default to tokio-epoll-uring in CI tests & on Linux (#7077)

All of production is using it now as of
https://github.com/neondatabase/aws/pull/1121

The change in `flaky_tests.py` resets the flakiness detection logic.

The alternative would have been to repeat the choice of io engine in
each test name, which would junk up the various test reports too much.

---------

Co-authored-by: Alexander Bayandin <alexander@neon.tech>
This commit is contained in:
Christian Schwarz
2024-03-11 15:35:59 +01:00
committed by GitHub
parent b4972d07d4
commit 2b0f3549f7
4 changed files with 19 additions and 8 deletions

View File

@@ -46,9 +46,12 @@ def pytest_generate_tests(metafunc: Metafunc):
metafunc.parametrize("pg_version", pg_versions, ids=map(lambda v: f"pg{v}", pg_versions))
# A hacky way to parametrize tests only for `pageserver_virtual_file_io_engine=tokio-epoll-uring`
# And do not change test name for default `pageserver_virtual_file_io_engine=std-fs` to keep tests statistics
if (io_engine := os.getenv("PAGESERVER_VIRTUAL_FILE_IO_ENGINE", "")) not in ("", "std-fs"):
# A hacky way to parametrize tests only for `pageserver_virtual_file_io_engine=std-fs`
# And do not change test name for default `pageserver_virtual_file_io_engine=tokio-epoll-uring` to keep tests statistics
if (io_engine := os.getenv("PAGESERVER_VIRTUAL_FILE_IO_ENGINE", "")) not in (
"",
"tokio-epoll-uring",
):
metafunc.parametrize("pageserver_virtual_file_io_engine", [io_engine])
# For performance tests, parametrize also by platform