Skip to content

Fix flaky test_gather_dep_one_worker_always_busy #6554

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
crusaderky merged 2 commits into from
Jun 10, 2022
Merged

Fix flaky test_gather_dep_one_worker_always_busy #6554

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
40a5f00
Fix flaky test_gather_dep_one_worker_always_busy
crusaderky Jun 9, 2022
07860b3
tweak
crusaderky Jun 9, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 6 additions & 3 deletions distributed/tests/test_worker.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1939,6 +1939,7 @@ def f(x):
assert not C.instances


@pytest.mark.slow
@gen_cluster(client=True)
async def test_gather_dep_one_worker_always_busy(c, s, a, b):
# Ensure that both dependencies for H are on another worker than H itself.
Expand Down Expand Up @@ -1966,12 +1967,14 @@ async def test_gather_dep_one_worker_always_busy(c, s, a, b):
assert b.tasks[g.key].state in ("flight", "fetch")

with pytest.raises(asyncio.TimeoutError):
await h.result(timeout=0.5)
await h.result(timeout=0.8)

story = b.story("busy-gather")
# 1 busy response straight away, followed by 1 retry every 150ms for 500ms.
# 1 busy response straight away, followed by 1 retry every 150ms for 800ms.
# The requests for b and g are clustered together in single messages.
assert 3 <= len(story) <= 7
# We need to be very lax in measuring as PeriodicCallback+network comms have been
# observed on CI to occasionally lag behind by several hundreds of ms.
assert 2 <= len(story) <= 8

async with Worker(s.address, name="x") as x:
# We "scatter" the data to another worker which is able to serve this data.
Expand Down