Amazon DMS release notes - Amazon Database Migration Service
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Amazon DMS release notes

Following, you can find release notes for current and previous versions of Amazon Database Migration Service (Amazon DMS).

Amazon DMS doesn't differentiate between major and minor versions when you enable Automatic version upgrade for your replication instance. DMS automatically upgrades the replication instance's version during the maintenance window if the version is deprecated.

Note that to upgrade your replication instance's version manually (using the API or CLI) from version 3.4.x to 3.5.x, you must set the AllowMajorVersionUpgrade parameter to true. For information about the AllowMajorVersionUpgrade parameter, see ModifyReplicationInstance in the DMS API documentation.

Note

The current default engine version for Amazon DMS is 3.5.1.

The following table shows the following dates for active DMS versions:

  • The version's release date

  • The date after which you can't create new instances with the version

  • The date when DMS automatically updates instances of that version (the EOL date)

Version Release date No new instance date EOL date
3.5.3 May 17, 2024 Aug 31, 2025 Oct 31, 2025
3.5.2 Oct 29, 2023 Mar 30, 2025 Apr 29, 2025
3.5.1 Jun 30, 2023 Nov 30, 2024 Jan 30, 2025
3.4.7 May 31, 2022 Jul 30, 2024 Aug 29, 2024
3.4.6 Nov 30, 2021 May 26, 2024 June 27, 2024

Amazon Database Migration Service 3.5.3 release notes

New features in Amazon DMS 3.5.3
New feature or enhancement Description
Enhanced PostgreSQL source endpoint for Babelfish support Amazon DMS has enhanced its PostgreSQL source endpoint to support Babelfish datatypes. For more information, see Using a PostgreSQL database as an Amazon DMS source.
Support for S3 Parquet as a source Amazon DMS supports S3 Parquet as a source. For more information, see Using Amazon S3 as a source for Amazon DMS
Support for PostgreSQL 16.x Amazon DMS supports PostgreSQL version 16.x. For more information, see Using a PostgreSQL database as an Amazon DMS source and Using a PostgreSQL database as a target for Amazon Database Migration Service.

Amazon DMS version 3.5.3 includes the following resolved issues:

Issues resolved in the DMS 3.5.3 release dated 17-May-2024
Resolved Issue Description
Data Validation override function Fixed an issue for the data validation feature where DMS would not honor source filtering when a rule action was set to override-validation-function in table mappings.
MySQL source CDC errors Fixed an issue for MySQL as a source where CDC migration would fail with UTF16 encoding.
Data validation collation differences Fixed an issue for the data validation feature where DMS would not properly apply the HandleCollationDiff task setting when column filtering was used.
Data validation task hanging. Fixed an issue for the data validation feature where the DMS task would hang with a target is null" error.
Task failures in PostgreSQL to PostgreSQL replication. Fixed an issue for PostgreSQL to PostgreSQL migrations where a DMS task would fail while inserting LOB data into the target during CDC replication.
Data loss with PostgreSQL as a source Fixed an issue for PostgreSQL as a source where data loss occurring in certain edge-case scenarios.
MySQL 5.5 source CDC errors Fixed an issue for MySQL as a source where CDC replication would fail with MySQL verion 5.5.
Oracle source IOT table issue. Fixed an issue for Oracle as a source where DMS wouldn't replicate UPDATE statements correctly for IOT tables with supplemental logging enabled on all columns.
MySQL source LOBS Fixed an issue for MySQL to Redshift migrations where the DMS task would fail due to LOBs exceeding the maximum size allowed by Redshift.
Validation issue with SkipLobColumns Fixed an issue for the data validation feature where the DMS task would fail with SkipLobColumns = true when a primary key was on the last column in the source table.
Skip validation where the unique key is null Fixed an issue for the data validation feature where DMS doesn't skip rows with null unique keys properly.
Data validation improvements for Oracle COLLATE operator. Fixed an issue for the data validation feature where the validation would fail with a syntax error on Oracle versions prior to 12.2.
Error handling during full load Fixed an issue for PostgreSQL as a target where the task would hang during the full-load phase after a table error caused by invalid data.
Revalidation of CDC validation-only tasks Enhanced the data validation feature to allow revalidation on a CDC validation-only task.
S3 as a target CdcMaxBatchInterval Out of Memory issue Fixed an issue for S3 as a target where the DMS task would fail with an Out of Memory condition with CdcMaxBatchInterval set.
Oracle source driver Upgraded the DMS Oracle source driver from v12.2 to v19.18.
LOB truncation warning with SQL Server source Enhanced logging for SQL Server as a source to show warnings on LOB truncation during CDC.
Oracle binary reader enhancements Enhanced the Oracle source binary reader to support the following:
  • Big Endian platform

  • Parallel DML hints with HCC compression

  • Advanced Oracle Compressions with Golden Gate enabled

Amazon Database Migration Service 3.5.2 release notes

New features in Amazon DMS 3.5.2
New feature or enhancement Description
Redshift data validation Amazon DMS now supports validating data in Redshift targets.
Support for Microsoft SQL Server version 2022 as a source and target. Amazon DMS now supports using Microsoft SQL Server version 2022 as a source and target.
IBM Db2 LUW as a target Amazon DMS now supports IBM Db2 LUW as a target. Using Amazon DMS, you can now perform live migrations from IBM Db2 LUW to IBM Db2 LUW.

Amazon DMS version 3.5.2 includes the following resolved issues:

Issues resolved in the DMS 3.5.2 maintenance release dated 29-Apr-2024
Resolved Issue Description
IBM Db2 target segmented full load Added support for segmented full load with IBM Db2 as a target.
Amazon Timestream as a target settings Enhanced the handling of invalid timestamp settings and unsupported table operations for Timestream as a target.
Task crash with column filter Fixed an issue where a task was crashing while using a filter on a column that DMS added dynamically using a transformation rule.
Logging transaction swap file reading Added logging to show when DMS is reading from transaction swap files.
S3 as a target with CdcInsertsAndUpdates Fixed an issue for S3 as a target where a task would crash when CdcInsertsAndUpdates is true and PreserveTransactions is true.
Source filter negative operators Fixed an issue where the source filter-operator when set to a negative operator had incorrect behavior if the same column had a transformation rule defined.
Added logging for when DMS pauses reading from the source Enhanced logging to show when DMS temporarily pauses reading from the source to improve performance.
Source filters with escaped characters Fixed an issue for source filters where DMS applies escaped characters to newly created tables during CDC.
PostgreSQL as a target, incorrectly replicated deletes Fixed an issue for PostgreSQL as a target where DMS replicates deletes as null values.
Oracle as a source logging improvements Enhanced logging for Oracle as a source to remove extraneous error codes.
Improved logging of XMLTYPE limitations Improved logging for Oracle as a source to show DMS' lack of support for full LOB mode for the XMLTYPE data type.
MySQL data loss Fixed an issue for MySQL as a target where corrupted column metadata could cause task crashes or data loss.
Filter applied to a new column Fixed an issue during full load where DMS ignores a filter that a transformation rule adds to a new column.
S3 as a target: Validation issue Fixed an issue for S3 as a target where data validation would fail while migrating multiple tables with different validation partitioning definitions.
CDC-only task crash Fixed an issue for CDC-only tasks where the task would crash when TaskRecoveryTableEnabled is true.
MySQL to MariaDB incompatible collations Fixed an issue for MySQL to MariaDB migrations where DMS doesn't migrate MySQL v8 tables with tf8mb4_0900_ai_ci collation.
Task crashes with BatchApplyEnabled Fixed an issue for the Batch Apply feature where the task would fail under certain conditions.
Non UTF-8 characters in Amazon DocumentDB Added support for non UTF-8 characters for Amazon DocumentDB endpoints.
Batch Apply task crash Fixed an issue for the Batch Apply feature where the DMS task crashes while replicating large transactions.
Db2 transaction rollback handling Fixed an issue for Db2 as a source where DMS would replicate an INSERT to the target, despite being rolled back on the source.
Validation with source filters Fixed an issue where validation was not respecting source filters.

Amazon Database Migration Service 3.5.1 release notes

The following table shows the new features and enhancements introduced in Amazon Database Migration Service (Amazon DMS) version 3.5.1.

New feature or enhancement Description
Support for PostgreSQL 15.x Amazon DMS version 3.5.1 supports PostgreSQL version 15.x. For more information, see Using PostgreSQL as a source and Using PostgreSQL as a target.
Support for Amazon DocumentDB Elastic Clusters with sharded collections Amazon DMS version 3.5.1 supports Amazon DocumentDB Elastic Clusters with sharded collections. For more information, see Using Amazon DocumentDB as a target for Amazon Database Migration Service.
Redshift Serverless as a Target Support for using Amazon Redshift Serverless as a target endpoint. For more information, see Using an Amazon Redshift database as a target for Amazon Database Migration Service.
Babelfish Endpoint Settings Enhanced PostgreSQL target endpoint settings for providing Babelfish support. For more information, see Using a PostgreSQL database as a target for Amazon Database Migration Service.
Oracle Source Open Transactions Amazon DMS 3.5.1 improves the methodology of handling open transactions when starting a CDC-Only task from the Start Position for an Oracle source. For more information, see OpenTransactionWindow in the Endpoint settings when using Oracle as a source for Amazon DMS section.
Amazon Timestream as a Target Support for using Amazon Timestream as a target endpoint. For more information, see Using Amazon Timestream as a target for Amazon Database Migration Service.

