How do I troubleshoot AWS DMS connectivity failures when I replicate from Oracle as source databases?

Resolution

If you have connectivity failures when you replicate from Oracle as a source database, then use AWS DMS diagnostic support Amazon Machine Image (AMI) to troubleshoot the network-related issues.

Troubleshoot latency

If your database experiences full load and Change Data Capture (CDC) when you migrate large objects (LOBs), then AWS DMS might experience poor performance. LOB data requires two-phase migration with additional round trips. If AWS DMS experiences high-network latency and extra LOB lookup operations, then the CDC experiences latency spikes.

To identify log entries that show LOB value retrieval from your source database to target database, specify the logging actions in the JSON file.

Example:

[TARGET_APPLY]T: Going to fill lob data for columns in table table_schema.table_name.

Note: Replace table_schema with your table schema and table_name with your table name.

Troubleshoot endpoint disconnections

If AWS DMS experiences frequent endpoint disconnections, then the AWS DMS task activates retry loops. If your source database connections drop, then you get the following error messages in AWS DMS task logs:

"[SOURCE_CAPTURE]E: ORA-03113: end-of-file on communication channel Process ID: 9999 Session ID: 777 Serial number: 12341"

"[SOURCE_CAPTURE]I: Disconnected Oracle database"

"[SOURCE_CAPTURE]E: Oracle communication failure"

"[TASK_MANAGER]E: Task 'ABCDEFGHIJ12345' encountered a recoverable error, retry attempt # n"

If you still have disconnections, then launch the AWS DMS diagnostic support AMI and perform the network diagnostic troubleshooting tests.

Troubleshoot table data

If your tasks get stuck when you retrieve table metadata and you used a wildcard (%) in table mapping, then the metadata capture fails and you get the following error message:

"[METADATA_MANAGE]W: Failed to prepare get capture list statement"

To resolve this error, launch the AMI in the same region and virtual private cloud (VPC), and within the same subnet as your replication instance to simulate the same networking environment as the AWS DMS replication instance.