPDATE t1 SET ...
UPDATE query aborts and error message comes out on T2: ERROR:
canceling statement due to statement timeout
A side effect of this is, statement timeout works correctly (rather
than statement timeout value * # of node).
Note that to initiate T1 by directory connecting to slave node will
not work (still results in kind mismatch error). But I don't think
this could happend in the real world.
Also refactor error detecting modules (detect_deadlock_error and so
on).
2009-07-21 18:48 t-ishii
* main.c: Check read(2) error in health_check(). This should detect
case described in message below, i.e. postmaster is blocked by
SIGSTOP.
Subject: [Pgpool-general] Healt Check issues From: Vctrjsn@gmail.com
To: pgpool-general@pgfoundry.org Date: Sun, 05 Jul 2009 18:01:18
+0000
2009-07-21 10:43 t-ishii
* pool.h, pool_proto_modules.c: Check frontend connection while
waiting for response from backend for extended protocol (Execute()).
Replace synchronize() with wait_for_query_response(). Also tweak
pool.h so that it includes cancel packet's mazic proto number.
2009-07-21 09:30 t-ishii
* pool_process_query.c: Do not start transaction if the command is
CLUSTER with no argument
2009-07-20 21:21 t-ishii
* child.c, pool.h, pool_process_query.c, pool_proto_modules.c,
pool_proto_modules.h: Check frontend connection while waiting for
response from backend. This is done in wait_for_query_response() by
writing a dummy parameter status packet to pgpool client every 1
second. If the write fails, a cancel packet is sent to backend to
free backend. This is usefull to prevent pgpool child from waiting
forever for backend response. Pgpool client will diconnect to pgpool
if the query takes too long. This is typical use case for Web
applications.
2009-07-20 21:16 t-ishii
* pool_process_query.c: Fix bug with SimpleForwardToFrontend().
Clients using the extend protocol send 'H'(Flush) after each
command. Problem is, SimpleForwardToFrontend does not immediately
forward 'C'(Command Complete) packet to frontend. So the client
sends 'H' and wait for 'C'. On the other hand backend receives 'H'
but reply back nothing since Command complete has nothing to be
returned. Both side stuck.
The fix is, forwarding 'C' packet immediately after receiving it
from backend. To prevent similar bugs happen in the future, I put
pool_flush(frontend) after calling SimpleForwardToFrontend in
pool_process_query() (other places already do this).
The extra pool_flush() might harm performance. So I did some tests
using pgbench -S, and found nothing has been changed regarding
performance after patching.
Bug report from Gavin Sherry along with patches for
SimpleForwardToFrontend().
2009-07-20 21:06 t-ishii
* pool.h, pool_process_query.c: Add pool_set_timeout() to control
pool_check_fd()'s behavior. This can be used for various places.
The first user is reset_backend() to fix following scenario: Do
COPY_FROM(for example pgbench -i) Cancel pgbench pgpool waits for
response from backend while sending reset queries and hang. Backend
waits for data coming from pgpool forever.
BTW, I don't think PostgreSQL's befaviro for COPY_FROM is quite
correct since the doc stats:"Receipt of any other non-copy message
type constitutes an error that will abort the copy-in state as
described above" If my understanding correct, COPY_FROM should abort
copy-in state if it receives ABORT (issued in reset_backend). It
seems PostgreSQL is not implemented as stated in the doc...
2009-07-15 15:09 t-ishii
* pool_process_query.c: Fix bug with prepared statement
deallocation:
Subject: [Pgpool-general] Pgpool probelms From: Nimesh Satam
<nimesh.satam@gmail.com> To: pgpool-general@pgfoundry.org Cc: Nimesh
Satam <nimesh.zedo@gmail.com> Date: Wed, 17 Jun 2009 11:31:37 +0530
This was caused a bug in reset_backend().
if (qcnt >= qn + prepared_list.cnt) {
reset_prepared_list(&prepared_list);
return 2; } send_deallocate(backend, &prepared_list, qcnt - qn);
return 1;
Problem is, send_deallocate -> SimpleQuery("DEALLOCATE") calls
del_prepared_list, which will decrement prepared_list.cnt. In short,
incrementing of qcnt *and* decrementing of prepared_list.cnt happen
at the same time. So deallocate is only called 1/2 times than it
should be.
