Cinder unittest failures have deep stracebacks due to osprofiler

[Clark Boylan]

Logs come from http://logs.openstack.org/95/176095/1/check/gate-cinder-python27/747e8f0/console.html#_2015-04-22_15_27_20_571

When cinder unittests fail they can have very deep stracebacks with a large portion of the traceback being nested osprofiler calls. I think there are possibly two bugs here.

First why is osprofiler enabled when running unittests? Shouldn't we be testing non profiled code with an optional opt in for profiling to do debugging? Mostly thinking that this osprofiler output is noise for typical debug cases and is really only needed when you want that profiling to be done.

Second would it be possible to have osprofiler better annotate its wrapping so that we can see what has been wrapped and where calls are made when we get the stacktraces? as is its a bit of a mess.

[Clark Boylan (cboylan)]

I intended on pasting the logs here for historical reasons and failed. Here they are:

2015-04-22 15:27:20.570 | Captured stderr:
2015-04-22 15:27:20.570 | ~~~~~~~~~~~~~~~~
2015-04-22 15:27:20.571 | Traceback (most recent call last):
2015-04-22 15:27:20.571 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/eventlet/hubs/hub.py", line 457, in fire_timers
2015-04-22 15:27:20.571 | timer()
2015-04-22 15:27:20.571 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/eventlet/hubs/timer.py", line 58, in __call__
2015-04-22 15:27:20.571 | cb(*args, **kw)
2015-04-22 15:27:20.571 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/eventlet/greenthread.py", line 214, in main
2015-04-22 15:27:20.571 | result = function(*args, **kwargs)
2015-04-22 15:27:20.571 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/osprofiler/profiler.py", line 105, in wrapper
2015-04-22 15:27:20.571 | return f(*args, **kwargs)
2015-04-22 15:27:20.571 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/osprofiler/profiler.py", line 105, in wrapper
2015-04-22 15:27:20.572 | return f(*args, **kwargs)
2015-04-22 15:27:20.572 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/osprofiler/profiler.py", line 105, in wrapper
2015-04-22 15:27:20.572 | return f(*args, **kwargs)
2015-04-22 15:27:20.572 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/osprofiler/profiler.py", line 105, in wrapper
2015-04-22 15:27:20.572 | return f(*args, **kwargs)
2015-04-22 15:27:20.572 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/osprofiler/profiler.py", line 105, in wrapper
2015-04-22 15:27:20.572 | return f(*args, **kwargs)
2015-04-22 15:27:20.572 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/osprofiler/profiler.py", line 105, in wrapper
2015-04-22 15:27:20.572 | return f(*args, **kwargs)
2015-04-22 15:27:20.572 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/osprofiler/profiler.py", line 105, in wrapper
2015-04-22 15:27:20.572 | return f(*args, **kwargs)
2015-04-22 15:27:20.573 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/osprofiler/profiler.py", line 105, in wrapper
2015-04-22 15:27:20.573 | return f(*args, **kwargs)
2015-04-22 15:27:20.573 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/osprofiler/profiler.py", line 105, in wrapper
2015-04-22 15:27:20.573 | return f(*args, **kwargs)
2015-04-22 15:27:20.573 | File "/home/jenkins/workspace/gate-cinder-python27/.tox/py27/local/lib/python2.7/site-packages/osprofiler/profiler.py", line 105, in wrapper
2015-04-22 1...

[Eric Harney (eharney)]

I agree that we shouldn't be wrapping code with the profiler by default for unit tests.

[Yuriy Nesenenko (ynesenenko)]

This bug could not be reproduced.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/266288

[OpenStack Infra (hudson-openstack)]

Change abandoned by Boris Pavlovic (<email address hidden>) on branch: master
Review: https://review.openstack.org/266288

[Dina Belova (dbelova)]

Was partially fixed with https://github.com/openstack/cinder/commit/f55db194876e08ef4671d31c52b8665fb540c00e

[Sean McGinnis (sean-mcginnis)]

I think the fix referenced in #6 addresses this. Closing.