-
Notifications
You must be signed in to change notification settings - Fork 382
/
api_data.py
1972 lines (1892 loc) · 105 KB
/
api_data.py
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
# This file is a python script that describes the WiredTiger API.
class Method:
def __init__(self, config):
# Deal with duplicates: with complex configurations (like WT_SESSION::create), it's simpler
# to deal with duplicates once than manually as configurations are defined.
self.config = []
lastname = None
for c in sorted(config):
if '.' in c.name:
raise "Bad config key '%s'" % c.name
if c.name == lastname:
continue
lastname = c.name
self.config.append(c)
class Config:
def __init__(self, name, default, desc, subconfig=None, **flags):
self.name = name
self.default = default
self.desc = desc
self.subconfig = subconfig
self.flags = flags
# Comparators for sorting.
def __eq__(self, other):
return self.name == other.name
def __ne__(self, other):
return self.name != other.name
def __lt__(self, other):
return self.name < other.name
def __le__(self, other):
return self.name <= other.name
def __gt__(self, other):
return self.name > other.name
def __ge__(self, other):
return self.name >= other.name
common_runtime_config = [
Config('app_metadata', '', r'''
application-owned metadata for this object'''),
Config('assert', '', r'''
declare timestamp usage''',
type='category', subconfig= [
Config('commit_timestamp', 'none', r'''
this option is no longer supported, retained for backward compatibility''',
choices=['always', 'key_consistent', 'never', 'none'], undoc=True),
Config('durable_timestamp', 'none', r'''
this option is no longer supported, retained for backward compatibility''',
choices=['always', 'key_consistent', 'never', 'none'], undoc=True),
Config('read_timestamp', 'none', r'''
if set, check that timestamps are \c always or \c never used on reads with this table,
writing an error message if the policy is violated. If the library was built in
diagnostic mode, drop core at the failing check''',
choices=['always', 'never', 'none']),
Config('write_timestamp', 'off', r'''
if set, check that timestamps are used consistently with the configured
\c write_timestamp_usage option for this table, writing an error message if the policy
is violated. If the library was built in diagnostic mode, drop core at the failing
check''',
choices=['off', 'on'], undoc=True),
]),
Config('verbose', '[]', r'''
this option is no longer supported, retained for backward compatibility''',
type='list', choices=['write_timestamp'], undoc=True),
Config('write_timestamp_usage', 'none', r'''
describe how timestamps are expected to be used on table modifications. The choices
are the default, which ensures that once timestamps are used for a key, they are always
used, and also that multiple updates to a key never use decreasing timestamps and
\c never which enforces that timestamps are never used for a table. (The \c always,
\c key_consistent, \c mixed_mode and \c ordered choices should not be used, and are
retained for backward compatibility.)''',
choices=['always', 'key_consistent', 'mixed_mode', 'never', 'none', 'ordered']),
]
# Metadata shared by all schema objects
common_meta = common_runtime_config + [
Config('collator', 'none', r'''
configure custom collation for keys. Permitted values are \c "none" or a custom collator
name created with WT_CONNECTION::add_collator'''),
Config('columns', '', r'''
list of the column names. Comma-separated list of the form <code>(column[,...])</code>.
For tables, the number of entries must match the total number of values in \c key_format
and \c value_format. For colgroups and indices, all column names must appear in the
list of columns for the table''',
type='list'),
]
source_meta = [
Config('source', '', r'''
set a custom data source URI for a column group, index or simple table. By default,
the data source URI is derived from the \c type and the column group or index name.
Applications can create tables from existing data sources by supplying a \c source
configuration''',
undoc=True),
Config('type', 'file', r'''
set the type of data source used to store a column group, index or simple table.
By default, a \c "file:" URI is derived from the object name. The \c type configuration
can be used to switch to a different data source, such as LSM or an extension configured
by the application'''),
]
format_meta = common_meta + [
Config('key_format', 'u', r'''
the format of the data packed into key items. See @ref schema_format_types for details.
By default, the key_format is \c 'u' and applications use WT_ITEM structures to manipulate
raw byte arrays. By default, records are stored in row-store files: keys of type \c 'r'
are record numbers and records referenced by record number are stored in column-store
files''',
type='format', func='__wt_struct_confchk'),
Config('value_format', 'u', r'''
the format of the data packed into value items. See @ref schema_format_types for details.
