MRS New Minor Version available

MRS (MapReduce Service) is used for analyzing and processing large amounts of data, especially for big data applications. It enables the creation and management of Hadoop clusters in the cloud and supports various big data components such as Hadoop, Spark, HBase, Kafka, and Storm. MRS provides a platform for running programs based on these components and allows for monitoring and management of clusters through the MapReduce Service.  

This new minor version upgrade 3.5.0-LTS includes the following new features, new changes and fixed bugs.

New Features:

• The Doris engine is now available in the SQL inspector function. That means, SQL inspector blocks and intercepts large and slow SQL statements in Doris based on preset and custom inspection policies.
• The system can detects and intercept slow query statements in Doris. That means, Doris slow query detection is added, improving the O&M monitoring capability of Doris.
• HBase supports table data synchronization using the Spark BulkLoad tool. To quickly synchronize Hive or Spark table data to HBase tables, you can use the Spark BulkLoad tool. It also allows you to import full or incremental data in ORC/PAQUET format.
• The Hive component of a normal cluster supports cascading authorization.
• Hive intercepts SQL statements that read the data written by themselves. That means, you can intercept SQL statements that read the data written by themselves in Hive.
• FlinkSQL supports the MultiJoin operator. Joining wide tables with Flink's FULL OUTER JOIN operator increases backend pressure and slows performance. You can use the MultiJoin operator to join wide tables, doubling the computing performance.
• HDFS supports asynchronous deletion. That means, the HDFS asynchronous deletion feature is used to delete large directories. Deleting blocks asynchronously with traffic control can effectively shorten the continuous lock period.
• YARN can detect large tasks to prevent the disks from being used up. That means, YARN scans local temporary files (such as shuffle files) and key HDFS directories for large Hive, HetuEngine, and Spark jobs. It reports events when jobs consume excessive storage resources (local disks or key HDFS directories).
• HetuEngine supports Oracle and GBase data sources. That means, HetuEngine allows you to configure, access, and query Oracle and GBase data sources.
• Now Keytab files can be downloaded to a remote node. You can do this now by:
  1. Log in to MRS Manager.
  2. On the System > User page, choose More > Download Authentication Credential.
  3. Select a download path and click OK. You can download Keytab files to a local path, the active OMS node, or a specified node.

Further information can be found in MapReduce Service documentation area of the Help Center.