Skip to content

Add DmsReloadTablesOperator - #70569

Merged
o-nikolas merged 8 commits into
apache:mainfrom
AlejandroMorgante:add-dms-reload-tables-operator
Jul 30, 2026
Merged

Add DmsReloadTablesOperator#70569
o-nikolas merged 8 commits into
apache:mainfrom
AlejandroMorgante:add-dms-reload-tables-operator

Conversation

@AlejandroMorgante

Copy link
Copy Markdown
Contributor

Add DmsReloadTablesOperator to reload selected target tables for an active AWS DMS replication task.

The operator wraps the AWS DMS ReloadTables API and supports:

  • data-reload and validate-only reload options.
  • Synchronous and deferrable completion waits.
  • Configurable waiter delay and maximum attempts.
  • Templated replication task, table, and reload-option arguments.

The DMS system test now creates a full-load-and-cdc task with PostgreSQL logical replication enabled, waits for the initial load, and exercises a deferrable table reload against AWS.

Testing:

  • System test validated end-to-end against a real AWS environment.
  • DMS unit tests: 131 passed.

References:


Was generative AI tooling used to co-author this PR?
  • Yes — Codex (GPT-5)

Generated-by: Codex (GPT-5) following the guidelines

@AlejandroMorgante AlejandroMorgante changed the title Add AWS DMS table reload operator Add DmsReloadTablesOperator Jul 28, 2026
Comment thread providers/amazon/src/airflow/providers/amazon/aws/hooks/dms.py Outdated
Comment thread providers/amazon/src/airflow/providers/amazon/aws/operators/dms.py Outdated
@AlejandroMorgante
AlejandroMorgante marked this pull request as draft July 28, 2026 21:45
@AlejandroMorgante
AlejandroMorgante marked this pull request as ready for review July 28, 2026 21:45
Airflow users need to rerun full loads for selected tables in active replication tasks without implementing custom boto3 tasks.
AWS no longer offers dms.t3.micro in us-east-1, preventing the system test from reaching the operator under test.
Real AWS validation requires a full-load-and-CDC task, whose PostgreSQL source must expose logical changes and commit its seed data before replication starts.
AWS can report Before load while a table awaits its full load, so the completion waiter should classify that state explicitly as transient.
The provider documentation spellcheck rejects the conjugated form, preventing the Amazon documentation job from completing.
A one-to-one wrapper expands the public hook surface without adding reusable behavior.
Both DMS reload strategies are asynchronous, so wait_for_completion should provide consistent semantics for validate-only as well as data-reload.
Data validation requires a database target, and end-to-end coverage should include validate-only completion as well as data reload completion.
@AlejandroMorgante
AlejandroMorgante force-pushed the add-dms-reload-tables-operator branch from aba60d6 to 5f708ab Compare July 29, 2026 16:43
@o-nikolas
o-nikolas merged commit ffe077e into apache:main Jul 30, 2026
88 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants