You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@storm.apache.org by "Rick Kellogg (JIRA)" <ji...@apache.org> on 2015/10/09 02:47:26 UTC
[jira] [Updated] (STORM-351) multilang python process fall into
endless loop
[ https://issues.apache.org/jira/browse/STORM-351?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Rick Kellogg updated STORM-351:
-------------------------------
Component/s: storm-multilang
> multilang python process fall into endless loop
> -----------------------------------------------
>
> Key: STORM-351
> URL: https://issues.apache.org/jira/browse/STORM-351
> Project: Apache Storm
> Issue Type: Bug
> Components: storm-multilang
> Affects Versions: 0.9.3
> Environment: storm 0.9.3-incubating
> Reporter: DashengJu
> Assignee: DashengJu
> Priority: Blocker
> Fix For: 0.9.3
>
>
> 1. steps to reproduce
> 1) write a topology with a python bolt, run the topology on storm; then there will be two process for the bolt: the worker(java process for ShellBolt), python process.
> 2)kill -9 the worker(java process for ShellBolt);
> 2. expected behavior
> the worker exit and the python process exist
> 3. actual, incorrect behavior
> the worker exit, but the python process never exist and fall into endless loop
> 4. analyse
> in storm.py,read tuple from stdin with follow function:
> def readMsg():
> msg = ""
> while True:
> line = sys.stdin.readline()[0:-1]
> if line == "end":
> break
> msg = msg + line + "\n"
> return json_decode(msg[0:-1])
> when sys.stdin is closed, EOF is encountered, readline() return None, so readMsg fall into endless loop.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)