ambari-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Dmitro Lisnichenko" <dlysniche...@hortonworks.com>
Subject Review Request 36771: Provide traceback patch to debug hanging agents
Date Fri, 24 Jul 2015 10:09:42 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/36771/
-----------------------------------------------------------

Review request for Ambari, Andrew Onischuk and Sumit Mohanty.


Bugs: AMBARI-12522
    https://issues.apache.org/jira/browse/AMBARI-12522


Repository: ambari


Description
-------

Patch tries to load and register traceback handler if it is available, and skips if not. Also
it fixes binding signal handlers twice during agent start.
To install faulthandler under Centos 6 (faulthandler is not included to default distribution
of Python 2.x), we have to perform:
yum install python-devel gcc -y
# install setup tools
curl https://bitbucket.org/pypa/setuptools/raw/bootstrap/ez_setup.py | python -
# install pip
curl https://raw.github.com/pypa/pip/master/contrib/get-pip.py | python -
easy_install faulthandler
If faulthandler module is available, agent posts Registered faulthandler to agent out file.
After that, we start agent and can dump tracebacks for all running threads like that:
# kill -USR1 `cat /var/run/ambari-agent/ambari-agent.pid`
# cat /var/log/ambari-agent/ambari-agent.out


Diffs
-----

  ambari-agent/src/main/python/ambari_agent/HeartbeatHandlers.py f9d0e14 
  ambari-agent/src/main/python/ambari_agent/main.py 0971cbb 
  ambari-agent/src/test/python/ambari_agent/TestMain.py 6218b59 

Diff: https://reviews.apache.org/r/36771/diff/


Testing
-------

test_invalid (TestFileSystem.TestFileSystem) ... ok

----------------------------------------------------------------------
Ran 410 tests in 6.028s

OK

Process finished with exit code 0


Thanks,

Dmitro Lisnichenko


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message