Issue | Description |
|---|---|
DBMI-25116 | For application ingestion and replication jobs and database ingestion and replication jobs that have an Amazon Redshift target, if a source column of a string datatype contains NULL data, the values are written as empty strings ("") to the target. (October 2025) |
Issue | Description |
|---|---|
DBMI-26804 | Application ingestion and replication jobs and database ingestion and replication jobs that have an Amazon S3 target might fail if the AWS region of the S3 bucket is AWS GovCloud (US-West). The following error message is issued: Cannot create enum from region description AWS GovCloud (US-West) value (April 2026) |
Issue | Description |
|---|---|
DBMI-29670 | For application ingestion and replication jobs and database ingestion and replication jobs that have a Databricks target and use Parquet file staging, the infa_operation_time audit column might be written with only 3 decimal places of sub-second precision (milliseconds) instead of the 6 decimal places (microseconds) available in the source journal timestamps. (July 2026) |
DBMI-28330 | For application ingestion and replication jobs and database ingestion and replication jobs that have a Databricks target and use volume staging, when the initial file upload attempt fails because of network connectivity issues and leaves partially written files, the subsequent retry attempts might fail. The failure occurs because the retry logic does not overwrite existing partial files by default, resulting in an error. The error indicates that the file already exists in staging and overwriting it is not enabled. (April 2026) |
DBMI-27152 | For application ingestion and replication tasks and database ingestion and replication tasks that have a Databricks target, empty strings ("") in a source string column are incorrectly replicated as NULL to the Databricks target. (April 2026) |
DBMI-26154 | Application ingestion and replication jobs and database ingestion and replication jobs that have a Databricks target might fail during the COPY INTO operation if a source table begins with a special character. (April 2026) |
DBMI-22028 | Application ingestion and replication jobs and database ingestion and replication jobs that have a Databricks Unity Catalog target might fail with an error related to invalid Min Workers or Max Workers values. The job directly parses the values without handling invalid or malformed input. (April 2026) |
Issue | Description |
|---|---|
DBMI-26101,DBMI-26373 | For application ingestion and replication or a database ingestion and replication tasks that use the Databricks target, the volume generated automatically as the staging environment is not deleted along with the "temp_jobid" directory when the task is in stopped and failed state. (October 2025) |
Issue | Description |
|---|---|
DBMI-29723 | When configuring an application ingestion and replication task or a database ingestion and replication task that has a Google BigQuery target, an option to specify a separate billing project ID at the task level is missing in the task definition. As a result, all BigQuery query costs might be billed to the project defined in the connection, making it impossible to separate FinOps costs across multiple BigQuery projects. (July 2026) |
DBMI-24787 | If you configured a proxy server but then tried to bypass it for Secure Agent communication with Google BigQuery targets by setting task_container.jvm.http.nonProxyHosts property at agent level, ingestion and replications jobs that have Google BigQuery targets might fail. The jobs fail because communication still goes through the proxy, and communications from the proxy are blocked a firewall. (October 2025) |
Issue | Description |
|---|---|
DBMI-25753 | For application ingestion and replication tasks and database ingestion and replication tasks that use a single topic for all writes to the Kafka target, the restart logic examines all Kafka topics instead of only the required topic, which might cause the task to fail if access to all topics is not allowed. (October 2025) |
Issue | Description |
|---|---|
DBMI-29281 | For application ingestion and replication tasks and database ingestion and replication tasks that have a Microsoft Azure Synapse Analytics target, when an incremental load job encounters a connection timeout while executing a CREATE EXTERNAL FILE FORMAT statement, the job might fail immediately instead of retrying the timed-out query. Connection timeout errors might not be included in the retry handling logic for Microsoft Azure Synapse Analytics target operations. (July 2026) |
DBMI-28958 | For application ingestion and replication tasks and database ingestion and replication tasks, when a connection reset occurs immediately after a DROP TABLE statement is executed against the Microsoft Azure Synapse Analytics target, the job might fail with error 2714 "There is already an object named '...' in the database" upon reconnection and retry. The retry logic might only retry the CREATE TABLE statement rather than the full DROP-then-CREATE sequence, so the create attempt might collide with the object that the earlier DROP has not yet fully propagated across Synapse compute nodes. (July 2026) |
