Open pont-us opened 1 year ago
Another one, from https://github.com/dcs4cop/xcube/actions/runs/6467629870/job/17559781773:
FAILED test/webapi/places/test_routes.py::PlacesRoutesTest::test_places - urllib3.exceptions.MaxRetryError:
HTTPConnectionPool(host='localhost', port=33397): Max retries exceeded with url: /places (Caused by
NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7f6942170a10>: Failed to establish a new connection:
[Errno 111] Connection refused'))
And another from https://github.com/dcs4cop/xcube/actions/runs/6467629870/job/17562750121:
FAILED test/webapi/ows/stac/test_routes.py::StacRoutesTest::test_fetch_catalog_collections - urllib3.exceptions.MaxRetryError:
HTTPConnectionPool(host='localhost', port=53383): Max retries exceeded with url: /ogc/collections (Caused by
NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7f09cc4ae090>: Failed to establish a new connection:
[Errno 111] Connection refused'))
https://github.com/dcs4cop/xcube/actions/runs/6467629870/job/17571143347
FAILED test/webapi/datasets/test_routes.py::DatasetsRoutesTest::test_fetch_datasets - urllib3.exceptions.MaxRetryError: HTTPConnectionPool(host='localhost', port=39147): Max retries exceeded with url: /datasets (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7fb970811990>: Failed to establish a new connection: [Errno 111] Connection refused'))
FAILED test/webapi/s3/test_routes.py::S3RoutesNewTest::test_fetch_get_s3_object - urllib3.exceptions.MaxRetryError: HTTPConnectionPool(host='localhost', port=54325): Max retries exceeded with url: /s3/datasets/demo.zarr/.zattrs (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7fb98c58f950>: Failed to establish a new connection: [Errno 111] Connection refused'))
Not sure if this one is related, but this test run on AppVeyor produced
ERROR test/core/zarrstore/test_generic.py::CommonS3ZarrStoreTest::test_it - TimeoutError: timed out
Reopened as per @TonioF's comment. The PR addresses the current problems, but probably new related issues will appear, therefore the issue stays open.
https://github.com/dcs4cop/xcube/pull/918#pullrequestreview-1829460478
Describe the bug
Recently, CI unit test suite jobs have been producing increasingly frequent intermittent test failures due to time-outs and excessive retries. The problem seems to occur most frequently on GitHub, but occasionally also on AppVeyor. This issue was prompted by this GitHub actions run which produced the following error:
We can use this issue to collect further instances of the problem.
So far our fix for such errors has been "re-run the job and hope it goes away", which it generally does, but this is turning into something of a time sink.
To Reproduce Steps to reproduce the behavior: keep re-running the GitHub unittest job until an error occurs.
Expected behavior All tests should pass reliably on every CI run.
Additional context With luck, this might be fairly easily fixable by tweaking some back-off / time-out / retry parameters in
ServerTestCase
or similar.