mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-11-24 00:22:29 +00:00
1016 lines
35 KiB
Python
1016 lines
35 KiB
Python
from time import sleep
|
|
import pytest
|
|
import re
|
|
import os.path
|
|
from helpers.cluster import ClickHouseCluster
|
|
from helpers.test_tools import TSV, assert_eq_with_retry
|
|
|
|
|
|
cluster = ClickHouseCluster(__file__)
|
|
|
|
main_configs = [
|
|
"configs/remote_servers.xml",
|
|
"configs/replicated_access_storage.xml",
|
|
"configs/replicated_user_defined_sql_objects.xml",
|
|
"configs/backups_disk.xml",
|
|
"configs/lesser_timeouts.xml", # Default timeouts are quite big (a few minutes), the tests don't need them to be that big.
|
|
]
|
|
|
|
user_configs = [
|
|
"configs/allow_database_types.xml",
|
|
"configs/zookeeper_retries.xml",
|
|
]
|
|
|
|
node1 = cluster.add_instance(
|
|
"node1",
|
|
main_configs=main_configs,
|
|
user_configs=user_configs,
|
|
external_dirs=["/backups/"],
|
|
macros={"replica": "node1", "shard": "shard1"},
|
|
with_zookeeper=True,
|
|
)
|
|
|
|
node2 = cluster.add_instance(
|
|
"node2",
|
|
main_configs=main_configs,
|
|
user_configs=user_configs,
|
|
external_dirs=["/backups/"],
|
|
macros={"replica": "node2", "shard": "shard1"},
|
|
with_zookeeper=True,
|
|
stay_alive=True, # Necessary for the "test_stop_other_host_while_backup" test
|
|
)
|
|
|
|
|
|
node3 = cluster.add_instance(
|
|
"node3",
|
|
main_configs=main_configs,
|
|
user_configs=user_configs,
|
|
external_dirs=["/backups/"],
|
|
macros={"replica": "node3", "shard": "shard1"},
|
|
with_zookeeper=True,
|
|
)
|
|
|
|
|
|
@pytest.fixture(scope="module", autouse=True)
|
|
def start_cluster():
|
|
try:
|
|
cluster.start()
|
|
yield cluster
|
|
finally:
|
|
cluster.shutdown()
|
|
|
|
|
|
@pytest.fixture(autouse=True)
|
|
def drop_after_test():
|
|
try:
|
|
yield
|
|
finally:
|
|
node1.query("DROP TABLE IF EXISTS tbl ON CLUSTER 'cluster3' SYNC")
|
|
node1.query("DROP TABLE IF EXISTS tbl2 ON CLUSTER 'cluster3' SYNC")
|
|
node1.query("DROP DATABASE IF EXISTS mydb ON CLUSTER 'cluster3' SYNC")
|
|
node1.query("DROP USER IF EXISTS u1, u2 ON CLUSTER 'cluster3'")
|
|
|
|
|
|
backup_id_counter = 0
|
|
|
|
|
|
def new_backup_name():
|
|
global backup_id_counter
|
|
backup_id_counter += 1
|
|
return f"Disk('backups', '{backup_id_counter}')"
|
|
|
|
|
|
def get_path_to_backup(backup_name):
|
|
name = backup_name.split(",")[1].strip("')/ ")
|
|
return os.path.join(node1.cluster.instances_dir, "backups", name)
|
|
|
|
|
|
def test_replicated_table():
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' ("
|
|
"x UInt8, y String"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY x"
|
|
)
|
|
|
|
node1.query("INSERT INTO tbl VALUES (1, 'Don''t')")
|
|
node2.query("INSERT INTO tbl VALUES (2, 'count')")
|
|
node1.query("INSERT INTO tbl SETTINGS async_insert=true VALUES (3, 'your')")
|
|
node2.query("INSERT INTO tbl SETTINGS async_insert=true VALUES (4, 'chickens')")
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' tbl")
|
|
|
|
backup_name = new_backup_name()
|
|
|
|
# Make backup on node 1.
|
|
node1.query(
|
|
f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name} SETTINGS replica_num=1"
|
|
)
|
|
|
|
# Drop table on both nodes.
|
|
node1.query(f"DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
# Restore from backup on node2.
|
|
node2.query(f"RESTORE TABLE tbl ON CLUSTER 'cluster' FROM {backup_name}")
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' tbl")
|
|
|
|
assert node2.query("SELECT * FROM tbl ORDER BY x") == TSV(
|
|
[[1, "Don\\'t"], [2, "count"], [3, "your"], [4, "chickens"]]
|
|
)
|
|
|
|
assert node1.query("SELECT * FROM tbl ORDER BY x") == TSV(
|
|
[[1, "Don\\'t"], [2, "count"], [3, "your"], [4, "chickens"]]
|
|
)
|
|
|
|
|
|
def test_empty_replicated_table():
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' ("
|
|
"x UInt8, y String"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY x"
|
|
)
|
|
|
|
backup_name = new_backup_name()
|
|
|
|
# Make backup on node 1.
|
|
node1.query(
|
|
f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name} SETTINGS replica_num=1"
|
|
)