Amazon DMS version 3.5.1 includes the following resolved issues:

Resolved Issue Description
Oracle as a source growing inactive sessions Fixed an issue for Oracle source where CDC-only tasks had continuously growing inactive sessions, resulting in the following exception: ORA-00020: maximum number of processes exceeded on the source database.
Replicating UPDATE changes to DocumentDB Fixed an issue for DocumentDB as a target where UPDATE statements were not properly replicated in some scenarios.
Validation-only task Improved error handling for the data validation feature to properly fail the task when data validation is disabled for validation-only tasks.
Redshift replication after connection termination Fixed an issue for Redshift target where the DMS task would not retry applying changes on the target when the target has ParallelApplyThreads set greater than zero after connection termination, which would result in data loss.
MySQL text to mediumtext replication Fixed an issue for MySQL to MySQL replication of mediumtext data types with full-LOB mode.
CDC task not replicating with rotated secret Fixed an issue for DMS tasks with BatchApplyEnabled set to true where DMS would stop replicating data after Secrets Manager rotated the password.
MongoDB/DocumentDB segmentation issue Fixed an issue for MongoDB / DocDB source where range segmentation would not work properly when the primary key column contained a large value.
Oracle data validation of unbound numeric values Fixed an issue for Oracle target where DMS would recognize a value of unbound data type NUMERIC as a STRING during data validation.
SQL Server data validation Fixed an issue for SQL Server endpoints where DMS data validation constructed an invalid SQL statement.
MongoDB Auto-Segementation Improved the functionality of automatic partitioning of data when migrating documents in parallel from MongoDB as a source.
Amazon S3 Apache Parquet format Fixed an issue so that Apache Parquet files written to S3 as a target can be viewed with Python with Apache Arrow C++.
PostgreSQL as a source DDL handling Fixed an issue with PostgreSQL source where unsupported DDL operations were not properly ignored.
PostgreSQL timestamptz data error Fixed an issue with PostgreSQL to PostgreSQL migrations where timestamp with time zone data was not migrated properly with Batch apply enabled during CDC.
Oracle to PostgreSQL validation failure Fixed an issue with Oracle to PostgreSQL migrations where data validation was failing for the NUMERIC(38,30) data type.
Oracle extended data type error Fixed an issue with Oracle source where extended varchar data type was being truncated.
Combining filter operators Fixed an issue for the column filtering functionality where the null column operator could not be combined with other types of operators.
CDC latency resulting from excessive logging. Fixed an issue with PostgreSQL source where excessive logging of pglogical plug-in warnings was causing source CDC latency.
Bi-directional replication handling of Create Table DDL Fixed an issue for bi-directional replication from PostgreSQL to PostgreSQL where the Create Table DDL change was not replicated properly.
CDC failure while using filters Fixed an issue for the filtering feature where CDC replication was failing.
Certificate authority hostname validation for Kafka endpoints Enhanced the functionality of Kafka endpoints by adding the option to disable hostname validation of the certificate authority (SslEndpointIdentificationAlgorithm).
IBM Db2 LUW validation Fixed an issue where Db2 LUW source date, timestamp, and time data types were not handled properly during data validation.
S3 validation Fixed an issue with Db2 LUW to S3 migrations where the validation fuctionality was not handling the timestamp(0) datatype properly.
DMS task restart failure Fixed an issue with PostgreSQL source where the Amazon DMS task was failing to restart and could not consume relational events when using the pglogical plugin.
SQL Server validation of HIERARCHY data type Fixed an issue for SQL Server source where validation of HIERARCHY data type would fail.
SQL Server strings with control characters Fixed an issue for SQL Server soruce where strings with control characters were not replicated correctly.
Redshift with Secrets Manager Fixed an issue with Redshift target where testing the endpoint would fail when using Secrets Manager.
MySQL ParallelLoadThreads setting inconsistency Fixed an issue with MySQL target where the ParallelLoadThreads setting was not properly retained after task settings changes.
Error with PostgreSQL to Oracle data type mapping Fixed an issue with PostgreSQL to Oracle migrations where the task would fail when replicating from data type TEXT to data type VARCHAR2(2000).
Oracle to PostgreSQL data validation Fixed an issue with Oracle to PostgreSQL migrations where data validation reported false positives when NULL characters were replicated as SPACE characters.
SQL Server source in AlwaysOn configuration Fixed an issue with SQL Server source in AlwaysOn confuguration where the Amazon DMS task would fail when the replica name didn't match the actual server name exactly.
Oracle source endpoint test failure Fixed an issue with Oracle source where the Amazon DMS endpoint connection test would fail due to insufficient priveleges while retrieving the Oracle session ID (SID).
CDC not picking up new tables Fixed an issue with CDC-only tasks where tables created on the source after the task was started were not replicated in some cases.
Open transactions in Oracle as a source Improved the methodology of handling open transactions when starting a CDC-Only task from the Start Position for an Oracle source.
Missing data issue Fixed an issue of missing data when resuming a task if it was stopped after cached changes were applied (StopTaskCachedChangesApplied option set to true). This issue could occur rarely if Amazon DMS persists cached changes to the Amazon DMS replication instance disk due to a high volume of changes on the source.
Data validation issue on extended datatype Fixed an issue for PostgreSQL to Oracle data validation where validation was failing for extended data types.
Data validation issue on inconsistent character encoding Fixed an issue for SQL Server to PostgreSQL data validation where validation was failing when character encoding was inconsistent between source and target.
Data validation issue ORA-01455 Fixed an issue where an ORA-01455 error occurs during validation when a PostgreSQL integer maps to an Oracle number(10).
SQL Server IDENTITY support Fixed an issue for SQL Server to SQL Server data replication where migrating identity columns fails when the target column had the IDENTITY property.
Character set issue with ALTER statements Fixed an issue for MySQL to MySQL replication where Amazon DMS changes the character set to UTF16 when migrating an ALTER statement during CDC.
PostgreSQL to Redshift Spatial data type support Added support for the spatial datatype when migrating from PostgreSQL to Amazon Redshift.
GZIP compression of .parquet files Fixed an issue where Amazon DMS fails to generate .parquet files with GZIP compression with S3 as a target.
MongoDB/DocDB source migration Fixed an issue where Amazon DMS doesn't migrate some of the partitions from a MongoDB source.
Table statistics issue Fixed an issue where table statistics were not shown when at least one of the tasks on the replication instance contained more than 1001 tables.
Table suspended for IBM Db2 LUW versions 10.1.0 and lower Fixed an issue for Db2 LUW source where table migration was suspended with the error TYPESTRINGUNITS is not valid when the source database version is 10.1.0 or lower.
MongoDB partitioning issue Fixed an issue for MongoDB/DocDB where one or more segments of the source partition are missing.
MongoDB partitioning issue Fix an issue where segmentation based on a column with the NumberLong() type fails due to type conversion bug.
MongoDB partitioning issue Improved autosegmentation performance for large data sets with MongoDB as a source.
MongoDB driver version Downgraded the MongoDB driver to 1.20.0 to continue support for MongoDB versions 3.6 and lower.
Amazon S3 Apache Parquet timestamp datatype Fixed an issue for Amazon S3 parquet target. Amazon DMS now sets the format parameter isAdjustedToUTC to true to match the behavior in previous versions of Amazon DMS.
Amazon Redshift as a target copy command Fixed an issue for Amazon Redshift as a target where the copy command failed for large tables when copying data from Amazon S3 to Amazon Redshift.
PostgreSQL geometry datatypes Fixed an issue for PostgreSQL to PostgreSQL migrations where migration failed on large geometry datatypes.
Oracle to PostgreSQL XML Fixed an issue where migration added an extra space on XML when replicating from Oracle to PostgreSQL.
Updating target checkpoint in supported engines Amazon DMS now updates the target checkpoint in the awsdms_txn_state table in the target database.
MongoDB/DocDB records sent to wrong collection Fixed an issue for MongoDB/DocDB where data gets sent to the wrong target collection.
Oracle source new table selection with EscapeCharacter endpoint setting Fixed an issue for Oracle source where Amazon DMS was only picking up new tables for replication when the task was stopped and resumed while the EscapeCharacter endpoint setting was set.
CDC recovery checkpoint Fixed an inconsistency in the CDC recovery checkpoint observed between the target datastore and Amazon DMS console.
CDC validation-only tasks Fixed an issue with CDC validation-only tasks where the task would not fail, even if all tables in the task encountered failures.
Validation behavior with source or target connection issues Fixed an issue with data validation where Amazon DMS would suspend tables on the source or target when the connection dropped.
Oracle to PostgreSQL data validation false positives Fixed an issue with Oracle to PostgreSQL data validation where Amazon DMS reported false positives. This is because the differences in the representation of source NULL characters on the target were not accounted for with text-based datatypes other than VARCHAR.
Oracle to PostgreSQL data truncation Fixed an issue with Oracle as a source and PostgreSQL as a target where Amazon DMS was truncating data for NVARCHAR columns with the Oracle NLS_NCHAR_CHARACTERSET setting set to AL16UTF16.
Data validation error Fixed an issue with data validation where an unable to create where filter clause error was thrown when both source filtering and an add-column transformation rule were in use.
Redshift target error handling Fixed an issue with Redshift as a target where error handling did not function as configured when the CDC task had the ParallelApplyThreads task setting set to a value greater than zero.
Oracle as a source communication failure Fixed an issue with Oracle as a source where the task remained in the RUNNING state, but was unable to migrate any data after a communication failure.
CDC table suspended with column filters Fixed an issue with Full Load + CDC tasks where a table would be suspended during the CDC phase when column filters were applied.
S3 as a target data validation failure for special characters Fixed an issue with S3 target data validation where the task would fail if the table name included a special character other than an underscore.
MongoDB source Full Load and CDC failure Fixed an issue with MongoDB as a source where a Full Load + CDC task would fail during handling of cache events while migrating a large collection.
Upgrade issue with BatchApplyEnabled set to true Fixed an issue where a task with the BatchApplyEnabled task setting set to true would fail after migrating from Amazon DMS version 3.4.6 to 3.5.1 in some cases.
SQL Server AlwaysOn source with case-sensitive collation Fixed an issue with SQL Server AlwaysOn as a source where a task would fail with case-sensitive collation.
MySQL source task hanging Fixed an issue with MySQL as a source where a task would hang instead of failing when the source was not properly configured.
S3 source full load task failure Fixed an issue with S3 as a source where a task would fail on resume after upgrading from Amazon DMS version 3.4.6 or 3.4.7 to version 3.5.1.
PostgreSQL source with CaptureDDLs set to false Fixed an issue with PostgreSQL as a source where DDLs were not properly handled with the CaptureDDLs endpoint setting set to false.
Oracle source task crash during resume Fixed an issue with Oracle as a source where a task would crash on resume due to incorrect data in the column name.
MySQL source LOB lookup failure Fixed an issue with MySQL as a source where an LOB lookup would fail when the ParallelApplyThreads task setting was set to a value greater than zero.
SQL Server source illogical LSN error Fixed an issue with SQL Server as a source where a task would fail with an illogical LSN sequencing state error error after upgrading from Amazon DMS version 3.4.7 to version 3.5.1.
PostgreSQL source with pglogical Fixed an issue with PostgreSQL as a source where a task using the pglogical plugin would fail when the task was stopped, a table was removed from selection rules, the task was resumed, and changes were made to the removed table.
Aurora MySQL incorrect recovery checkpoint. Fixed an issue for Aurora MySQL as a source where an incorrect recovery checkpoint would be saved as a result of an Aurora failover or Aurora source stop and start.
SQL Server as a source task crash. Fixed an issue for SQL Server as a source where a task would crash when SafeguardPolicy was set to RELY_ON_SQL_SERVER_REPLICATION_AGENT.
Incorrect data type casting with MySQL as a target Fixed an issue for MySQL as a target there where CDC replication would fail as a result of incorrect data type casting in the batch-apply phase.
Task failure with CaptureDDLs set to false for PostgreSQL as a source. Fixed an issue for PostgreSQL as a source where a task would fail due to a DDL being treated as a DML when the CaptureDDLs endpoint setting was set to false.
MongoDB empty collection crash Fixed an issue for MongoDB as a source where the task would crash due to an empty collection.
Redshift as a target full load task crash Fixed an issue for Redshift as a target where a task would crash during the full load phase when the recovery checkpoint control table was enabled.
S3 to S3 - no data movement. Fixed an issue for S3 to S3 replication where Amazon DMS would not replicate the data if the bucketFolder was not specified.
CDC latency with GlueCatalogGeneration set to true Fixed an issue for S3 as a target where excessive latency would occur when GlueCatalogGeneration was set to true.
Oracle as a target data truncation Fixed an issue with Oracle as a target where Amazon DMS truncates data in VARCHAR2 columns.
PostgreSQL underscore wildcard behavior Fixed an issue for PostgreSQL as a source where the behavior of the '_' wildcard in the selection rules was not working as documented.
PostgreSQL as a source empty WAL header issue. Fixed an issue for PostgreSQL as a source where the task would fail due to an empty WAL header received from the replication slot.
MySQL or MariaDB as a source with compressed binary logs Fixed an issue for MySQL and MariaDB as sources where a proper error message was not emitted when Amazon DMS detected BINLOG compression.
S3 data validation special characters Improved S3 data validation to handle special characters in primary and non-primary key columns.
Misleading task log entries with Redshift as a target. Fixed an issue for Redshift as a target where misleading entires were present in the task log reporting batch-apply statement failures on UPDATES and DELETES.
SQL Server to S3 migration task crash. Fixed an issue for SQL Server to S3 migrations where the task would crash while applying cached changes.
Missing data on batch-apply errors. Fixed an issue for the batch-apply feature where an error in applying a batch would result in missing data.