Bug analysis and fix provided by Akio Ishida.
2009-07-14 17:39 t-ishii
* sql/pgpool-recovery/pgpool-recovery.c: PostgreSQL 8.4 needs to
include utils/binutils.h
2009-05-24 17:43 t-ishii
* sql/pgpool-recovery/Makefile: Support pre 8.0 PostgreSQL which
does not have pg_config --pgxs option. Patch provided by Enrico
Pirozzi with slight modification by Tatsuo Ishii
2009-05-12 10:01 t-ishii
* doc/pgpool-en.html: Fix description of client_idle_limit
2009-05-02 17:48 t-ishii
* pool_process_query.c: Fix is_select_query() to test "sql"
parameter is NULL. As of 2.2, pgpool sets Portal->sql_string to
NULL for SQL command PREPARE. Usually this is ok, since in most
cases SQL command EXECUTE follows anyway. Problem is, some
applications mix PREPARE with extended protocol command "EXECUTE"
and so on. Execute() seems to think this never happens but it is not
real. Someday we should extract actual query string from
PrepareStmt->query and set it to Portal->sql_string.
2009-05-02 17:45 t-ishii
* pool_stream.c: Fix pool_flush() so that it ignores error on
frontend connection. This is needed to continue processing with
backend, otherwise we risk data incositency.
2009-05-02 17:41 t-ishii
* child.c: Call pool_send_frontend_exits() before pool_discard_cp()
when resetting connection fails. This will avoid "unexpected EOF on
client connection" error on when pgpool client exits abnormaly
2009-05-01 09:37 t-ishii
* pool_proto_modules.c: Add missing patches for this:
date: 2009/04/08 11:12:11; author: t-ishii; state: Exp; lines:
+15 -14 In master slave mode, sometimes DEALLOCATE fails. This is
caused by that the first PREPARE was not executed on the slave.
Patch contributed by Toshihiro Kitagawa.
2009-04-08 20:11 t-ishii
* pool_proto_modules.c: In master slave mode, sometimes DEALLOCATE
fails. This is caused by that the first PREPARE was not executed on
the slave. Patch contributed by Toshihiro Kitagawa.
2009-04-03 22:43 t-ishii
* doc/: pgpool-en.html, pgpool-ja.html: Add documents about
configuration reloading.
2009-04-03 22:30 t-ishii
* pool_proto_modules.c: Fix SimpleQuery() to properly check
extended query related commands. Before it ignores them if query is
not NULL. This is problematic if frontend == NULL, which could
happen with queries defined in reset_query_list. The bug had been
there since 2.1 BTW.
2009-03-23 10:14 devrim
* pgpool.spec, redhat/pgpool.init:
- Own /usr/share/pgpool-II directory.
- Fix pid file path in init script, per pgcore #81.
- Fix spec file -- we don't use short_name macro in pgcore spec
file.
- Create pgpool pid file directory, per pgcore #81.
- Fix stop/start routines, also improve init script a bit.
- Install conf files to a new directory (/etc/pgpool-II), and get
rid of sample conf files.
2009-03-23 10:13 devrim
* redhat/pgpool.conf.sample.patch: Add a patch which configures
sample conf file to match RPM and Red Hat defaults.
2009-03-14 12:39 t-ishii
* pool_process_query.c: Fix need_insert_lock so that it returns
false if the protocol version is 2. insert_lock has not been working
for V2 protocol anyway, so this is ok.
2009-03-09 00:00 gleu
* main.c, pgpool.8.in, doc/pgpool-en.html: Update usage function,
doc and man page.
2009-03-07 18:19 t-ishii
* pool_process_query.c: Add a comment to need_insert_lock() if the
query being used to check if the table has SERIAL column is valid
for PostgreSQL 7.3 to 8.3.
2009-03-07 08:49 t-ishii
* pool_process_query.c: Make excessive message level in
read_kind_from_backend from LOG to DEBUG.