|
|
|
|
# Drop table on both nodes.
|
|
node1.query(f"DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
# Restore from backup on node2.
|
|
node1.query(f"RESTORE TABLE tbl ON CLUSTER 'cluster' FROM {backup_name}")
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' tbl")
|
|
|
|
assert node1.query("SELECT * FROM tbl") == ""
|
|
assert node2.query("SELECT * FROM tbl") == ""
|
|
|
|
|
|
def test_replicated_database():
|
|
node1.query(
|
|
"CREATE DATABASE mydb ON CLUSTER 'cluster' ENGINE=Replicated('/clickhouse/path/','{shard}','{replica}')"
|
|
)
|
|
|
|
node1.query(
|
|
"CREATE TABLE mydb.tbl(x UInt8, y String) ENGINE=ReplicatedMergeTree ORDER BY x"
|
|
)
|
|
|
|
node2.query("SYSTEM SYNC DATABASE REPLICA mydb")
|
|
|
|
node1.query("INSERT INTO mydb.tbl VALUES (1, 'Don''t')")
|
|
node2.query("INSERT INTO mydb.tbl VALUES (2, 'count')")
|
|
node1.query("INSERT INTO mydb.tbl VALUES (3, 'your')")
|
|
node2.query("INSERT INTO mydb.tbl VALUES (4, 'chickens')")
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' mydb.tbl")
|
|
|
|
# Make backup.
|
|
backup_name = new_backup_name()
|
|
node1.query(
|
|
f"BACKUP DATABASE mydb ON CLUSTER 'cluster' TO {backup_name} SETTINGS replica_num=2"
|
|
)
|
|
|
|
# Drop table on both nodes.
|
|
node1.query("DROP DATABASE mydb ON CLUSTER 'cluster' SYNC")
|
|
|
|
# Restore from backup on node2.
|
|
node1.query(f"RESTORE DATABASE mydb ON CLUSTER 'cluster' FROM {backup_name}")
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' mydb.tbl")
|
|
|
|
assert node1.query("SELECT * FROM mydb.tbl ORDER BY x") == TSV(
|
|
[[1, "Don\\'t"], [2, "count"], [3, "your"], [4, "chickens"]]
|
|
)
|
|
|
|
assert node2.query("SELECT * FROM mydb.tbl ORDER BY x") == TSV(
|
|
[[1, "Don\\'t"], [2, "count"], [3, "your"], [4, "chickens"]]
|
|
)
|
|
|
|
|
|
def test_different_tables_on_nodes():
|
|
node1.query(
|
|
"CREATE TABLE tbl (`x` UInt8, `y` String) ENGINE = MergeTree ORDER BY x"
|
|
)
|
|
node2.query("CREATE TABLE tbl (`w` Int64) ENGINE = MergeTree ORDER BY w")
|
|
|
|
node1.query(
|
|
"INSERT INTO tbl VALUES (1, 'Don''t'), (2, 'count'), (3, 'your'), (4, 'chickens')"
|
|
)
|
|
node2.query("INSERT INTO tbl VALUES (-333), (-222), (-111), (0), (111)")
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query(f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}")
|
|
|
|
node1.query("DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
node2.query(f"RESTORE TABLE tbl ON CLUSTER 'cluster' FROM {backup_name}")
|
|
|
|
assert node1.query("SELECT * FROM tbl") == TSV(
|
|
[[1, "Don\\'t"], [2, "count"], [3, "your"], [4, "chickens"]]
|
|
)
|
|
assert node2.query("SELECT * FROM tbl") == TSV([-333, -222, -111, 0, 111])
|
|
|
|
|
|
def test_backup_restore_on_single_replica():
|
|
node1.query(
|
|
"CREATE DATABASE mydb ON CLUSTER 'cluster' ENGINE=Replicated('/clickhouse/path/','{shard}','{replica}')"
|
|
)
|
|
node1.query(
|
|
"CREATE TABLE mydb.test (`name` String, `value` UInt32) ENGINE = ReplicatedMergeTree ORDER BY value"
|
|
)
|
|
node1.query("INSERT INTO mydb.test VALUES ('abc', 1), ('def', 2)")
|
|
node1.query("INSERT INTO mydb.test VALUES ('ghi', 3)")
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query(f"BACKUP DATABASE mydb TO {backup_name}")
|
|
|
|
node1.query("DROP DATABASE mydb SYNC")
|
|
|
|
# Cannot restore table because it already contains data on other replicas.
|
|
expected_error = "already contains some data"
|
|
assert expected_error in node1.query_and_get_error(
|
|
f"RESTORE DATABASE mydb FROM {backup_name}"
|
|
)
|
|
|
|
# Can restore table with structure_only=true.
|
|
node1.query(
|
|
f"RESTORE DATABASE mydb FROM {backup_name} SETTINGS structure_only=true"
|
|
)
|
|
|
|
node1.query("SYSTEM SYNC REPLICA mydb.test")
|
|
assert node1.query("SELECT * FROM mydb.test ORDER BY name") == TSV(
|
|
[["abc", 1], ["def", 2], ["ghi", 3]]
|
|
)
