Issue with CDC while replicating partition tables from Aurora PostgreSQL to Snowflake

[atk]

Source and Target Endpoint Details:

Source endpoint: 

AWS Aurora Cloud for PostgreSQL

Tables in PostgreSQL are both partitioned and non-partitioned.

Target Endpoint:

Snowflake on AWS

Issue Description:

We observed that the Qlik Replicate task consistently completes the full load but fails intermittently during CDC, specifically when reading metadata from PostgreSQL’s WAL (Write-Ahead Log).
The logs indicate that WAL OID resolution issues and object ambiguity caused by duplicate table names across schemas are preventing Replicate from capturing CDC changes.

We escalated this to the Qlik vendor support team. Their initial analysis suggests this may be a potential Replicate product defect, and they have recommended opening a formal support ticket for deeper investigation. 

[john_wang]

Hello @atk ,

Seems we’re not aware of any such defect. Could you please help by doing the following:

1. Set source_capture to Verbose and reproduce the issue.
2. Upload the task Diagnostics Package.
3. Decrypt the verbose task log file and attach it to the support ticket as well.

Thank you for your cooperation.

John.

[atk]

Hi @john_wang 

Support case #00437802 was created which have logs and diagnostic package attached.

00437802

[john_wang]

Thank you for your cooperation! @atk Our support team will keep you posted.