VVR 8.0.5 — the new engine version of Ververica Runtime (VVR) for Realtime Compute for Apache Flink — was released on January 4, 2024. This release is an enterprise-level Flink engine built on Apache Flink 1.17.2 and includes connector updates, performance improvements, and bug fixes.
This version is rolling out as a canary release. If new features are not yet available in your fully managed Flink environment, the upgrade has not reached your account. To request an early upgrade, submit a ticket. To track rollout progress, check the latest announcement on the Realtime Compute for Apache Flink console homepage.
Overview
VVR 8.0.5 delivers updates across multiple connectors and resolves defects inherited from Apache Flink 1.17.2 — including fixes from the Apache Flink community — along with engine-specific stability improvements.
After the canary release completes, upgrade your deployments to this version. For instructions, see Upgrade the engine version of deployments.
Features
| Feature | What changed | References |
|---|---|---|
| MySQL CDC: specify start offset or timestamp | Reading from a MySQL Change Data Capture (CDC) source table now supports a start offset or timestamp, directly locating the read position and increasing data reading speed. | MySQL connector |
| MongoDB connector: schema sync and auto column addition for source tables | When the MongoDB connector is used for a source table, the CREATE TABLE AS or CREATE DATABASE AS statement synchronizes schema changes from the source table to the result table and automatically adds columns. | CREATE TABLE AS statement, CREATE DATABASE AS statement |
| MongoDB connector: dimension table support | The MongoDB connector can now be used for dimension tables, enabling join queries. | MongoDB connector (public preview) |
| MongoDB catalog support | Register metadata using a MongoDB catalog to infer the schema of a collection automatically, eliminating the need to write DDL statements to create a MongoDB source table in SQL. | Manage MongoDB catalogs (public preview) |
| JDBC connector: JSONB and UUID type support | The JDBC connector now supports data of the JSONB and UUID extension types. | JDBC connector |
| Hologres connector: SSL-encrypted transmission | Data reading and writing now uses SSL-encrypted transmission for more secure communication. | Hologres connector |
| Hologres connector: bulkload mode | Writing data in bulkload mode significantly reduces the workload of writing to a Hologres result table. | Hologres connector |
| Hologres connector: deduplication during batch processing | When sdkMode is set to jdbc or jdbc_fixed, deduplication is available during batch data processing, improving data consistency and database performance while reducing network overhead. | Hologres connector |
Hologres connector: jdbcBinlogSlotName parameter removed | In Hologres V2.1 or later, the jdbcBinlogSlotName parameter no longer needs to be configured. This parameter specified the slot name of a binary log source table in JDBC mode. | Hologres connector, Use the Hologres connector to consume Hologres data in real time |
| New PolarDB for Oracle 1.0 connector | The PolarDB for Oracle 1.0 connector is now available. Use it to write data to PolarDB for PostgreSQL (Compatible with Oracle). | PolarDB for Oracle 1.0 connector |
| StarRocks connector: JSON type support | StarRocks result tables now support data of the JSON type. | None |
ApsaraMQ for RocketMQ: partitionField parameter | Configure the partitionField parameter on an ApsaraMQ for RocketMQ result table to specify a field as the partition key column. Data with the same hash value is written to the same partition. | ApsaraMQ for RocketMQ connector |
| Elasticsearch connector: V8.X support | The Elasticsearch connector now supports writing to Elasticsearch result tables of V8.X. | Elasticsearch connector |
Elasticsearch connector: ignoreKeywordSuffix parameter | When the Elasticsearch connector is used for a dimension table, set ignoreKeywordSuffix to true to ignore the .keyword suffix for fields of the TEXT data type. | Elasticsearch connector |
| MySQL connector: null value handling | Null values in data written to a MySQL result table during updates can now be ignored. | MySQL connector |