Skip to main content

​
Key Features

​
Added

  • Added support for key vaults to manage secrets outside of connection strings and scripts.
    • Added a local vault and a read-only system vault that exposes selected host environment variables.
    • Added vault references to individual connection properties using {vault:item} or {vault:vaultid/item}.
    • Added the [vault(vaultid/item)] function to reference vault items in event scripts.
  • Added support for multiple schedule rules per job and for scheduling a recurring re-sync within a job. A recurring re-sync is useful for sources that do not report deletes, such as Dynamics 365.
  • Added Sync MCP server tools for finding Sync resources, including workspaces, connections, jobs, transformations, pipelines, and their logs. Also added a check_health tool. See the Remote MCP Server documentation.
  • Added SAP Gateway as a change data capture (CDC) source.
  • Added support for capturing changes per member of a multi-member physical file in DB2 for i CDC. Select an SQL alias that points at a member to capture only that member.
  • Added global search to find connections, pipelines, jobs, transformations, and events by name, along with the GET /api.rsc/v2/search endpoint.
  • Added CDC engine endpoints to the Admin API V2 to get the status of, start, stop, and reset the engine independently of the CDC job.
  • Added the TruncateTableMode replicate option, which clears destination tables with TRUNCATE TABLE instead of DELETE. This avoids transaction log growth and command timeouts on large tables.
  • Added the Force Stop action to end a hung Replicate Agent when a graceful cancel does not complete.
  • Added the cdata.app.db.schema property to sync.properties to choose the schema that holds the application database tables.
  • Enhanced the API connector.
    • Added CSV as a format option.
    • Added the API Key and OAuthPKCE authentication schemes.

​
Changed

  • Improved the job and task status display. Job status now shows a rollup with progress details, and task run details now show the run type, last checkpoint, start position, and the number of records inserted, merged, and deleted.
  • Improved CDC event processing performance.
  • Improved initial snapshot performance for parallel partitioned reads by using INSERT instead of MERGE when writing to the destination.
  • Improved CDC resilience on low disk space. Sync now monitors disk space and automatically stops and resumes CDC engines to help prevent corrupted staging files. As a result, the stagemaxsize property is deprecated.
  • Improved memory management. Added the MemoryBudgetPercentage setting, which sets the share of memory for Sync and the Replicate Agent combined, and the optional ReplicateAgentMaxHeap setting. The default heap cap of the Sync JVM is now lower to leave room for the Replicate Agent.
  • Microsoft SQL Server CDC now collapses multiple changes to the same key within a transaction (such as a DELETE and INSERT pair) into a single net-effect row.
  • Sync sessions no longer expire while you are actively editing.
  • Improved single sign-on (SSO) just-in-time (JIT) provisioning. Users who match no mapped group are now denied access when no default role is configured.
  • Added the ability to upgrade Delta and Iceberg (LakeBridge) connectors from the UI.
  • Breaking change: The status reported for a successful pipeline changed from SUCCEED to SUCCESS to match jobs, transformations, and tasks.
  • Behavior change: DB2 for i CDC no longer trims string values. Leading and trailing spaces are now preserved as stored in the source.
  • Behavior change: Selecting the bare physical file in DB2 for i CDC now captures only the default member (the member that shares the file name).

​
Fixed

  • Fixed an issue where pipelines could get stuck in the Canceling state after cancellation.
  • Fixed an issue where Error Only alerts were not sent for pipelines.
  • Fixed an issue where the Jobs and Transformations pages could not load with a Microsoft SQL Server application database after upgrading to 26.3.
  • Fixed an issue where importing sync_status failed with a primary key violation when migrating from a Derby application database.
  • Fixed an issue where row failures from Salesforce Bulk API v1 were not recorded for reverse ETL jobs.
  • Fixed an issue where the History tab showed no records because a saved filter was silently reapplied.
  • Fixed slow loading of the Job History page when filtering, sorting, and paging through large histories.
  • Fixed an issue where the POST /api.rsc/v2/users/generate-token endpoint ignored the requested expiry time.
  • Fixed an issue where SSO could be enabled even when Save & Test reported errors.
  • Fixed an issue where connections with the same name in two workspaces were moved between workspaces after a Git pull.
  • Fixed an issue where DB2 i CDC jobs did not report the Unique Key as the PK in the UI.