By default, the value_format is \c 'u' and applications use a WT_ITEM structure to
manipulate raw byte arrays. Value items of type 't' are bitfields, and when configured
with record number type keys, will be stored using a fixed-length store''',
type='format', func='__wt_struct_confchk'),
]
lsm_config = [
Config('lsm', '', r'''
options only relevant for LSM data sources''',
type='category', subconfig=[
Config('auto_throttle', 'true', r'''
Throttle inserts into LSM trees if flushing to disk isn't keeping up''',
type='boolean'),
Config('bloom', 'true', r'''
create Bloom filters on LSM tree chunks as they are merged''',
type='boolean'),
Config('bloom_bit_count', '16', r'''
the number of bits used per item for LSM Bloom filters''',
min='2', max='1000'),
Config('bloom_config', '', r'''
config string used when creating Bloom filter files, passed to WT_SESSION::create'''),
Config('bloom_hash_count', '8', r'''
the number of hash values per item used for LSM Bloom filters''',
min='2', max='100'),
Config('bloom_oldest', 'false', r'''
create a Bloom filter on the oldest LSM tree chunk. Only supported if Bloom filters
are enabled''',
type='boolean'),
Config('chunk_count_limit', '0', r'''
the maximum number of chunks to allow in an LSM tree. This option automatically
times out old data. As new chunks are added old chunks will be removed. Enabling
this option disables LSM background merges''',
type='int'),
Config('chunk_max', '5GB', r'''
the maximum size a single chunk can be. Chunks larger than this size are not
considered for further merges. This is a soft limit, and chunks larger than this
value can be created. Must be larger than chunk_size''',
min='100MB', max='10TB'),
Config('chunk_size', '10MB', r'''
the maximum size of the in-memory chunk of an LSM tree. This limit is soft, it is
possible for chunks to be temporarily larger than this value. This overrides the
\c memory_page_max setting''',
min='512K', max='500MB'),
Config('merge_custom', '', r'''
configure the tree to merge into a custom data source''',
type='category', subconfig=[
Config('prefix', '', r'''
custom data source prefix instead of \c "file"'''),
Config('start_generation', '0', r'''
merge generation at which the custom data source is used (zero indicates no
custom data source)''',
min='0', max='10'),
Config('suffix', '', r'''
custom data source suffix instead of \c ".lsm"'''),
]),
Config('merge_max', '15', r'''
the maximum number of chunks to include in a merge operation''',
min='2', max='100'),
Config('merge_min', '0', r'''
the minimum number of chunks to include in a merge operation. If set to 0 or 1 half
the value of merge_max is used''',
max='100'),
]),
]
tiered_config = [
Config('tiered_storage', '', r'''
configure a storage source for this table''',
type='category', subconfig=[
Config('name', 'none', r'''
permitted values are \c "none" or a custom storage source name created with
WT_CONNECTION::add_storage_source. See @ref custom_storage_sources for more
information'''),
Config('auth_token', '', r'''
authentication string identifier'''),
Config('bucket', '', r'''
the bucket indicating the location for this table'''),
Config('bucket_prefix', '', r'''
the unique bucket prefix for this table'''),
Config('cache_directory', '', r'''
a directory to store locally cached versions of files in the storage source. By
default, it is named with \c "-cache" appended to the bucket name. A relative
directory name is relative to the home directory'''),
Config('local_retention', '300', r'''
time in seconds to retain data on tiered storage on the local tier for faster
read access''',
min='0', max='10000'),
Config('object_target_size', '0', r'''
this option is no longer supported, retained for backward compatibility''',
min='0', undoc=True),
Config('shared', 'false', r'''
enable sharing tiered tables across other WiredTiger instances.''',
type='boolean'),
]),
]
tiered_tree_config = [
Config('bucket', '', r'''
the bucket indicating the location for this table'''),
Config('bucket_prefix', '', r'''
the unique bucket prefix for this table'''),
Config('cache_directory', '', r'''
a directory to store locally cached versions of files in the storage source. By default,
it is named with \c "-cache" appended to the bucket name. A relative directory name
is relative to the home directory'''),
]
file_runtime_config = common_runtime_config + [
Config('access_pattern_hint', 'none', r'''
It is recommended that workloads that consist primarily of updates and/or point queries
specify \c random. Workloads that do many cursor scans through large ranges of data
should specify \c sequential and other workloads should specify \c none. The option leads
to an appropriate operating system advisory call where available''',
choices=['none', 'random', 'sequential']),
Config('cache_resident', 'false', r'''
do not ever evict the object's pages from cache. Not compatible with LSM tables; see
@ref tuning_cache_resident for more information''',
type='boolean'),
Config('log', '', r'''
the transaction log configuration for this object. Only valid if \c log is enabled in
::wiredtiger_open''',
type='category', subconfig=[
Config('enabled', 'true', r'''
if false, this object has checkpoint-level durability''',
type='boolean'),
]),
Config('os_cache_max', '0', r'''
maximum system buffer cache usage, in bytes. If non-zero, evict object blocks from
the system buffer cache after that many bytes from this object are read or written into
the buffer cache''',
min=0),
Config('os_cache_dirty_max', '0', r'''
maximum dirty system buffer cache usage, in bytes. If non-zero, schedule writes for
dirty blocks belonging to this object in the system buffer cache after that many bytes
from this object are written into the buffer cache''',
min=0),
]
# Per-file configuration
file_config = format_meta + file_runtime_config + tiered_config + [
Config('block_allocation', 'best', r'''
configure block allocation. Permitted values are \c "best" or \c "first"; the \c "best"
configuration uses a best-fit algorithm, the \c "first" configuration uses a
first-available algorithm during block allocation''',
choices=['best', 'first',]),
