[ARROW-14219] [R] [CI] DuckDB valgrind failure

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Fix Version/s: 6.0.0
• Component/s: Continuous Integration, R
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/18873
• Assignee: Jonathan Keane
• Reporter: Jonathan Keane
• Created: 05/Oct/21 12:32
• Updated: 11/Jan/23 08:38
• Resolved: 05/Oct/21 19:21

Description

We're getting an error now that we are testing DuckDB:

https://dev.azure.com/ursacomputing/crossbow/_build/results?buildId=12649&view=logs&j=0da5d1d9-276d-5173-c4c4-9d4d4ed14fdb&t=d9b15392-e4ce-5e4c-0c8c-b69645229181

==3189==    by 0x31023018: duckdb_moodycamel::ProducerToken::ProducerToken<std::unique_ptr<duckdb::Task, std::default_delete<duckdb::Task> >, duckdb_moodycamel::ConcurrentQueueDefaultTraits>(duckdb_moodycamel::ConcurrentQueue<std::unique_ptr<duckdb::Task, std::default_delete<duckdb::Task> >, duckdb_moodycamel::ConcurrentQueueDefaultTraits>&) (concurrentqueue.h:3612)
==3189==    by 0x3101A3A4: duckdb::QueueProducerToken::QueueProducerToken(duckdb::ConcurrentQueue&) (task_scheduler.cpp:39)
==3189==    by 0x31023552: std::unique_ptr<duckdb::QueueProducerToken, std::default_delete<duckdb::QueueProducerToken> > duckdb::make_unique<duckdb::QueueProducerToken, duckdb::ConcurrentQueue&>(duckdb::ConcurrentQueue&) (helper.hpp:22)
==3189==    by 0x30FFA807: duckdb::TaskScheduler::CreateProducer() (task_scheduler.cpp:110)
==3189==    by 0x30FF6FC8: duckdb::Executor::Initialize(duckdb::PhysicalOperator*) (executor.cpp:34)
==3189== 
==3189== 984 (448 direct, 536 indirect) bytes in 4 blocks are definitely lost in loss record 1,191 of 4,641
==3189==    at 0x483BE63: operator new(unsigned long) (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==3189==    by 0x30F288C6: std::unique_ptr<duckdb::PreparedStatement, std::default_delete<duckdb::PreparedStatement> > duckdb::make_unique<duckdb::PreparedStatement, char const*>(char const*&&) (helper.hpp:22)
==3189==    by 0x30E7FDF9: duckdb::ClientContext::Prepare(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) (client_context.cpp:376)
==3189==    by 0x30E85BC8: duckdb::Connection::Prepare(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) (connection.cpp:78)
==3189==    by 0x302BF4E4: duckdb::RApi::Prepare(SEXPREC*, SEXPREC*) (statement.cpp:61)
==3189==    by 0x4942217: R_doDotCall (dotcode.c:604)
==3189==    by 0x494D3C9: do_dotcall (dotcode.c:1284)
==3189==    by 0x498CA4E: Rf_eval (eval.c:843)
==3189==    by 0x4992DB8: do_set (eval.c:2982)
==3189==    by 0x498C7F5: Rf_eval (eval.c:815)
==3189==    by 0x499183B: do_begin (eval.c:2530)
==3189==    by 0x498C7F5: Rf_eval (eval.c:815)
==3189== 
==3189== LEAK SUMMARY:
==3189==    definitely lost: 448 bytes in 4 blocks
==3189==    indirectly lost: 536 bytes in 4 blocks
==3189==      possibly lost: 1,055 bytes in 3 blocks
==3189==    still reachable: 334,880,826 bytes in 71,040 blocks
==3189==                       of which reachable via heuristic:
==3189==                         newarray           : 4,264 bytes in 1 blocks
==3189==         suppressed: 0 bytes in 0 blocks
==3189== Reachable blocks (those to which a pointer was found) are not shown.
==3189== To see them, rerun with: --leak-check=full --show-leak-kinds=all

We should report these to the DuckDB folks + ensure that we are not running these tests on cran.

Issue Links

• links to: GitHub Pull Request #11321