[Pulp-list] debugging rhel5 pulp unit tests
jason.dobies at redhat.com
Fri Jun 22 18:10:51 UTC 2012
On 06/22/2012 01:41 PM, wes hayutin wrote:
> On 06/22/2012 01:28 PM, John Matthews wrote:
>> I think we may have introduced something on our v2 tests that is
>> making python on rhel5 unhappy.
>> I setup an env similar to Wes and verified I was seeing the same issue
>> with v2 tests but not with v1 tests.
>> From v2 tests we are seeing the below:
>> Memory Handler Error:
>> File "/usr/lib64/python2.4/logging/__init__.py", line 674, in close
>> del _handlers[self]
>> KeyError:<nose.plugins.logcapture.MyMemoryHandler instance at
>> Then I checked out pulp_v1 and ran the unit tests on the same box, the
>> logging errors no longer showed up.
>> Most of the tests ran correct, there were a few failures related to
>> certs, but nothing related to logging.
>> It may be worth investigating this more to determine what did we do in
>> v2 to cause these python patches to be required.
> Thats actually a legit rhel5 python issue
I'm with John, this wasn't always an issue for us, so it stands to
reason that we can fix this in our code. I also want to be 100% sure
that the issue is in our tests only and isn't going to manifest itself
on the server itself.
Can you figure out what changed in the test logging config between v1
and v2 and make sure that it's only a test case thing and then just
clean up our tests? It's going to be a maintenance headache if we have
to patch every RHEL5 box we need to use.
>> ----- Original Message -----
>>> Ok.. there are still some errors and failures, but its at least sane
>>> Attached are my results and some patches I found to fix python..
>>> Let me know if you think I should apply them to the rhel5 slave
>>> Pulp-list mailing list
>>> Pulp-list at redhat.com
Freenode: jdob @ #pulp
http://pulpproject.org | http://blog.pulpproject.org
More information about the Pulp-list