ignite-3/RELEASE_NOTES.txt at main · apache/ignite-3 · GitHub

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
Apache Ignite 3 Release Notes
===========================

Apache Ignite In-Memory Distributed Database 3.1.0
-----------------------------------------------------------

Ignite:
* Fixed different bugs related to Metastorage compaction.
* Fixed bugs and numerous improvements during huge number of tables creation.
* Fixed deadlock in data streamer.
* Fixed different bugs related to B+Tree.
* Fixed index creation failure after node restart.
* Table based replication is deprecated, current implementation relies on zone replication.
* Numerous compatibility tests are appended.
* Fixed a data corruption in "aimem" and "aipersist" engines when inserting data entries with certain very specific payload sizes.
* Examples are expanded.
* Numerous logging related changes.
* Fixed data streamer with receiver tuples support.
* Throttling metrics are introduced.
* Table metrics are introduced.
* Transaction metrics are introduced.
* JVM and Os metrics are introduced.
* SQL related metrics are introduced.
* Data region metrics are introduced.
* Cluster topology metrics introduced.
* Fixed race between increasing partition generation and performing write operations to it.
* Fixed allow unaligned memory access on Aarch64.
* Support configuration deprecation\rename\deletion.
* Fixed do not log 'Not leader' as ERROR on leader miss.
* Fixed relax mapper requirements to allow unmapped columns.
* Optimize implicit RO transaction flow.
* Fixed rebalancing via raft snapshot does not work after low watermark update.
* Fixed error saving FreeList metadata causing checkpointer to crash.
* Fixed do not allow MG repair to be run when MG has majority.
* Fixed RAFT messages incompatibility.
* Fixed commit/rollback from client is executed in network thread.
* Fixed sensitive data gets logged when doing CMG disaster recovery.
* Add Cluster API.
* Fixed catalog compaction hangs and holds network thread.
* Fixed reduce message verbosity and stacktrace printing for node join issues.
* Extended public CLI API for partition restart with cleanup functionality.
* Extended public REST API for partition restart with cleanup functionality.
* Fixed heap lock manager behave incorrectly in case of overflow.
* Fixed log jvm info to the server log at startup.
* Fixed catalog compaction never happens after zone was dropped with all its tables.
* Fixed open telemetry exporter doesn't work if it was disabled and then enabled while cluster running.
* Fixed incorrect tuple comparison in B-tree prefix.
* Fixed huge log message with assignment.
* Fixed extend Compute API with supporting qualified names.
* Fixed extend IgniteCatalog API with supporting qualified names.
* Fixed add the ability to rollback any transaction by ID.
* Fixed log node configuration on start.
* Fixed it is possible to call enlist on a rollbacked transaction and cause a resource leak.

SQL:
* Partition pruning introduced.
* Partition awareness introduced.
* Fixed different bugs related to Hash Join algorithm.
* Updated Calcite version to 1.40.
* Fixed NullPointerException when building multi-bounds for a hash index.
* Fixed query in explicit RO tx fails when one replica is not available.
* Fixed result of script parsing with single statement is cached incorrectly.
* Fixed read column with valid quoted identifier using client kv view.
* Fixed failure upon creating a table with default TIMESTAMP column.
* Support CURRENT_USER function.
* Fixed Time, Timestamp, Timestamp with local time zone and insert/update with cast operation ignores target precision.
* Optimize row values conversion by sql engine to/from internal representation.
* Optimize sql row serialization.
* Migrate to calcite's UUID.
* Fixed switch sql request processing to another thread pool.
* Changed explain output.
* Fixed improve OrderedMergePublisher performance.
* Fixed query optimizer hangs while preparing UNION with huge number of tables.
* Fixed increase maximum precision for VARBINARY and VARCHAR types.
* Fixed query with ORDER BY returns unsorted result.
* Fixed freezes in MergeJoin[type=RIGHT].
* Fixed similar CAST expressions produce completely different results.
* Numerous different system views changes.
* Fixed MergeJoin is failed with IndexOutOfBoundsException.
* Fixed unable to create SCHEMA and TABLE in the same script.
* Fixed explain plan show incorrect collation for join.
* Fixed NPE when querying system.locks table.
* Fixed query plan does not use the most optimal table index.
* Fixed partition pruning, support UUID literals.
* Fixed correlated sub-query in ON classes of join cannot be rewritten.
* Fixed remove Limit if there is Sort with fetch.
* Fixed drop cached SQL plans through CLI and REST.
* Fixed Time, Timestamp, Timestamp LTZ. Cast Format. Sub-millsecond fractional format is broken.
* Fixed TIMESTAMP_WITH_LOCAL_TIME_ZONE interval arithmetic.
* Fixed batching DDL statement for scripts.
* System Views implemented.
* Fixed update cached query plans when related statistics changed.
* Fixed improve row count estimation for all join types.

C++:
* Support C++ code compilation with the C++23 standard.
* Document C++ Heartbeats feature.
* Implement heartbeats.
* ODBC 3.0: Add SSL/TLS support.
* Fixed remove unixODBC dependency.
* Implement transaction timeouts.

Python:
* Fixed DB API Driver 3: Wheels created by GitHub Actions are failing on import.
* DB API Driver 3: Add a macOS support.
* DB API Driver 3: Add SSL/TLS support.
* Fixed gradlew platforms:updateVersion is broken.

cli:
* Add sorting in the list of metrics for cli by name.
* Fixed cannot get partitions status using AI3 CLI.

Java thin client:
* Fixed switch sql request processing to another thread pool.
* Fixed client connector produces excessive logs when SSL connection fails.

Ignite .NET:
* Fixed UpsertAll failures on large collections.
* Fixed compute executor for jobs executions with different Ignite versions.
* Improve xmldoc for ColumnType.
* Fixed IndexOutOfRangeException in data streamer with receiver on empty key.
* Fixed incomplete marshaller support in Data Streamer Receiver.
* Fixed error logged on graceful disconnect.
* Fixed update NuGet README with platform compute info.
* Fixed extend .NET Client API with supporting qualified names.
* Add platform job execution public APIs.
* Platform job executor appended.
* Fixed update protocol to support platform job execution.
* Platform streamer receiver implemented.
* Implemented essential ADO.NET classes.
* Implemented APIs to cancel / release resources.

JDBC:
* Fixed JDBC doesn't know schemas except a PUBLIC one.
* Fixed improve test coverage for JDBC transactions.

Migration tools:
* Implement Migration Tools for data migration from Ignite 2 to Ignite 3.
* Fixed windows startup scripts.
* Fixed support client credentials in migration tools.
* Fixed docker version for migration tools tests.
* Fixed improve migration tools cache migartion error messages.
* Fixed upgrade Ignite 2 dependecy in the Migration Tools modules.
* Fixed support Table Schemas.

Apache Ignite In-Memory Distributed Database 3.0.0
-----------------------------------------------------------

This is the first release of Apache Ignite 3.
This release features completely reworked source code compared to Apache Ignite 2,
built on up-to-date technologies and designed from scratch to answer modern data storage and computing challenges.

The main feature set of Ignite 3 In-Memory Distributed Database includes:

* Apache Calcite-based SQL engine
* OpenAPI-compatible REST API
* Dynamic configuration with HOCON support
* Improved cluster management and control tools
* Improved transaction protocol
* Consensus-based architecture with Raft algorithm
* Streamlined table and schema management