mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-11-18 05:32:52 +00:00
168 lines
5.9 KiB
Python
168 lines
5.9 KiB
Python
import time
|
|
import pytest
|
|
import logging
|
|
|
|
from helpers.cluster import ClickHouseCluster
|
|
from helpers.test_tools import assert_eq_with_retry
|
|
from helpers.network import PartitionManager
|
|
from helpers.corrupt_part_data_on_disk import corrupt_part_data_by_path
|
|
|
|
|
|
def fill_node(node):
|
|
node.query_with_retry(
|
|
"""
|
|
CREATE TABLE IF NOT EXISTS test(n UInt32)
|
|
ENGINE = ReplicatedMergeTree('/clickhouse/tables/test', '{replica}')
|
|
ORDER BY n PARTITION BY n % 10;
|
|
""".format(
|
|
replica=node.name
|
|
)
|
|
)
|
|
|
|
|
|
cluster = ClickHouseCluster(__file__)
|
|
|
|
node_1 = cluster.add_instance("replica1", with_zookeeper=True)
|
|
node_2 = cluster.add_instance("replica2", with_zookeeper=True)
|
|
node_3 = cluster.add_instance("replica3", with_zookeeper=True)
|
|
|
|
|
|
@pytest.fixture(scope="module")
|
|
def start_cluster():
|
|
try:
|
|
cluster.start()
|
|
fill_node(node_1)
|
|
fill_node(node_2)
|
|
# the third node is filled after the DETACH query
|
|
yield cluster
|
|
|
|
except Exception as ex:
|
|
print(ex)
|
|
|
|
finally:
|
|
cluster.shutdown()
|
|
|
|
|
|
def check_data(nodes, detached_parts):
|
|
for node in nodes:
|
|
print(
|
|
"> Replication queue for",
|
|
node.name,
|
|
"\n> table\treplica_name\tsource_replica\ttype\tposition\n",
|
|
node.query_with_retry(
|
|
"SELECT table, replica_name, source_replica, type, position FROM system.replication_queue"
|
|
),
|
|
)
|
|
|
|
node.query_with_retry("SYSTEM SYNC REPLICA test")
|
|
|
|
print("> Checking data integrity for", node.name)
|
|
|
|
for i in range(10):
|
|
assert_eq_with_retry(
|
|
node,
|
|
"SELECT count() FROM test WHERE n % 10 == " + str(i),
|
|
"0\n" if i in detached_parts else "10\n",
|
|
)
|
|
|
|
assert_eq_with_retry(
|
|
node,
|
|
"SELECT count() FROM system.parts WHERE table='test'",
|
|
str(10 - len(detached_parts)) + "\n",
|
|
)
|
|
|
|
res: str = node.query("SELECT * FROM test ORDER BY n")
|
|
|
|
for other in nodes:
|
|
if other != node:
|
|
logging.debug(
|
|
f"> Checking data consistency, {other.name} vs {node.name}"
|
|
)
|
|
assert_eq_with_retry(other, "SELECT * FROM test ORDER BY n", res)
|
|
|
|
|
|
# 1. Check that ALTER TABLE ATTACH PART|PARTITION does not fetch data from other replicas if it's present in the
|
|
# detached/ folder.
|
|
# 2. Check that ALTER TABLE ATTACH PART|PARTITION downloads the data from other replicas if the detached/ folder
|
|
# does not contain the part with the correct checksums.
|
|
def test_attach_without_fetching(start_cluster):
|
|
# Note here requests are used for both PARTITION and PART. This is done for better test diversity.
|
|
# The partition and part are used interchangeably which is not true in most cases.
|
|
# 0. Insert data on two replicas
|
|
node_1.query("INSERT INTO test SELECT * FROM numbers(100)")
|
|
|
|
check_data([node_1, node_2], detached_parts=[])
|
|
|
|
# 1.
|
|
# This part will be fetched from other replicas as it would be missing in the detached/ folder and
|
|
# also attached locally.
|
|
node_1.query("ALTER TABLE test DETACH PART '0_0_0_0'")
|
|
# This partition will be just fetched from other replicas as the checksums won't match
|
|
# (we'll manually break the data).
|
|
node_1.query("ALTER TABLE test DETACH PARTITION 1")
|
|
# This partition will be just fetched from other replicas as the part data will be corrupted with one of the
|
|
# files missing.
|
|
node_1.query("ALTER TABLE test DETACH PARTITION 2")
|
|
|
|
check_data([node_1, node_2], detached_parts=[0, 1, 2])
|
|
|
|
# 2. Create the third replica
|
|
fill_node(node_3)
|
|
|
|
# 3. Break the part data on the second node to corrupt the checksums.
|
|
# Replica 3 should download the data from replica 1 as there is no local data.
|
|
# Replica 2 should also download the data from 1 as the checksums won't match.
|
|
logging.debug("Checking attach with corrupted part data with files missing")
|
|
|
|
to_delete = node_2.exec_in_container(
|
|
[
|
|
"bash",
|
|
"-c",
|
|
"cd {p} && ls *.bin".format(
|
|
p="/var/lib/clickhouse/data/default/test/detached/2_0_0_0"
|
|
),
|
|
],
|
|
privileged=True,
|
|
)
|
|
logging.debug(f"Before deleting: {to_delete}")
|
|
|
|
node_2.exec_in_container(
|
|
[
|
|
"bash",
|
|
"-c",
|
|
"cd {p} && rm -fr *.bin".format(
|
|
p="/var/lib/clickhouse/data/default/test/detached/2_0_0_0"
|
|
),
|
|
],
|
|
privileged=True,
|
|
)
|
|
|
|
node_1.query("ALTER TABLE test ATTACH PARTITION 2")
|
|
check_data([node_1, node_2, node_3], detached_parts=[0, 1])
|
|
|
|
# 4. Break the part data on the second node to corrupt the checksums.
|
|
# Replica 3 should download the data from replica 1 as there is no local data.
|
|
# Replica 2 should also download the data from 1 as the checksums won't match.
|
|
print("Checking attach with corrupted part data with all of the files present")
|
|
|
|
corrupt_part_data_by_path(
|
|
node_2, "/var/lib/clickhouse/data/default/test/detached/1_0_0_0"
|
|
)
|
|
|
|
node_1.query("ALTER TABLE test ATTACH PARTITION 1")
|
|
check_data([node_1, node_2, node_3], detached_parts=[0])
|
|
|
|
# 5. Attach the first part and check if it has been fetched correctly.
|
|
# Replica 2 should attach the local data from detached/.
|
|
# Replica 3 should download the data from replica 2 as there is no local data and other connections are broken.
|
|
print("Checking attach with valid checksums")
|
|
|
|
with PartitionManager() as pm:
|
|
# If something goes wrong and replica 2 wants to fetch data, the test will fail.
|
|
pm.partition_instances(node_2, node_1, action="REJECT --reject-with tcp-reset")
|
|
pm.partition_instances(node_1, node_3, action="REJECT --reject-with tcp-reset")
|
|
|
|
node_1.query("ALTER TABLE test ATTACH PART '0_0_0_0'")
|
|
|
|
check_data([node_1, node_2, node_3], detached_parts=[])
|