[ARROW-404] [Python] Closing an HdfsClient while there are still open file handles results in a crash

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.2.0
• Component/s: Python
• External issue URL: https://github.com/apache/arrow/issues/16052
• Assignee: Wes McKinney
• Reporter: Wes McKinney
• Created: 05/Dec/16 20:04
• Updated: 11/Jan/23 07:08
• Resolved: 09/Dec/16 05:50

Description

Reproducible as follows:

In [3]: con = io.HdfsClient.connect(...)

In [4]: f = con.open(hdfs_file_path)

In [5]: con = None

In [6]: f = None

The attempt to invoke CloseFile on the internal file handle in f causes the JVM to segfault. f should retain a reference to its parent con so this does not occur

Issue Links

• is related to: ARROW-410 [C++] Add Flush method to arrow::io::OutputStream