Logging tests are broken on stable/liberty with error "AttributeError: 'module' object has no attribute 'RFCSysLogHandler'"

[Matt Riedemann]

Seen here:

http://logs.openstack.org/periodic-stable/periodic-murano-python27-liberty/7df5a09/console.html#_2016-03-03_06_14_56_445

2016-03-03 06:14:55.863 | {3} test_log.SysLogHandlersTestCase.test_standard_format [0.000723s] ... FAILED
2016-03-03 06:14:55.863 |
2016-03-03 06:14:55.864 | Captured traceback:
2016-03-03 06:14:55.864 | ~~~~~~~~~~~~~~~~~~~
2016-03-03 06:14:55.864 | Traceback (most recent call last):
2016-03-03 06:14:55.864 | File "/home/jenkins/workspace/periodic-murano-python27-liberty/murano/tests/unit/test_log.py", line 34, in setUp
2016-03-03 06:14:55.864 | self.rfclogger = handlers.RFCSysLogHandler(address='/dev/log',
2016-03-03 06:14:55.864 | AttributeError: 'module' object has no attribute 'RFCSysLogHandler'
2016-03-03 06:14:55.864 |
2016-03-03 06:14:56.445 | {1} db.migration.test_migrations.TestMigrationsPostgresql.test_walk_versions [8.103970s] ... ok
2016-03-03 06:14:56.445 |
2016-03-03 06:14:56.445 | Captured stderr:
2016-03-03 06:14:56.445 | ~~~~~~~~~~~~~~~~
2016-03-03 06:14:56.445 | Traceback (most recent call last):
2016-03-03 06:14:56.445 | File "/usr/lib/python2.7/logging/__init__.py", line 872, in emit
2016-03-03 06:14:56.445 | stream.write(fs % msg)
2016-03-03 06:14:56.446 | File "/usr/lib/python2.7/StringIO.py", line 213, in write
2016-03-03 06:14:56.446 | _complain_ifclosed(self.closed)
2016-03-03 06:14:56.446 | File "/usr/lib/python2.7/StringIO.py", line 40, in _complain_ifclosed
2016-03-03 06:14:56.446 | raise ValueError, "I/O operation on closed file"
2016-03-03 06:14:56.446 | ValueError: I/O operation on closed file
2016-03-03 06:14:56.446 | Logged from file provision.py, line 71

And that goes into an infinite loop it looks like.

This started on 3/1:

http://logstash.openstack.org/#dashboard/file/logstash.json?query=message%3A%5C%22AttributeError%3A%20'module'%20object%20has%20no%20attribute%20'RFCSysLogHandler'%5C%22&from=7d

So I'm guessing this broke it:

https://github.com/openstack/murano/commit/0cfcf423ad50906a5a211f169c18599b8fcbce8f

[Matt Riedemann (mriedem)]

I was wrong, the regression was actually fixed by https://review.openstack.org/#/c/286856/, not caused by it.