Config('allocation_size', '4KB', r'''
the file unit allocation size, in bytes, must be a power of two; smaller values decrease
the file space required by overflow items, and the default value of 4KB is a good choice
absent requirements from the operating system or storage device''',
min='512B', max='128MB'),
Config('block_compressor', 'none', r'''
configure a compressor for file blocks. Permitted values are \c "none" or a custom
compression engine name created with WT_CONNECTION::add_compressor. If WiredTiger
has builtin support for \c "lz4", \c "snappy", \c "zlib" or \c "zstd" compression,
these names are also available. See @ref compression for more information'''),
Config('checksum', 'on', r'''
configure block checksums; the permitted values are \c on, \c off, \c uncompressed and
\c unencrypted. The default is \c on, in which case all block writes include a checksum
subsequently verified when the block is read. The \c off setting does no checksums,
the \c uncompressed setting only checksums blocks that are not compressed, and the
\c unencrypted setting only checksums blocks that are not encrypted. See @ref
tune_checksum for more information.''',
choices=['on', 'off', 'uncompressed', 'unencrypted']),
Config('dictionary', '0', r'''
the maximum number of unique values remembered in the Btree row-store leaf page value
dictionary; see @ref file_formats_compression for more information''',
min='0'),
Config('encryption', '', r'''
configure an encryptor for file blocks. When a table is created, its encryptor is not
implicitly used for any related indices or column groups''',
type='category', subconfig=[
Config('name', 'none', r'''
Permitted values are \c "none" or a custom encryption engine name created with
WT_CONNECTION::add_encryptor. See @ref encryption for more information'''),
Config('keyid', '', r'''
An identifier that identifies a unique instance of the encryptor. It is stored in
clear text, and thus is available when the WiredTiger database is reopened. On the
first use of a (name, keyid) combination, the WT_ENCRYPTOR::customize function is
called with the keyid as an argument'''),
]),
Config('format', 'btree', r'''
the file format''',
choices=['btree']),
Config('huffman_key', 'none', r'''
This option is no longer supported, retained for backward compatibility'''),
Config('huffman_value', 'none', r'''
configure Huffman encoding for values. Permitted values are \c "none", \c "english",
\c "utf8<file>" or \c "utf16<file>". See @ref huffman for more information'''),
Config('ignore_in_memory_cache_size', 'false', r'''
allow update and insert operations to proceed even if the cache is already at
capacity. Only valid in conjunction with in-memory databases. Should be used with caution -
this configuration allows WiredTiger to consume memory over the configured cache limit''',
type='boolean'),
Config('internal_key_truncate', 'true', r'''
configure internal key truncation, discarding unnecessary trailing bytes on internal keys
(ignored for custom collators)''',
type='boolean'),
Config('internal_page_max', '4KB', r'''
the maximum page size for internal nodes, in bytes; the size must be a multiple of the
allocation size and is significant for applications wanting to avoid excessive L2 cache
misses while searching the tree. The page maximum is the bytes of uncompressed data,
that is, the limit is applied before any block compression is done''',
min='512B', max='512MB'),
Config('internal_item_max', '0', r'''
This option is no longer supported, retained for backward compatibility''',
min=0, undoc=True),
Config('internal_key_max', '0', r'''
This option is no longer supported, retained for backward compatibility''',
min='0'),
Config('key_gap', '10', r'''
This option is no longer supported, retained for backward compatibility''',
min='0'),
Config('leaf_key_max', '0', r'''
the largest key stored in a leaf node, in bytes. If set, keys larger than the specified
size are stored as overflow items (which may require additional I/O to access).
The default value is one-tenth the size of a newly split leaf page''',
min='0'),
Config('leaf_page_max', '32KB', r'''
the maximum page size for leaf nodes, in bytes; the size must be a multiple of the
allocation size, and is significant for applications wanting to maximize sequential data
transfer from a storage device. The page maximum is the bytes of uncompressed data,
that is, the limit is applied before any block compression is done. For fixed-length
column store, the size includes only the bitmap data; pages containing timestamp
information can be larger, and the size is limited to 128KB rather than 512MB''',
min='512B', max='512MB'),
Config('leaf_value_max', '0', r'''
the largest value stored in a leaf node, in bytes. If set, values larger than the
specified size are stored as overflow items (which may require additional I/O to
access). If the size is larger than the maximum leaf page size, the page size is
temporarily ignored when large values are written. The default is one-half the size of
a newly split leaf page''',
min='0'),
Config('leaf_item_max', '0', r'''
This option is no longer supported, retained for backward compatibility''',
min=0, undoc=True),
Config('memory_page_image_max', '0', r'''
the maximum in-memory page image represented by a single storage block. Depending on
compression efficiency, compression can create storage blocks which require significant
resources to re-instantiate in the cache, penalizing the performance of future point
updates. The value limits the maximum in-memory page image a storage block will need. If
set to 0, a default of 4 times \c leaf_page_max is used''',
min='0'),
Config('memory_page_max', '5MB', r'''
the maximum size a page can grow to in memory before being reconciled to disk. The
specified size will be adjusted to a lower bound of <code>leaf_page_max</code>, and an
upper bound of <code>cache_size / 10</code>. This limit is soft - it is possible for
pages to be temporarily larger than this value. This setting is ignored for LSM trees,
see \c chunk_size''',
min='512B', max='10TB'),
Config('prefix_compression', 'false', r'''
configure prefix compression on row-store leaf pages''',