2009-03-06 18:35 gleu
* main.c: Add support for long options name in pgpool command.
2009-03-01 22:22 t-ishii
* doc/pgpool-ja.html: Fix release note for 2.2
2009-02-28 15:29 t-ishii
* Makefile.in: Update Makefile.am
2009-02-28 15:25 t-ishii
* Makefile.am: Add doc/load_balance.png, doc/load_balance.odp
2009-02-28 13:45 t-ishii
* ChangeLog, NEWS, TODO (tags: V2_2_3, V2_2_2, V2_2_1), configure
(utags: V2_2_0): Version 2.2
2009-02-28 13:40 t-ishii
* configure.in (tags: V2_2_0): Version 2.2
2009-02-24 11:20 t-ishii
* pool_auth.c (tags: V2_2_3, V2_2_2, V2_2_1), pool_process_query.c
(utags: V2_2_0): Prpare for receiving parameter status packets.
If postgresql.conf was changed and reloaded, PostgreSQL sends
parameter status packets asynchronously even those changed
parameters are not one of session_authorization etc., which is
marked as GUC_REPORT in guc.c(GUC_REPORT means parameter status
packet is sent if its value is changed).
After receiving those packets, they are discarded. I think this is
harmless since most of changes are not affected to the frontend
behavior.
2009-02-21 19:00 t-ishii
* TODO, configure, ChangeLog (utags: V2_2_0_RC2): pgpool-II 2.2 RC2
2009-02-21 18:56 t-ishii
* configure.in (tags: V2_2_0_RC2): pgpool 2.2 RC2
2009-02-21 18:52 t-ishii
* doc/load_balance.png (tags: V2_2_5, V2_2_4, V2_2_3, V2_2_2,
V2_2_1, V2_2_0, V2_2_0_RC2): A load balance logic explanation
figure
2009-02-21 18:47 t-ishii
* doc/: pgpool-en.html, pgpool-ja.html (utags: V2_2_0, V2_2_0_RC2):
Add comment about rsync and autovacuum.
2009-02-20 10:34 t-ishii
* pool_process_query.c (tags: V2_2_0_RC2): Fix
pool_send_frontend_exits() so that it checks the connection to
backend if it actually exists. This is needed, for example, in the
final stae of online recovery. pgpool child holds 1 connection to a
backend. After failback one more backend added but there's no
connection to the backend.
2009-02-17 18:50 t-ishii
* pool_process_query.c: Do not start internal transaction if CREATE
INDEX CONCURRENTLY
2009-02-17 15:16 devrim
* pgpool.spec (tags: V2_2_0, V2_2_0_RC2): Update to 2.2 (Download
URL is broken, since we don't know yet the pgfoundry download URL
yet)
2009-02-15 14:35 t-ishii
* configure (tags: V2_2_0_RC1): run autoconf
2009-02-15 14:32 t-ishii
* configure.in (tags: V2_2_0_RC1): 2.2 RC1
2009-02-15 14:31 t-ishii
* ChangeLog (tags: V2_2_0_RC1): Update Changelog
2009-02-15 14:29 t-ishii
* TODO (tags: V2_2_0_RC1): Mention that two items were resolved
2009-02-15 14:26 t-ishii
* pgpool.conf.sample (tags: V2_2_5, V2_2_4, V2_2_3, V2_2_2, V2_2_1,
V2_2_0, V2_2_0_RC2), doc/pgpool-en.html, doc/pgpool-ja.html
(utags: V2_2_0_RC1): Reflect the fact that DISCARD ALL holds
exclusive lock on pg_listener, thus will degrade performance.