Amazon Database Migration Service 3.5.0 Beta release notes

Important

Amazon DMS 3.5.0 is a beta version of the replication instance engine. Amazon DMS supports this version the same as all previous releases. But we recommend that you test Amazon DMS 3.5.0 Beta before using it for production purposes.

The following table shows the new features and enhancements introduced in Amazon Database Migration Service (Amazon DMS) version 3.5.0 Beta.

New feature or enhancement Description
Time Travel for Oracle and Microsoft SQL Server You can now use Time Travel in all Amazon Regions with DMS-supported Oracle, Microsoft SQL Server, and PostgreSQL source endpoints, and DMS-supported PostgreSQL and MySQL target endpoints.
S3 validation Amazon DMS now supports validating replicated data in Amazon S3 target endpoints. For information about validating Amazon S3 target data, see Amazon S3 target data validation.
Glue Catalog Integration Amazon Glue is a service that provides simple ways to categorize data, and consists of a metadata repository known as Amazon Glue Data Catalog. You can now integrate an Amazon Glue Data Catalog with your Amazon S3 target endpoint and query Amazon S3 data through other Amazon services such as Amazon Athena. For more information, see Using Amazon Glue Data Catalog with an Amazon S3 target for Amazon DMS.
Parallel apply for DocumentDB as a target Using DocumentDB as the target with new ParallelApply* task settings, Amazon DMS now supports a maximum of 5000 records per second during CDC replication. For more information, see Using Amazon DocumentDB as a target for Amazon Database Migration Service.
Customer Centric Logging You can now examine and manage task logs more effectively with Amazon DMS version 3.5.0. For information about viewing and managing Amazon DMS task logs, see Viewing and managing Amazon DMS task logs.
SASL_PLAIN mechanism for Kafka target endpoints You can now use SASL_PLAIN authentication to support Kafka MSK target endpoints.
Replication of XA transactions in MySQL You can now use XA transactions on your MySQL DMS source. Prior to DMS 3.5.0, DML changes applied as part of XA transactions weren’t replicated correctly.
Oracle Extended Data Types Amazon DMS now supports replication of extended data types in Oracle version 12.2 and higher.
Db2 LUW PureScale Environment Amazon DMS now supports replication from a Db2 LUW PureScale environment. This functionality is only supported using the Start processing changes from source change position option.
SQL Server source with READ_COMMITTED_SNAPSHOT option When using a Microsoft SQL Server source database with the READ_COMMITTED_SNAPSHOT option set to TRUE, you can replicate DML changes correctly by setting the forceDataRowLookup connection attribute.

Amazon DMS 3.5.0 includes the following resolved issues:

