I've also tried reproducing this issue in an lxc container on an s390x
machine to no avail. The test that is failing is
-----------
@pytest.mark.xfail(CI_ON_MACOS, reason='continuous integration on macOS fails')
def test_wait_publishes_periodically(bus):
"""Test that wait publishes each tick."""
callback = unittest.mock.MagicMock()
bus.subscribe('main', callback)
def set_start():
time.sleep(0.05)
bus.start()
threading.Thread(target=set_start).start()
bus.wait(bus.states.STARTED, interval=0.01, channel='main')
assert callback.call_count > 3
-------------
The bus.wait method is supposed to send an event on `channel` every
`interval` seconds until the bus has reached specified state (here
`bus.states.STARTED`). The test simply checks that if we spin up a
thread that waits `0.05` seconds and then transitions the bus to
`STARTED` we'll receive more than 3 events. In the failing tests we
instead receive 3.
As noted above, the codepaths in this test do not in anyway touch the
changes in the SRU (or cheroot) at all.
It seems to me that this test is flaky by its nature (it's marked as
xfail on macos CI) and that rather than this being an actual regression,
the patches in the SRU have changed the timing enough of the test run to
hit this flakiness if the machine is running slowly enough.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2143920
Title:
[SRU] Uncaught SSL errors can crash worker threads
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/2143920/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
