forked from crate/crate
-
Notifications
You must be signed in to change notification settings - Fork 0
/
CHANGES.txt
3888 lines (2400 loc) · 119 KB
/
CHANGES.txt
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
=================
Changes for Crate
=================
Unreleased
==========
- Fix: The ``minimum_should_match`` fulltext search option was ignored
which lead to incorrect results for queries with multiple terms when
``minimum_should_match`` was explicitly set to a non-default value.
- Fixed an issue that causes the `try_cast` function to fail while casting
the WKT representation of a point with invalid coordinates to geo_point.
- All operations on closed partitions result in error now.
- Fixed an issue that could cause `COPY <table> TO DIRECTORY` statements to
fail
- BREAKING: Removed deprecated Plugin interface methods.
- Fixed an issue that might cause queries to fail with a `Same bucket of a
page set more than once` error.
- Deprecated the ``CrateClient`` class
You should start using the PostgreSQL JDBC driver instead.
- Extended the postgres wire protocol to support describe for queries with
result sets. This removes the requirement to set "prepareThreshold=0" when
using the JDBC client.
- Fixed an issue that could cause arrays with null entries to cause a NPE
during the generation of another error message.
- Fix: An issue on self joins when columns from both relations are
used in where clause.
E.g. FROM table t1, table t2 WHERE t1.col = 'foo' and t2.col = 'bar';
- Added support for SET LOCAL|SESSION statements.
- Fix: ``ANY`` operator fails in the ``WHERE`` clause if a subscript is
used as an expression.
E.g. 'value' = any(obj_array[1]['a']['b'])
- Support ``SELECT`` queries without the ``FROM`` clause. Such queries are
executed against the ``sys.cluster`` table.
- Fix: ``WHERE`` clauses with a ``NOT`` operator where the column
reference resolves to ``NULL`` now return the correct results:
``NOT NULL`` -> ``NULL`` -> no match
- The PostgreSQL wire protocol is enabled by default.
- Fix: The ``count`` aggregation function counts null values.
E.g. count(cast(null as string))
- Fixed evaluation of ``NOT NULL`` (=> ``NULL``) in ``WHERE`` clause
- Fix: Prevented usage of ``MATCH`` predicates on columns of both relations
of a join if they cannot be applied to each of them separately.
eg: FROM t1, t2 WHERE match(t1.txt, 'foo') OR match(t2.txt, 'bar');
- Fixed an issue which could cause ``EXPLAIN`` statements to fail
2016/09/23 0.56.0
=================
- BREAKING: Column names returned on queries does not include the table and schema
names anymore, i.e., ``select doc.t1.a from t1`` will return column name ``a``
instead of ``doc.t1.a``.
- BREAKING: Moved some `sys.checks` into a new `sys.node_checks` table as
those checks were actually node specific.
Added the ability to acknowledge failing checks on this new
`sys.node_checks` table.
- Updated crate-admin to ``0.20.0`` which contains following changes:
- Added a new console option to allow users to display the error trace
when an error occurs.
- Added translation for tutorial plugin. This enables support for i18n for
any admin ui plugin.
- Fix: aggregations on no-match JOIN queries incorrectly returned empty result
Now these queries fail because aggregations are not supported on joins
- Updated crash to 0.18.0 which contains following changes:
- Don't show tables from ``pg_catalog`` if ``\dt`` command is used.
- Make the ``\check`` command to display both, node and cluster checks.
Allow ``node`` and ``cluster`` arguments for the command to display
node and cluster checks, respectively.
- Added support for outer joins: LEFT, RIGHT and FULL.
- Added `percentile` aggregate function.
- Fixed an issue that could cause joins on 3+ tables to produce the wrong
result.
- Fix: Inserting in tables which have a generated column with
`current_timestamp` failed.
- Added the sys.summits table. The table contains the information about
mountains in the Alps higher than 2000 meters, which are used as node names.
- Replaced the existing list of random node names with the list of mountains
in the Alps.
- Remove limitation that prevented usage of columns from multiple tables
in the same order by expression.
eg: FROM t1, t2, t3 ORDER BY t1.x - t2.y + t3.z
- F501: Added SQL conformance info to ``information_schema.sql_features``
- Updated crate-admin to ``0.19.0`` which contains following changes:
- Display failing node checks on overview page and allow to dismiss them.
- Added German translation.
This change also enables the possibility for further i18n.
