[jira] [Comment Edited] (FLINK-25464) Python tests hangs on install dependencies

["Huang Xingbo (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/FLINK-25464?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17469239#comment-17469239 ] 

Huang Xingbo edited comment on FLINK-25464 at 1/5/22, 12:03 PM:
----------------------------------------------------------------

For a long time, all the log information about downloaded and installed by python has not been printed out if there is no failure during this progress. In the Azure laboratory network environment, this progress of installation takes a longer time. Once the total time exceeds 15 minutes, it will be killed although it doesn't hang up. My repair plan is to turn on log information (the downloaded infos is about 300 lines). 
The reason why python3.8 only hangs is that as I analyzed earlier, there is a python package of grpcio-tools under python3.8 that is installed from the source code rather than installation from whl package. This process takes about 5min, which lead to the total installation in the python3.8 environment is easier to exceed 15 minutes. I will create another JIRA and PR to upgrade grpcio-tools in the master branch.


was (Author: hxbks2ks):
For a long time, all the log information about downloaded and installed by python has not been printed out if there is no failure during this progress. In the Azure laboratory network environment, this progress of installation takes a longer time. Once the total time exceeds 15 minutes, it will be killed although it doesn't hang up. My repair plan is to open log information (the downloaded infos is about 300 lines). 
The reason why python3.8 only hangs is that as I analyzed earlier, there is a python package of grpcio-tools under python3.8 that is installed from the source code rather than installation from whl package. This process takes about 5min, which lead to the total installation in the python3.8 environment is easier to exceed 15 minutes. I will create another JIRA and PR to upgrade grpcio-tools in the master branch.

> Python tests hangs on install dependencies
> ------------------------------------------
>
>                 Key: FLINK-25464
>                 URL: https://issues.apache.org/jira/browse/FLINK-25464
>             Project: Flink
>          Issue Type: Bug
>          Components: API / Python, Build System / Azure Pipelines
>    Affects Versions: 1.13.5, 1.14.2
>            Reporter: Yun Gao
>            Assignee: Huang Xingbo
>            Priority: Critical
>              Labels: test-stability
>             Fix For: 1.15.0, 1.13.6, 1.14.3
>
>
> {code:java}
> Dec 25 04:35:54 py38-cython create: /__w/1/s/flink-python/.tox/py38-cython
> Dec 25 04:35:58 py38-cython installdeps: -rdev/dev-requirements.txt, pytest, apache-flink-libraries
> Dec 25 04:51:00 ==============================================================================
> Dec 25 04:51:00 Process produced no output for 900 seconds.
> Dec 25 04:51:00 ==============================================================================
> Dec 25 04:51:00 ==============================================================================
> Dec 25 04:51:00 The following Java processes are running (JPS)
> Dec 25 04:51:00 ==============================================================================
> Picked up JAVA_TOOL_OPTIONS: -XX:+HeapDumpOnOutOfMemoryError
> Dec 25 04:51:00 137834 Jps
> Dec 25 04:51:00 ==============================================================================
> Dec 25 04:51:00 Printing stack trace of Java process 137834
> Dec 25 04:51:00 ==============================================================================
> Picked up JAVA_TOOL_OPTIONS: -XX:+HeapDumpOnOutOfMemoryError
> 137834: No such process
> Dec 25 04:51:00 Killing process with pid=725 and all descendants
> ./flink-python/dev/lint-python.sh: line 580:  2770 Terminated              $TOX_PATH -c $FLINK_PYTHON_DIR/tox.ini --recreate 2>&1
>       2771                       | tee -a $LOG_FILE
> /__w/1/s/tools/ci/watchdog.sh: line 113:   725 Terminated              $cmd
> Dec 25 04:51:00 Process exited with EXIT CODE: 143.
> Dec 25 04:51:00 Trying to KILL watchdog (720).
> Dec 25 04:51:00 Searching for .dump, .dumpstream and related files in '/__w/1/s'
> The STDIO streams did not close within 10 seconds of the exit event from process '/bin/bash'. This may indicate a child process inherited the STDIO streams and has not yet exited.
> ##[error]Bash exited with code '143'.
> {code}
> https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=28588&view=logs&j=821b528f-1eed-5598-a3b4-7f748b13f261&t=6bb545dd-772d-5d8c-f258-f5085fba3295&l=23407



--
This message was sent by Atlassian Jira
(v8.20.1#820001)