|
|
|
|
# Can restore table with allow_non_empty_tables=true.
|
|
node1.query("DROP DATABASE mydb SYNC")
|
|
node1.query(
|
|
f"RESTORE DATABASE mydb FROM {backup_name} SETTINGS allow_non_empty_tables=true"
|
|
)
|
|
|
|
node1.query("SYSTEM SYNC REPLICA mydb.test")
|
|
assert node1.query("SELECT * FROM mydb.test ORDER BY name") == TSV(
|
|
[["abc", 1], ["abc", 1], ["def", 2], ["def", 2], ["ghi", 3], ["ghi", 3]]
|
|
)
|
|
|
|
|
|
def test_table_with_parts_in_queue_considered_non_empty():
|
|
node1.query(
|
|
"CREATE DATABASE mydb ON CLUSTER 'cluster' ENGINE=Replicated('/clickhouse/path/','{shard}','{replica}')"
|
|
)
|
|
node1.query(
|
|
"CREATE TABLE mydb.test (`x` UInt32) ENGINE = ReplicatedMergeTree ORDER BY x"
|
|
)
|
|
node1.query("INSERT INTO mydb.test SELECT number AS x FROM numbers(10000000)")
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query(f"BACKUP DATABASE mydb TO {backup_name}")
|
|
|
|
node1.query("DROP DATABASE mydb SYNC")
|
|
|
|
# Cannot restore table because it already contains data on other replicas.
|
|
expected_error = "already contains some data"
|
|
assert expected_error in node1.query_and_get_error(
|
|
f"RESTORE DATABASE mydb FROM {backup_name}"
|
|
)
|
|
|
|
|
|
def test_replicated_table_with_not_synced_insert():
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' ("
|
|
"x UInt32"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY x"
|
|
)
|
|
|
|
node1.query("INSERT INTO tbl VALUES (111)")
|
|
node2.query("INSERT INTO tbl VALUES (222)")
|
|
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' tbl")
|
|
node1.query("SYSTEM STOP REPLICATED SENDS ON CLUSTER 'cluster' tbl")
|
|
|
|
node1.query("INSERT INTO tbl VALUES (333)")
|
|
node2.query("INSERT INTO tbl VALUES (444)")
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query(f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}")
|
|
|
|
node1.query(f"DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
node1.query(f"RESTORE TABLE tbl ON CLUSTER 'cluster' FROM {backup_name}")
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' tbl")
|
|
|
|
assert node1.query("SELECT * FROM tbl ORDER BY x") == TSV([111, 222, 333, 444])
|
|
assert node2.query("SELECT * FROM tbl ORDER BY x") == TSV([111, 222, 333, 444])
|
|
|
|
|
|
def test_replicated_table_with_not_synced_merge():
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' ("
|
|
"x UInt32"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY x"
|
|
)
|
|
|
|
node1.query("SYSTEM STOP MERGES ON CLUSTER 'cluster' tbl")
|
|
|
|
node1.query("INSERT INTO tbl VALUES (111)")
|
|
node1.query("INSERT INTO tbl VALUES (222)")
|
|
|
|
node2.query("SYSTEM SYNC REPLICA tbl")
|
|
|
|
node2.query("SYSTEM START MERGES tbl")
|
|
node2.query("OPTIMIZE TABLE tbl FINAL")
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query(f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}")
|
|
|
|
node1.query(f"DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
node1.query(f"RESTORE TABLE tbl ON CLUSTER 'cluster' FROM {backup_name}")
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' tbl")
|
|
|
|
assert node1.query("SELECT * FROM tbl ORDER BY x") == TSV([111, 222])
|
|
assert node2.query("SELECT * FROM tbl ORDER BY x") == TSV([111, 222])
|
|
|
|
|
|
def test_replicated_table_restored_into_bigger_cluster():
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' ("
|
|
"x UInt32"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY x"
|
|
)
|
|
|
|
node1.query("INSERT INTO tbl VALUES (111)")
|
|
node2.query("INSERT INTO tbl VALUES (222)")
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query(f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}")
|
|
|
|
node1.query("DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
node1.query(f"RESTORE TABLE tbl ON CLUSTER 'cluster3' FROM {backup_name}")
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster3' tbl")
|
|
|
|
assert node1.query("SELECT * FROM tbl ORDER BY x") == TSV([111, 222])
|
|
assert node2.query("SELECT * FROM tbl ORDER BY x") == TSV([111, 222])
|
|
assert node3.query("SELECT * FROM tbl ORDER BY x") == TSV([111, 222])
|
|
|
|
|
|
def test_replicated_table_restored_into_smaller_cluster():
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' ("
|
|
"x UInt32"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY x"
|
|
)
|
|
