[ARROW-15676] [R] New memory leaks in valgrind nightly build

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Duplicate
• Component/s: R
• External issue URL: https://github.com/apache/arrow/issues/31130
• Assignee: Unassigned
• Reporter: Dewey Dunnington
• Created: 14/Feb/22 13:56
• Updated: 11/Jan/23 11:38
• Resolved: 14/Feb/22 14:08

Description

The 'valgrind' nightly build has been failing for R since February 1 (the January 26 build passed). The error that currently causes the failure is:

==10264== 12,959 (256 direct, 12,703 indirect) bytes in 1 blocks are definitely lost in loss record 1,755 of 3,746
==10264==    at 0x483BE63: operator new(unsigned long) (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==10264==    by 0xF165862: allocate (new_allocator.h:115)
==10264==    by 0xF165862: allocate (alloc_traits.h:460)
==10264==    by 0xF165862: __allocate_guarded<std::allocator<std::_Sp_counted_ptr_inplace<arrow::MergedGenerator<arrow::dataset::EnumeratedRecordBatch>::State, std::allocator<arrow::MergedGenerator<arrow::dataset::EnumeratedRecordBatch>::State>, (__gnu_cxx::_Lock_policy)2> > > (allocated_ptr.h:97)
==10264==    by 0xF165862: __shared_count<arrow::MergedGenerator<arrow::dataset::EnumeratedRecordBatch>::State, std::allocator<arrow::MergedGenerator<arrow::dataset::EnumeratedRecordBatch>::State>, std::function<arrow::Future<std::function<arrow::Future<arrow::dataset::EnumeratedRecordBatch>()> >()>, int&> (shared_ptr_base.h:680)
==10264==    by 0xF165862: __shared_ptr<std::allocator<arrow::MergedGenerator<arrow::dataset::EnumeratedRecordBatch>::State>, std::function<arrow::Future<std::function<arrow::Future<arrow::dataset::EnumeratedRecordBatch>()> >()>, int&> (shared_ptr_base.h:1371)
==10264==    by 0xF165862: shared_ptr<std::allocator<arrow::MergedGenerator<arrow::dataset::EnumeratedRecordBatch>::State>, std::function<arrow::Future<std::function<arrow::Future<arrow::dataset::EnumeratedRecordBatch>()> >()>, int&> (shared_ptr.h:408)
==10264==    by 0xF165862: allocate_shared<arrow::MergedGenerator<arrow::dataset::EnumeratedRecordBatch>::State, std::allocator<arrow::MergedGenerator<arrow::dataset::EnumeratedRecordBatch>::State>, std::function<arrow::Future<std::function<arrow::Future<arrow::dataset::EnumeratedRecordBatch>()> >()>, int&> (shared_ptr.h:860)
==10264==    by 0xF165862: make_shared<arrow::MergedGenerator<arrow::dataset::EnumeratedRecordBatch>::State, std::function<arrow::Future<std::function<arrow::Future<arrow::dataset::EnumeratedRecordBatch>()> >()>, int&> (shared_ptr.h:876)
==10264==    by 0xF165862: MergedGenerator (async_generator.h:996)
==10264==    by 0xF165862: std::function<arrow::Future<arrow::dataset::EnumeratedRecordBatch> ()> arrow::MakeMergedGenerator<arrow::dataset::EnumeratedRecordBatch>(std::function<arrow::Future<std::function<arrow::Future<arrow::dataset::EnumeratedRecordBatch> ()> > ()>, int) (async_generator.h:1185)
==10264==    by 0xF14A2DF: arrow::dataset::(anonymous namespace)::MakeScanNode(arrow::compute::ExecPlan*, std::vector<arrow::compute::ExecNode*, std::allocator<arrow::compute::ExecNode*> >, arrow::compute::ExecNodeOptions const&) (scanner.cc:794)
==10264==    by 0xF10BF13: std::_Function_handler<arrow::Result<arrow::compute::ExecNode*> (arrow::compute::ExecPlan*, std::vector<arrow::compute::ExecNode*, std::allocator<arrow::compute::ExecNode*> >, arrow::compute::ExecNodeOptions const&), arrow::Result<arrow::compute::ExecNode*> (*)(arrow::compute::ExecPlan*, std::vector<arrow::compute::ExecNode*, std::allocator<arrow::compute::ExecNode*> >, arrow::compute::ExecNodeOptions const&)>::_M_invoke(std::_Any_data const&, arrow::compute::ExecPlan*&&, std::vector<arrow::compute::ExecNode*, std::allocator<arrow::compute::ExecNode*> >&&, arrow::compute::ExecNodeOptions const&) (invoke.h:60)
==10264==    by 0xEFBDCB4: std::function<arrow::Result<arrow::compute::ExecNode*> (arrow::compute::ExecPlan*, std::vector<arrow::compute::ExecNode*, std::allocator<arrow::compute::ExecNode*> >, arrow::compute::ExecNodeOptions const&)>::operator()(arrow::compute::ExecPlan*, std::vector<arrow::compute::ExecNode*, std::allocator<arrow::compute::ExecNode*> >, arrow::compute::ExecNodeOptions const&) const (std_function.h:622)
==10264==    by 0xEFBC18F: arrow::compute::MakeExecNode(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, arrow::compute::ExecPlan*, std::vector<arrow::compute::ExecNode*, std::allocator<arrow::compute::ExecNode*> >, arrow::compute::ExecNodeOptions const&, arrow::compute::ExecFactoryRegistry*) (exec_plan.h:329)
==10264==    by 0xEFB5F6E: MakeExecNodeOrStop(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, arrow::compute::ExecPlan*, std::vector<arrow::compute::ExecNode*, std::allocator<arrow::compute::ExecNode*> >, arrow::compute::ExecNodeOptions const&) (compute-exec.cpp:52)
==10264==    by 0xEFB7426: ExecNode_Scan(std::shared_ptr<arrow::compute::ExecPlan> const&, std::shared_ptr<arrow::dataset::Dataset> const&, std::shared_ptr<arrow::compute::Expression> const&, std::vector<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::allocator<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > > >) (compute-exec.cpp:157)
==10264==    by 0xEF3DFE1: _arrow_ExecNode_Scan (arrowExports.cpp:1621)
==10264==    by 0x4942B4F: R_doDotCall (dotcode.c:610)
==10264==    by 0x49B2C97: bcEval (eval.c:7695)
==10264==    by 0x498DBC8: Rf_eval (eval.c:748)

First failure: https://dev.azure.com/ursacomputing/crossbow/_build/results?buildId=19221&view=logs&j=0da5d1d9-276d-5173-c4c4-9d4d4ed14fdb&t=d9b15392-e4ce-5e4c-0c8c-b69645229181&l=20910

Most recent failure: https://dev.azure.com/ursacomputing/crossbow/_build/results?buildId=19759&view=logs&j=0da5d1d9-276d-5173-c4c4-9d4d4ed14fdb&t=d9b15392-e4ce-5e4c-0c8c-b69645229181&l=21089

Issue Links

• duplicates: ARROW-15656 [C++] [R] Make valgrind builds slightly quicker