Issues resolved in Amazon DMS 3.5.0 launched on 17-March-2023
Topic Resolution
Oracle—comparing special case for string that was converted from numeric Fixed an issue for Oracle source where filtering rules weren't working as expected for a numeric column when data type transformation to string existed for the same column.
On-premises SQL Server AG enhancements Improved efficiency of connection handling with SQL Server source in AlwaysOn configuration by eliminating unnecessary connections to replicas that aren't used by DMS.
SQL Server HIERARCHYID internal conversion Fixed an issues with SQL Server Source where HIERARCHYID data type was replicated as VARCHAR(250) instead of HIERARCHYID to SQL Server target.
S3 target move task fix Fixed an issue when moving a task with an S3 target would take a very long time, appear frozen or never complete.
SASL Plain mechanism of Kafka Introduced support for SASL Plain authentication method for Kafka MSK target endpoint.
Parallel Load/Apply fails due to _type parameter with Opensearch 2.x Fixed an issue for Opensearch 2.x target where parallel load or parallel apply would fail due to lack of support for _type parameter.
Support table mapping filter with mixed operators Removed a limitation where only one filter could be applied on a column.
S3, Kinesis, Kafka endpoints — alter-based lob columns migration in CDC phase Fixed an issue for Kinesis, Kafka and S3 targets where data in LOB columns added during CDC wasn't replicated.
MongoDB driver upgrade Upgraded the MongoDB driver to v1.23.2.
Kafka driver update Upgraded the Kafka driver from 1.5.3 to 1.9.2.
S3 endpoint setting was not working properly Fixed an issue for S3 target where the AddTrailingPaddingCharacter endpoint setting was not working when data contained the character specified as the delimiter for the S3 target.
Kinesis target task would crash Fixed an issue for Kinesis target where a task would crash when PK value was empty and detailed debug was enabled.
When S3 targets' column names were shifted by one position Fixed an issue for an S3 target where column names where shifted by one position when AddColumnName was set to true and TimestampColumnName was set to "".
Improved logging LOB truncation warning Improved warning logging for LOB truncation for SQL Server source to include the select statement used to retrieve the LOB.
Add Fatal error to avoid DMS task crashes if TDE password is wrong. Introduced meaningful error message and eliminated task crash issue in situations where DMS task was failing with no error message due incorrect TDE password for Oracle as a source.
Allows PostgreSQL CTAS(Create table as selected) DDL's migration during CDC. Removed limitations of DMS not being able to replicate PostgreSQL CTAS (create table as selected) DDLs during CDC.
Fix pg_logical task crash when table columns are dropped in CDC. Fixed an issue for PostgreSQL source with S3 target where columns were misaligned on the target when support for LOBs was disabled and LOBs were present.
Fix memory leak in MySQL connection handling Fixed an issue for MySQL source where task memory consumption was increasing continuously.
Oracle source endpoint setting – ConvertTimestampWithZoneToUTC Set this attribute to true to convert the timestamp value of 'TIMESTAMP WITH TIME ZONE' and 'TIMESTAMP WITH LOCAL TIME ZONE' columns to UTC. By default the value of this attribute is 'false' and data is replicated using the source database timezone.
Oracle source - DataTruncationErrorPolicy to SUSPEND_TABLE not working Fixed an issue for Oracle source with S3 target where tables were not suspended while the DataTruncationErrorPolicy task setting was set to SUSPEND_TABLE.
SQL Server fail on long schema/table while building query clause Fixed an issues for SQL Server source where task would fail or become unresponsive when selection rule contained comma separated list of tables.
Secret Manager authentication with MongoDB endpoint Fixed an issue for MongoDB and DocumentDB endpoints where secret manager based authentication wasn't working.
DMS truncating the data during CDC for a multi-byte varchar column when NLS_NCHAR_CHARACTERSET is set to UTF8 Fixed an issues for Oracle source with Oracle target where data was being truncated for multi-byte VARCHAR columns with NLS_NCHAR_CHARACTERSET set to UTF8.
filterTransactionsOfUser ECA for Oracle LogMiner Added an Extra Connection Attribute (ECA) filterTransactionsOfUser to allow DMS to ignore transactions from a specified user when replicating from Oracle using LogMiner.
SQL Server Setting recoverable error when lsn missing from backup Fixed an issue for SQL Server where a task would not fail on missing LSN.

Amazon Database Migration Service 3.4.7 release notes

The following table shows the new features and enhancements introduced in Amazon Database Migration Service (Amazon DMS) version 3.4.7.

New feature or enhancement Description
Support Babelfish as a target

Amazon DMS now supports Babelfish as a target. Using Amazon DMS, you can now migrate live data from any Amazon DMS supported source to a Babelfish, with minimal downtime.

For more information, see Using Babelfish as a target for Amazon Database Migration Service.

Support IBM Db2 z/OS databases as a source for full load only

Amazon DMS now supports IBM Db2 z/OS databases as a source. Using Amazon DMS, you can now perform live migrations from Db2 mainframes to any Amazon DMS supported target.

For more information, see Using IBM Db2 for z/OS databases as a source for Amazon DMS.

Support SQL Server read replica as a source

Amazon DMS now supports SQL Server read replica as a source. Using Amazon DMS, you can now perform live migrations from SQL Server read replica to any Amazon DMS supported target.

For more information, see Using a Microsoft SQL Server database as a source for Amazon DMS.

Support EventBridge DMS events

Amazon DMS supports managing event subscriptions using EventBridge for DMS events.

For more information, see Working with Amazon EventBridge events and notifications in Amazon Database Migration Service.

Support VPC source and target endpoints

Amazon DMS now supports Amazon Virtual Private Cloud (VPC) endpoints as sources and targets. Amazon DMS can now connect to any Amazon service with VPC endpoints when explicitly defined routes to the services are defined in their Amazon DMS VPC.

Note

Upgrades to Amazon DMS versions 3.4.7 and higher require that you first configure Amazon DMS to use VPC endpoints or to use public routes. This requirement applies to source and target endpoints for Amazon S3, Amazon Kinesis Data Streams, Amazon Secrets Manager, Amazon DynamoDB, Amazon Redshift, and Amazon OpenSearch Service.

For more information, see Configuring VPC endpoints as Amazon DMS source and target endpoints.

New PostgreSQL version PostgreSQL version 14.x is now supported as a source and as a target.
Support Aurora Serverless v2 as a target

Amazon DMS now supports Aurora Serverless v2 as a target. Using Amazon DMS, you can now perform live migrations to Aurora Serverless v2.

For information about supported Amazon DMS targets, see Targets for data migration.

New IBM Db2 for LUW versions

Amazon DMS now supports IBM Db2 for LUW versions 11.5.6 and 11.5.7 as a source. Using Amazon DMS, you can now perform live migrations from the latest versions of IBM DB2 for LUW.

For information about Amazon DMS sources, see Sources for data migration.

For information about supported Amazon DMS targets, see Targets for data migration.

Amazon DMS 3.4.7 includes the following new or changed behavior and resolved issues:

  • You can now use a date format from the table definition to parse a data string into a date object when using Amazon S3 as a source.

  • New table statistics counters are now available: AppliedInserts, AppliedDdls, AppliedDeletes, and AppliedUpdates.

  • You can now choose the default mapping type when using OpenSearch as a target.

  • The new TrimSpaceInChar endpoint setting for Oracle, PostgreSQL, and SQLServer sources allows you to specify whether to trim data on CHAR and NCHAR data types.

  • The new ExpectedBucketOwner endpoint setting for Amazon S3 prevents sniping when using S3 as a source or target.

  • For RDS SQL Server, Azure SQL Server, and self-managed SQL Server — DMS now provides automated setup of MS-CDC on all tables selected for a migration task that are with or without a PRIMARY KEY, or with a unique index considering the enablement priority for MS-REPLICATION on self-managed SQL Server tables with PRIMARY KEY.

  • Added support for replication of Oracle Partition and sub-partition DDL Operations during Oracle homogenous migrations.

  • Fixed an issue where a data validation task crashes with a composite primary key while using Oracle as a source and target.

  • Fixed an issue with correctly casting a varying character type to a boolean while the target column was pre-created as a boolean when using Redshift as a target.

  • Fixed an issue that was causing data truncation for varchar data types migrated as varchar(255) due to a known ODBC issue when using PostgreSQL as a target.

  • Fixed an issue where Parallel Hint for the DELETE operation wasn't respected with BatchApplyEnabled set to true and BatchApplyPreserveTransaction set to false when using Oracle as a target.

  • The new AddTrailingPaddingCharacter endpoint setting for an Amazon S3 adds padding on string data when using S3 as a target.

  • The new max_statement_timeout_seconds task setting extends the default timeout of endpoint queries. This setting is currently used by MySQL endpoint metadata queries.

  • When using PostgreSQL as a target, fixed an issue where a CDC task wasn't properly utilizing the error handling task settings.

  • Fixed an issue where DMS was unable to correctly identify Redis mode for a Redis Enterprise instance.

  • Extended the support of includeOpForFullLoad extra connection attribute (ECA) for the S3 target parquet format.

  • Introduced a new PostgreSQL endpoint setting migrateBooleanAsBoolean. When this setting is set to true for a PostgreSQL to Redshift migration, a boolean will be migrated as varchar(1). When it is set to false, a boolean is migrated as varchar(15), which is the default behavior.

  • When using SQL Server source, fixed a migration issue with datetime datatype. This fix addresses the issue of inserting Null when precision is in milliseconds.

  • For PostgresSQL source with PGLOGICAL, fixed a migration issue when using pglogical and removing a field from the source table during the CDC phase, where the value after the removed field wasn't migrated to the target table.

  • Fixed a SQL Server Loopback migration issue with Bidirectional replication getting repeated records.

  • Added a new ECA mapBooleanAsBoolean for PostgresSQL as a source. Using this extra connection attribute , you can override default data type mapping of a PostgresSQL Boolean to a RedShift Boolean data type.

  • Fixed a migration issue when using SQL Server as source that addresses the ALTER DECIMAL/NUMERIC SCALE not replicating to targets.

  • Fixed connection issue with SQL Server 2005.

  • As of November 25, 2022, with DMS 3.4.7 you can convert database schemas and code objects using DMS Schema Conversion, and discover databases in your network environment that are good candidates for migration using DMS Fleet Advisor.

  • As of November 25, 2022, DMS Studio is retired.

  • As of January 31, 2023, DMS Schema Conversion supports Aurora MySQL and Aurora PostgreSQL as a target data provider.

