[jira] [Comment Edited] (ARROW-7038) [Python] Reading from HDFS after ctrl+c(SIGTERM) causes python hangs

["Kevin Jung (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/ARROW-7038?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16967246#comment-16967246 ] 

Kevin Jung edited comment on ARROW-7038 at 11/5/19 7:24 AM:
------------------------------------------------------------

[~wesm]  It is still present when I upgrade pyarrow to 0.15.1. Please refer these screen shots. 

  !arrow version.png!

 [HDFS test|[https://github.com/JaebooJung/temp_repo/blob/master/test.jpg]]

In that case, only CTRL+Z can exit python REPL.

 [Normal test|[https://github.com/JaebooJung/temp_repo/blob/master/test_normal.png]]

This is a normal case. Is there a any SIGTERM signal handler registered when I use HDFS as a filesystem?


was (Author: kallsu):
[~wesm]  It is still present when I upgraded pyarrow to 0.15.1. Please refer these screen shots. 

  !arrow version.png!

 [HDFS test|[https://github.com/JaebooJung/temp_repo/blob/master/test.jpg]]

In that case, only CTRL+Z can exit python REPL.

 [Normal test|[https://github.com/JaebooJung/temp_repo/blob/master/test_normal.png]]

This is a normal case. Is there a any SIGTERM signal handler registered when I use HDFS as a filesystem?

> [Python] Reading from HDFS after ctrl+c(SIGTERM) causes python hangs
> --------------------------------------------------------------------
>
>                 Key: ARROW-7038
>                 URL: https://issues.apache.org/jira/browse/ARROW-7038
>             Project: Apache Arrow
>          Issue Type: Bug
>          Components: Python
>         Environment: RHEL 7.4
> python 3.6.7
> pyarrow 0.11.1
> hadoop 2.7.2
>            Reporter: Kevin Jung
>            Priority: Major
>         Attachments: arrow version.png
>
>
> Python always hangs when it reads the parquet file from hdfs twice after sending ctrl+c.
> To reproduce,
> {code:java}
> import pyarrow.parquet as pq
> pq.read_table('hdfs://localhost:9000/path/to/parquet')
> # Ctrl+C here
> pq.read_table('hdfs://localhost:9000/path/to/parquet')
> # hangs{code}
> There's the weird thing ctrl+c at line 3 doesn't produce KeyboardInterrupt even though using the local file produces it at all.
> For example,
> {code:java}
> import pyarrow.parquet as pq
> pq.read_table('/home/path/to/localfile')
> # doing Ctrl+C here prints out 'KeyboardInterrupt'
> pq.read_table('/home/path/to/localfile')
> # run properly{code}
> Is this a sort of bug from JNI?
>  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)