shuke987 opened a new pull request, #65461: URL: https://github.com/apache/doris/pull/65461
### What problem does this PR solve? Issue Number: None Related PR: #64559 Problem Summary: The TPC-DS SF1000 DPHyp query85 case checks an exact physical plan shape. It already pins the enumeration-count budget with `dphyper_limit=2600`, but it leaves the independent wall-clock `join_reorder_time_limit` at its 1000 ms default. `PlanReceiver` stops full enumeration when either limit is exceeded. When that happens, `JoinOrderJob` invokes `GraphSimplifier` and re-enumerates a reduced candidate space, which can select a different valid join order and fail the exact shape golden. The TeamCity timing split matches that boundary: | Build | Shape result | Time from `Execute tag` to generated output | First differing join | |---|---|---:|---| | [991071](http://43.132.222.7:8111/viewLog.html?buildId=991071&buildTypeId=Doris_DorisRegression_P0Regression) | PASS | 864 ms | expected `web_page` join | | [989238](http://43.132.222.7:8111/viewLog.html?buildId=989238&buildTypeId=Doris_DorisRegression_P0Regression) | FAILURE | 1378 ms | `customer_address` join | | [989267](http://43.132.222.7:8111/viewLog.html?buildId=989267&buildTypeId=Doris_DorisRegression_P0Regression) | FAILURE | 1351 ms | `cd1` join | | [989379](http://43.132.222.7:8111/viewLog.html?buildId=989379&buildTypeId=Doris_DorisRegression_P0Regression) | FAILURE | 1444 ms | `customer_address` join | | [991088](http://43.132.222.7:8111/buildConfiguration/Doris_DorisRegression_P0Regression/991088?buildTab=tests&expandedTest=build%3A%28id%3A991088%29%2Cid%3A2000003392) | FAILURE | 1361 ms | `date_dim` join | The four genuine join-order failures all crossed the approximately one-second boundary. The current failing PR #65434 changes only file-cache code, so the failure is not caused by its product diff. Separate qualifier-only failures from PRs #65346 and #65264 are not part of this diagnosis. This PR pins `join_reorder_time_limit=10000` for query85 only. It keeps `dphyper_limit=2600`, the SQL, and the expected golden unchanged, so the test continues to validate the intended DPHyp plan while avoiding load-dependent fallback. Validation: - `git diff --check` - Verified the patch changes only the query85 session setup. - Local single-suite execution was not run because no Doris regression environment was allocated for this task. - `run buildall` and the exact P0 case will provide CI validation on the patched head. ### Release note None ### Check List (For Author) - Test - [x] Regression test - [ ] Unit Test - [ ] Manual test (add detailed scripts or steps below) - [ ] No need to test or manual test. Explain why: - [ ] This is a refactor/code format and no logic has been changed. - [ ] Previous test can cover this change. - [ ] No code files have been changed. - [ ] Other reason - Behavior changed: - [x] No. - [ ] Yes. - Does this need documentation? - [x] No. - [ ] Yes. ### Check List (For Reviewer who merge this PR) - [ ] Confirm the release note - [ ] Confirm test cases - [ ] Confirm document - [ ] Add branch pick label -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