type='boolean'),
Config('prefix_compression_min', '4', r'''
minimum gain before prefix compression will be used on row-store leaf pages''',
min=0),
Config('split_deepen_min_child', '0', r'''
minimum entries in a page to consider deepening the tree. Pages will be considered for
splitting and deepening the search tree as soon as there are more than the configured
number of children''',
type='int', undoc=True),
Config('split_deepen_per_child', '0', r'''
entries allocated per child when deepening the tree''',
type='int', undoc=True),
Config('split_pct', '90', r'''
the Btree page split size as a percentage of the maximum Btree page size, that is,
when a Btree page is split, it will be split into smaller pages, where each page is
the specified percentage of the maximum Btree page size''',
min='50', max='100'),
]
# File metadata, including both configurable and non-configurable (internal)
file_meta = file_config + [
Config('checkpoint', '', r'''
the file checkpoint entries'''),
Config('checkpoint_backup_info', '', r'''
the incremental backup durable information'''),
Config('checkpoint_lsn', '', r'''
LSN of the last checkpoint'''),
Config('id', '', r'''
the file's ID number'''),
Config('readonly', 'false', r'''
the file is read-only. All methods that modify a file are disabled. See @ref
readonly for more information''',
type='boolean'),
Config('tiered_object', 'false', r'''
this file is a tiered object. When opened on its own, it is marked as readonly and may
be restricted in other ways''',
type='boolean', undoc=True),
Config('version', '(major=0,minor=0)', r'''
the file version'''),
]
lsm_meta = file_config + lsm_config + [
Config('last', '0', r'''
the last allocated chunk ID'''),
Config('chunks', '', r'''
active chunks in the LSM tree'''),
Config('old_chunks', '', r'''
obsolete chunks in the LSM tree'''),
]
tiered_meta = file_meta + tiered_config + [
Config('flush_time', '0', r'''
indicates the time this tree was flushed to shared storage or 0 if unflushed'''),
Config('flush_timestamp', '0', r'''
timestamp at which this tree was flushed to shared storage or 0 if unflushed'''),
Config('last', '0', r'''
the last allocated object ID'''),
Config('oldest', '1', r'''
the oldest allocated object ID'''),
Config('tiers', '', r'''
list of data sources to combine into a tiered storage structure''',
type='list'),
]
tier_meta = file_meta + tiered_tree_config
# Objects need to have the readonly setting set and bucket_prefix.
# The file_meta already contains those pieces.
object_meta = file_meta + [
Config('flush_time', '0', r'''
indicates the time this object was flushed to shared storage or 0 if unflushed'''),
Config('flush_timestamp', '0', r'''
timestamp at which this object was flushed to shared storage or 0 if unflushed'''),
]
table_only_config = [
Config('colgroups', '', r'''
comma-separated list of names of column groups. Each column group is stored separately,
keyed by the primary key of the table. If no column groups are specified, all columns
are stored together in a single file. All value columns in the table must appear in
at least one column group. Each column group must be created with a separate call to
WT_SESSION::create using a \c colgroup: URI''',
type='list'),
]
index_only_config = [
Config('extractor', 'none', r'''
configure a custom extractor for indices. Permitted values are \c "none" or an extractor
name created with WT_CONNECTION::add_extractor'''),
Config('immutable', 'false', r'''
configure the index to be immutable -- that is, the index is not changed by any update to
a record in the table''',
type='boolean'),
]
colgroup_meta = common_meta + source_meta
index_meta = format_meta + source_meta + index_only_config + [
Config('index_key_columns', '', r'''
number of public key columns''',
type='int', undoc=True),
]
table_meta = format_meta + table_only_config
# Connection runtime config, shared by conn.reconfigure and wiredtiger_open
connection_runtime_config = [
Config('block_cache', '', r'''
block cache configuration options''',
type='category', subconfig=[
Config('cache_on_checkpoint', 'true', r'''
cache blocks written by a checkpoint''',
type='boolean'),
Config('cache_on_writes', 'true', r'''
cache blocks as they are written (other than checkpoint blocks)''',
type='boolean'),
Config('enabled', 'false', r'''
enable block cache''',
type='boolean'),
Config('blkcache_eviction_aggression', '1800', r'''
seconds an unused block remains in the cache before it is evicted''',
min='1', max='7200'),
Config('full_target', '95', r'''
the fraction of the block cache that must be full before eviction will remove
unused blocks''',
min='30', max='100'),
Config('size', '0', r'''
maximum memory to allocate for the block cache''',
min='0', max='10TB'),
Config('hashsize', '32768', r'''
number of buckets in the hashtable that keeps track of blocks''',
min='512', max='256K'),
Config('max_percent_overhead', '10', r'''
maximum tolerated overhead expressed as the number of blocks added and removed as
percent of blocks looked up; cache population and eviction will be suppressed if
the overhead exceeds the threshold''',
min='1', max='500'),
Config('nvram_path', '', r'''
the absolute path to the file system mounted on the NVRAM device'''),
Config('percent_file_in_dram', '50', r'''
bypass cache for a file if the set percentage of the file fits in system DRAM
(as specified by block_cache.system_ram)''',
min='0', max='100'),
Config('system_ram', '0', r'''
the bytes of system DRAM available for caching filesystem blocks''',
min='0', max='1024GB'),
Config('type', '', r'''
cache location: DRAM or NVRAM'''),
]),
Config('cache_size', '100MB', r'''
maximum heap memory to allocate for the cache. A database should configure either
\c cache_size or \c shared_cache but not both''',
min='1MB', max='10TB'),
Config('cache_max_wait_ms', '0', r'''
the maximum number of milliseconds an application thread will wait for space to be
available in cache before giving up. Default will wait forever''',
min=0),
Config('cache_stuck_timeout_ms', '300000', r'''
the number of milliseconds to wait before a stuck cache times out in diagnostic mode.
