mirror of
https://github.com/scylladb/scylladb.git
synced 2026-05-22 07:42:16 +00:00
d18eb9479f
In `ks_prop_defs::as_ks_metadata(...)` a default initial tablets count
is set to 0, when tablets are enabled and the replication strategy
is NetworkReplicationStrategy.
This effectively sets _uses_tablets = false in abstract_replication_strategy
for the remaining strategies when no `tablets = {...}` options are specified.
As a consequence, it is possible to create vnode-based keyspaces even
when tablets are enforced with `tablets_mode_for_new_keyspaces`.
The patch sets a default initial tablets count to zero regardless of
the chosen replication strategy. Then each of the replication strategy
validates the options and raises a configuration exception when tablets
are not supported.
All tests are altered in the following way:
+ whenever it was correct, SimpleStrategy was replaced with NetworkTopologyStrategy
+ otherwise, tablets were explicitly disabled with ` AND tablets = {'enabled': false}`
Fixes https://github.com/scylladb/scylladb/issues/25340
Closes scylladb/scylladb#25342
75 lines
3.1 KiB
Python
75 lines
3.1 KiB
Python
#
|
|
# Copyright (C) 2023-present ScyllaDB
|
|
#
|
|
# SPDX-License-Identifier: LicenseRef-ScyllaDB-Source-Available-1.1
|
|
#
|
|
|
|
import asyncio
|
|
import logging
|
|
import time
|
|
import pytest
|
|
|
|
from cassandra.query import SimpleStatement # type: ignore
|
|
from cassandra.cluster import ConsistencyLevel # type: ignore
|
|
from cassandra.protocol import WriteTimeout # type: ignore
|
|
|
|
from test.pylib.manager_client import ManagerClient
|
|
from test.cluster.util import wait_for_token_ring_and_group0_consistency, new_test_keyspace, reconnect_driver
|
|
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@pytest.mark.skip_mode(mode='release', reason='error injections are not supported in release mode')
|
|
async def test_hints_manager_shutdown_hang(manager: ManagerClient) -> None:
|
|
"""Reproducer for #8079"""
|
|
s1 = await manager.server_add(config={
|
|
'error_injections_at_startup': ['decrease_hints_flush_period'],
|
|
}, property_file={"dc": "dc1", "rack": "rack1"})
|
|
s2 = await manager.server_add(property_file={"dc": "dc1", "rack": "rack2"})
|
|
await wait_for_token_ring_and_group0_consistency(manager, time.time() + 30)
|
|
|
|
cql = manager.get_cql()
|
|
|
|
logger.info("Create keyspace and table")
|
|
async with new_test_keyspace(manager, "with replication = {'class': 'NetworkTopologyStrategy', 'replication_factor': 2}") as ks:
|
|
await cql.run_async(f"create table {ks}.t (pk int primary key)")
|
|
|
|
logger.info(f"Stop {s2}")
|
|
await manager.server_stop(s2.server_id)
|
|
|
|
logger.info("Write data with small timeout")
|
|
# We're using a small timeout for the insert so it's not unexpected that it would fail on slow
|
|
# CI machines. To avoid flakiness we disable the test in debug mode (as well as release since
|
|
# it requires an error injection - so it will run only in dev mode) and we retry the write 10 times.
|
|
passed = False
|
|
for _ in range(10):
|
|
try:
|
|
await cql.run_async(SimpleStatement(f"insert into {ks}.t (pk) values (0) using timeout 500ms",
|
|
consistency_level=ConsistencyLevel.ONE))
|
|
except WriteTimeout:
|
|
logger.info("write timeout, retrying")
|
|
else:
|
|
passed = True
|
|
break
|
|
|
|
if not passed:
|
|
pytest.fail("Write timed out on each attempt")
|
|
|
|
# The write succeeded but a background task was left to finish the write to the other node
|
|
# (which is dead but the first node didn't mark it as dead yet).
|
|
# The background task will timeout shortly because of 'using timeout' in the statement.
|
|
# This will cause a hint to get created.
|
|
# The hints manager starts sending the hint soon after (hint flushing happens every
|
|
# ~1 second with the error injection).
|
|
logger.info("Sleep")
|
|
await asyncio.sleep(2)
|
|
|
|
logger.info(f"Stop {s1} gracefully")
|
|
await manager.server_stop_gracefully(s1.server_id)
|
|
|
|
# For dropping the keyspace
|
|
await asyncio.gather(*[manager.server_start(s.server_id) for s in [s1, s2]])
|
|
await reconnect_driver(manager)
|