2009-02-10 10:11 t-ishii
* pool_process_query.c (tags: V2_2_0_RC1): Fix
is_internal_transaction_needed() so that DROP DATABASE does not
start internal transaction
2009-02-07 21:57 t-ishii
* configure (tags: V2_2_0_BETA2): Run autoconf
2009-02-07 21:57 t-ishii
* configure.in (tags: V2_2_0_BETA2): Update version
2009-02-07 21:52 t-ishii
* NEWS (tags: V2_2_0_RC2, V2_2_0_RC1, V2_2_0_BETA2): Update NEWS
2009-02-07 00:30 t-ishii
* pool_process_query.c (tags: V2_2_0_BETA2): Add pid_file_name to
show pool_status
2009-02-05 17:24 t-ishii
* doc/pgpool-ja.html (tags: V2_2_0_BETA2): Fix typo
2009-02-05 09:33 t-ishii
* ChangeLog (tags: V2_2_0_BETA2): Run cvs2cl.pl -t --fsf
2009-02-05 09:30 t-ishii
* doc/pgpool-ja.html: Add more fix to release info
2009-02-05 09:26 t-ishii
* recovery.c (tags: V2_2_3, V2_2_2, V2_2_1, V2_2_0, V2_2_0_RC2,
V2_2_0_RC1, V2_2_0_BETA2): Fix online recovery to wait for
failback done before allowing to accept connections from clients. It
was supposed to work like this but actually was not since the day 0
when online recovery was born. Without the fix there could be
potential data inconsistency among DB nodes.
2009-02-05 09:24 t-ishii
* main.c (tags: V2_2_0, V2_2_0_RC2, V2_2_0_RC1, V2_2_0_BETA2):
Fix/add comments
2009-02-01 19:13 t-ishii
* NEWS: Add new item
2009-02-01 19:11 t-ishii
* pool_process_query.c: Start internal transaction if necessary.
Before we only started transaction INSERT, UPDATE, DELETE and
SELECT. Obviously we need to start transaction other than these, for
example CREATE TABLE. To avoid an overhead to look up large number
of statements (currently 70), we use bsearch() now.
2009-01-31 22:17 t-ishii
* doc/pgpool-ja.html: Update release note
2009-01-30 17:15 t-ishii
* test/jdbc/README.euc_jp (tags: V2_2_5, V2_2_4, V2_2_3, V2_2_2,
V2_2_1, V2_2_0, V2_2_0_RC2, V2_2_0_RC1, V2_2_0_BETA2): Add
comment how to add CLASSPATH
2009-01-30 09:06 y-mori
* pool_rewrite_query.c, pool_rewrite_query.h (tags: V2_2_5, V2_2_4)
(utags: V2_2_0, V2_2_0_BETA2, V2_2_0_RC1, V2_2_0_RC2, V2_2_1,
V2_2_2, V2_2_3): Add comment
2009-01-28 15:48 t-ishii
* NEWS: Add more items
2009-01-27 10:04 t-ishii
* COPYING (tags: V2_2_5, V2_2_4, V2_2_3, V2_2_2, V2_2_1, V2_2_0,
V2_2_0_RC2, V2_2_0_RC1, V2_2_0_BETA2): Update copyright year
2009-01-25 20:13 t-ishii
* NEWS, TODO (tags: V2_2_0_BETA2), configure, configure.in,
version.h (tags: V2_2_5, V2_2_4, V2_2_3, V2_2_2, V2_2_1, V2_2_0,
V2_2_0_RC2, V2_2_0_RC1, V2_2_0_BETA2), doc/pgpool-en.html (tags:
V2_2_0_BETA2), doc/pgpool-ja.html (utags: V2_2_0_BETA1): Prepare
2.2 beta1
2009-01-25 19:13 t-ishii
* main.c, pool.h (tags: V2_2_2, V2_2_1, V2_2_0, V2_2_0_RC2,
V2_2_0_RC1, V2_2_0_BETA2), pool_config.l (tags: V2_2_5, V2_2_4,
V2_2_3, V2_2_2, V2_2_1, V2_2_0, V2_2_0_RC2, V2_2_0_RC1,
V2_2_0_BETA2), pgpool.conf.sample (tags: V2_2_0_BETA2),
pool_config.c (tags: V2_2_3, V2_2_2, V2_2_1, V2_2_0, V2_2_0_RC2,
V2_2_0_RC1, V2_2_0_BETA2) (utags: V2_2_0_BETA1): Add
pid_file_name directive
2009-01-25 13:32 t-ishii
* ChangeLog (tags: V2_2_0_BETA1): ChangeLog is now generated by
using cvs2cl.pl -t --fsf
2009-01-25 09:09 t-ishii
* pool_rewrite_query.c (tags: V2_2_0_BETA1): Remove unused static
function check_whereClause() Update Copyright year
2009-01-24 20:34 t-ishii
* pool.h: Include limit.h. This seems to be necessary Fedra box.