|
|
node1.query("INSERT INTO tbl VALUES (111)")
|
|
node2.query("INSERT INTO tbl VALUES (222)")
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query(f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}")
|
|
|
|
node1.query("DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
node1.query(f"RESTORE TABLE tbl ON CLUSTER 'cluster1' FROM {backup_name}")
|
|
assert node1.query("SELECT * FROM tbl ORDER BY x") == TSV([111, 222])
|
|
|
|
|
|
def test_replicated_database_async():
|
|
node1.query(
|
|
"CREATE DATABASE mydb ON CLUSTER 'cluster' ENGINE=Replicated('/clickhouse/path/','{shard}','{replica}')"
|
|
)
|
|
|
|
node1.query("CREATE TABLE mydb.tbl(x UInt8) ENGINE=ReplicatedMergeTree ORDER BY x")
|
|
|
|
node1.query(
|
|
"CREATE TABLE mydb.tbl2(y String) ENGINE=ReplicatedMergeTree ORDER BY y"
|
|
)
|
|
|
|
node2.query("SYSTEM SYNC DATABASE REPLICA mydb")
|
|
|
|
node1.query("INSERT INTO mydb.tbl VALUES (1)")
|
|
node1.query("INSERT INTO mydb.tbl VALUES (22)")
|
|
node2.query("INSERT INTO mydb.tbl2 VALUES ('a')")
|
|
node2.query("INSERT INTO mydb.tbl2 VALUES ('bb')")
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' mydb.tbl")
|
|
|
|
backup_name = new_backup_name()
|
|
[id, status] = node1.query(
|
|
f"BACKUP DATABASE mydb ON CLUSTER 'cluster' TO {backup_name} ASYNC"
|
|
).split("\t")
|
|
|
|
assert status == "CREATING_BACKUP\n" or status == "BACKUP_CREATED\n"
|
|
|
|
assert_eq_with_retry(
|
|
node1,
|
|
f"SELECT status, error FROM system.backups WHERE id='{id}'",
|
|
TSV([["BACKUP_CREATED", ""]]),
|
|
)
|
|
|
|
node1.query("DROP DATABASE mydb ON CLUSTER 'cluster' SYNC")
|
|
|
|
[id, status] = node1.query(
|
|
f"RESTORE DATABASE mydb ON CLUSTER 'cluster' FROM {backup_name} ASYNC"
|
|
).split("\t")
|
|
|
|
assert status == "RESTORING\n" or status == "RESTORED\n"
|
|
|
|
assert_eq_with_retry(
|
|
node1,
|
|
f"SELECT status, error FROM system.backups WHERE id='{id}'",
|
|
TSV([["RESTORED", ""]]),
|
|
)
|
|
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' mydb.tbl")
|
|
|
|
assert node1.query("SELECT * FROM mydb.tbl ORDER BY x") == TSV([1, 22])
|
|
assert node2.query("SELECT * FROM mydb.tbl2 ORDER BY y") == TSV(["a", "bb"])
|
|
|
|
|
|
# By default `backup_restore_keeper_value_max_size` is 1 MB, but in this test we'll set it to 50 bytes just to check it works.
|
|
def test_keeper_value_max_size():
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' ("
|
|
"x UInt32"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY x"
|
|
)
|
|
|
|
node1.query("INSERT INTO tbl VALUES (111)")
|
|
node2.query("INSERT INTO tbl VALUES (222)")
|
|
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' tbl")
|
|
node1.query("SYSTEM STOP REPLICATED SENDS ON CLUSTER 'cluster' tbl")
|
|
|
|
node1.query("INSERT INTO tbl VALUES (333)")
|
|
node2.query("INSERT INTO tbl VALUES (444)")
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query(
|
|
f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}",
|
|
settings={"backup_restore_keeper_value_max_size": 50},
|
|
)
|
|
|
|
node1.query(f"DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
node1.query(f"RESTORE TABLE tbl ON CLUSTER 'cluster' FROM {backup_name}")
|
|
node1.query("SYSTEM SYNC REPLICA ON CLUSTER 'cluster' tbl")
|
|
|
|
assert node1.query("SELECT * FROM tbl ORDER BY x") == TSV([111, 222, 333, 444])
|
|
assert node2.query("SELECT * FROM tbl ORDER BY x") == TSV([111, 222, 333, 444])
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"interface, on_cluster", [("native", True), ("http", True), ("http", False)]
|
|
)
|
|
def test_async_backups_to_same_destination(interface, on_cluster):
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' ("
|
|
"x UInt8"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY x"
|
|
)
|
|
|
|
nodes = [node1, node2]
|
|
node1.query("INSERT INTO tbl VALUES (1)")
|
|
|
|
backup_name = new_backup_name()
|
|
on_cluster_part = "ON CLUSTER 'cluster'" if on_cluster else ""
|
|
|
|
# Multiple backups to the same destination.
|
|
ids = []
|
|
for node in nodes:
|
|
if interface == "http":
|
|
res, err = node.http_query_and_get_answer_with_error(
|
|
f"BACKUP TABLE tbl {on_cluster_part} TO {backup_name} ASYNC"
|
|
)
|
|
else:
|
|
res, err = node.query_and_get_answer_with_error(
|
|
f"BACKUP TABLE tbl {on_cluster_part} TO {backup_name} ASYNC"
|
|
)
