branch-4.1: [fix](filecache) fix flaky be UT for LRU dump (pick#65427)

[freemandealer]

What problem does this PR solve?

Issue Number: None

Related PR: None

Problem Summary: BlockFileCacheTest.version3_add_remove_restart relied on sleeping for the background LRU dump thread after replaying remove records. If the background dump did not run before the restart phase, stale LRU tail records could be restored and the restarted cache could report non-empty queues. The test now explicitly dumps LRU queues after replaying the remove logs, so persisted LRU state is synchronized before the restart check without changing file cache runtime code.

Release note

None

Check List (For Author)

• Test:
  • Unit Test: DORIS_TOOLCHAIN=clang DISABLE_BE_JAVA_EXTENSIONS=ON ENABLE_INJECTION_POINT=ON ENABLE_CACHE_LOCK_DEBUG=0 ENABLE_PCH=0 sh run-be-ut.sh --run --filter=BlockFileCacheTest.version3_add_remove_restart
  • Style check: git diff --check -- be/test/io/cache/block_file_cache_test_meta_store.cpp
• Behavior changed: No
• Does this need documentation: No

[hello-stephen]

Thank you for your contribution to Apache Doris.
Don't know what should be done next? See How to process your PR.

Please clearly describe your PR:

1. What problem was fixed (it's best to include specific error reporting information). How it was fixed.
2. Which behaviors were modified. What was the previous behavior, what is it now, why was it modified, and what possible impacts might there be.
3. What features were added. Why was this function added?
4. Which code was refactored and why was this part of the code refactored?
5. Which functions were optimized and what is the difference before and after the optimization?

[freemandealer]

run buildall

[freemandealer]

/review

[github-actions]

Automated review summary for PR 65425 at 7c0c876.

I found no blocking or inline-worthy issues in this patch.

Critical checkpoints:

• Scope/head: verified the live PR still points to head 7c0c876 on base 7126cf6, with one changed file: be/test/io/cache/block_file_cache_test_meta_store.cpp.
• Existing review context: no existing inline review comments or replies were present, so there were no duplicate issues to suppress.
• Test behavior: the changed test now calls cache.dump_lru_queues(true) after replay_lru_logs_once() has drained the pending removal records and before the restart phase. The related implementation synchronously dumps the shadow LRU queues under _dump_lru_queues_mtx, so this matches the persisted restart state the test is validating and removes the previous timing dependency on the background dump interval.
• Config/session compatibility: the modified test already saves and restores the affected file-cache config values through Defer, and this patch does not touch runtime configuration propagation or production behavior.
• Validation: git diff --check FETCH_HEAD..HEAD -- be/test/io/cache/block_file_cache_test_meta_store.cpp passed. I did not run the targeted BE UT in this runner because thirdparty/installed and thirdparty/installed/bin/protoc are absent.
• User focus: no additional user-provided review focus was specified.

Subagent conclusions:

• optimizer-rewrite: found no optimizer/rewrite, semantic-equivalence, join, aggregate, or parallel plan-path issue; convergence round 1 returned NO_NEW_VALUABLE_FINDINGS.
• tests-session-config: found no regression-test, expected-output, session/config propagation, compatibility, or basic style issue; convergence round 1 returned NO_NEW_VALUABLE_FINDINGS.

Overall opinion: the patch is a narrowly scoped deterministic test fix and looks good from static review.

[hello-stephen]

BE Regression && UT Coverage Report

Increment line coverage 100% (0/0) 🎉

Increment coverage report
Complete coverage report

Category	Coverage
Function Coverage	72.06% (26934/37375)
Line Coverage	55.16% (287291/520847)
Region Coverage	52.69% (240143/455781)
Branch Coverage	53.75% (103558/192649)