hazelcast/hazelcast v5.1.5

on GitHub

This document lists the new features, enhancements, fixed issues and, removed or deprecated features for Hazelcast Platform 5.1.x releases. The numbers in the square brackets refer to the issues in Hazelcast's GitHub repositories.

==== 5.1.5 ====

Fixes

• Fixed a memory leak due to incomplete clean-up of backup replica sync operations. [#22767]
• Fixed an issue where NullPointerException was thrown around the CREATE JOB statement which is using Kafka Sink connector when Kafka has no records yet. Now, it produces an appropriate log message. [#22697]
• Fixed an issue where the map entries recovered from persistence were not expiring after their time-to-live durations. [#22558]
• Fixed an issue where replication over WAN was failing on the source cluster members, when there are multiple batch publishers configured in a single WAN replication. [#22500]

==== 5.1.4 ====

Enhancements

• Removed the hazelcast-jet-cdc-mysql module from the Hazelcast Platform Enterprise distributions due to licensing.
  You need to manually download this module if you have the Enterprise edition. See Capture Changes from MySQL for details. [#5312]
• The Kubernetes discovery plugin has been enhanced to support the service token expiration. The token is time-bound and its content
  is refreshed periodically in the recent Kubernetes versions. A new token reader has been implemented to get the token value periodically. [#22394]
• You can now provide comma-separated lists to give multiple labels for the Kubernetes Discovery Plugin configurations service-label-name, service-label-value, pod-label-name, and pod-label-value. [#22401]

==== 5.1.3 ====

Enhancements

• Improved connection handling. [#21641]

Fixes

• When the hazelcast.socket.buffer.direct property is enabled, clusters no longer throw exceptions in response to text protocols such as the REST API. [#21700]

==== 5.1.2 ====

Fixes

• Fixed an issue where a cluster could not be formed when security is enabled, various client permissions are set, and multiple members
  are started simultaneously. [#21510]
• Fixed an issue where a cluster was unresponsive when you perform a health check to see the members are in the safe state;
  cluster members were hanging in the REPLICA_NOT_SYNC state during such health checks. [#21208]
• Fixed an issue where the list of members in the cluster was reset to an empty list when the UUID of a cluster changes after its restart:
  this was causing startup failures since Hazelcast could not manage the events due to the empty member list after a restart. [#21178]
• Fixed an issue where the statistics like puts and removals were not increasing when these operations are executed through Transactional interface. [#21107]
• Fixed a data race in SingleProtocolEncoder; while one method of this interface is called from the input thread,
  another one is called from the output thread which was causing the race. [#20994]
• Fixed an issue where the automatic module name in hazelcast-5.x.jar could not be detected using Gradle. The reason was
  /META-INF/MANIFEST.MF not being the first or second entry in the JAR file; now this manifest file is the second entry. [#20976]

==== 5.1.1 ====

Fixes

• Fixed an issue where the partition migrations were failing when there is a single map in the cluster with Merkle trees enabled. [#20929]
• Fixed a potential deadlock during partition migrations and inability to make progress while performing graceful shutdown with persistence enabled. [#20813]

==== 5.1 ====

New Features

• Tiered Storage (BETA): Introduced the Tiered Storage feature which is an extension to Hazelcast Map
  assigning data to various types of storage media based on access patterns. It is an Hazelcast Enterprise feature
  and in BETA state. See https://docs.hazelcast.com/hazelcast/5.1/tiered-storage/overview.
• Persistence for Dynamic Configuration: You can now reload the updates made in configuration dynamically (at runtime)
  after a cluster restart. See https://docs.hazelcast.com/hazelcast/5.1/configuration/dynamic-config.

