How to deal with "RobotRemote is already in use" error?

we are facing a weird error while running Robot Framework tests of different packages sequentially; it seems some process is left behind:

Running tests:
  Tear down in 0.000 seconds.
  Set up Traceback (most recent call last):
  File "/home/travis/build/plonegovbr/portal.buildout/eggs/zope.testing-3.9.7-py2.7.egg/zope/testing/testrunner/", line 366, in run_layer
    setup_layer(options, layer, setup_layers)
  File "/home/travis/build/plonegovbr/portal.buildout/eggs/zope.testing-3.9.7-py2.7.egg/zope/testing/testrunner/", line 628, in setup_layer
    setup_layer(options, base, setup_layers)
  File "/home/travis/build/plonegovbr/portal.buildout/eggs/zope.testing-3.9.7-py2.7.egg/zope/testing/testrunner/", line 633, in setup_layer
  File "/home/travis/build/plonegovbr/portal.buildout/eggs/", line 73, in setUp
    portal._setObject(id_, globals()[id_]())
  File "/home/travis/build/plonegovbr/portal.buildout/eggs/Zope2-2.13.24-py2.7.egg/OFS/", line 326, in _setObject
    v = self._checkId(id)
  File "/home/travis/build/plonegovbr/portal.buildout/eggs/Products.CMFCore-2.2.9-py2.7.egg/Products/CMFCore/", line 213, in _checkId
    return superCheckId(self, id, allow_dup)
  File "/home/travis/build/plonegovbr/portal.buildout/eggs/Zope2-2.13.24-py2.7.egg/OFS/", line 116, in checkValidId
    'The id "%s" is invalid - it is already in use.' % id)
BadRequest: The id "RobotRemote" is invalid - it is already in use.

does anybody seen this error before? is there any workaround?

Do you know when did this start? I recall this being an issue with test layer ordering. Somehow robot remote layer tear down is not called or it was unable to delete that (or it reappears because of rollback in Plone sandbox layer). It was weird enough that I did not just add code to skip that issue in robot remote layer setup.

I don't know because we started testing these packages sequentially yesterday.

as you can see above in the log, there is a thread left behind in some test:

  Set up plone.testing.z2.ZServer in 0.502 seconds.
  Set up in 0.000 seconds.
    1/10 (10.0%)The following test left new threads behind:
Criar nova capa (test_capa.robot)
New thread(s): [<_DummyThread(Dummy-1, started daemon 139809298974464)>]

I've seen this happening before with collective.cover's RF tests.

what can I do to kill it?