- Added plugin for unicast host discovery on Azure (ARM mode)
- Optimized sys.nodes queries
- Queries on the `sys.nodes` table will now short circuit in case some nodes
in the cluster are unresponsive. All columns of the unresponsive nodes will
contain null values except `id` and `name`.
- Added support for subselect statements that can be rewritten as a single query.
- Fetch operations are now threaded per shard instead of per node.
- Added initial experimental support for the postgres wire protocol. This is
disabled by default.
- Improved the error message of `COPY FROM` statements which fail due to
invalid entries in the files being imported.
- Implemented basic conditional functions
(``coalesce()``, ``greatest()``, ``least()``, ``nullif()``).
- Implemented NOT NULL column constraint.
- Added high/low disk watermark node checks exposed via `sys.node_checks`.
- Leverage new Lucene queries for geo point queries (``distance``, ``within``,
``match`` using ``within``). Newly created tables with geo_point columns
automatically use the new queries whereas previously created tables still
use the backwards compatible old queries.
- Implemented OPTIMIZE table statement which optimizes tables and partitions
by merging the segments of which each table and partition consists.
- Updated crash to 0.17.0 which contains following changes:
- Added support for OPTIMIZE table statement.
- All statements will respond with an affected rows count now.
- Added a configuration file that stores default values for Crash. The file is
placed inside the user's application config folder.
A path to a custom configuration file can be specified using the ``--config``
command line argument.
- Dropped Python 2.6 support
- Added support for client certificates
- It is now possible to disable autocompletion using ``--no-autocomplete``
command line argument and to toggle autocomplete using the ``\autocomplete``
crash command.
- Fixed "read file" command (``\r``) for legacy Python 2.7
- Fixed result info command output for whitespace other than space
characters in queries.
- Fixed call of `sysinfo` command when launching the shell.
- Deprecated usage of COPY TO without specifying a DIRECTORY. This will prevent a user from defining an output path
that conflicts with multiple shards.
- Optimize count aggregations by using custom data type to avoid boxing of primitives into Objects.
2016/09/19 0.55.5
=================
- Fix: Dropping a partition while inserting by primary key into same
partition should not raise an exception but returns a 0 row count instead.
- Fix: JAR conflict if HDFS plugin with Hadoop 2.x is used.
- Fix: Now it is also possible to restore a snapshot of a single partition
of a partitioned table in a non-existent table.
This requires that the snapshot has been created with version ``0.55.5``
or greater!
- Fix: Join conditions were not properly validated and an expression
with a not yet defined relation could be used.
- Fix: Setting of a default schema was not possible by using the HTTP REST
endpoint.
- Fix: Error was thrown in case of primary key bulk delete statement
with null bulk arguments.
- Fix: Return 0 rows instead of an error when executing select queries
that try to match null for primary key columns.
eg: select * from t where pk in (null);
- Fix: Allow select statements on columns of multiple nested object array
- Fix: do not allow to create tables in read-only system schemas
- Fix: `COPY FROM` statements fail if null values are inserted into partitioned
tables.
- Fix: type was not set correctly when creating a token filter
with default name and no type
- Fix: prevent creating `char filters` with missing properties
2016/08/16 0.55.4
=================
- Fixed a NPE in ordered select statements.
- Fix: Inserting a duplicate primary key lead to an update on
replicas even though it resulted in a DuplikateKeyException
(introduced with version 0.55.0).
- Fix: Querying a partitioned table by a single primary key raised an
exception instead of responding whith a zero row count.
- Improved the performance of sorted join queries
2016/08/05 0.55.3
=================
- Fixed a race condition which could cause select statements to get stuck
- Fixed an issue which prevents java client from running explain
statements.
- Fixed an issue which caused `current_timestamp` as part of a generated
expression to not work correctly.
This fix will only affect new tables. Tables which already contain a
generated column which uses `current_timestamp` will need to be recreated.
- Improved error handling when grouping by invalid data types.
- Fixed an issue when ordering by a function on a joined query.
- Fixed an issue which could lead to wrong Error messages due to a
race condition.
- Fix: Arrays of geo_point can be used in insert/update
statements as literals as well as bind arguments.
- Fixed a read isolation issue which could cause joins involving
system and blob tables to return inconsistent results.
- On Hunspell dictionary configuration, the parameter
indices.analysis.hunspell.dictionary.location has been removed and <path.conf>/hunspell
is always used.
- Fixed a bug that prevented join and sort on blob tables.
