[C++][CI] gcc-16 MinGW failures - remaining fixes (follow-up to #49930)

[tadeja]

Describe the bug, including details regarding any error messages, version, and platform.

Update:
Diagnostic CI run confirms a single upstream root cause for all failures listed here.
Filed upstream as GCC #125312 (diagnosis: #49961). Fix in upstream MSYS2 msys2/MINGW-packages#29502

Umbrella issue for follow-up to closed #49930/merged PR #49931 (a temporary workaround!).
The MSYS gcc 15.2 -> 16.1 update post 2026 April 30 broke the MINGW64 MinGW job. #49931 pins gcc-libs to 15.2 to mask/workaround it. This issue tracks the actual fixes so the pin step can be reverted (later with #49945).
Sub-issues will be opened for individual failures as triage progresses.

First failure on main:

The following tests FAILED:
	 41 - arrow-async-utility-test (Exit code 0xc0000374)   arrow-tests unittest
	 62 - arrow-dataset-dataset-writer-test (SEGFAULT)      arrow_dataset unittest
	 65 - arrow-dataset-file-test (Failed)                  arrow_dataset unittest
Error: Process completed with exit code 8.

Another CI run from PR 49462

Empirical isolation (PR #49931 commit fb4d7a0, pins removed, MinIO bumped) - MINGW64 CI run:

The following tests FAILED:
	 41 - arrow-async-utility-test (Exit code 0xc0000374)   arrow-tests unittest
	 44 - arrow-threading-utility-test (Timeout)            arrow-tests unittest
	 62 - arrow-dataset-dataset-writer-test (Failed)        arrow_dataset unittest
	 65 - arrow-dataset-file-test (Failed)                  arrow_dataset unittest
	 76 - arrow-s3fs-test (Timeout)                         arrow-tests filesystem unittest

CLANG64 in the same run: 100% tests passed. Confirmed gcc-16 is the trigger (not the gcc-libs cascade or aws-sdk-cpp/MinIO behavior).

---

Per-test status:

Root cause: GCC 125312, diagnosis: #49961, verifying CI run: https://github.com/apache/arrow/actions/runs/25916033144/job/76180927231#step:14:494

• arrow-async-utility-test was: Heap corruption 0xc0000374 in PushGenerator.Stress (sometimes passed - on 2nd attempt CI log there, failed 3 times there)

• arrow-threading-utility-test was: SegFault log1 or Timeout 300.00 sec log2 (race signature)

• arrow-dataset-dataset-writer-test was: std::bad_weak_ptr in BatchWriteConcurrent

• arrow-dataset-file-test was: SegFault, same race as arrow-dataset-dataset-writer-test

• (intermittent) arrow-acero-hash-join-node-test was: HashJoin.Random (parallel = true) doesn't emit the expected number of batches. Ex. CI failed logs here1 and here2

• (Update: NOT a distinct root cause) arrow-flight-test was: gRPC's memory_quota.cc:280 CHECK failure in LargeBatch tests, also at memory_quota.cc:281, byte mismatches vary across retries . See in Example log1, log2 and new log3 - race from gRPC's shared_ptr and buggy __exchange_and_add in libstdc++-6.dll. Needs the mingw-w64-x86_64-gcc-libs 15.2 pin (or better the MSYS2 rebuild, see gcc: Fix atomicity of reference counters in std::{shared,weak}_ptr msys2/MINGW-packages#29502) as gRPC.dll was pre-compiled by MSYS2 against the broken DLL
  -- (earlier) original [C++] Linux Flight/FlightSQL test randomly segfaults after gRPC version bump #49767 describes non-MinGW Windows (MSVC/ODBC) that might still be a separate gRPC issue (bump in GH-48926: [C++] Upgrade Abseil/Protobuf/GRPC/Google-Cloud-CPP bundled versions #48964).

(fixed)

• arrow-json-test : MultipleChunksParallel crash - Fixed by GH-49272: [C++][CI] Fix intermittent segfault in arrow-json-test with MinGW #49462
• arrow-s3fs-test : _fut.Wait() timeout in GetFileInfoGeneratorStress. Observed failing and passing. Same thread-pool class as others; the SDK-side MissingContentMD5 mode was separately resolved by GH-49930: [CI][C++] Pin MinGW MSYS2 packages to unblock CI #49931's MinIO bump.

---

This issue closes when

• GCC 125312 is fixed upstream or sooner with MSYS2 publishing related fix ( gcc: Fix atomicity of reference counters in std::{shared,weak}_ptr msys2/MINGW-packages#29502 - merged 2026 May 15! )
• all affected tests pass on the current upstream MSYS2 packages
• at that point GH-49948: [CI][C++] Revert PR 49931 (Pin MinGW MSYS2 packages) but keep bumped minIO version #49945 reverts GH-49930: [CI][C++] Pin MinGW MSYS2 packages to unblock CI #49931's temporary pin step (- name: Pin MSYS2 packages)

Component(s)

Continuous Integration, C++

[tadeja]

Just a short overview here of CI logs from #49945

XX - test : Previous run 2026 May 07, Last run 1st attempt 2026 May 08, 2nd attempt 2026 May 08, 3rd attempt 2026 May 10 (only with debug enabled)
41 - arrow-async-utility-test : pass, heap corruption, heap corruption, heap corruption
44 - arrow-threading-utility-test : Timeout, Timeout, SegFault, Timeout, SegFault/0xc0000374
62 - arrow-dataset-dataset-writer-test : Failed, Failed, Failed, SegFault
65 - arrow-dataset-file-test : Failed, Failed, Failed, SegFault
76 - arrow-s3fs-test : pass, pass, pass, pass
78 - arrow-flight-test : Failed, Failed, Failed, Failed

Edit newer run 2026 May 12
46 - arrow-csv-test (SEGFAULT) : failure happens in AsyncReaderTests.Stress -