Default will wait for 5 minutes, 0 will wait forever''',
min=0),
Config('cache_overhead', '8', r'''
assume the heap allocator overhead is the specified percentage, and adjust the cache
usage by that amount (for example, if there is 10GB of data in cache, a percentage of
10 means WiredTiger treats this as 11GB). This value is configurable because different
heap allocators have different overhead and different workloads will have different
heap allocation sizes and patterns, therefore applications may need to adjust this
value based on allocator choice and behavior in measured workloads''',
min='0', max='30'),
Config('checkpoint', '', r'''
periodically checkpoint the database. Enabling the checkpoint server uses a session
from the configured \c session_max''',
type='category', subconfig=[
Config('log_size', '0', r'''
wait for this amount of log record bytes to be written to the log between each
checkpoint. If non-zero, this value will use a minimum of the log file size.
A database can configure both log_size and wait to set an upper bound for checkpoints;
setting this value above 0 configures periodic checkpoints''',
min='0', max='2GB'),
Config('wait', '0', r'''
seconds to wait between each checkpoint; setting this value above 0 configures
periodic checkpoints''',
min='0', max='100000'),
]),
Config('chunk_cache', '', r'''
chunk cache configuration options''',
type='category', subconfig=[
Config('capacity', '10GB', r'''
maximum memory or storage to use for the chunk cache''',
min='0', max='100TB'),
Config('chunk_cache_evict_trigger', '90', r'''
chunk cache percent full that triggers eviction''',
min='0', max='100'),
Config('chunk_size', '1MB', r'''
size of cached chunks''',
min='512KB', max='100GB'),
Config('device_path', '', r'''
the absolute path to the file system or a block device used as cache location'''),
Config('enabled', 'false', r'''
enable chunk cache''',
type='boolean'),
Config('hashsize', '1024', r'''
number of buckets in the hashtable that keeps track of objects''',
min='64', max='1048576'),
Config('type', '', r'''
cache location: DRAM or FILE (file system or block device)'''),
]),
Config('debug_mode', '', r'''
control the settings of various extended debugging features''',
type='category', subconfig=[
Config('corruption_abort', 'true', r'''
if true and built in diagnostic mode, dump core in the case of data corruption''',
type='boolean'),
Config('checkpoint_retention', '0', r'''
adjust log removal to retain the log records of this number of checkpoints. Zero
or one means perform normal removal.''',
min='0', max='1024'),
Config('cursor_copy', 'false', r'''
if true, use the system allocator to make a copy of any data returned by a cursor
operation and return the copy instead. The copy is freed on the next cursor
operation. This allows memory sanitizers to detect inappropriate references to
memory owned by cursors.''',
type='boolean'),
Config('cursor_reposition', 'false', r'''
if true, for operations with snapshot isolation the cursor temporarily releases any page
that requires force eviction, then repositions back to the page for further operations.
A page release encourages eviction of hot or large pages, which is more likely to
succeed without a cursor keeping the page pinned.''',
type='boolean'),
Config('eviction', 'false', r'''
if true, modify internal algorithms to change skew to force history store eviction
to happen more aggressively. This includes but is not limited to not skewing newest,
not favoring leaf pages, and modifying the eviction score mechanism.''',
type='boolean'),
Config('log_retention', '0', r'''
adjust log removal to retain at least this number of log files.