- Adapted behaviour of stats.enabled, stats.jobs_log_size and stats.operations_log_size settings
to match behaviour of other cluster wide settings like for example minimum_master_nodes.
- Fix: Do not fetch table stats when node is not fully available yet.
This could have led to NullPointerExceptions during long recovery
process on node start.
- Fixed an issue which prevented running join queries on empty partitioned
tables.
- Fix: the ``hostname`` column of the ``sys.nodes`` table returns a hostname
instead of an IP address.
- Fixed an issue that caused select statements to get stuck if used as part of
a bulk request. Now a proper error is raised.
- Updated Elasticsearch to 2.3.4
- Fix: the ``within`` scalar function now allows querying geo_shapes within
geo_shapes.
2016/07/07 0.55.2
=================
- Fix: selecting ``os['timestamp']`` multiple times in same ``sys.nodes``
statement could result in different values.
- Fix: selecting ``os``, ``network``, ``fs`` or ``process`` from ``sys.nodes``
in very quick succession or multiple times in same select statement
could lead to wrong results for percentual values from the underlying sigar
library.
- Fixed performance issues with arithmetic operators and functions.
- Fixed duplicate key error handling for multiple row inserts. Instead of
responding with an error the number of affected row is adjusted.
- Fixed a NPE that occurred when inserting null values to a generated column.
2016/06/28 0.55.1
=================
- Fix: Failures on large bulk operations could lead to an
stackoverflow error and unresponsive cluster.
- Updated crash to 0.16.2 which contains following changes:
- Fix: crash_standalone could crash because it contained non-ascii character
but no encoding was declared
- Fixed "read file" command (``\r``) for legacy Python 2.7
- Fixed result info command output for whitespace other than space
characters in queries.
- Fixed call of `sysinfo` command when launching the shell. stackoverflow error and unresponsive cluster.
- Fix: Ordered queries with limit 0 failed
- Fixed error handling while preparing execution plans.
- Fix: Number of bulk results were equal to the total number of items
on bulk inserts/updates, ignoring possible multi-value bulk inserts/updates.
- Fixed path to sigar lib in crate.bat for Windows
- Fixed proper validation of the ``blocks.*`` table settings at all
SQL operations.
- Fix: Allow select statements with more than one level nested arrays in table columns.
eg: SELECT categories['items']['id'] FROM assets
- Fixed an edge case that could cause sorted join queries to fail.
- Fix: Allow concurrent deletion of partition without raising an exception.
- Throw a more meaningful error if an attempt is made to order by an invalid field in a query with aggregations.
- Fixed cpu used bug. It was returning -1, even if cpu info loaded from sigar.
- Update crate-admin to 0.18.1 which contains following changes:
- Resolve path to ``/_sql`` endpoint correctly when Crate is served at a
different location than ``/``
- Note that 0.55.0 didn't include 0.18 but 0.17, although it was mentioned
in the changelog. For a list of changes from 0.17 to 0.18 see the 0.55.0
release section.
- Fix in Cloud AWS plugin: Use the current region of the instance if no EC2
region is specified in the configuration.
- Fixed a validation issue that allowed setting a primary key column value to
null
2016/05/30 0.55.0
=================
- BREAKING: Units are now required for time or byte based settings in the
`crate.yml` configuration file. Nodes won't start otherwise.
- BREAKING: The `gateway.local.sync` setting has been renamed to
`translog.sync_interval`.
- BREAKING: It is no longer possible to specify the location of the config
file with the `CONF_FILE` environment variable, or the `-Des.config`,
`-Des.default.config`, or `-Delasticsearch.config` parameters.
The config file must now be named `crate.yml` and must be located in the
default `config/` directory, in the directory specified in the `CONF_DIR`
environmental variable, or the one specified in the `-Des.path.home`
parameter.
- BREAKING: Crate will die now on all plugin loading errors.
- BREAKING: Column names may not contain the `.` character anymore.
- BREAKING: Tables are not refreshed automatically anymore before and after
a `DELETE-BY-QUERY` operation is applied.
- Upgraded Elasticsearch to 2.3.3
- Implemented trigonometric scalar functions
(sin(), cos(), tan(), asin(), acos() & atan()).