Enhancements

SQL Engine

• Added support of the following statements, operators, and functions:
  ** CREATE VIEW
  ** DROP VIEW
  ** EXPLAIN
  ** CREATE INDEX
  ** UNION
  ** UNION ALL
  ** EXISTS
  ** NOT EXISTS
  ** RIGHT JOIN
  ** Sub-queries and VALUES() execution in JOIN arguments
  ** JSON_ARRAY, JSON_OBJECT
  ** SHOW VIEWS
  ** Added support for stream aggregation (HOP and TUMBLE functions)
  [#20413], [#20342], [#20268], [#20042], [#19894], [#19850], [#19810], [#19768], [#19742], [#19729], [#19650], [#19589]

• Added JSON as a supported type for the SQL engine with functions like JSON_QUERY and JSON_VALUE;
  they return a JSON object/array and a primitive value, respectively. For an overview of JSON in SQL, see https://docs.hazelcast.com/hazelcast/5.1/sql/working-with-json. [#19303]

Distribution

• Hazelcast command line interface is now available within the Hazelcast distribution package; previously
  it needed to be installed separately. [#20062]

Cloud Discovery Plugins

• Added EndpointSlices support for the Kubernetes discovery plugin; now the EndpointSlices API usage is default,
  and the old Endpoints API is not used as a fallback if EndpointSlices are not available. [#19643]

Serialization

• Added zero-configuration support for the Java records in compact serialization. [#20724]
• Updated the names of methods and fields in the Compact Serialization API to make them
  non-Java centric. [#20257]
• Added the declarative configuration elements for Compact Serialization. [#20016]
• Introduced the FieldKind class to get the field types for portable and compact serializations;
  previously it was FieldType. [#19517]

Data Structures

• Added the support of include-value parameter for MultiMap item listeners; you can set it to true
  if you want the item event to contain the item values. [#18815]

Configuration

• Added support of the jar scheme in the JCache URI configuration. Without this support,
  when, for example, hazelcast-client.zip is provided in Spring boot application properties file and packaged
  as a boot JAR, then the Hazelcast caching provider was failing to load the file. [#20543]
• Introduced a system property for allowing you to audit that all the Hazelcast instances
  running in your environment have the instance tracking file name set correctly in the configuration.
  See the note in https://docs.hazelcast.com/hazelcast/5.1/maintain-cluster/monitoring#instance-tracking. [#19928]
• Introduced the module integrity checker configuration property. Integrity checker is a component
  which verifies that the executable contains all the required META-INF/services.
  https://github.com/hazelcast/hazelcast/pull/19971[#19971]

Logging

• Added support of customizing the logging pattern and specifying configuration files via environment
  variables. We also added support for JsonTemplateLayout for which the JSON template can be customized. [#20528]

Management Center

• Added the data-access-enabled property for the Management Center configuration on the member side.
  This allows you to enable or disable the access of Management Center to the data stored by the data structures in the cluster.
  Management Center can't access the data if at least one member has the data access disabled. Its default value is true (enabled). [#19991]
• Added the console-enabled property for the Management Center configuration on the member side.
  This allows you to disable the console feature on Management Center as it supports lots of operations and it's not subject
  to the regular access restrictions; it can read from/write to any data structure even if Management Center is restricted
  to access those via client permissions. [#19718]

Licensing

• Hazelcast now shows the possible solutions in the terminal, if the license key is not set while starting a Hazelcast Enterprise cluster. #4614
• Added the ability to persist the renewal of a Hazelcast license key when this change is made without shutting down the cluster.
  Renewing a license key has already been possible using the REST API; https://docs.hazelcast.com/hazelcast/5.1/deploy/updating-license-rest.
  This enhancement makes such license renewals to be persisted so that when a member or the cluster is restarted, you will not need to
  re-update the license. [#20446]

