[C++][parquet][hadoop]memory leak when read parquet file from hadoop

[asfimport]

when I use hdfs interface under arrow/io folder to access and read parquets files, it will lead memory leak. some of example codes blow:

 

std::shared_ptrarrow::io::HadoopFileSystem hdfs = nullptr;

...//set hdfs conf var;

msg = arrow::io::HadoopFileSystem::Connect(&conf, &hdfs);

...//check is connect hdfs or not;

std::shared_ptrarrow::io::HdfsReadableFile file_reader;

...//set file_path var;

arrow::Status ret = hdfs->OpenReadable(file_path, &file_reader);

...//check open success or not;

std::unique_ptrparquet::ParquetFileReader parquet_reader;

parquet_reader = parquet::ParquetFileReader::Open(file_reader);

...//through parquet_reader to get data;

 

And I check when I read parquet file in local path,

static std::unique_ptr OpenFile(
const std::string& path, bool memory_map = true,
const ReaderProperties& props = default_reader_properties(),
std::shared_ptr metadata = NULLPTR);

use OpenFile API, it is OK, will not lead memory leak;

 

is there any wrong step when I use hdfs API? or exist some issues. 

any quetions can reply me, thanks.

Environment: linux
Reporter: yzr

_{Note: This issue was originally created as ARROW-10650. Please see the migration documentation for further details.}

[thisisnic]

This issue has been closed due to having no activity since 2020 or earlier. Feel free to reopen if this has been closed in error.