- Updated crate-admin to 0.18.0 which contains following changes:
- Cluster check include now clickable links
- Load additional links in menu dynamically from news feed
- Fixed an issue which caused rendering errors in the node list if the cluster
contains multiple nodes with the same name
note: changes the url to the detail view of single nodes
- Added UTM tokens to links that link to the Crate.IO blog
- Updated news feed URL
- Label the master node in the cluster node list
- Fixed a bug which prevented complete deletion of tables in custom
schemas.
- Improved the `KILL` behaviour for `COPY`, `UPDATE` and `DELETE` statements.
- Fixed a `ClassCastException` that occurred if the left side of a `IN`
predicate was a function.
- Added links to Cluster check documentation.
- Removed the `creationTime` property from `SQL(Bulk)Request` and changed the
`duration` property on `SQL(Bulk)Response` to return the server-side
duration instead of the full round-trip duration.
- The default number of shards is calculated dynamically upon the table
creation.
- Deprecated camelCase names on `CREATE ANALYZER`.
- `+HeapDumpOnOutOfMemoryError` is no longer set by default.
It can still be set using the JAVA_OPTS environment variable.
- Made sure orphaned contexts are cleaned up eventually after a coordinator
node dies.
- Load ``crate.yml`` explicitly and do not allow any other extension than
``yml`` for the configuration file.
- Restrict creation of tables having the primary key constraint within
an array column type or its children.
- Renamed all scalar function names to lowercase (CamelCase names
were transformed to lowercase using `_` as word boundary)
- Fixed a NPE at the `substr` scalar function when using column
references as from/to arguments containing null values.
- NOTE: Upgrading to this Crate version is only supported from >= 0.52
- Removed the `bulk.partition_creation_timeout` setting.
- Removed internal migration code that ensured schema backward compatibility
to Crate versions <= 0.45
- Improved the speed of global aggregation queries which involve more than 1
shard per node.
- The log level of loggers can now be set at runtime via the ``SET`` statement.
- Improved validation of geo point data type; now it is not possible any more
to insert invalid latitude or longitude in array literals.
- Nodes can be configured now to allow only SQL statements resulting
in read operations using new setting ``node.sql.read_only``.
- Improved bulk insert/update and copy-from performance when replicas
are configured.
- Fixed planner issues that could cause INSERT INTO with a JOIN in the
sub-query to fail or get stuck.
- Removed the ``jobs.keep_alive_timeout`` setting and all related logic.
This means it is no longer possible to enable automatic job termination.
- Added a ``node_filters`` option to ``COPY FROM``.
- Add support for ``COPY FROM`` to ``EXPLAIN``.
- The ``COPY FROM`` statement now also accepts an array of strings as URI.
- Added support for schemes supported by the used JVM's `URL` implementation
for ``COPY FROM``. These usually include``http``, ``https``, ``ftp`` and
``jar``.
- Added support for the ``unnest`` table function.
- Implemented cluster check for installed java version.
- removed wrong behaving CrateClient feature which allowed to load
client settings from a configuration file.
- Crate will now recognize settings defined by crate plugins.
- Implemented the EXPLAIN command for SELECT statements.
- Removed the limitation that the WHERE clause couldn't contain
certain expressions in `DELETE` statements
- Exposed setting the default schema via HTTP header.
- Added `lower`` and `upper`` string functions.
2016/05/12 0.54.9
=================
- Improved retry logic of bulk inserts/updates in order to prevent stack
overflows caused by recursive retry calls that could lead to hanging updates
and imports using ``copy from``.
- Fixed a race condition that could lead to relative write operations (`x = x
+ 1`) to be executed twice.
- In certain cases settings provided via ``-Des.`` system properties could
be overridden by the ``crate.yml`` provided in ``$CRATE_HOME/config``
- Fixed an issue that prevented generated columns to work correctly
with arrays.
- Restrict creation of tables having the primary key constraint within
an array column type or its children.
with arrays.
- Fixed an issue that could lead to the `internal_action_long_timeout` and
`recovery_activity_timeout` settings to be reset to a wrong default value on
a setting update
- Fixed an issue that could lead to stuck queries if a node dropped out of the
cluster.
- Fix: Changing the column policy on a partitioned table throws an
exception if more than 1 partition exists.
- Fix: Altering a setting on an empty partitioned table is changing
the setting on all exising tables.
- Fix: Throw proper error if a column is used twice in primary key constraint.
- Fix: Improved error handling to make sure JOIN queries don't get stuck on
shard failures.
- Fix bug which could cause the MATCH predicate in WHERE clause for
partitioned tables to fail. This happens in cases when a column ident of the
MATCH predicate is not a primary key column.
