cql-pytest: recognize when Scylla crashes

Before this patch, if Scylla crashes during some test in cql-pytest, all
tests after it will fail because they can't connect to Scylla - and we can
get a report on hundreds of failures without a clear sign of where the real
problem was.

This patch introduces an autouse fixture (i.e., a fixture automatically
used by every test) which tries to run a do-nothing CQL command after each
test.  If this CQL command fails, we conclude that Scylla crashed and
report the test in which this happened - and exist pytest instead of failing
a hundred more tests.

Fixes #8080

Signed-off-by: Nadav Har'El <nyh@scylladb.com>
Message-Id: <20210304132804.1527977-1-nyh@scylladb.com>
This commit is contained in:
Nadav Har'El
2021-03-04 15:28:04 +02:00
committed by Avi Kivity
parent 1226fc755f
commit acfa180766

View File

@@ -64,6 +64,19 @@ def cql(request):
)
return cluster.connect()
# A function-scoped autouse=True fixture allows us to test after every test
# that the CQL connection is still alive - and if not report the test which
# crashed Scylla and stop running any more tests.
@pytest.fixture(scope="function", autouse=True)
def cql_test_connection(cql, request):
yield
try:
# We want to run a do-nothing CQL command. "use system" is the
# closest to do-nothing I could find...
cql.execute("use system")
except:
pytest.exit(f"Scylla appears to have crashed in test {request.node.parent.name}::{request.node.name}")
# "test_keyspace" fixture: Creates and returns a temporary keyspace to be
# used in tests that need a keyspace. The keyspace is created with RF=1,
# and automatically deleted at the end. We use scope="session" so that all