IBM Big Replicate V2.14

IBM Big Replicate V2.14 and add-ons expand feature set and compatibility for selected offerings, further simplifying data replication and data migration of critical business data and applications

Big Replicate provides replication technology that gives consistent data across all business application environments, regardless of geographic location, data platform architecture, or cloud storage provider.

Purpose-built to maintain data consistency across platforms and locations, and for changing data at petabyte scale, it enables deployment of clusters on any combination of Hadoop distributions, Hadoop Compatible File Systems (HCFS), and cloud object storage.

New in Big Replicate V2.14
The following offerings are enhanced with support for selected Hadoop distributions that are based on Apache Hadoop 3 (CDH 6, 6.1, 6.2, and HDP 3.0 and 3.1) and additional capabilities:

• Big Replicate for Hadoop, V2.14 enables active-active data replication between Hadoop clusters, including between cross-distribution clusters.
• Big Replicate for Object Stores, V2.14 enables the active replication of data from Hadoop data clusters to object storage targets. It enables Hadoop users to replicate data and metadata efficiently, simply, and confidently.
• Big Replicate for Hive, V4.0 is an add-on to Big Replicate for Hadoop that enables the active-active replication of the Hive metastore, which stores the metadata for all tables and databases that are created in Hive. Replicate Hive metadata as it changes in any cluster, regardless of where and when changes are made. All applications have access to the same Hive tables wherever they are required.
• Big Replicate for Backup V4.0 is an add-on feature to Big Replicate. This add-on enhances environment by enabling backup and data restoration capabilities to help eliminate administrative error or malicious activity losing data in a replicated environment, retain data that have been modified, and restore data that have been modified in error

Big Replicate for Hadoop, V2.14 and Big Replicate for Object Stores, V2.14 include following new features:

• Writer selection. It is a mechanism by which the writer role that is associated with a specific Fusion server for each distributed state machine used by replication rules can be safely transferred between Fusion servers. It applies to two scenarios. The first one is a graceful handover from one node to another with minimal disruption in live operation. The second one allows for a new writer to take over responsibility for an old writer that is inoperable.
  • Visibility of writer role status is present throughout the user interface, including:
    • On the node tab
    • Writer status
    • Reassign writer
    • Revoke writer
    • Track progress of reassigns
    • Track progress of revokes
    • On the replication rule tab
    • Reassign writer
    • Reassign for a replication rule that has no writer
    • Track progress of reassigns
  • When a Fusion server stop is initiated from the user interface, writer roles for any affected replication rules are reassigned to other nodes automatically. A Fusion server that is stopped by other means does not have its writer roles reassigned automatically
• In-place update support. Big Replicate 2.14 adopts a versioned communication mechanism between clients and the server processes for cross-version compatibility. This helps minimize system disruption during upgrades that follow the V2.14.0 release.
• Path mapping. Previous versions of Big Replicate were limited to matching locations, where a replication rule assumed that the same base directory was used in each zone. The V2.14 release adds the capability to replicate content between directories that are named differently in each zone. This feature introduces changes to the API and user interface, providing the capability to select different locations in each zone for a given replication rule.
• Per-zone username translation. Username translations that apply to operations performed in other zones when replicated can be defined in the Username Translation section of the user interface. These translations include the originating zone, so that different patterns of usernames that apply in each zone can be accounted for during replication.
• Fast bypass (disable replication) option. The fast bypass mechanism uses the presence of a special bypass directory under the replicated exchange directory of the replicated file system to signal that clients should bypass coordination, as a measure to promote continued cluster availability ahead of replication data integrity. The V2.14 release adds the capability to enable, disable, and check the status of this mechanism from the user interface.

For IBM Big Replicate version mapping to WANdisco Fusion releases, see Big Replicate Mapping.

Planned availability date: September 3, 2019