(Warning: this option can remove log files required for recovery if no checkpoints
have yet been done and the number of log files exceeds the configured value. As
WiredTiger cannot detect the difference between a system that has not yet checkpointed
and one that will never checkpoint, it might discard log files before any checkpoint is
done.) Ignored if set to 0''',
min='0', max='1024'),
Config('realloc_exact', 'false', r'''
if true, reallocation of memory will only provide the exact amount requested. This
will help with spotting memory allocation issues more easily.''',
type='boolean'),
Config('realloc_malloc', 'false', r'''
if true, every realloc call will force a new memory allocation by using malloc.''',
type='boolean'),
Config('rollback_error', '0', r'''
return a WT_ROLLBACK error from a transaction operation about every Nth operation
to simulate a collision''',
min='0', max='10M'),
Config('slow_checkpoint', 'false', r'''
if true, slow down checkpoint creation by slowing down internal page processing.''',
type='boolean'),
Config('stress_skiplist', 'false', r'''
Configure various internal parameters to encourage race conditions and other issues
with internal skip lists, e.g. using a more dense representation.''',
type='boolean'),
Config('table_logging', 'false', r'''
if true, write transaction related information to the log for all operations, even
operations for tables with logging turned off. This additional logging information
is intended for debugging and is informational only, that is, it is ignored during
recovery''',
type='boolean'),
Config('update_restore_evict', 'false', r'''
if true, control all dirty page evictions through forcing update restore eviction.''',
type='boolean'),
]),
Config('error_prefix', '', r'''
prefix string for error messages'''),
Config('eviction', '', r'''
eviction configuration options''',
type='category', subconfig=[
Config('threads_max', '8', r'''
maximum number of threads WiredTiger will start to help evict pages from cache. The
number of threads started will vary depending on the current eviction load. Each
eviction worker thread uses a session from the configured session_max''',
min=1, max=20),
Config('threads_min', '1', r'''
minimum number of threads WiredTiger will start to help evict pages from
cache. The number of threads currently running will vary depending on the
current eviction load''',
min=1, max=20),
]),
Config('eviction_checkpoint_target', '1', r'''
perform eviction at the beginning of checkpoints to bring the dirty content in cache
to this level. It is a percentage of the cache size if the value is within the range of
0 to 100 or an absolute size when greater than 100. The value is not allowed to exceed
the \c cache_size. Ignored if set to zero.''',
min=0, max='10TB'),
Config('eviction_dirty_target', '5', r'''
perform eviction in worker threads when the cache contains at least this much dirty
content. It is a percentage of the cache size if the value is within the range of 1 to
100 or an absolute size when greater than 100. The value is not allowed to exceed the
\c cache_size and has to be lower than its counterpart \c eviction_dirty_trigger''',
min=1, max='10TB'),
Config('eviction_dirty_trigger', '20', r'''
trigger application threads to perform eviction when the cache contains at least this much
dirty content. It is a percentage of the cache size if the value is within the range of
1 to 100 or an absolute size when greater than 100. The value is not allowed to exceed
the \c cache_size and has to be greater than its counterpart \c eviction_dirty_target.
This setting only alters behavior if it is lower than eviction_trigger''',
min=1, max='10TB'),
Config('eviction_target', '80', r'''
perform eviction in worker threads when the cache contains at least this much content. It
is a percentage of the cache size if the value is within the range of 10 to 100 or
an absolute size when greater than 100. The value is not allowed to exceed the \c
cache_size and has to be lower than its counterpart \c eviction_trigger''',
min=10, max='10TB'),
Config('eviction_trigger', '95', r'''
trigger application threads to perform eviction when the cache contains at least this
much content. It is a percentage of the cache size if the value is within the range of
10 to 100 or an absolute size when greater than 100. The value is not allowed to exceed
the \c cache_size and has to be greater than its counterpart \c eviction_target''',
min=10, max='10TB'),
Config('eviction_updates_target', '0', r'''
perform eviction in worker threads when the cache contains at least this many bytes of
updates. It is a percentage of the cache size if the value is within the range of 0 to 100
or an absolute size when greater than 100. Calculated as half of \c eviction_dirty_target
by default. The value is not allowed to exceed the \c cache_size and has to be lower
than its counterpart \c eviction_updates_trigger''',
min=0, max='10TB'),
Config('eviction_updates_trigger', '0', r'''
trigger application threads to perform eviction when the cache contains at least this
many bytes of updates. It is a percentage of the cache size if the value is within
the range of 1 to 100 or an absolute size when greater than 100\. Calculated as half
of \c eviction_dirty_trigger by default. The value is not allowed to exceed the \c
cache_size and has to be greater than its counterpart \c eviction_updates_target. This
setting only alters behavior if it is lower than \c eviction_trigger''',
min=0, max='10TB'),
Config('extra_diagnostics', '[]', r'''
enable additional diagnostics in WiredTiger. These additional diagnostics include
diagnostic assertions that can cause WiredTiger to abort when an invalid state
is detected.
Options are given as a list, such as
<code>"extra_diagnostics=[out_of_order,visibility]"</code>.
