Skip to content

[DSIP-95][API] Complete the functionality of using dependencies in the complement data#18003

Open
det101 wants to merge 25 commits intoapache:devfrom
det101:DSIP-95
Open

[DSIP-95][API] Complete the functionality of using dependencies in the complement data#18003
det101 wants to merge 25 commits intoapache:devfrom
det101:DSIP-95

Conversation

@det101
Copy link
Copy Markdown
Contributor

@det101 det101 commented Feb 27, 2026

Was this PR generated or assisted by AI?

close #17748

Purpose of the pull request

Brief change log

Verify this pull request

This pull request is code cleanup without any test coverage.

(or)

This pull request is already covered by existing tests, such as (please describe tests).

(or)

This change added tests and can be verified as follows:

(or)

Pull Request Notice

Pull Request Notice

If your pull request contains incompatible change, you should also add it to docs/docs/en/guide/upgrade/incompatible.md

Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Implements downstream workflow triggering for complement/backfill runs in the API layer, adding support for “trigger dependent workflows” behavior and accompanying unit tests.

Changes:

  • Implemented doBackfillDependentWorkflow to fetch downstream workflow definitions and trigger backfill runs for them.
  • Added visited-code tracking intended to prevent self/cyclic triggering and duplicate downstream triggers.
  • Added BackfillWorkflowExecutorDelegateTest with basic scenarios for downstream triggering and filtering.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File Description
dolphinscheduler-api/src/main/java/org/apache/dolphinscheduler/api/executor/workflow/BackfillWorkflowExecutorDelegate.java Adds dependent workflow backfill triggering logic and wiring for lineage + workflow definition lookup.
dolphinscheduler-api/src/test/java/org/apache/dolphinscheduler/api/executor/workflow/BackfillWorkflowExecutorDelegateTest.java Adds unit tests for the new dependent backfill triggering logic (single-hop scenarios).

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copy link
Copy Markdown
Member

@SbloodyS SbloodyS left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You didn't fill in the content according to the PR template, please fix it.

Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.


💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

luxl and others added 6 commits March 25, 2026 16:51
…endent workflow

- Problem 1: Replace single-level self-dep check with ThreadLocal visited set
  to detect and skip indirect circular dependencies (A→B→A), preventing
  StackOverflowError when allLevelDependent=true
- Problem 2: Set startNodes=null for downstream workflows; upstream task node
  codes are not valid in a different workflow definition
- Add tests for OFFLINE skip, not-found skip, and startNodes null assertion

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…tion

ThreadLocal is unnecessary here since the call chain is synchronous and
private. Passing visitedCodes as a parameter is simpler, clearer, and
avoids ThreadLocal lifecycle management.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Align Java source and test formatting with project spotless rules to keep style checks consistent.

Made-with: Cursor
Ensure dependent backfill triggering is evaluated per parallel date chunk, and batch-load downstream workflow definitions to avoid N+1 queries. Add regression coverage for parallel mode visited-code isolation.

Made-with: Cursor
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

Comments suppressed due to low confidence (1)

dolphinscheduler-api/src/main/java/org/apache/dolphinscheduler/api/executor/workflow/BackfillWorkflowExecutorDelegate.java:119

  • In parallel mode, expectedParallelismNumber can be 0 (the validator only rejects values < 0). If it is 0, splitDateTime(listDate, expectedParallelismNumber) will divide by zero and throw ArithmeticException. Treat 0 the same as null (default to listDate.size()), or explicitly guard against <= 0 before calling splitDateTime.
        final BackfillWorkflowDTO.BackfillParamsDTO backfillParams = backfillWorkflowDTO.getBackfillParams();
        Integer expectedParallelismNumber = backfillParams.getExpectedParallelismNumber();

        List<ZonedDateTime> listDate = backfillParams.getBackfillDateList();
        if (expectedParallelismNumber != null) {
            expectedParallelismNumber = Math.min(listDate.size(), expectedParallelismNumber);
        } else {
            expectedParallelismNumber = listDate.size();
        }