DBMI-19405 | Tests of a Microsoft Azure Synapse Analytics Database Ingestion connection for an application ingestion and replication task or database ingestion and replication task fail with the following error if the user name specified in the connection properties includes a special character such as @ and the target database uses AAD authentication: Cannot instantiate datasource because of error: Failed to initialize pool: Cannot open server 'noconline.onmicrosoft.com' requested by the login. The login failed. ClientConnectionId: <identifier> This problem occurs when the Synapse Analytics connector package is missing some required .jar files. These .jar files have now been added in the connector package at package-AzureDWGen2MI.xxx\package\dw\thirdparty\informatica.azuredwgen2mi:
(April 2026) |
Issue | Description |
|---|---|
DBMI-25163 | For application ingestion and replication jobs and database ingestion and replication jobs that have a SQL Server target, if more than three DML operations are performed on primary key column values of the same row in a CDC cycle, the job might fail with a primary key constraint error. (October 2025) |
Issue | Description |
|---|---|
DBMI-29443 | For application ingestion and replication jobs and database ingestion and replication jobs that have an Oracle target, when a combined load job fails or completes the initial load phase, Oracle database connections might not be released, which might lead to connection pool exhaustion and ORA-00020 "maximum number of processes exceeded" errors in the Oracle alert log. Connections might persist until the Oracle database is restarted. (July 2026) |
Issue | Description |
|---|---|
DBMI-29970 | When an application ingestion and replication job or a database ingestion and replication job that has a Snowflake target encounters HTTP 503 errors during merge apply operations or while executing LIST stage queries, the job might fail instead of retrying, even when retry codes are configured via the writerCustomRetryCodes custom property. The retry logic does not cover the autoCommitOff path or the stage initialization path, leaving those failure points unprotected. (July 2026) |
DBMI-29745 | Application ingestion and replication jobs and database ingestion and replication jobs that have a Snowflake target and use the Superpipe option might intermittently fail with "JDBC driver encountered communication error. Message: HTTP status=503" errors during CDC merge operations, even when writerCustomRetryCodes=503,200015 is configured. Certain internal code paths might bypass the retry mechanism, causing job failures on transient Snowflake service unavailability. (July 2026) |
DBMI-29594 | For application ingestion and replication jobs and database ingestion and replication jobs that have a Snowflake target and use the Superpipe option, when a Snowflake Streaming channel is invalidated due to insert throttling, a deadlock in the InputHelper shutdown sequence might prevent the CDC job from stopping or restarting promptly. The job might continue reporting a "Running" status for several hours after the failure while no data is being processed. (July 2026) |
DBMI-29563 | For application ingestion and replication jobs and database ingestion and replication jobs that have a Snowflake target and use the Superpipe option, when a table-specific resync is triggered, the operation might incorrectly close all open streaming channels instead of only the channel corresponding to the target table being resynced. This might cause unnecessary channel invalidations and disrupt replication for all other tables in the job. (July 2026) |
DBMI-29448 | For application ingestion and replication and database ingestion and replication combined load jobs that have a Snowflake target and use the Superpipe option, when the Soft Delete apply mode is enabled, the initial load phase might fail with "java.lang.IndexOutOfBoundsException: Index 128 out of bounds for length 128" for tables that have exactly 128 source columns. The error might not occur when the job runs in Standard Apply mode. (July 2026) |
DBMI-28191 | For application ingestion and replication and database ingestion and replication jobs that have a Snowflake target and use the Superpipe option, when the job is restarted, the generated recovery table might contain duplicate entries which are not detected at startup. (April 2026) |
DBMI-26132 | Application ingestion and replication and database ingestion and replication combined load jobs might fail with the followng error: BacklogReader does not implement insertMarkerWithAnnotations() (October 2025) |
DBMI-26085 | If multiple application ingestion and replication tasks or database ingestion and replication tasks that have differently configured Snowflake target properties are deployed nearly simultaneously, the Data Ingestion and Replication writer might use the incorrect target properties from the other task. In this situation, the deployment might fail or unexpected behavior related to the target table definitions can occur. (October 2025) |
DBMI-25090 | After the May 2025 release, application or database ingestion and replication jobs that have a Snowflake Data Cloud target and use Authorization Code authorization might fail with the following error when the access token that's generated from user interface expires: :caused by: net.snowflake.client.jdbc.SnowflakeSQLLoggedException: JDBC driver internal error: Unsupported authenticator type: OAUTH The error occurs because of the JDBC driver upgrade in the May release. (October 2025) |