Other Enhancements

• Added safe-serialization property to protect MulticastDiscoveryStrategy against Java deserialization attacks. [#20728]
• Upgraded the Elasticsearch client in Hazelcast's elasticsearch-7 Jet engine connector.
  Since the client license was changed from Apache 2.0 to Elastic 2.0 in Elasticsearch 7.11, we also added the new
  https://github.com/hazelcast/hazelcast/blob/master/extensions/elasticsearch/elasticsearch-7/ELASTIC-LICENSE-README.md[license] to explain the changes. [#20427]
• Updated log4j2 dependency version to 2.17.1 in Hazelcast's root pom.xml. [#20184]
• With the introduction of Tiered Storage feature, we've improved the partition migration system to support chunked partition migrations.
  By this way, the data is sent as a stream of chunks while the partitions are migrated upon member failures or new member additions.
  Previously, at most 250 MB was recommended as the maximum size for a partition; however, if a partition is backed by Tiered Storage,
  this size may go beyond this limit and may cause out of memory errors during the migrations of partitions having large data. Chunked
  partition migrations eliminate these errors. [#20005]
• The hz-start script now accepts absolute paths when providing the Hazelcast configuration file's location. [#19908]
• JSON strings can now work with paging predicate queries. [#19880]
• You can now check if Hazelcast is started properly in the Docker environment simply by using
  a curl command, e.g., curl -f http://hazelcast-host:5701/hazelcast/health/ready. [#19664]
• Hazelcast's memcached implementation was interpreting the number values and parameters
  for incr and decr wrongly (numbers were being converted into byte arrays instead of decimals).
  This has been fixed by making these commands' implementations strictly follow the
  memcached protocol specification. [#19653]
• Since the name of Hot Restart Persistence feature changed to Persistence, the prefix for its
  metrics also has been changed from "hot-restart" to "persistence". [#19543]
• Aligned the Near Cache and query cache behaviors when the in-memory format is OBJECT:
  ** Improved query cache methods to eliminate extra deserialization
  ** Added serializeKeys flag to query cache configuration to align the behavior with Near Cache. [#20265]
• Improved the speed of connection by a member when it joins the cluster, by removing the unnecessary
  sleep statements in the code. [#18932]

Fixes

• Fixed the NullPointerException issue when running the SELECT statement
  if there is a mapping for a map with partitioned indexes. [#20601]

• Fixed an issue where a single SQL query having a mix of JSON string and HazelcastJsonValue for the INSERT statement
  was not working. [#20303]

• Fixed various issues when using hostnames in Hazelcast's network and WAN Replication configurations.
  With this fix, you can seamlessly use hostnames wherever the IP addresses of the members are used. [#20014], [#15722]

• Fixed an issue where the hazelcast.yaml file was ignored when it is the only configuration file present in the
  Hazelcast setup; during startup it was looking only for the hazelcast.xml file and producing an error message saying that
  the configuration does not exist even though there is the yaml configuration file. Now it automatically uses hazelcast.yaml
  when hazelcast.xml is not available. [#20003]

• Fixed an issue where the Hazelcast command line interfaces commands were outputting incorrect command names
  when you want to see their usages using the --help argument. For example, the command hz-start --help was outputting
  the following:

  Usage: hazelcast-start [-d]
  -d, --daemon Starts Hazelcast in daemon mode

instead of the following:

Usage: hz-start [-d]
-d, --daemon Starts Hazelcast in daemon mode

• Fixed an issue where querying a map with SELECT (SQL) was failing when the data has compact serialization
  and the cluster has more than one member (with the class not being on the classpath). [#19952]
• In Kubernetes environment, when the health check endpoint was taking too long to respond, the
  Hazelcast members were considered to be unresponsive and terminated; this issue has been fixed. [#19829]
• Fixed an issue where the command hz-stop --help was not displaying the help but executing
  the hz-stop command. [#19749]
• When you both enable the persistence and automatic removal of stale data in the configuration,
  member startup failures were occurring. This has been fixed by adding the auto-remove-stale-data
  element to the configuration schema. [#19683]
• Fixed an issue where the totalPublishes statistics for the Reliable Topic data structure
  were always generated as 0. [#19642]
• Fixed an issue where some Spring XML configuration elements having values as property placeholders
  were not working when Hazelcast is upgraded to a newer version. [#19629]
• Fixed an issue where the totalPublishes statistics for the Reliable Topic data structure
  were always generated as 0. [#19555]
• Fixed an issue where the serialization was failing when the object has enum fields, or it is an enum itself. [#19314]

Removed/Deprecated Features

• Removed the elasticsearch-5 module from Hazelcast distributions since the version 5.0 Elasticsearch has passed its end of life date; see its https://www.elastic.co/guide/en/elasticsearch/reference/5.0/release-notes-5.0.0.html[release notes]. [#20458]
• Deprecated the log(LogEvent logEvent) method in the ILogger class (com.hazelcast.logging.ILogger).

Contributors

We would like to thank the contributors from our open source community
who worked on this release:

• Lenny Primak
• Chelsea31
• Tomasz Gaweda
• Katha Patel