Apache Flink 1.4.2 发布，通用数据处理平台

Apache Flink 1.4 系列的第二个 bug 修复版本已发布。此版本包含 10 多个修补程序和对 Flink 1.4.1 的小改进。官方强烈建议所有用户升级到 Flink 1.4.2。

下面的内容包含了所有修补程序的详细列表：

Sub-task

• [FLINK-6321] - RocksDB state backend Checkpointing is not working with KeyedCEP.
• [FLINK-7756] - RocksDB state backend Checkpointing (Async and Incremental) is not working with CEP.

Bug

• [FLINK-8423] - OperatorChain#pushToOperator catch block may fail with NPE
• [FLINK-8451] - CaseClassSerializer is not backwards compatible in 1.4
• [FLINK-8520] - CassandraConnectorITCase.testCassandraTableSink unstable on Travis
• [FLINK-8621] - PrometheusReporterTest.endpointIsUnavailableAfterReporterIsClosed unstable on Travis
• [FLINK-8692] - Mistake in MyMapFunction code snippet
• [FLINK-8735] - Add savepoint migration ITCase that covers operator state
• [FLINK-8741] - KafkaFetcher09/010/011 uses wrong user code classloader
• [FLINK-8772] - FlinkKafkaConsumerBase partitions discover missing a log parameter
• [FLINK-8791] - Fix documentation on how to link dependencies
• [FLINK-8798] - Make commons-logging a parent-first pattern
• [FLINK-8849] - Wrong link from concepts/runtime to doc on chaining

改进

• [FLINK-8202] - Update queryable section on configuration page
• [FLINK-8574] - Add timestamps to travis logging messages
• [FLINK-8576] - Log message for QueryableState loading failure too verbose
• [FLINK-8652] - Reduce log level of QueryableStateClient.getKvState() to DEBUG

Task

• [FLINK-8308] - Update yajl-ruby dependency to 1.3.1 or higher

更新

<dependency>
  <groupId>org.apache.flink</groupId>
  <artifactId>flink-java</artifactId>
  <version>1.4.2</version>
</dependency>
<dependency>
  <groupId>org.apache.flink</groupId>
  <artifactId>flink-streaming-java_2.11</artifactId>
  <version>1.4.2</version>
</dependency>
<dependency>
  <groupId>org.apache.flink</groupId>
  <artifactId>flink-clients_2.11</artifactId>
  <version>1.4.2</version>
</dependency>

二进制包下载地址：http://flink.apache.org/downloads.html

Apache Flink 是高效和分布式的通用数据处理平台。Apache Flink 声明式的数据分析开源系统，结合了分布式 MapReduce 类平台的高效，灵活的编程和扩展性。同时在并行数据库发现查询优化方案。