- Fix: While restoring a snapshot with multiple indices inside,
selecting concrete indices can result in a wrong index being restored.
- Fix: Insert by query into a table with nested primary keys or
nested partitioned by columns defined did not work.
2016/04/08 0.54.8
=================
- Updated crate/elasticsearch to fix an issue where the `minAvailability`
setting was not read from config file
- Fixed an issue that caused `CREATE TABLE` statements to ignore PRIMARY KEY
constraints on nested columns
- Fixed an issue that caused queries to get stuck if an invalid regex is used.
- Fixed an issue that caused INSERT-FROM-QUERY statements to fail if the
target table contained a generated column as primary key
- Fixed an issue that caused delete queries to fail if they contain
a expression on the left side of a comparison, e.g. 2 + 1 > x
- Fix: ``path.conf`` setting was ignored when set as command line argument
- Fixed error handling of concurrent update/insert and delete
statements on partitioned tables.
- Improved the error message of ON DUPLICATE KEY statements that attempt to
update null objects using a subscript notation.
- Fixed issue that caused certain operations not showing up at all
in operations_log while other marked as finished even before they actually
finished
- Fixed an issue that caused `timestamp_column = <exact_value>` not to match
in DELETE queries
- Added support for set to array and array to set conversions
- Improved the error message of insert into target/source column size
miss-match
- Create Table statements now validate that the PRIMARY KEY constraint doesn't
contain columns with unsupported data types
- Fixed an issue that could cause INSERT INTO with an INNER JOIN in the
sub-query to fail.
- Avoid running KeepAliveTimers every ms if ``jobs.keep_alive_timeout`` is set
to zero
2016/03/11 0.54.7
=================
- Clarified error messages on unsupported join operations
- Removed timestamp type guessing within objects and within the PARTITION
clause of the RESTORE SNAPSHOT statement.
- Made equality operator for ``object`` types work correctly for different
inner numeric values.
- Fix: Querying system tables could fail if connected to a client
node and cluster contains another client node.
- Upgraded the included groovy dependency from 2.4.0 to 2.4.4 for security
fixes
- Conversion from (double) arrays to geo_point is now more lenient and allows
integer values inside the array. This was done since some encoders tend to
convert float/double values to integers.
2016/02/25 0.54.6
=================
- Fixed a NPE that occurred if a table was created using the ES API with a
mapping that uses ``copy_to`` but does not define ``_meta.indices``.
- Fixed a NPE that could occur if a system table was joined with a regular
user table.
- Fixed an issue that could cause queries with ORDER BY on partitioned tables
to get stuck if multiple shards were on the same node and a single shard
couldn't produce the whole result.
2016/02/15 0.54.5
=================
- Updated crate-admin to 0.17.1 which contains following changes:
- Fix: updated BlenderPro webfont
This fixes CSS glitches in Firefox 44 which misinterpreted height of webfont.
- Fix: The recent update of the shard calculation for relocating shards caused
problems with partitioned tables showing incorrect table health.
- Fix: Reading an undefined property while cluster is offline caused
Javascript error.
- Fix: Rename cluster check indicator on statusbar to ``Checks``.
- Improved display of recovery percentage in table view
- Fix: relocating shards where previously displayed as "underreplicated" shards
causing a yellow cluster state while relocating was in progress
- Fix: change in calculation of recovery percentage broke calculations for
partitioned tables causing table partitions not showing up
- Fix: now also include unassigned shards (as 0% recovered) in calculation of
recovery percentage which is displayed per table.
Previously it could show '100% recovery done' even if there were still
unassigned shards.
- Updated crash to 0.16.0 which contains following changes:
- The order of the keys / columns of the ``json`` output format is now
deterministic and reflects the order of the columns in the executed query.
- `Ctrl-c` no longer results in a KeyboardInterrupt exception if invoked while
a query is being executed.
- Added initial auto-completion support for the internal ``\format`` and
``\r`` commands.
- Added a new ``\r`` command which can be used to read and execute statements
from a file.
- Fixed an error that occurred under python 3 if one of the built-in commands
that require an argument were called without argument.
- Fixed a NullPointerException that occured if `= ANY ()` was used with a
literal array containing null values.
- Fixed an issue that caused array-literals with null items to result in an
error.