  • As of March 6, 2023, you can generate right sized target recommendations for your source databases with DMS Fleet Advisor.

  • As of March 6, 2023, Amazon DMS supports the Amazon managed policy that allows publishing metric data points to Amazon CloudWatch.

Issues resolved in the DMS 3.4.7 maintenance release dated 5-May-2023
Topic Resolution
PostgreSQL source task failure Fixed an issue for PostgreSQL source where tasks were failing when exceeding the maximum allowed DDL operations in a single event.
PostgreSQL source data validation false positives Fixed an issue for PostgreSQL source with Oracle target where incorrect casting of timestamp field resulted in false positive data validation errors.
MySQL source error handling Fixed an issues for a MySQL source where the DMS task wasn't failing when the next BIN log was unavailable.
MySQL source ROTATE_EVENT logging Improved logging for MySQL source related to ROTATE_EVENT – included the BIN log name being read.
Data validation timeout issue Fixed an issue for Data Validation feature where the executeTimeout endpoint setting wasn't being respected for queries related to data validation.
PostgreSQL target parallel full load issue Fixed an issue for PostgreSQL target where segmented (parallel) full load was failing due to "connection down" error.
DMS task move issue Fixed an issue for S3 target where a DMS task move operation was taking a very long time or never completing.
PostgreSQL source duplicate record issue Fixed an issue for PostgreSQL source where a DMS task would throw errors related to duplicates on the target after a task stop and resume.
Oracle target data validation false positives Fixed an issue for Oracle target where data validation would report false positive errors due to incorrectly replicated timezone for timestamp fields.
Issues resolved in the DMS 3.4.7 maintenance release dated 22-February-2023
Topic Resolution
SQL Server AG replicas as a source Added support for SQL Server source in AlwaysOn configuration where the listener TCP port differed from replica TCP port.
Data loss with Amazon Redshift as a target Fixed an issue for Redshift target where in some rare cases unexpected Redshift restart could have caused missing data on target.
SQL Server source safeguard support Fixed an issue for SQL Server source where the DMS task could fail with an error indicating inability to read transaction log backups when Endpoint Setting "SafeguardPolicy": "EXCLUSIVE_AUTOMATIC_TRUNCATION" is specified.
Data validation task failure for Oracle as a source Fixed an issue for Oracle source where DMS task could fail on data validation due to incorrectly identified Primary Key values.
Kinesis before image data issue Fixed an issue for streaming targets (Kinesis, Kafka) where the "EnableBeforeImage" task setting was working only for character data types.
Time Travel log files Fixed an issue for the Time Travel feature where DMS was creating zero byte time travel log files when the source is idle.
Issues resolved in the DMS 3.4.7 maintenance release dated 16-December-2022
Topic Resolution
BatchApplyEnabled Fixed an issue of excessive logging when BatchApplyEnabled is set to True.
New MongoDB endpoint setting–FullLoadNoCursorTimeout MongoDB endpoint setting FullLoadNoCursorTimeout specifies NoCursorTimeout for the full load cursor. NoCursorTimeout is a MongoDB connection setting that prevents the server from closing the cursor if idle.
MongoDB—Filter function for single column segmentation New filter function improves performance for migrating MongoDB databases using a single column for segmentation.
MongoDB to Redshift When migrating from MongoDB to Redshift, if the MongoDB collection has binary data type, fixed an issue where DMS was not creating the target table on Redshift.
New MongoDB SocketTimeoutMS connection attribute New MongoDB SocketTimeoutMS extra connection attribute configures the connection timeout for MongoDB clients in units of milliseconds. If the value is less than or equal to zero, then the MongoDB client default is used.
Fixed issue causing an Amazon Kinesis task to crash When migrating to Amazon Kinesis Data Streams as a target, fixed an issue handling null values if a primary key wasn't present in the table.
Oracle NULL PK/UK data validation supported Removed the limitation that data validation of NULL PK/UK values aren't supported.
Oracle to Amazon S3 When migrating from Oracle to Amazon S3, fixed an issue where a few records were incorrectly migrated as NULL.
Oracle Standby When using Oracle Standby as a source, added the ability for DMS to handle open transactions.
Oracle to Oracle migration with SDO_GEOMETRY spatial data type When migrating from Oracle to Oracle, fixed an issue where the task failed if the table had an SDO_GEOMERY column present in the DDL.
Oracle as source When using Oracle as a source, fixed an issue where DMS occasionally skips an Oracle redo log sequence number.
Oracle as source—missing archive/online redo logs When using Oracle as a source, fixed an issue so that the DMS task fails when archive logs are missing.
Fixed—DMS occasionally skips Oracle standby redo log When using Oracle as a source, fixed an issue where DMS occasionally skips an Oracle redo log sequence number.
Fixed—Oracle to Oracle spatial datatypes not replicating during CDC When replicating from Oracle to Oracle, fixed an issue where spatial datatypes were not replicating during CDC.
Oracle as target When using Oracle as a target, fixed an issue where the target apply was failing with an ORA-01747 error.
Amazon S3—Fixed reload table data loss When using Amazon S3 as a target, fixed an issue where a table reload operation wasn't generating CDC files.
Fixed—SQL Server Always On context initialization in case primary server as a source When using SQL Server Always On as a source, fixed an issue to not initialize Availability Groups (AG) if the source is primary and AlwaysOnSharedSyncedBackupIsEnabled is set to true.
Updated SQL Server endpoint setting When a source endpoint is SQL Server Always On Availability Group and is a secondary replica, fixed an issue where the replication task fails if AlwaysOnSharedSynchedBackupsIsEnabled is set to True.
PostgreSQL as source Fixed an issue where CDC fails to migrate delete/update operations on the PostgreSQL source, which was introduced in 3.4.7 in supporting mapBooleanAsBoolean.

Amazon Database Migration Service 3.4.6 release notes

The following table shows the new features and enhancements introduced in Amazon Database Migration Service (Amazon DMS) version 3.4.6.

New feature or enhancement Description
Amazon DMS Time Travel Amazon DMS introduces Time Travel, a feature granting customers flexibility on their logging capabilities, and enhancing their troubleshooting experience. With Time Travel, you can store and encrypt Amazon DMS logs using Amazon S3, and view, download and obfuscate the logs within a certain time frame.
Support Microsoft Azure SQL Managed Instance as a source

Amazon DMS now supports Microsoft Azure SQL Managed Instance as a source. Using Amazon DMS, you can now perform live migrations from Microsoft Azure SQL Managed Instance to any Amazon DMS supported target.

For information about Amazon DMS sources, see Sources for data migration.

For information about supported Amazon DMS targets, see Targets for data migration.

Support Google Cloud SQL for MySQL as a source

Amazon DMS now supports Google Cloud SQL for MySQL as a source. Using Amazon DMS, you can now perform live migrations from Google Cloud SQL for MySQL to any Amazon DMS supported target.

For information about Amazon DMS sources, see Sources for data migration.

For information about supported Amazon DMS targets, see Targets for data migration.

Support parallel load for partitioned data to S3 Amazon DMS now supports parallel load for partitioned data to Amazon S3, improving the load times for migrating partitioned data from supported database engine source data to Amazon S3. This feature creates Amazon S3 sub-folders for each partition of the table in the database source, allowing Amazon DMS to run parallel processes to populate each sub-folder.
Support multiple Apache Kafka target topics in a single task Amazon DMS now supports Apache Kafka multi-topic targets with a single task. Using Amazon DMS, you can now replicate multiple schemas from a single database to different Apache Kafka target topics using the same task. This eliminates the need to create multiple separate tasks in situations where many tables from the same source database need to be migrated to different Kafka target topics.

The issues resolved in Amazon DMS 3.4.6 include the following:

  • Fixed an issue where columns from UPDATE statements were populated to incorrect columns if the primary key column is not the first column when using Amazon S3 as a target with CSV format.

  • Fixed an issue where Amazon DMS tasks might crash when using the pglogical plugin with NULL values in BYTEA columns under limited LOB mode when using PostgreSQL as a source.

  • Fixed an issue where Amazon DMS tasks might crash when a large number of source tables are deleted when using PostgreSQL as a source.

  • Improved Amazon S3 date-based folder partitioning by introducing a new Amazon S3 setting DatePartitionTimezone to allow partitioning on non-UTC dates.

  • Supported the mapping between data type TIMESTAMP WITH TIME ZONE from sources to TIMESTAMPTZ when using Redshift as a target

  • Improved the performance of CDC for tasks without wildcard selection rules when using MongoDB or Amazon DocumentDB as a source.

  • Fixed an issue where schema names with underscore wildcard and length less than 8 were not captured by Amazon DMS tasks when using Db2 LUW as a source.

  • Fixed an issue where Amazon DMS instances ran out of memory under large data volume when using OpenSearch Service as a target.

  • Improved the performance of data validation by supporting full load validation only tasks.

  • Fixed an issue where Amazon DMS tasks failed to resume after forced failover when using Sybase as a source.

  • Fixed an issue where Amazon DMS sent warning Invalid BC timestamp was encountered in column incorrectly.

