BladePipe 1.11.0: Support MongoDB two-way sync.
跳到主要内容

BladePipe 1.11.0

Release Date: October 10, 2026

Version: 1.11.0

New Pipelines​

  • TDSQL PostgreSQL > Oracle / TDSQL PostgreSQL: Full migration, incremental sync, data verification, data correction, and schema migration.
  • SQL Server / Oracle > TDSQL PostgreSQL: Full migration, incremental sync, data verification, data correction, and schema migration.
  • MongoDB - MongoDB: Two-way sync.

New Features​

  • SAP HANA > PostgreSQL / MySQL now supports incremental sync with scheduled scans.
  • SQL Server > Oracle now supports incremental sync with scheduled scans.
  • StarRocks > Kafka now supports incremental sync with scheduled scans.
  • Scheduled full migration jobs can now run once.
  • Schema verification now supports MySQL, PolarDB for MySQL, Oracle, PostgreSQL, SQL Server, StarRocks, Doris, SelectDB, and TiDB tables.
  • MySQL incremental sync now supports strong typing for better type compatibility.
  • MySQL Binlog Dump now supports SSL certificate verification. Capability negotiation and handshake timeout issues were also fixed.
  • PostgreSQL source jobs can now subscribe to multiple schemas in one database.
  • Iceberg can now use Hive Metastore to manage catalog metadata.
  • You can now split selected databases and tables into separate jobs. The original job settings are updated at the same time.
  • Job settings now use switches for Boolean values and options for enum values. This helps prevent incorrect edits.
  • For deployments outside containers, clients can now be upgraded or downgraded through a dedicated RSocket channel. SSH access and keys are not needed.
  • Schema migration now supports the COLLATE attribute on OceanBase for MySQL targets.
  • MongoDB full reads now support cursors. This prevents full sync failures or missing data when _id values have mixed types. (Parameter: cursorRead)

Improvements​

  • Improved Oracle source LogMiner reads when source lag is high. BladePipe now adjusts the number of logs registered for each redo thread. (Parameter: oraLogminerV2Enable)
  • Improved sync of Oracle source BLOB and CLOB values during transaction rollback, savepoint rollback, and long transactions. (Parameters: oraLogminerV2Enable, oraLmLobEnable)
  • Oracle source jobs now show a chart of LogMiner mining time to help track query performance.
  • Oracle target schema migration now creates a UNIQUE INDEX for a primary key to ensure uniqueness.
  • Improved filtering of Online DDL temporary tables on MySQL incremental sources. This prevents jobs with expression-based subscriptions from syncing DML from temporary tables by mistake.
  • Adjusted default fullRingBufferSize and increRingBufferSize settings for large jobs to reduce memory use.
  • Fixed schema migration failures on older OceanBase for MySQL versions when metadata queries returned duplicate primary key columns.
  • Worker alerts can now be found in alert logs.
  • Added a check that partition settings match when the same Kafka topic maps to different tables.
  • Improved the steps for resetting positions in jobs with multiple positions.
  • Updated the job details page to make key information on the right easier to see.
  • Made editing users, data sources, and job settings more consistent.

Bug Fixes​

  • Fixed schema migration issue of identity columns from Oracle 12+ to Oracle 12+.
  • Fixed lost Oracle source transactions that started before the saved position but committed after it during the first startup. (Parameters: oraLogminerV2Enable, oraV2StartupReplay)
  • Fixed shifted sync columns caused by hidden columns on Oracle sources.
  • Fixed incorrect schema and sync columns after an Oracle table was unsubscribed and subscribed again.
  • Fixed an invalid IGNORE_ROW_ON_DUPKEY_INDEX index hint when composite primary key columns and unique index columns had different orders on Oracle.
  • Fixed half-open connections that could occur when MySQL, PolarDB for MySQL, or OceanBase for MySQL sources rebuilt parsing metadata.
  • Fixed a case-sensitive MySQL issue where a missing table could use columns from another table with the same name in a different case.
  • Fixed the parsing of longitude and latitude axis order for GEOMETRY in strong typing mode on MySQL 8.0+.
  • MySQL target write templates for spatial columns now use ST_GeomFromText consistently.
  • Fixed migration of the MySQL SET type to PostgreSQL.
  • Fixed data overflow when syncing MySQL unsigned BIGINT and high-precision DECIMAL values to SQL Server.
  • Fixed duplicate columns caused by DDL metadata replay in incremental ANTLR ddlParserMode for MySQL and OceanBase for MySQL in Binlog mode.
  • Fixed incorrect ddlhistory position-range filtering during metadata recovery for MySQL, PolarDB for MySQL, PolarDB-X, and OceanBase for MySQL in Binlog mode.
  • Fixed an old timestamp being kept after a file-position reset for MySQL, PolarDB for MySQL, PolarDB-X, and OceanBase for MySQL in Binlog mode. This could make the read position differ from the metadata recovery boundary.
  • Fixed startup after a file-position reset for MySQL, PolarDB for MySQL, PolarDB-X, and OceanBase for MySQL in Binlog mode. BladePipe now reads the timestamp from the reset position, so recovery uses the correct table schema and columns.
  • Fixed backticks in metadata names for MySQL, PolarDB for MySQL, PolarDB-X, and OceanBase for MySQL in Binlog mode. They could cause snapshot checks to be skipped and an incorrect table schema to be saved.
  • Fixed topic rename mapping from Dameng and Spanner to Kafka. Messages now go to the configured target topic.
  • Fixed regular expression filters not working during Redis full migration.
  • Fixed loop-prevention tables with schema names not being found in PostgreSQL two-way sync, or being filtered out too early by a whitelist.
  • Fixed type conversion and bit-width errors when writing strongly typed BIT values to PostgreSQL targets.
  • Fixed an extra SET NOT NULL clause in DDL generated when changing a column type on PostgreSQL incremental targets.
  • Fixed time offsets and dates when syncing strongly typed PostgreSQL TIME WITH TIME ZONE values to StarRocks.
  • Fixed MongoDB document field removal ($unset) not being synced.
  • Fixed updates failing when a specified primary key was not used for MongoDB to relational database sync.
  • Fixed missing SSL connection options for ClickHouse and options that were incompatible with the driver version.
  • Fixed repeated scans in StarRocks full jobs caused by LARGEINT primary key overflow.
  • Fixed out-of-memory errors in StarRocks full scans by enabling streaming reads.
  • Fixed missing task JVM GC metrics caused by G1 GC metric names not being recognized.
  • Fixed unstable partition-key calculation for binary primary keys such as RAW. Rows with the same primary key could be sent to different execution queues.
  • Fixed child job merge errors when the main job had not finished restarting after a subscription change.
  • Fixed Sidecar alerts being sent every minute when automatic scheduling was off.
  • Fixed the page allowing a custom target table name for pipelines that do not support creating target tables.
  • Fixed virtual-column primary keys becoming null when strong typing mode was enabled for incremental sync.