Open Stack

On 11/25/2014 03:33 PM, Steve Baker wrote:
> On 26/11/14 05:05, Matthew Treinish wrote:
>> On Tue, Nov 25, 2014 at 10:28:52AM -0500, Sean Dague wrote:
>>> So as I was waiting for other tests to return, I started looking through
>>> our existing test lists.
>>>
>>> gate-tempest-dsvm-neutron-heat-slow has been slowly evaporating, and I'm
>>> no longer convinced that it does anything useful (and just burns test
>>> nodes).
>>>
>>> The entire output of the job is currently as follows:
>>>
>>> 2014-11-25 14:43:13.801 | heat-slow runtests: commands[0] | bash
>>> tools/pretty_tox.sh
>>> (?=.*\[.*\bslow\b.*\])(^tempest\.(api|scenario)\.orchestration)
>>> --concurrency=4
>>> 2014-11-25 14:43:21.313 | {1}
>>> tempest.scenario.orchestration.test_server_cfn_init.CfnInitScenarioTest.test_server_cfn_init
>>>
>>> ... SKIPPED: Skipped until Bug: 1374175 is resolved.
>>> 2014-11-25 14:47:36.271 | {0}
>>> tempest.api.orchestration.stacks.test_neutron_resources.NeutronResourcesTestJSON.test_created_network
>>>
>>> [0.169736s] ... ok
>>> 2014-11-25 14:47:36.271 | {0}
>>> tempest.api.orchestration.stacks.test_neutron_resources.NeutronResourcesTestJSON.test_created_resources
>>>
>>> [0.000508s] ... ok
>>> 2014-11-25 14:47:36.291 | {0}
>>> tempest.api.orchestration.stacks.test_neutron_resources.NeutronResourcesTestJSON.test_created_router
>>>
>>> [0.019679s] ... ok
>>> 2014-11-25 14:47:36.313 | {0}
>>> tempest.api.orchestration.stacks.test_neutron_resources.NeutronResourcesTestJSON.test_created_router_interface
>>>
>>> [0.020597s] ... ok
>>> 2014-11-25 14:47:36.564 | {0}
>>> tempest.api.orchestration.stacks.test_neutron_resources.NeutronResourcesTestJSON.test_created_server
>>>
>>> [0.250788s] ... ok
>>> 2014-11-25 14:47:36.580 | {0}
>>> tempest.api.orchestration.stacks.test_neutron_resources.NeutronResourcesTestJSON.test_created_subnet
>>>
>>> [0.015410s] ... ok
>>> 2014-11-25 14:47:44.113 |
>>> 2014-11-25 14:47:44.113 | ======
>>> 2014-11-25 14:47:44.113 | Totals
>>> 2014-11-25 14:47:44.113 | ======
>>> 2014-11-25 14:47:44.114 | Run: 7 in 0.478173 sec.
>>> 2014-11-25 14:47:44.114 |  - Passed: 6
>>> 2014-11-25 14:47:44.114 |  - Skipped: 1
>>> 2014-11-25 14:47:44.115 |  - Failed: 0
>>> 2014-11-25 14:47:44.115 |
>>> 2014-11-25 14:47:44.116 | ==============
>>> 2014-11-25 14:47:44.116 | Worker Balance
>>> 2014-11-25 14:47:44.116 | ==============
>>> 2014-11-25 14:47:44.117 |  - Worker 0 (6 tests) => 0:00:00.480677s
>>> 2014-11-25 14:47:44.117 |  - Worker 1 (1 tests) => 0:00:00.001455s
>>>
>>> So we are running about 1s worth of work, no longer in parallel (as
>>> there aren't enough classes to even do parallel runs).
>>>
>>> Given the emergence of the heat functional job, and the fact that this
>>> is really not testing anything any more, I'd like to propose we just
>>> remove it entirely at this stage and get the test nodes back.
>>>
>>
> +1, I think its time to go. NeutronResourcesTestJSON needs to be ported
> to a functional test, but I don't think killing heat-slow needs to wait
> for that to happen.

Awesome. The change is proposed here -
https://review.openstack.org/#/c/137123/

	-Sean

-- 
Sean Dague
http://dague.net