Choosing \c all enables all assertions. When WiredTiger is compiled with
\c HAVE_DIAGNOSTIC=1 all assertions are enabled and cannot be reconfigured
''',
type='list', choices=[
"all", "checkpoint_validate", "cursor_check", "disk_validate", "eviction_check",
"generation_check", "hs_validate", "key_out_of_order", "log_validate", "prepared",
"slow_operation", "txn_visibility"]),
Config('file_manager', '', r'''
control how file handles are managed''',
type='category', subconfig=[
Config('close_handle_minimum', '250', r'''
number of handles open before the file manager will look for handles to close''',
min=0),
Config('close_idle_time', '30', r'''
amount of time in seconds a file handle needs to be idle before attempting to close
it. A setting of 0 means that idle handles are not closed''',
min=0, max=100000),
Config('close_scan_interval', '10', r'''
interval in seconds at which to check for files that are inactive and close them''',
min=1, max=100000),
]),
Config('history_store', '', r'''
history store configuration options''',
type='category', subconfig=[
Config('file_max', '0', r'''
the maximum number of bytes that WiredTiger is allowed to use for its history store
mechanism. If the history store file exceeds this size, a panic will be triggered. The
default value means that the history store file is unbounded and may use as much
space as the filesystem will accommodate. The minimum non-zero setting is 100MB.''',
# !!! Must match WT_HS_FILE_MIN
min='0')
]),
Config('io_capacity', '', r'''
control how many bytes per second are written and read. Exceeding the capacity results
in throttling.''',
type='category', subconfig=[
Config('total', '0', r'''
number of bytes per second available to all subsystems in total. When set,
decisions about what subsystems are throttled, and in what proportion, are made
internally. The minimum non-zero setting is 1MB.''',
min='0', max='1TB'),
]),
Config('json_output', '[]', r'''
enable JSON formatted messages on the event handler interface. Options are given as a
list, where each option specifies an event handler category e.g. 'error' represents
the messages from the WT_EVENT_HANDLER::handle_error method.''',
type='list', choices=['error', 'message']),
Config('lsm_manager', '', r'''
configure database wide options for LSM tree management. The LSM manager is started
automatically the first time an LSM tree is opened. The LSM manager uses a session
from the configured session_max''',
type='category', subconfig=[
Config('worker_thread_max', '4', r'''
Configure a set of threads to manage merging LSM trees in the database. Each worker
thread uses a session handle from the configured session_max''',
min='3', # !!! Must match WT_LSM_MIN_WORKERS
max='20'), # !!! Must match WT_LSM_MAX_WORKERS
Config('merge', 'true', r'''
merge LSM chunks where possible''',
type='boolean')
]),
Config('operation_timeout_ms', '0', r'''
if non-zero, a requested limit on the number of elapsed real time milliseconds
application threads will take to complete database operations. Time is measured from the
start of each WiredTiger API call. There is no guarantee any operation will not take
longer than this amount of time. If WiredTiger notices the limit has been exceeded, an
operation may return a WT_ROLLBACK error. The default of 0 is to have no limit''',
min=0),
Config('operation_tracking', '', r'''
enable tracking of performance-critical functions. See @ref operation_tracking for
more information''',
type='category', subconfig=[
Config('enabled', 'false', r'''
enable operation tracking subsystem''',
type='boolean'),
Config('path', '"."', r'''
the name of a directory into which operation tracking files are written. The
directory must already exist. If the value is not an absolute path, the path
is relative to the database home (see @ref absolute_path for more information)'''),
]),
Config('shared_cache', '', r'''
shared cache configuration options. A database should configure either a cache_size
or a shared_cache not both. Enabling a shared cache uses a session from the configured
session_max. A shared cache can not have absolute values configured for cache eviction
settings''',
type='category', subconfig=[
Config('chunk', '10MB', r'''
the granularity that a shared cache is redistributed''',
min='1MB', max='10TB'),
Config('name', 'none', r'''
the name of a cache that is shared between databases or \c "none" when no shared
cache is configured'''),
Config('quota', '0', r'''
maximum size of cache this database can be allocated from the shared cache. Defaults
to the entire shared cache size''',
type='int'),
Config('reserve', '0', r'''
amount of cache this database is guaranteed to have available from the shared
cache. This setting is per database. Defaults to the chunk size''',
type='int'),
Config('size', '500MB', r'''
maximum memory to allocate for the shared cache. Setting this will update the value
if one is already set''',
min='1MB', max='10TB')
]),
Config('statistics', 'none', r'''
Maintain database statistics, which may impact performance. Choosing "all" maintains
all statistics regardless of cost, "fast" maintains a subset of statistics that are
relatively inexpensive, "none" turns off all statistics. The "clear" configuration
resets statistics after they are gathered, where appropriate (for example, a cache size
statistic is not cleared, while the count of cursor insert operations will be cleared).
When "clear" is configured for the database, gathered statistics are reset each time a
statistics cursor is used to gather statistics, as well as each time statistics are logged
using the \c statistics_log configuration. See @ref statistics for more information''',
type='list',
choices=['all', 'cache_walk', 'fast', 'none', 'clear', 'tree_walk']),
Config('timing_stress_for_test', '', r'''
enable code that interrupts the usual timing of operations with a goal of uncovering
race conditions and unexpected blocking. This option is intended for use with internal
stress testing of WiredTiger.''',
type='list', undoc=True,
choices=[
'aggressive_sweep', 'backup_rename', 'checkpoint_evict_page', 'checkpoint_handle',
'checkpoint_slow', 'checkpoint_stop', 'compact_slow', 'evict_reposition',
'failpoint_eviction_fail_after_reconciliation',
'failpoint_history_store_delete_key_from_ts', 'history_store_checkpoint_delay',
'history_store_search', 'history_store_sweep_race', 'prepare_checkpoint_delay',
'sleep_before_read_overflow_onpage', 'split_1', 'split_2', 'split_3', 'split_4', 'split_5',
'split_6', 'split_7', 'split_8', 'tiered_flush_finish']),
Config('verbose', '[]', r'''
enable messages for various subsystems and operations. Options are given as a list,
where each message type can optionally define an associated verbosity level, such as
<code>"verbose=[evictserver,read:1,rts:0]"</code>. Verbosity levels that can be provided
include <code>0</code> (INFO) and <code>1</code> through <code>5</code>, corresponding to
(DEBUG_1) to (DEBUG_5).''',
type='list', choices=[
'api',
'backup',
'block',
'block_cache',
'checkpoint',
'checkpoint_cleanup',
'checkpoint_progress',
'chunkcache',
'compact',
'compact_progress',
'error_returns',
'evict',
'evict_stuck',
'evictserver',
'fileops',
'generation',
'handleops',
'history_store',
'history_store_activity',
'log',
'lsm',
'lsm_manager',
'metadata',
'mutex',
'out_of_order',
'overflow',
'read',
'reconcile',
'recovery',
'recovery_progress',
'rts',
'salvage',
'shared_cache',
'split',
'temporary',
'thread_group',
'tiered',
'timestamp',
'transaction',
'verify',
'version',
'write']),
]