|
|
|
|
# The second backup to the same destination is expected to fail. It can either fail immediately or after a while.
|
|
# If it fails immediately we won't even get its ID.
|
|
if not err:
|
|
ids.append(res.split("\t")[0])
|
|
|
|
ids_for_query = "[" + ", ".join(f"'{id}'" for id in ids) + "]"
|
|
|
|
for node in nodes:
|
|
assert_eq_with_retry(
|
|
node,
|
|
f"SELECT status FROM system.backups WHERE id IN {ids_for_query} AND status == 'CREATING_BACKUP'",
|
|
"",
|
|
)
|
|
|
|
num_created_backups = sum(
|
|
[
|
|
int(
|
|
node.query(
|
|
f"SELECT count() FROM system.backups WHERE id IN {ids_for_query} AND status == 'BACKUP_CREATED'"
|
|
).strip()
|
|
)
|
|
for node in nodes
|
|
]
|
|
)
|
|
|
|
num_failed_backups = sum(
|
|
[
|
|
int(
|
|
node.query(
|
|
f"SELECT count() FROM system.backups WHERE id IN {ids_for_query} AND status == 'BACKUP_FAILED'"
|
|
).strip()
|
|
)
|
|
for node in nodes
|
|
]
|
|
)
|
|
|
|
# Only one backup should succeed.
|
|
if (num_created_backups != 1) or (num_failed_backups != len(ids) - 1):
|
|
for node in nodes:
|
|
print(
|
|
node.query(
|
|
f"SELECT status, error FROM system.backups WHERE id IN {ids_for_query}"
|
|
)
|
|
)
|
|
|
|
assert num_created_backups == 1
|
|
assert num_failed_backups == len(ids) - 1
|
|
|
|
# Check that the succeeded backup is all right.
|
|
node1.query("DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
node1.query(f"RESTORE TABLE tbl FROM {backup_name}")
|
|
assert node1.query("SELECT * FROM tbl") == "1\n"
|
|
|
|
|
|
def test_required_privileges():
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' ("
|
|
"x UInt8"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY x"
|
|
)
|
|
|
|
node1.query("INSERT INTO tbl VALUES (100)")
|
|
|
|
node1.query("CREATE USER u1")
|
|
node1.query("GRANT CLUSTER ON *.* TO u1")
|
|
|
|
backup_name = new_backup_name()
|
|
expected_error = "necessary to have grant BACKUP ON default.tbl"
|
|
assert expected_error in node1.query_and_get_error(
|
|
f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}", user="u1"
|
|
)
|
|
|
|
node1.query("GRANT BACKUP ON tbl TO u1")
|
|
node1.query(f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}", user="u1")
|
|
|
|
node1.query(f"DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
expected_error = "necessary to have grant INSERT, CREATE TABLE ON default.tbl2"
|
|
assert expected_error in node1.query_and_get_error(
|
|
f"RESTORE TABLE tbl AS tbl2 ON CLUSTER 'cluster' FROM {backup_name}", user="u1"
|
|
)
|
|
|
|
node1.query("GRANT INSERT, CREATE TABLE ON tbl2 TO u1")
|
|
node1.query(
|
|
f"RESTORE TABLE tbl AS tbl2 ON CLUSTER 'cluster' FROM {backup_name}", user="u1"
|
|
)
|
|
|
|
assert node2.query("SELECT * FROM tbl2") == "100\n"
|
|
|
|
node1.query(f"DROP TABLE tbl2 ON CLUSTER 'cluster' SYNC")
|
|
node1.query("REVOKE ALL FROM u1")
|
|
|
|
expected_error = "necessary to have grant INSERT, CREATE TABLE ON default.tbl"
|
|
assert expected_error in node1.query_and_get_error(
|
|
f"RESTORE ALL ON CLUSTER 'cluster' FROM {backup_name}", user="u1"
|
|
)
|
|
|
|
node1.query("GRANT INSERT, CREATE TABLE ON tbl TO u1")
|
|
node1.query(f"RESTORE ALL ON CLUSTER 'cluster' FROM {backup_name}", user="u1")
|
|
|
|
assert node2.query("SELECT * FROM tbl") == "100\n"
|
|
|
|
|
|
def test_system_users():
|
|
node1.query("CREATE USER u1 SETTINGS custom_a=123")
|
|
node1.query("GRANT SELECT ON tbl TO u1")
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query("CREATE USER u2 SETTINGS allow_backup=false")
|
|
node1.query("GRANT CLUSTER ON *.* TO u2")
|
|
|
|
expected_error = "necessary to have grant BACKUP ON system.users"
|
|
assert expected_error in node1.query_and_get_error(
|
|
f"BACKUP TABLE system.users ON CLUSTER 'cluster' TO {backup_name}", user="u2"
|
|
)
|
|
|
|
node1.query("GRANT BACKUP ON system.users TO u2")
|
|
node1.query(
|
|
f"BACKUP TABLE system.users ON CLUSTER 'cluster' TO {backup_name}", user="u2"
|
|
)
|
|
|
|