- Added `routing_state` column to `sys.shards` table in order to fix the
underreplicated shards calculation in admin ui
- Fix: exporting a table with a generated partitioned column using `copy to`
caused an Exception
2016/01/22 0.54.4
=================
- Job prepare phases now run on their own threads in order to prevent
deadlocks in netty event loop.
- Improved reliability of queries on system tables even if cluster is
overloaded and/or doc table queries are stuck.
- Fewer retries are done when a shard is not fully recovered and
queried in order to safe threads.
- Fix: The recovery expected nodes check incorrectly included client nodes
instead of just data and master nodes.
2016/01/18 0.54.3
=================
- Fix: Minimum master nodes settings check. Number of minimum master nodes is
validated against eligible master nodes instead of all nodes in the cluster.
- Fixed a race condition which could cause a join on more than 2 tables to
produce an incorrect result.
- Fix: Comparing generated partition columns using a <= or >=
operator resulted in a NPE exception
- fix: don't quote column in response columns
- Added Amazon S3 support for Snapshot and Restore
- Made repository types and their settings pluggable,
so Crate plugins may add additional repository types
2016/01/08 0.54.2
=================
- Fixed bug that prevented ``insert into`` to respect ``blocks.read_only``
setting of partitioned tables
- ``primary_key`` column has been removed from ``show columns`` result.
- fix: ``show columns`` command returns an empty list if the table contains
no constraints.
2015/12/23 0.54.1
=================
- Improved error message on type cast failure
- Fixed: Columns of type array and analyzed columns can now be selected in
statements that contain joins
2015/12/16 0.54.0
=================
- Fix: Plugin modules are now instantiated with settings
- Added support for 2+ table joins
- Fix: All insert from subquery statements are now executed without ``limit``
- Updated crate-admin to 0.17.0 which includes following changes:
- Display current shard recovery status on table info page. It shows the
percentage of done recovery.
- Updated calculation of underreplicated shards/records to work with the
change in Crate server where shards in sys.shards table have a more fine
granular state now and relocating shards are listed, too.
- Added ``show schemas``, ``show tables``, and ``show columns``
commands.
- Updated crash to 0.15.0 which contains following changes:
- Added new format ``json_row`` which prints a dataset as json row-by-row
in separate lines.
- Added the ``sysinfo`` command which queries the ``sys`` tables for
system and cluster information. The command requires Crate >= 0.54.0
- Added filtering support to ``COPY TO`` statement
- Fix: disabled type detection for timestamps
- Added support for ``Generated columns`` (Feature T175 ANSI SQL:2006)
- fixed an issue which caused dangling readers after executing a self-join
- Added ``char_length``, ``bit_length`` and ``octet_length`` scalar
functions.
- The ``jobs.keep_alive_timeout`` value can now be configured and the default has been
changed to 0 (deactivated) from ``5m``
- Allow creating the CrateClient with settings and servers at the same time
- Expose shard recovery information in sys.shards table
- Expose relocating shards in sys.shards table
Note that relocating shards are shown additionally to their already assigned
source shards.
- Removed the limitation that all selected columns must also appear in the
ORDER BY clause for joins.
- Added support for explicit INNER JOIN .. ON syntax.
- added os and jvm infos to sys.nodes table
- sort columns in ``sys`` and `` information_schema`` tables
alphabetically when selecting all columns with ``*``
2015/12/02 0.53.1
=================
- Updated Crash to 0.14.3, which includes the following change:
- Python2.6/2.7 output now prints unicode strings correctly when using
csv output format
- Improved the speed of group by queries which contain tables with more
than 1 shard per node.
- Updated crate-admin to 0.16.1 which includes following changes:
- fix: column headers where not shown in result table on console page
when two or more columns with the same name were selected
- Fix: insert into a partitioned primary key column of type long
could fail under some circumstances
- Fix: missing characters when concatenating string under some
rare circumstances
- Internal optimization: Avoid writing partition values to the source.
- Updated crate certificates in the ssl truststore.
- Fixed an issue that could cause bulk update requests to fail if the
``bulkArgs`` contained only one item
2015/11/16 0.53.0
=================
- Updated crate-admin to 0.16.0 which includes following changes:
- The table list is now fully searchable and sorted by table name within
its schemas.
- fix: charts are not correctly displayed in Microsoft Edge browser.
- UDC: identify anonymous user with cluster id.
Both user id and cluster id are anonymous traits.
Note: You can disable UDC via the `udc.enabled` setting on the Crate server.