Issues resolved in the DMS 3.4.6 maintenance release include the following:

  • Fixed an issue of a task crashing when bulk apply mode is enabled when using Oracle as the source and target.

  • Fixed an issue so that a full load task properly uses the ExecuteTimeout endpoint setting with PostgreSQL as source.

  • Fixed an issue with migrating Array data type columns when the task is set to limited LOB mode while using PostgreSQL as a source.

  • Fixed an issue with migrating timestamps with time zone before 1970-01-01 when using PostgreSQL as a source.

  • Fixed an issue where DMS was treating an empty string as null during replication when using SQL Server as a source and target.

  • Fixed an issue to honor session read and write timeout endpoint settings when using MySQL source/target.

  • Fixed an issue where a DMS CDC task was downloading full load related files when using Amazon S3 as a source.

  • Fixed a log crashing issue when CdcInsertsAndUpdates and PreserveTransactions are both set to true when using Amazon S3 as a target.

  • Fixed an issue where a task crashed when the ParallelApply* feature is enabled, but some tables don't have a default primary key when using Amazon Kinesis Data Streams as a source.

  • Fixed an issue where an error wasn't given for an incorrect StreamArn when using Amazon Kinesis Data Streams as a source.

  • Fixed an issue where a primary key value as an empty string would cause a task to crash when using OpenSearch as a target.

  • Fixed an issue where too much disk space was used by data validation.

Issues resolved in the DMS 3.4.6 maintenance release dated 13-December-2022
Topic Resolution
SAP ASE odbc driver Fixed an issue for SAP ASE as a source so that the ODBC driver can support character sets.

Sqlserver datetime primary key bug for lob lookup

Fixed an issue for SQL Server as a source where LOB lookup was not working properly, when primary key has a datetime datatype, with precision in milliseconds.

SQL Server to Redshift–'datetimeoffset' mapped to 'timestamptz'

For migrations from SQL Server to Redshift, improved mapping so that the SQL Server 'datetimeoffset' format is mapped to the Redshift 'timestamptz' format.

Data Validation - SkipLobColumns is True

Fixed an issue where DMS task crashes when SkipLobColumns is True, there is a LOB on the source, the Primary Key is on the last column, and a data difference is detected by validation.

Data Validation with MySQL as source

Fixed an issue for MySQL as a source with data validation enabled, where a DMS task crash occurs while using a table that has a composite unique key that has null values.

MySQL as source

Fixed an issue for MySQL as a source, where a table is getting suspended with Overflow error when the columns are altered to add precision.

Upgrade MySQL ODBC driver to 8.0.23

Fixed an issue for MySQL as a source, where the collation ‘utf8mb4_0900_bin' was incompatible with the mysql driver used by DMS.

MySQL–support DDL changes for partitioned tables

Introduced a new MySQL endpoint setting skipTableSuspensionForPartitionDdl to allow the user to skip table suspension for partition DDL changes during CDC, so that DMS can now support DDL changes for partitioned MySQL tables.

MongoDB to Redshift migration

Fixed an issue for MongoDB to Redshift migrations, where DMS fails to create the target table on Redshift if the MongoDB collection has binary data type.

Redshift target–Time Travel Segfault in Bulk Apply

Fixed an issue for Redshift as a target, where DMS task crashes with BatchApplyEnabled set to true.

Redshift as target

Fixed an issue for Redshift as a target, where with parallel-load set to type=partitions-auto, parallel segments were writing bulk CSV files to the same table directory and interfering with each other.

Redshift as target

Fixed an issue with Redshift as a target, where during CDC the target column is of type boolean while the source is of type character varying.

Redshift as target

Improved the task log to identify a DDL change that fails to replicate to Redshift as a target.

Data validation with PostgreSQL

Fixed an issue for validation with PostgreSQL, where the validation fails when boolean data types are present.

PostgreSQL as source

Fixed an issue for PostgreSQL as a source, so that the full load uses the ExecuteTimeout field in Extra connection attributes.

PostgreSQL as source

Fixed an issue for PostgreSQL as a source, so that a task will fail if it's reading LSNs which are greater than the requested task resume LSN for more than 60 min to indicate that the is a problem with the replication slot being used.

PostgreSQL as a source–timestamptz before 1970-01-01

Fixed an issue for PostgreSQL as a source, where timestamptz before 1970-01-01 were not migrated correctly during CDC.

PostgreSQL as source

Fixed an issue for PostgreSQL as a source, where DMS was truncating character varying datatype values during CDC.

PostgreSQL as a source–resuming stopped task

Fixed an issue for PostgreSQL as a source where resuming a previously stopped task replay misses one or more transactions during CDC.

Amazon S3 as target

Fixed an issue for S3 as a target, where the resultant CSV file header is off by one column when AddColumnName is true and TimestampColumnName is "".

Amazon S3 as source–memory usage behavior in full load phase for task

Fixed an issue for S3 as a source, where a DMS task in full load was only releasing the used memory after the entire table was loaded to the target database.

Amazon S3 as target–table reload operation

Fixed an issue for S3 as a target, where a table reload operation missed generating CDC files.

Amazon Database Migration Service 3.4.5 release notes

The following table shows the new features and enhancements introduced in Amazon Database Migration Service (Amazon DMS) version 3.4.5.

New feature or enhancement Description
Support for Redis as a target Amazon DMS now supports Redis as a target. Using Amazon DMS, you can now migrate live data from any Amazon DMS supported source to a Redis data store, with minimal downtime. For information about Amazon DMS targets, see Targets for data migration.
Support for MongoDB 4.2 and 4.4 as sources Amazon DMS now supports MongoDB 4.2 and 4.4 as sources. Using Amazon DMS, you can now migrate data from MongoDB 4.2 and 4.4 clusters to any Amazon DMS supported target including Amazon DocumentDB (with MongoDB compatibility), with minimal downtime. For information about Amazon DMS sources, see Sources for data migration.
Support for multiple databases using MongoDB as a source Amazon DMS now supports migrating multiple databases in one task using MongoDB as a source. Using Amazon DMS, you can now group multiple databases of a MongoDB cluster, and migrate them using one database migration task. You can migrate to any Amazon DMS supported target, including Amazon DocumentDB (with MongoDB compatibility), with minimal downtime.
Support for automatic segmentation using MongoDB or Amazon DocumentDB (with MongoDB compatibility) as a source Amazon DMS now supports automatic segmentation using MongoDB or Amazon DocumentDB as a source. Using Amazon DMS, you can configure database migration tasks to segment the collection of a MongoDB or DocumentDB cluster automatically. You can then migrate the segments in parallel to any Amazon DMS supported target, including Amazon DocumentDB, with minimal downtime.
Amazon Redshift full load performance improvement Amazon DMS now supports using parallel threads when using Amazon Redshift as a target during full load. By taking advantage of the multithreaded full load task settings, you can improve the performance of your initial migration from any Amazon DMS supported source to Amazon Redshift. For information about Amazon DMS targets, see Targets for data migration.

The issues resolved in Amazon DMS 3.4.5 include the following:

  • Fixed an issue where data could potentially be missing or duplicated after resuming, when using PostgreSQL as a source with high transaction concurrency.

  • Fixed an issue where database migration tasks fail with error Could not find relation id … when using PostgreSQL as a source, with the pglogical plugin enabled.

  • Fixed an issue where VARCHAR columns are not replicated correctly when using PostgreSQL as a source and Oracle as a target.

  • Fixed an issue where delete operations are not properly captured when the primary key is not the first column in the table definition, when using PostgreSQL as a source.

  • Fixed an issue where database migration tasks miss LOB updates in a special metadata setting when using MySQL as a source.

  • Fixed an issue where TIMESTAMP columns are treated as DATETIME in full LOB mode when using MySQL version 8 as a source.

  • Fixed an issue where database migration tasks fail when parsing NULL DATETIME records when using MySQL 5.6.4 and higher as a source.

  • Fixed an issue where database migration tasks get stuck after encountering a Thread is exiting error when using Amazon Redshift as a target with parallel apply.

  • Fixed an issue where data could potentially be lost, when database migration tasks disconnect with a Amazon Redshift target endpoint during batch-apply CDC.

  • Improved the performance of full load by calling ACCEPTINVCHARS when using Amazon Redshift as a target.

  • Fixed an issue where duplicated records are replicated when reverting from one-by-one mode to parallel apply mode using Amazon Redshift as a target.

  • Fixed an issue where database migration tasks do not switch Amazon S3 object ownership to bucket owner with cannedAclForObjects=bucket_owner_full_control when using Amazon S3 as a target.

  • Improved Amazon DMS by supporting multiple archive destinations with ECA additionalArchivedLogDestId when using Oracle as a source.

  • Fixed an issue where database migration tasks fail with error OCI_INVALID_HANDLE while updating a LOB column in full LOB mode.

  • Fixed an issue where NVARCHAR2 columns are not migrated properly during CDC when using Oracle as a source.

  • Improved Amazon DMS by enabling SafeguardPolicy when using RDS for SQL Server as a source.

  • Fixed an issue where database migration tasks report error on rdsadmin when using a non-RDS SQL Server source.

