Known Issues for Oracle Database Migration Service

Learn about the issues you may encounter when using DMS and how to work around them.

Migrations

PRGC-1024 thrown on resuming a paused migration job with Oracle RDS source

Issue: An offline or online migration job, migrating a database from Oracle RDS source, fails with the following error upon resuming the migration:

PRGC-1024 : failed to create a reverse connection NAT IP for private endpoint "ocid1.privateendpoint.oc1.phx.aaaaa<.....>" and customer IP address "null"

Solution: Do not use the pause and resume features during a migration job with an Oracle RDS source.

ORA-29280 thrown during Amazon RDS migration with S3 bucket storage

Issue: Amazon RDS migration with S3 bucket storage fails with ORA-29280 if you specify the export directory name in lower case.

Solution: Update the configuration to specify the export directory name in upper case.

PRGG-1043 : No heartbeat table entries were found for Oracle GoldenGate Replicat process

Issue: A migration job can report error PRGG-1043: No heartbeat table entries were found for Oracle GoldenGate Replicat process process_name due to one of the following causes:

  1. Initialization parameter job_queue_processes was set to zero in the source or target database.

    Solution: Run the following statements on the database: 

    show parameter job_queue_processes;
alter system set job_queue_processes=100 scope=both;
exec dbms_scheduler.set_scheduler_attribute('SCHEDULER_DISABLED','FALSE');

  2. Scheduled job GG_UPDATE_HEARTBEATS is not active in the source database.

    Solution: See Oracle GoldenGate documentation Understanding Heartbeat Table End-To-End Replication Flow for details about database scheduler job GG_UPDATE_HEARTBEATS.

  3. The server hosting Oracle GoldenGate deployments has a different time zone than the source database.

    Solution: First, do one of the following solutions:

    • Modify the time zone for the server hosting Oracle GoldenGate deployments, OR

    • Use the web UI for the Oracle GoldenGate deployment to add Extract parameter TRANLOGOPTIONS SOURCE_OS_TIMEZONE and restart Extract.

      For example, if the source database time zone is UTC-5, then set parameter TRANLOGOPTIONS SOURCE_OS_TIMEZONE -5. For more information, see TRANLOGOPTIONS in Reference for Oracle GoldenGate.

    Then, ensure that the DST_PRIMARY_TT_VERSION property in the source database is up to date.

GoldenGate Extract does not populate global_name correctly for Oracle Database 12.1.0.2

Issue: In case of online logical migration, this replication issue will occur if DB_GLOBAL_NAME was modified for a 12.1.0.2 source database.

Cause: SCN-based versioning of DB_GLOBAL_NAME was introduced in Oracle Database 12.2 along with the GSBA table function. In Oracle Database 12.2, LogMiner starts to track the DDL that changes the DB_GLOBAL_NAME and stores the DB_GLOBAL_NAME in the system.logmnrc_gsba table. GSBA table function gets the db_global_name from the logmnrc_gsba cache table. When the table function was backported to Oracle Database 12.1.0.2, DB_GLOBAL_NAME versioning was not backported, and DB_GLOBAL_NAME is not stored in logmnrc_gsba table. The only data available is the initial DB_GLOBAL_NAME/PDB_GLOBAL_NAME in the LOGMNR_DICTIONARY$ table. The GSBA table function gets the DB_GLOBAL_NAME from system.logmnr_dictionary$. This is the cause of the incorrect PDB_GLOBAL_NAME provided by the table function. If the database is a CDB, the table function should get the PDB_GLOBAL_NAME out of the LOGMNR_DCTIONARY$ table.

Solution: The table function needed by Oracle GoldenGate is available in 12.1.0.2 DBBP210119, and the fix to change the query for DB_GLOBAL_NAME is being tracked by RDBMS bug 32770084.

ORA-39001: Invalid argument: Data Pump Export parallelism set to 0 for single-instance database

A migration job for a single-instance database will report error ORA-39001: Invalid argument when performing Data Pump Export.

API-Initiated Offline Migration Fails if DATAPUMPPARAMETERS Property is Empty

Issue: A REST API request without any values specified under dataPumpParameters parameter causes none of the "datapumpSettings"."dataPumpParameters" properties to be defined.

Here the issue is that the REST API request processor is not instantiating the payload sub-properties for the parent parameter class that are not defined.

This issue is not seen in a Database Migration UI requested migration job, as the payload submitted has all sub-properties defined with at least 'null'.

Solution: Specify the following property in the input API payload. 

"datapumpSettings": {
"dataPumpParameters":
{
"isCluster":true
}
}

Schema does not get created for object-level include rule

Issue: When creating an object-level include rule to select a specific object for inclusion, such as a table MyUser.MyTable, the user schema does not get created in the target database.

Solution: Create the schema manually prior to migration, and grant the necessary quota in required tablespaces.

Schema level exclude rules are rejected

Issue: When creating a schema-level exclude rule for a specific schema, such as Owner Name=”MySchema”, Object Name=”.*”, Object Type=”VIEW”, the rule causes a validation error:

“Specified DatabaseObject owner name parameter is not a general wildcard (".*")”.

Solution: Use Owner Name=”.*” to exclude the type for all schemas in the migration, and ensure that the object Owner Name value is entered in uppercase letters.

You enter the owner name in uppercase letters to exclude object types other than TABLE or ALL. If you use lowercase letters for the owner name, object types other than TABLE or ALL are ignored.

For example, if you use the following settings then the Object Type PROCEDURE is not excluded.

Action: Exclude

Owner Name=MySchema

Object Name=GREETINGS.*

Object Type=PROCEDURE

Cannot change Migration Type of existing migration

Issue: Changing the Migration Type from OFFLINE to ONLINE, or the reverse, for an existing migration fails.

Solution: Recreate the migration with the desired migration type.

Downloading Data Pump log fails before migration is finished

Issue: Trying to download the Data Pump logs for the “Export Initial Load” phase might fail if the migration is still ongoing.

Solution: Wait until after the migration is finished before downloading the logs.

IBM AIX database link not available

The IBM AIX source environment is now available in both online and offline modes. The database linking feature is not enabled for the online mode.

Issue:

The DBLINK feature is currently unavailable for the online mode of the IBM AIX source environment.

Registered Databases

FQDN host names are not supported in database connect strings

Issue: Fully qualified domain names are currently not supported in database connect strings. Only IP addresses are supported. When selecting a cloud database such as DOCS or Exacts is selected in the Database Migration console, a connect string with FQDN might be provided as default.

Solution: The user needs to determine the IP address for the database and replace the FQDN with the IP address.

Agents

Database Migration Agent goes into FAILED state after successful migration

Issue: The Database Migration agent might go into a FAILED state after being idle because of heartbeat timing issues. An agent is not shown in the Database Migration console when in a FAILED state, but it can be listed using the API or CLI.

Solution: Restart the agent when it goes into a FAILED state.

Documentation Accessibility

For information about Oracle's commitment to accessibility, visit the Oracle Accessibility Program website at http://www.oracle.com/pls/topic/lookup?ctx=acc&id=docacc.

Access to Oracle Support

Oracle customers that have purchased support have access to electronic support through My Oracle Support. For information, visit http://www.oracle.com/pls/topic/lookup?ctx=acc&id=info or visit http://www.oracle.com/pls/topic/lookup?ctx=acc&id=trs if you are hearing impaired.