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comment on lines +413 to +444
public void testDoParallelBackfillWorkflow_ShouldIsolateVisitedCodesAcrossChunks() {
long upstreamCode = 500L;
WorkflowDefinition upstreamWorkflow =
WorkflowDefinition.builder().code(upstreamCode).releaseState(ReleaseState.ONLINE).build();
List<ZonedDateTime> dates = Arrays.asList(
ZonedDateTime.parse("2026-02-01T00:00:00Z"),
ZonedDateTime.parse("2026-02-02T00:00:00Z"),
ZonedDateTime.parse("2026-02-03T00:00:00Z"));
BackfillWorkflowDTO.BackfillParamsDTO params = BackfillWorkflowDTO.BackfillParamsDTO.builder()
.runMode(RunMode.RUN_MODE_PARALLEL)
.backfillDateList(dates)
.expectedParallelismNumber(2)
.backfillDependentMode(ComplementDependentMode.ALL_DEPENDENT)
.allLevelDependent(true)
.executionOrder(ExecutionOrder.ASC_ORDER)
.build();
BackfillWorkflowDTO dto = BackfillWorkflowDTO.builder()
.workflowDefinition(upstreamWorkflow)
.backfillParams(params)
.build();
Set<Long> baseVisitedCodes = new HashSet<>(Collections.singleton(upstreamCode));
List<Set<Long>> visitedSnapshotPerChunk = new java.util.ArrayList<>();

doAnswer(invocation -> {
Set<Long> chunkVisited = invocation.getArgument(2);
visitedSnapshotPerChunk.add(new HashSet<>(chunkVisited));
chunkVisited.add(9000L + visitedSnapshotPerChunk.size());
return null;
}).when(backfillWorkflowExecutorDelegate).doBackfillDependentWorkflowForTesting(any(), any(), any());

List<Integer> result = backfillWorkflowExecutorDelegate.executeWithVisitedCodes(dto, baseVisitedCodes);

Copy link

Copilot AI Mar 25, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This test calls executeWithVisitedCodes, which will run the real doBackfillWorkflow and attempt to use registryClient / Clients.withService(IWorkflowControlClient) to contact a master. Since neither is mocked/stubbed in this test, it will fail with NPE or a ServiceException before exercising the visited-codes isolation assertions. Consider refactoring to unit-test the chunk visited-code cloning without invoking the master trigger, or add a test seam/mocking for the backfill trigger step.

Copilot uses AI. Check for mistakes.
Delete brittle reflection-based tests for private doBackfillDependentWorkflow; keep higher-level parallel visited-code isolation.

Made-with: Cursor
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 4 comments.


💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

luxl added 2 commits March 26, 2026 10:08
Use DependentWorkflowDefinition routing fields for dependent triggers, and force dependent runMode SERIAL when upstream is PARALLEL to traverse dependencies once per date chunk.

Made-with: Cursor
Cover all-level recursion gating, cycle skipping, and offline/missing downstream skipping.

Made-with: Cursor
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.


💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

det101 and others added 3 commits March 26, 2026 10:23
…/api/executor/workflow/BackfillWorkflowExecutorDelegate.java

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Group dependent definitions by workflowDefinitionCode and aggregate distinct startNodes to avoid skipping nodes when visitedCodes blocks duplicate entries.

Made-with: Cursor
Pass new HashSet instead of null for initial visitedCodes to simplify downstream handling.

Made-with: Cursor
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 4 comments.


💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

det101 and others added 5 commits March 26, 2026 10:59
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
…/api/executor/workflow/BackfillWorkflowExecutorDelegate.java

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
…/api/executor/workflow/BackfillWorkflowExecutorDelegate.java

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 4 comments.


💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

@sonarqubecloud
Copy link
Copy Markdown

Quality Gate Failed Quality Gate failed

Failed conditions
0.0% Coverage on New Code (required ≥ 60%)

See analysis details on SonarQube Cloud

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.

[DSIP-95][API] Complete the functionality of using dependencies in the complement data

3 participants