DLM Release Notes
Also available as:
PDF

What’s New in this Release

The DLM Service 1.5.0 release includes the following new features:

Summary Category Description
External Table Replication Hive Replication of External Tables to target clusters as external Hive tables, so that external tables can be used for running non-Hive workloads. This feature is supported for replication from source cluster HDP 2.6.5/3.1 to target cluster 3.1. For more information, see External Tables Replication in the DLM Administrator's Guide.
ACID Tables Replication Hive ACID tables replication involves full ACID and micro-managed table replication. ACID tables replication is supported only when source cluster is HDP 3.x. For more information, see ACID Tables Replication in the DLM Administrator's Guide.
Trusted Proxy DLM Engine Installing DLM using trusted proxy setup requires very less changes to the existing cluster. Unlike the earlier setup, it does not require enabling SSO for Ambari, Ranger, and Atlas. Hence it is less intrusive. For more information, see Trusted Proxy for DLM in the DLM Installation and Upgrade Guide.
Statistics Replication Replication DLM now supports statistics-based replication on ACID and non-ACID tables. For more information, see Statistics replication in the DLM Administrator's Guide.
Command-Line Interface (CLI) support for DLM Replication DLM now supports policy-based CLI operations. For more information, refer to DLM operations using Command-Line Interface in the DLM Administrator's Guide.
Replication progress status Replication Replication progress can be tracked at individual policy level based on the component that gets replicated. For more information, see Tracking replication progress in the DLM Administrator's Guide.
Preserve timestamp during replication Replication By default, DLM now supports timestamp which is preserved for on-premise to on-premise replication, if both source and target HDFS support access time updates (dfs.namenode.accesstime.precision). There is no option to configure this. Currently there is no support to preserve timestamp information for On-premise to Cloud replication.