ClickHouse/tests/integration/test_log_family_hdfs/test.py
2021-05-23 20:19:04 +00:00

60 lines
2.0 KiB
Python

import logging
import sys
import pytest
from helpers.cluster import ClickHouseCluster
from pyhdfs import HdfsClient
@pytest.fixture(scope="module")
def cluster():
try:
cluster = ClickHouseCluster(__file__)
cluster.add_instance("node",
main_configs=["configs/storage_conf.xml", "configs/config.d/log_conf.xml"],
with_hdfs=True)
logging.info("Starting cluster...")
cluster.start()
logging.info("Cluster started")
fs = HdfsClient(hosts='localhost')
fs.mkdirs('/clickhouse')
yield cluster
finally:
cluster.shutdown()
def assert_objects_count(cluster, objects_count, path='data/'):
fs = HdfsClient(hosts='localhost')
hdfs_objects = fs.listdir('/clickhouse')
assert objects_count == len(hdfs_objects)
@pytest.mark.parametrize(
"log_engine,files_overhead,files_overhead_per_insert",
[("TinyLog", 1, 1), ("Log", 2, 1), ("StripeLog", 1, 2)])
def test_log_family_hdfs(cluster, log_engine, files_overhead, files_overhead_per_insert):
node = cluster.instances["node"]
node.query("CREATE TABLE hdfs_test (id UInt64) ENGINE={} SETTINGS disk = 'hdfs'".format(log_engine))
node.query("INSERT INTO hdfs_test SELECT number FROM numbers(5)")
assert node.query("SELECT * FROM hdfs_test") == "0\n1\n2\n3\n4\n"
assert_objects_count(cluster, files_overhead_per_insert + files_overhead)
node.query("INSERT INTO hdfs_test SELECT number + 5 FROM numbers(3)")
assert node.query("SELECT * FROM hdfs_test order by id") == "0\n1\n2\n3\n4\n5\n6\n7\n"
assert_objects_count(cluster, files_overhead_per_insert * 2 + files_overhead)
node.query("INSERT INTO hdfs_test SELECT number + 8 FROM numbers(1)")
assert node.query("SELECT * FROM hdfs_test order by id") == "0\n1\n2\n3\n4\n5\n6\n7\n8\n"
assert_objects_count(cluster, files_overhead_per_insert * 3 + files_overhead)
node.query("TRUNCATE TABLE hdfs_test")
assert_objects_count(cluster, 0)
node.query("DROP TABLE hdfs_test")