Update copyright year
2009-01-22 18:29 y-mori
* pool_stream.c (tags: V2_2_1, V2_2_0, V2_2_0_RC2, V2_2_0_RC1,
V2_2_0_BETA2, V2_2_0_BETA1): Fix different function name was used
in the error message.
2009-01-22 18:16 y-mori
* pool_process_query.c (tags: V2_2_0_BETA1), pool_proto_modules.c
(tags: V2_2_0, V2_2_0_RC2, V2_2_0_RC1, V2_2_0_BETA2,
V2_2_0_BETA1), pool_rewrite_query.c, pool_rewrite_query.h (tags:
V2_2_0_BETA1): add the comment related to parallel query
2009-01-22 09:50 t-ishii
* child.c (tags: V2_2_1, V2_2_0, V2_2_0_RC2, V2_2_0_RC1,
V2_2_0_BETA2, V2_2_0_BETA1): Remove unnecessary function
declaration for child_exit()
Add comments to child_exit() which had no comments at all
Modify child_exit() so that it sends terminate mesagges before
exiting. This will prevent backend process from being idle waiting
for TCP keepalive event occurs. This is not good since a backend
process may hold locking on certain tables.
2009-01-22 09:45 t-ishii
* pool_process_query.c: Modify pool_send_frontend_exits() so that
it change fd to non block while sending terminate message (X) to
backend. This should prevent pgpool from being blocked while
sending data to socket.
2009-01-21 10:43 t-ishii
* pool_process_query.c: Add debug code and message
2009-01-18 22:54 t-ishii
* ps_status.c (tags: V2_2_3, V2_2_2, V2_2_1, V2_2_0, V2_2_0_RC2,
V2_2_0_RC1, V2_2_0_BETA2, V2_2_0_BETA1): Fix FreeBSD case. patch
provided by Jun Kuriyama.
2009-01-14 23:54 t-ishii
* pool_process_query.c: Fix excessive debug statement
2009-01-07 22:33 t-ishii
* pool_process_query.c: Add new argument to detect_error() so that
it can recognize the error class (E or N). per bug report from
Daniel: From: Daniel.Crespo@l-3com.com To:
<pgpool-general@pgfoundry.org> Date: Tue, 6 Jan 2009 12:50:08 -0500
The bug was caused by detect_postmaster_down_error() not detecting
receive crash shutdown error.
Enbug by: Subject: [Pgpool-committers] pgpool - pgpool-II: Fix
detect_error() so that it ignores NOTICE. From:
t-ishii@pgfoundry.org (User T-ishii) To:
pgpool-committers@pgfoundry.org Date: Tue, 30 Dec 2008 01:05:37
+0000 (UTC)
2009-01-07 22:27 t-ishii
* pool_connection_pool.c (tags: V2_2_2, V2_2_1, V2_2_0, V2_2_0_RC2,
V2_2_0_RC1, V2_2_0_BETA2, V2_2_0_BETA1), pool_proto_modules.c:
Update copyright year
2009-01-02 18:38 t-ishii
* pool_process_query.c: Emit pool_error when detects kind mismatch
error. Before we only sent error messages to client, but this will
leave only on client, which may or may not be kept. Leaving it on
pgpool log should leave on pgpool side at least.
2008-12-31 23:45 t-ishii
* child.c, main.c: Fix typos in comments. Patch contributed by Jun
Kuriyama.
2008-12-31 23:42 t-ishii
* pool_sema.c (tags: V2_2_2, V2_2_1, V2_2_0, V2_2_0_RC2,
V2_2_0_RC1, V2_2_0_BETA2, V2_2_0_BETA1): Enhance error message
when failed to create semaphores. Patch contributed by Jun
Kuriyama.
2008-12-31 19:25 t-ishii