node1.query("DROP USER u1")
|
|
|
|
expected_error = "necessary to have grant CREATE USER ON *.*"
|
|
assert expected_error in node1.query_and_get_error(
|
|
f"RESTORE TABLE system.users ON CLUSTER 'cluster' FROM {backup_name}", user="u2"
|
|
)
|
|
|
|
node1.query("GRANT CREATE USER ON *.* TO u2")
|
|
|
|
expected_error = "necessary to have grant SELECT ON default.tbl WITH GRANT OPTION"
|
|
assert expected_error in node1.query_and_get_error(
|
|
f"RESTORE TABLE system.users ON CLUSTER 'cluster' FROM {backup_name}", user="u2"
|
|
)
|
|
|
|
node1.query("GRANT SELECT ON tbl TO u2 WITH GRANT OPTION")
|
|
node1.query(
|
|
f"RESTORE TABLE system.users ON CLUSTER 'cluster' FROM {backup_name}", user="u2"
|
|
)
|
|
|
|
assert (
|
|
node1.query("SHOW CREATE USER u1") == "CREATE USER u1 SETTINGS custom_a = 123\n"
|
|
)
|
|
assert node1.query("SHOW GRANTS FOR u1") == "GRANT SELECT ON default.tbl TO u1\n"
|
|
|
|
|
|
def test_system_functions():
|
|
node1.query("CREATE FUNCTION linear_equation AS (x, k, b) -> k*x + b;")
|
|
|
|
node1.query("CREATE FUNCTION parity_str AS (n) -> if(n % 2, 'odd', 'even');")
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query(f"BACKUP TABLE system.functions ON CLUSTER 'cluster' TO {backup_name}")
|
|
|
|
node1.query("DROP FUNCTION linear_equation")
|
|
node1.query("DROP FUNCTION parity_str")
|
|
assert_eq_with_retry(
|
|
node2, "SELECT name FROM system.functions WHERE name='parity_str'", ""
|
|
)
|
|
|
|
node1.query(
|
|
f"RESTORE TABLE system.functions ON CLUSTER 'cluster' FROM {backup_name}"
|
|
)
|
|
|
|
assert node1.query(
|
|
"SELECT number, linear_equation(number, 2, 1) FROM numbers(3)"
|
|
) == TSV([[0, 1], [1, 3], [2, 5]])
|
|
|
|
assert node1.query("SELECT number, parity_str(number) FROM numbers(3)") == TSV(
|
|
[[0, "even"], [1, "odd"], [2, "even"]]
|
|
)
|
|
|
|
assert node2.query(
|
|
"SELECT number, linear_equation(number, 2, 1) FROM numbers(3)"
|
|
) == TSV([[0, 1], [1, 3], [2, 5]])
|
|
|
|
assert node2.query("SELECT number, parity_str(number) FROM numbers(3)") == TSV(
|
|
[[0, "even"], [1, "odd"], [2, "even"]]
|
|
)
|
|
|
|
assert_eq_with_retry(
|
|
node2,
|
|
"SELECT name FROM system.functions WHERE name='parity_str'",
|
|
"parity_str\n",
|
|
)
|
|
assert node2.query("SELECT number, parity_str(number) FROM numbers(3)") == TSV(
|
|
[[0, "even"], [1, "odd"], [2, "even"]]
|
|
)
|
|
|
|
|
|
def test_projection():
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' (x UInt32, y String) ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}') "
|
|
"ORDER BY y PARTITION BY x%10"
|
|
)
|
|
node1.query(f"INSERT INTO tbl SELECT number, toString(number) FROM numbers(3)")
|
|
|
|
node1.query("ALTER TABLE tbl ADD PROJECTION prjmax (SELECT MAX(x))")
|
|
node1.query(f"INSERT INTO tbl VALUES (100, 'a'), (101, 'b')")
|
|
|
|
assert (
|
|
node1.query(
|
|
"SELECT count() FROM system.projection_parts WHERE database='default' AND table='tbl' AND name='prjmax'"
|
|
)
|
|
== "2\n"
|
|
)
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query(f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}")
|
|
|
|
node1.query(f"DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
assert (
|
|
node1.query(
|
|
"SELECT count() FROM system.projection_parts WHERE database='default' AND table='tbl' AND name='prjmax'"
|
|
)
|
|
== "0\n"
|
|
)
|
|
|
|
node1.query(f"RESTORE TABLE tbl FROM {backup_name}")
|
|
|
|
assert node1.query("SELECT * FROM tbl ORDER BY x") == TSV(
|
|
[[0, "0"], [1, "1"], [2, "2"], [100, "a"], [101, "b"]]
|
|
)
|
|
|
|
assert (
|
|
node1.query(
|
|
"SELECT count() FROM system.projection_parts WHERE database='default' AND table='tbl' AND name='prjmax'"
|
|
)
|
|
== "2\n"
|
|
)
|
|
|
|
|
|
def test_replicated_table_with_not_synced_def():
|
|
node1.query(
|
|
"CREATE TABLE tbl ("
|
|
"x UInt8, y String"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY tuple()"
|
|
)
|
|
|
|
node2.query(
|
|
"CREATE TABLE tbl ("
|
|