- implement ``intersects`` scalar function
- fix: properly restrict partitioned by columns to primitive types
during analysis
- added functionality to index and match geo shapes
- restrict WITHIN function to only accept geo_point (and compatible types)
as first argument
- Added snapshot and restore functionality
- Changed crate launch scripts to no longer support extending the
CRATE_CLASSPATH environment variable
- Added initial limited cross and inner join support
- Updated crate/elasticsearch for improved logging on graceful shutdown errors
- Added ``array_cat``, ``array_unique`` and ``array_difference`` scalar
functions. Also modified ``concat`` function to resolve to ``array_cat`` when
``concat`` function is called with 2 array args, so the operator ``||`` will
work with arrays.
2015/11/11 0.52.2
=================
- Updated crate-admin to 0.15.3 which includes following changes:
- fix: REPLICATED DATA at the overview page showed incorrect percentage of
of replicated data when number of undereplicated records was higher than
number of total records
- Updated crash to 0.14.2 which contains following changes:
- Fixed an assertion error that could occur under python2.7
2015/11/04 0.52.1
=================
- Fixed an issue that caused orphaned partitions with an remaining alias to
not be recognized correctly
- Fix: clean up file descriptors in case of exceptions when retrieving blobs
over the REST api
- Fixed an issue that was causing like queries to fail if using a special regex
character
- Upgraded Elasticsearch to 1.7.3
- Fixed the issue which causes overloading of the first configured node in
the cluster when multiple clients connections are established
- Fixed a possible race condition that could lead to queries getting stuck
(Only affected SELECT queries with a limit/offset > 500000)
- Fix: Crate won't start properly if a plugin loading error occurs
(e.g. through a misconfigured plugin)
- Fixed an issue that was introduced in 0.52.0 which could cause a
FetchContext to remain open longer than necessary in an error case.
This could then lead to failures if an attempt was made to move shards to
which the FetchContext hold onto.
- Updated crate-admin to 0.15.2 which includes following changes:
- fixed a bug that displays partitions that are not related to the selected
table in the table-list.
- fixed a bug where cluster shows wrong state if the number of nodes gets
smaller than the minimum master nodes.
- fixed a bug that displays a zero value as NULL value
- fixed URL to download in dropdown that notifies about old Crate version
- fixed a visual bug that let an item in the node/table list partly disappear
if scrolling gets enabled
- Updated crash to version 0.14.1 which contains the following changes:
- Floating point values are no longer rounded to 6 places under python2.7 if
output using the tabular or mixed output format.
Now the behaviour for all output formats is the same and consistent with
the python3 output format.
Python2.6 will still display rounded values.
- Fixed the issue casuing parsing error while using internal ``non-sql``
crash commands with semicolon.
2015/10/13 0.52.0
=================
- The client will now accept IPv6 literals as server addresses
- Updated crate-admin to 0.15.0 which includes following changes:
- display cluster checks on overview page
- display relative Crate process CPU usage as bar chart and show number of
cores per node
- Allow quoted subscript expressions in SQL statement. This will allow
supporting various client applications on Windows that use ODBC.
- Added ``try_cast`` scalar function.
- Added cluster checks which are exposed via the ``sys.checks`` table.
- Added ``os_info`` column to ``sys.nodes`` table.
Currently the column only lists the number of available processors in the JVM
(which is usually equal to the number of cores of the CPU).
- Upgraded Elasticsearch to 1.7.2
This update changes the default value for delayed allocation of unassigned
shards (see: `https://crate.io/docs/en/latest/configuration.html#unassigned`_).
Also updated the upstream of EC2 discovery module
2015/10/09 0.51.6
=================
- Fix check if setting is settable/resettable on runtime for nested settings
- Exposed missing cluster gateway settings
- Throw an exception when someone attempts to change non-runtime changeable
settings
- Fixed a race condition that could cause COPY FROM operations to get stuck if
shards were to be relocated.
- Fixed a race condition that could cause the re-creation of an already
existing table partition and therefore potential dataloss when creating
a lot of partitions simultaniously using COPY FROM or bulk inserts.
2015/10/07 0.51.5
=================
- Updated crate-admin to 0.14.6 which includes following changes:
- fixed a visual bug that let an item in the node/table list partly disappear
if scrolling gets enabled
- fixed URL to download in dropdown that notifies about old Crate version
- fixed a bug that displays a zero value as NULL value
- Fixed a bug in SHOW CREATE TABLE statement that leads to a wrong table