  • Fixed an issue where data validation fails with UUID as the primary key in a partition setting when using SQL Server as a source.

  • Fixed an issue where full load plus CDC tasks might fail if the required LSN cannot be found in the database log when using Db2 LUW as a source.

  • Improved Amazon DMS by supporting custom CDC timestamps when using MongoDB as a source.

  • Fixed an issue where database migration tasks get stuck when stopping, using MongoDB as a source, when the MongoDB driver errors on endSessions.

  • Fixed an issue where Amazon DMS fails to update non-primary fields when using DynamoDB as a target

  • Fixed an issue where data validation reports false positive mismatches on CLOB and NCLOB columns.

  • Fixed an issue where data validation fails on whitespace-only records when using Oracle as a source.

  • Fixed an issue where database migration tasks crash when truncating a partitioned table.

  • Fixed an issue where database migration tasks fail when creating the awsdms_apply_exceptions control table.

  • Extended support of the caching_sha2_password authentication plugin when using MySQL version 8.

Amazon Database Migration Service 3.4.4 release notes

The following table shows the new features and enhancements introduced in Amazon DMS version 3.4.4.

New feature or enhancement Description
Support TLS encryption and TLS or SASL authentication using Kafka as a target Amazon DMS now supports TLS encryption and TLS or SASL authentication using Amazon MSK and on-premises Kafka cluster as a target. For more information on using encryption and authentication for Kafka endpoints, see Connecting to Kafka using Transport Layer Security (TLS).

The issues resolved in Amazon DMS 3.4.4 include the following:

  • Improved Amazon DMS logging on task failures when using Oracle endpoints.

  • Improved Amazon DMS task execution continues processing when Oracle source endpoints switch roles after Oracle Data Guard fail over.

  • Improved error handling treats ORA—12561 as a recoverable error when using Oracle endpoints.

  • Fixed an issue where EMPTY_BLOB() and EMPTY_CLOB() columns are migrated as null when using Oracle as a source.

  • Fixed an issue where Amazon DMS tasks fail to update records after add column DDL changes when using SQL Server as a source.

  • Improved PostgreSQL as a source migration by supporting the TIMESTAMP WITH TIME ZONE data type.

  • Fixed an issue where the afterConnectScript setting does not work during a full load when using PostgreSQL as a target.

  • Introduced a new mapUnboundedNumericAsString setting to better handle the NUMERIC date type without precision and scale when using PostgreSQL endpoints.

  • Fixed an issue where Amazon DMS tasks fail with “0 rows affected” after stopping and resuming the task when using PostgreSQL as a source.

  • Fixed an issue where Amazon DMS fails to migrate the TIMESTAMP data type with the BC suffix when using PostgreSQL as a source.

  • Fixed an issue where Amazon DMS fails to migrate the TIMESTAMP value “±infinity” when using PostgreSQL as a source.

  • Fixed an issue where empty strings are treated as NULL when using S3 as a source with the csvNullValue setting set to other values.

  • Improved the timestampColumnName extra connection attribute in a full load with CDC to be sortable during CDC when using S3 as a target.

  • Improved the handling of binary data types in hex format such as BYTE, BINARY, and BLOB when using S3 as a source.

  • Fixed an issue where deleted records are migrated with special characters when using S3 as a target.

  • Fixed an issue to handle empty key values when using Amazon DocumentDB (with MongoDB compatibility) as a target.

  • Fixed an issue where Amazon DMS fails to replicate NumberDecimal or Decimal128 columns when using MongoDB or Amazon DocumentDB (with MongoDB compatibility) as a source.

  • Fixed an issue to allow CDC tasks to retry when there is a fail over on MongoDB or Amazon DocumentDB (with MongoDB compatibility) as a source.

  • Added an option to remove the hexadecimal “0x” prefix to RAW data type values when using Kinesis, Kafka, or OpenSearch as a target.

  • Fixed an issue where validation fails on fixed length character columns when using Db2 LUW as a source.

  • Fixed an issue where validation fails when only the source data type or the target data type is FLOAT or DOUBLE.

  • Fixed an issue where validation fails on NULL characters when using Oracle as a source.

  • Fixed an issue where validation fails on XML columns when using Oracle as a source.

  • Fixed an issue where Amazon DMS tasks crash when there are nullable columns in composite keys using MySQL as a source.

  • Fixed an issue where Amazon DMS fails to validate both UNIQUEIDENTIFIER columns from SQL Server source endpoints and UUID columns from PostgreSQL target endpoints.

  • Fixed an issue where a CDC task does not use an updated source table definition after it is modified.

  • Improved Amazon DMS fail over to treat task failures caused by an invalid user name or password as recoverable errors.

  • Fixed an issue where Amazon DMS tasks fail because of missing LSNs when using RDS for SQL Server as a source.

Amazon Database Migration Service 3.4.3 release notes

The following table shows the new features and enhancements introduced in Amazon DMS version 3.4.3.

New feature or enhancement Description
New Amazon DocumentDB version Amazon DocumentDB version 4.0 is now supported as a source.
New MariaDB version MariaDB version 10.4 is now supported as both a source and target.
Support for Amazon Secrets Manager integration You can store the database connection details (user credentials) for supported endpoints securely in Amazon Secrets Manager. You can then submit the corresponding secret instead of plain-text credentials to Amazon DMS when you create or modify an endpoint. Amazon DMS then connects to the endpoint databases using the secret. For more information on creating secrets for Amazon DMS endpoints, see Using secrets to access Amazon Database Migration Service endpoints.
Larger options for C5 and R5 replication instances You can now create the following larger replication instance sizes: C5 sizes up to 96 vCPUs and 192 GiB of memory and R5 sizes up to 96 vCPUs and 768 GiB of memory.
Amazon Redshift performance improvement Amazon DMS now supports parallel apply when using Redshift as a target to improve the performance of on-going replication. For more information, see Multithreaded task settings for Amazon Redshift.

The issues resolved in Amazon DMS 3.4.3 include the following:

  • Fixed an issue where commit timestamp became “1970-01-01 00:00:00” for deferred events when using Db2 LUW as a source.

  • Fixed an issue where Amazon DMS tasks failed with an NVARCHAR column as primary key when using SQL Server as a source with Full LOB mode.

  • Fixed an issue of missing records during cached changes phase when using SQL Server as a source.

  • Fixed an issue where records were skipped after Amazon DMS tasks were resumed when using RDS for SQL Server as a source.

  • Fixed an issue where Amazon DMS ASSERTION logging component generates large logs for SQL Server.

  • Fixed an issue where data validation failed during CDC phase due to column parsing overflow when using MySQL as a source.

  • Fixed an issue where Amazon DMS tasks crashed due to a segmentation fault during data validation when using PostgreSQL as a target.

  • Fixed an issue where data validation failed on DOUBLE data type during CDC when using PostgreSQL as a source and a target.

  • Fixed an issue where records inserted by copy command were not replicated correctly when using PostgreSQL as a source and Redshift as a target.

  • Fixed a data loss issue during cached changes phase when using PostgreSQL as a source.

  • Fixed an issue which could potentially cause either data loss or record duplicates when using PostgreSQL as a source.

  • Fixed an issue where schemas with mixed cases failed to migrate with pglogical when using PostgreSQL as a source.

  • Fixed an issue where the Last Failure Message did not contain the ORA error when using Oracle as a source.

  • Fixed an issue where Amazon DMS tasks failed to build UPDATE statements when using Oracle as a target.

  • Fixed an issue where Amazon DMS tasks did not replicate data when using Oracle 12.2 as a source with ASM and Pluggable Database configuration.

  • Improved record parsing by preserving quotes to be compliant with RFC 4180 when using S3 as a source.

  • Improved the handling of timestampColumnName so that the column from Full Load is sortable with that from CDC.

  • By introducing a new endpoint setting MessageMaxBytes, fixed an issue where Amazon DMS tasks failed when there are LOB elements larger than 1MB .

  • Fixed an issue where Amazon DMS tasks crashed due to a segmentation fault when using Redshift as a target.

  • Improved error logging for Redshift test connection.

  • Fixed an issue where Amazon DMS did not transfer all documents from MongoDB to DocumentDB during Full Load.

  • Fixed an issue where Amazon DMS tasks reported fatal error when no tables were included in the table mapping rules.

  • Fixed an issue where schemas and tables created before restarting Amazon DMS tasks did not replicate to the target when using MySQL as a source.

  • Fixed an issue where wildcard escape [_] cannot escape wildcard “_” in exclude rule when using MySQL as a source.

  • Fixed an issue where column of data type UNSIGNED BIGINT did not replicate correctly when using MySQL as a source.

Amazon Database Migration Service 3.4.2 release notes

The following table shows the new features and enhancements introduced in Amazon DMS version 3.4.2.