"x UInt8, y String"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY tuple()"
|
|
)
|
|
|
|
node2.query("SYSTEM STOP REPLICATION QUEUES tbl")
|
|
node1.query("ALTER TABLE tbl MODIFY COLUMN x String")
|
|
|
|
# Not synced because the replication queue is stopped
|
|
assert node1.query(
|
|
"SELECT name, type FROM system.columns WHERE database='default' AND table='tbl'"
|
|
) == TSV([["x", "String"], ["y", "String"]])
|
|
assert node2.query(
|
|
"SELECT name, type FROM system.columns WHERE database='default' AND table='tbl'"
|
|
) == TSV([["x", "UInt8"], ["y", "String"]])
|
|
|
|
backup_name = new_backup_name()
|
|
node2.query(f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}")
|
|
|
|
node1.query("DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
# But synced after RESTORE anyway
|
|
node1.query(
|
|
f"RESTORE TABLE tbl ON CLUSTER 'cluster' FROM {backup_name} SETTINGS replica_num_in_backup=1"
|
|
)
|
|
assert node1.query(
|
|
"SELECT name, type FROM system.columns WHERE database='default' AND table='tbl'"
|
|
) == TSV([["x", "String"], ["y", "String"]])
|
|
assert node2.query(
|
|
"SELECT name, type FROM system.columns WHERE database='default' AND table='tbl'"
|
|
) == TSV([["x", "String"], ["y", "String"]])
|
|
|
|
node1.query("DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
node2.query(
|
|
f"RESTORE TABLE tbl ON CLUSTER 'cluster' FROM {backup_name} SETTINGS replica_num_in_backup=2"
|
|
)
|
|
assert node1.query(
|
|
"SELECT name, type FROM system.columns WHERE database='default' AND table='tbl'"
|
|
) == TSV([["x", "String"], ["y", "String"]])
|
|
assert node2.query(
|
|
"SELECT name, type FROM system.columns WHERE database='default' AND table='tbl'"
|
|
) == TSV([["x", "String"], ["y", "String"]])
|
|
|
|
|
|
def test_table_in_replicated_database_with_not_synced_def():
|
|
node1.query(
|
|
"CREATE DATABASE mydb ON CLUSTER 'cluster' ENGINE=Replicated('/clickhouse/path/','{shard}','{replica}')"
|
|
)
|
|
|
|
node1.query(
|
|
"CREATE TABLE mydb.tbl (x UInt8, y String) ENGINE=ReplicatedMergeTree ORDER BY tuple()"
|
|
)
|
|
|
|
node1.query("ALTER TABLE mydb.tbl MODIFY COLUMN x String")
|
|
|
|
backup_name = new_backup_name()
|
|
node2.query(f"BACKUP DATABASE mydb ON CLUSTER 'cluster' TO {backup_name}")
|
|
|
|
node1.query("DROP DATABASE mydb ON CLUSTER 'cluster' SYNC")
|
|
|
|
# But synced after RESTORE anyway
|
|
node1.query(
|
|
f"RESTORE DATABASE mydb ON CLUSTER 'cluster' FROM {backup_name} SETTINGS replica_num_in_backup=1"
|
|
)
|
|
assert node1.query(
|
|
"SELECT name, type FROM system.columns WHERE database='mydb' AND table='tbl'"
|
|
) == TSV([["x", "String"], ["y", "String"]])
|
|
assert node2.query(
|
|
"SELECT name, type FROM system.columns WHERE database='mydb' AND table='tbl'"
|
|
) == TSV([["x", "String"], ["y", "String"]])
|
|
|
|
node1.query("DROP DATABASE mydb ON CLUSTER 'cluster' SYNC")
|
|
|
|
node2.query(
|
|
f"RESTORE DATABASE mydb ON CLUSTER 'cluster' FROM {backup_name} SETTINGS replica_num_in_backup=2"
|
|
)
|
|
assert node1.query(
|
|
"SELECT name, type FROM system.columns WHERE database='mydb' AND table='tbl'"
|
|
) == TSV([["x", "String"], ["y", "String"]])
|
|
assert node2.query(
|
|
"SELECT name, type FROM system.columns WHERE database='mydb' AND table='tbl'"
|
|
) == TSV([["x", "String"], ["y", "String"]])
|
|
|
|
|
|
def has_mutation_in_backup(mutation_id, backup_name, database, table):
|
|
return (
|
|
os.path.exists(
|
|
os.path.join(
|
|
get_path_to_backup(backup_name),
|
|
f"shards/1/replicas/1/data/{database}/{table}/mutations/{mutation_id}.txt",
|
|
)
|
|
)
|
|
or os.path.exists(
|
|
os.path.join(
|
|
get_path_to_backup(backup_name),
|
|
f"shards/1/replicas/2/data/{database}/{table}/mutations/{mutation_id}.txt",
|
|
)
|
|
)
|
|
or os.path.exists(
|
|
os.path.join(
|
|
get_path_to_backup(backup_name),
|
|
f"shards/1/replicas/3/data/{database}/{table}/mutations/{mutation_id}.txt",
|
|
)
|
|
)
|
|
)
|
|
|
|
|
|
def test_mutation():