# wiredtiger_open and WT_CONNECTION.reconfigure compatibility configurations.
compatibility_configuration_common = [
Config('release', '', r'''
compatibility release version string'''),
]
connection_reconfigure_compatibility_configuration = [
Config('compatibility', '', r'''
set compatibility version of database. Changing the compatibility version requires
that there are no active operations for the duration of the call.''',
type='category', subconfig=compatibility_configuration_common)
]
wiredtiger_open_compatibility_configuration = [
Config('compatibility', '', r'''
set compatibility version of database. Changing the compatibility version requires
that there are no active operations for the duration of the call.''',
type='category', subconfig=
compatibility_configuration_common + [
Config('require_max', '', r'''
required maximum compatibility version of existing data files. Must be greater
than or equal to any release version set in the \c release setting. Has no effect
if creating the database.'''),
Config('require_min', '', r'''
required minimum compatibility version of existing data files. Must be less than
or equal to any release version set in the \c release setting. Has no effect if
creating the database.'''),
]),
]
# wiredtiger_open and WT_CONNECTION.reconfigure log configurations.
log_configuration_common = [
Config('archive', 'true', r'''
automatically remove unneeded log files (deprecated)''',
type='boolean', undoc=True),
Config('os_cache_dirty_pct', '0', r'''
maximum dirty system buffer cache usage, as a percentage of the log's \c file_max.
If non-zero, schedule writes for dirty blocks belonging to the log in the system buffer
cache after that percentage of the log has been written into the buffer cache without
an intervening file sync.''',
min='0', max='100'),
Config('prealloc', 'true', r'''
pre-allocate log files''',
type='boolean'),
Config('remove', 'true', r'''
automatically remove unneeded log files''',
type='boolean'),
Config('zero_fill', 'false', r'''
manually write zeroes into log files''',
type='boolean')
]
connection_reconfigure_log_configuration = [
Config('log', '', r'''
enable logging. Enabling logging uses three sessions from the configured session_max''',
type='category', subconfig=log_configuration_common)
]
wiredtiger_open_log_configuration = [
Config('log', '', r'''
enable logging. Enabling logging uses three sessions from the configured session_max''',
type='category', subconfig=
log_configuration_common + [
Config('enabled', 'false', r'''
enable logging subsystem''',
type='boolean'),
Config('compressor', 'none', r'''
configure a compressor for log records. Permitted values are \c "none" or a custom
compression engine name created with WT_CONNECTION::add_compressor. If WiredTiger
has builtin support for \c "lz4", \c "snappy", \c "zlib" or \c "zstd" compression,
these names are also available. See @ref compression for more information'''),
Config('file_max', '100MB', r'''
the maximum size of log files''',
min='100KB', # !!! Must match WT_LOG_FILE_MIN
max='2GB'), # !!! Must match WT_LOG_FILE_MAX
Config('force_write_wait', '0', r'''
enable code that interrupts the usual timing of flushing the log from the internal
log server thread with a goal of uncovering race conditions. This option is intended
for use with internal stress testing of WiredTiger.''',
min='1', max='60', undoc=True),
Config('path', '"."', r'''
the name of a directory into which log files are written. The directory must already
exist. If the value is not an absolute path, the path is relative to the database
home (see @ref absolute_path for more information)'''),
Config('recover', 'on', r'''
run recovery or fail with an error if recovery needs to run after an unclean
shutdown''',
choices=['error', 'on'])
]),
]
# wiredtiger_open and WT_CONNECTION.reconfigure statistics log configurations.
statistics_log_configuration_common = [
Config('json', 'false', r'''
encode statistics in JSON format''',
type='boolean'),
Config('on_close', 'false', r'''log statistics on database close''',
type='boolean'),
Config('sources', '', r'''
if non-empty, include statistics for the list of data source URIs, if they are open at the
time of the statistics logging. The list may include URIs matching a single data source
("table:mytable"), or a URI matching all data sources of a particular type ("table:")''',
type='list'),
Config('timestamp', '"%b %d %H:%M:%S"', r'''
a timestamp prepended to each log record. May contain \c strftime conversion specifications.
When \c json is configured, defaults to \c "%Y-%m-%dT%H:%M:%S.000Z"'''),
Config('wait', '0', r'''
seconds to wait between each write of the log records; setting this value above 0
configures statistics logging''',
min='0', max='100000'),
]
connection_reconfigure_statistics_log_configuration = [
Config('statistics_log', '', r'''
log any statistics the database is configured to maintain, to a file. See @ref
statistics for more information. Enabling the statistics log server uses a session from
the configured session_max''',
type='category', subconfig=