New feature or enhancement Description
Support for privately connecting your Amazon Virtual Private Cloud (Amazon VPC) to Amazon Database Migration Service (DMS) without requiring an internet gateway, NAT device, VPN connection, or Amazon Direct Connect connection. You can now connect to and access Amazon DMS from your Amazon VPC through a VPC interface endpoint that you create. This interface endpoint allows you to isolate all network activity of your Amazon DMS replication instance within the Amazon network infrastructure. By including a reference to this interface endpoint in all API calls to Amazon DMS using the Amazon CLI or an SDK, you ensure that all Amazon DMS activity remains invisible to the public Internet. For more information, see Infrastructure security in Amazon Database Migration Service.
Note

This feature is available using all supported Amazon DMS engine versions.

CDC date-based folder partitioning using Amazon S3 as a target

Amazon DMS now supports date-based folder partitioning when replicating data using S3 as a target. For more information, see Using date-based folder partitioning.

The issues resolved in Amazon DMS 3.4.2 include the following:

  • Added a STATUPDATE option when performing a migration using Redshift as a target.

  • Improved validation tasks by introducing a new setting. ValidQueryCdcDelaySecond delays the first validation query on both source and target endpoints to help reduce resource contention when migration latency is high.

  • Fixed an issue where Amazon DMS took a long time to start validation tasks.

  • Fixed an issue where empty records were generated when starting or stopping replication tasks using S3 as a target.

  • Fixed an issue where tasks got stuck after a full load completed.

  • Fixed an issue where tasks got stuck when a source table has data errors while using S3 as a source .

  • Fixed an issue where tasks got stuck while starting when the user account of the source endpoint is disabled.

  • Fixed an issue where tasks crashed when using PostgreSQL as a source with REPLICA IDENTITY FULL.

  • Fixed an issue where tasks missed transactions when using PostgreSQL as a source with pglogical plugin.

  • Fixed an issue when Amazon DMS didn't delete compressed source files when using Redshift as a target.

  • Fixed an issue where validation tasks reported false negatives when using MySQL as both source and target with data type BIGINT UNSIGNED.

  • Fixed an issue where validation tasks reported false positives when using SQL Server as a source with a primary key column as CHAR type.

  • Fixed an issue where Amazon DMS doesn't clear target objects when using start-replication to start replication tasks using S3 as a target.

  • Fixed several issues on data validation when using Db2 as a source.

  • Fixed an issue where validation tasks got stuck when using SQL Server as a source with VARCHAR column as primary key.

  • Added support for data type TIMESTAMP WITH TIMEZONE when using PostgreSQL as a source

Amazon Database Migration Service 3.4.1 Beta release notes

The following table shows the new features and enhancements introduced in Amazon DMS version 3.4.1 Beta.

New feature or enhancement Description
New MongoDB version

MongoDB version 4.0 is now supported as a source.

TLS 1.2 support for SQL Server

Amazon DMS now supports TLS 1.2 for SQL Server endpoints.

The issues resolved in Amazon DMS 3.4.1 Beta include the following:

  • Improved Oracle 19c TDE support.

  • Improved support of utf8mb4 character set and identity data type using Redshift as a target.

  • Improved replication task failure handling when using MySQL as a source and binary log is not present.

  • Improved data validation support on various data types and character sets.

  • Improved null value handling with a new endpoint setting IncludeNullAndEmpty when using Kinesis and Kafka as a target.

  • Improved error logging and handling when using Kafka as a target.

  • Improved DST time offset when using SQL Server as a source.

  • Fixed an issue where replication tasks try to create existing tables for Oracle as a target.

  • Fixed an issue where replication tasks get stuck after the database connection is killed when using Oracle as a source.

  • Fixed an issue where replication tasks failed to detect and reconnect to the new primary when using SQL Server as a source with AlwaysON setting.

  • Fixed an issue where replication tasks do not add a "D" for "OP" column under certain conditions for S3 as a target.

Amazon Database Migration Service 3.4.0 Beta release notes

The following table shows the new features and enhancements introduced in Amazon DMS version 3.4.0

New feature or enhancement Description
New MySQL version

Amazon DMS now supports MySQL version 8.0 as a source, except when the transaction payload is compressed.

TLS 1.2 support for MySQL

Amazon DMS now supports TLS 1.2 for MySQL endpoints.

New MariaDB version

Amazon DMS now supports MariaDB version 10.3.13 as a source.

Non-SysAdmin access to self-managed Microsoft SQL Server sources

Amazon DMS now supports access by non-SysAdmin users to on-premise and EC2-hosted SQL Server source endpoints.

Note

This feature is currently in Beta. If you want to try it out, contact Amazon support for more information.

CDC tasks and Oracle source tables created using CREATE TABLE AS

Amazon DMS now supports both full-load and CDC and CDC-only tasks running against Oracle source tables created using the CREATE TABLE AS statement.

The issues resolved in Amazon DMS 3.4.0 include the following:

  • Improved premigration task assessments. For more information, see Enabling and working with premigration assessments for a task.

  • Improved data validation for float, real, and double data types.

  • Improved Amazon Redshift as a target by better handling this error: "The specified key does not exist."

  • Supports multithreaded CDC load task settings, including ParallelApplyThreads, ParallelApplyBufferSize, and ParallelApplyQueuesPerThread, for Amazon OpenSearch Service (OpenSearch Service) as a target.

  • Improved OpenSearch Service as a target by supporting its use of composite primary keys.

  • Fixed an issue where test connection fails when using PostgreSQL as a source and the password has special characters in it.

  • Fixed an issue with using SQL Server as a source when some VARCHAR columns are truncated.

  • Fixed an issue where Amazon DMS does not close open transactions when using Amazon RDS SQL Server as a source. This can result in data loss if the polling interval parameter is set incorrectly. For more information on how to setup a recommended polling interval value, see Using a Microsoft SQL Server database as a source for Amazon DMS.

  • Fixed an issue for Oracle Standby as source where CDC tasks would stop unexpectedly when using Binary Reader.

  • Fixed an issue for IBM DB2 for LUW where the task failed with the message "The Numeric literal 0 is not valid because its value is out of range."

  • Fixed an issue for a PostgreSQL to PostgreSQL migration when a new column was added on the PostgreSQL source and the column was created with a different data type than the data type for which the column was originally created on the source.

  • Fixed an issue with a MySQL source the migration task stopped unexpectedly when it was unable to fetch binlogs.

  • Fixed an issue related to an Oracle target when BatchApply was being used.

  • Fixed an issue for MySQL and MariaDB when migrating the TIME data type.

  • Fixed an issue for an IBM DB2 LUW source where migrating tables with LOBs fail when the tables don't have a primary key or unique key.

Amazon Database Migration Service 3.3.4 release notes

The issues resolved in Amazon DMS 3.3.4 include the following:

  • Fixed an issue where transactions are dropped or duplicated when using PostgreSQL as a source.

  • Improved the support of using dollar sign ($) in schema names.

  • Fixed an issue where replication instances do not close open transactions when using RDS SQL Server as a source.

  • Fixed an issue where test connection fails when using PostgreSQL as a source and the password has special characters in it.

  • Improved Amazon Redshift as a target by better handling this error: "The specified key does not exist."

  • Improved data validation support on various data types and character sets.

  • Fixed an issue where replication tasks try to create existing tables for Oracle as a target.

  • Fixed an issue where replication tasks do not add a "D" for "OP" column under certain conditions for Amazon S3 as a target.

Amazon Database Migration Service 3.3.3 release notes

The following table shows the new features and enhancements introduced in Amazon DMS version 3.3.3.

New feature or enhancement Description
New PostgreSQL version

PostgreSQL version 12 is now supported as a source and target.

Support for composite primary key with Amazon OpenSearch Service as target

As of Amazon DMS 3.3.3, use of a composite primary key is supported by OpenSearch Service targets.

Support for Oracle extended data types

Oracle extended data types for both Oracle source and targets are now supported.

Increased number of Amazon DMS resources per account

The limit on the number of Amazon DMS resources you can create has increased. For more information, see Quotas for Amazon Database Migration Service.

The issues resolved in Amazon DMS 3.3.3 include the following:

  • Fixed an issue where a task crashes using a specific update statement with Parallel Apply in Amazon Kinesis.

  • Fixed an issue where a task crashes on the ALTER TABLE statement with Amazon S3 as a target.

  • Fixed an issue where values on polygon columns are truncated when using Microsoft SQL Server as a source.

  • Fixed an issue on Unicode converter of JA16SJISTILDE and JA16EUCTILDE when using Oracle as a source.

  • Fixed an issue where MEDIUMTEXT and LONGTEXT columns failed to migrate from MySQL to S3 comma-separated value (CSV) format.

  • Fixed an issue where boolean columns were transformed to incorrect types with Apache Parquet output.

  • Fixed an issue with extended varchar columns in Oracle.

  • Fixed an issue where data validation tasks failed due to certain timestamp combinations.

  • Fixed an issue with Sybase data definition language (DDL) replication.

  • Fixed an issue involving an Oracle Real Application Clusters (RAC) source crashing with Oracle Binary Reader.

  • Fixed an issue with validation for Oracle targets with schema names' case.

  • Fixed an issue with validation of IBM Db2 versions 9.7 and 10.

  • Fixed an issue for a task not stopping two times with StopTaskCachedChangesApplied and StopTaskCachedChangesNotApplied enabled.