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' ("
|
|
"x UInt8, y String"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY tuple()"
|
|
)
|
|
|
|
node1.query("INSERT INTO tbl SELECT number, toString(number) FROM numbers(5)")
|
|
|
|
node2.query("INSERT INTO tbl SELECT number, toString(number) FROM numbers(5, 5)")
|
|
|
|
node1.query("INSERT INTO tbl SELECT number, toString(number) FROM numbers(10, 5)")
|
|
|
|
node1.query("ALTER TABLE tbl UPDATE x=x+1 WHERE 1")
|
|
node1.query("ALTER TABLE tbl UPDATE x=x+1+sleep(3) WHERE 1")
|
|
node1.query("ALTER TABLE tbl UPDATE x=x+1+sleep(3) WHERE 1")
|
|
|
|
backup_name = new_backup_name()
|
|
node1.query(f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}")
|
|
|
|
assert not has_mutation_in_backup("0000000000", backup_name, "default", "tbl")
|
|
assert has_mutation_in_backup("0000000001", backup_name, "default", "tbl")
|
|
assert has_mutation_in_backup("0000000002", backup_name, "default", "tbl")
|
|
assert not has_mutation_in_backup("0000000003", backup_name, "default", "tbl")
|
|
|
|
node1.query("DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
|
|
node1.query(f"RESTORE TABLE tbl ON CLUSTER 'cluster' FROM {backup_name}")
|
|
|
|
|
|
def test_tables_dependency():
|
|
node1.query("CREATE DATABASE mydb ON CLUSTER 'cluster3'")
|
|
|
|
node1.query(
|
|
"CREATE TABLE mydb.src ON CLUSTER 'cluster' (x Int64, y String) ENGINE=MergeTree ORDER BY tuple()"
|
|
)
|
|
|
|
node1.query(
|
|
"CREATE DICTIONARY mydb.dict ON CLUSTER 'cluster' (x Int64, y String) PRIMARY KEY x "
|
|
"SOURCE(CLICKHOUSE(HOST 'localhost' PORT tcpPort() DB 'mydb' TABLE 'src')) LAYOUT(FLAT()) LIFETIME(0)"
|
|
)
|
|
|
|
node1.query(
|
|
"CREATE TABLE mydb.dist1 (x Int64) ENGINE=Distributed('cluster', 'mydb', 'src')"
|
|
)
|
|
|
|
node3.query(
|
|
"CREATE TABLE mydb.dist2 (x Int64) ENGINE=Distributed(cluster, 'mydb', 'src')"
|
|
)
|
|
|
|
node1.query("CREATE TABLE mydb.clusterfunc1 AS cluster('cluster', 'mydb.src')")
|
|
node1.query("CREATE TABLE mydb.clusterfunc2 AS cluster(cluster, mydb.src)")
|
|
node1.query("CREATE TABLE mydb.clusterfunc3 AS cluster(cluster, 'mydb', 'src')")
|
|
node1.query(
|
|
"CREATE TABLE mydb.clusterfunc4 AS cluster(cluster, dictionary(mydb.dict))"
|
|
)
|
|
node1.query(
|
|
"CREATE TABLE mydb.clusterfunc5 AS clusterAllReplicas(cluster, dictionary(mydb.dict))"
|
|
)
|
|
|
|
node3.query("CREATE TABLE mydb.clusterfunc6 AS cluster('cluster', 'mydb.src')")
|
|
node3.query("CREATE TABLE mydb.clusterfunc7 AS cluster(cluster, mydb.src)")
|
|
node3.query("CREATE TABLE mydb.clusterfunc8 AS cluster(cluster, 'mydb', 'src')")
|
|
node3.query(
|
|
"CREATE TABLE mydb.clusterfunc9 AS cluster(cluster, dictionary(mydb.dict))"
|
|
)
|
|
node3.query(
|
|
"CREATE TABLE mydb.clusterfunc10 AS clusterAllReplicas(cluster, dictionary(mydb.dict))"
|
|
)
|
|
|
|
backup_name = new_backup_name()
|
|
node3.query(f"BACKUP DATABASE mydb ON CLUSTER 'cluster3' TO {backup_name}")
|
|
|
|
node3.query("DROP DATABASE mydb")
|
|
|
|
node3.query(f"RESTORE DATABASE mydb ON CLUSTER 'cluster3' FROM {backup_name}")
|
|
|
|
node3.query("SYSTEM FLUSH LOGS ON CLUSTER 'cluster3'")
|
|
expect_in_logs_1 = [
|
|
"Table mydb.src has no dependencies (level 0)",
|
|
"Table mydb.dict has 1 dependencies: mydb.src (level 1)",
|
|
"Table mydb.dist1 has 1 dependencies: mydb.src (level 1)",
|
|
"Table mydb.clusterfunc1 has 1 dependencies: mydb.src (level 1)",
|
|
"Table mydb.clusterfunc2 has 1 dependencies: mydb.src (level 1)",
|
|
"Table mydb.clusterfunc3 has 1 dependencies: mydb.src (level 1)",
|
|
"Table mydb.clusterfunc4 has 1 dependencies: mydb.dict (level 2)",
|
|
"Table mydb.clusterfunc5 has 1 dependencies: mydb.dict (level 2)",
|
|
]
|
|
expect_in_logs_2 = [
|
|
"Table mydb.src has no dependencies (level 0)",
|
|
"Table mydb.dict has 1 dependencies: mydb.src (level 1)",
|
|
]
|
|
expect_in_logs_3 = [
|
|
"Table mydb.dist2 has no dependencies (level 0)",
|
|
"Table mydb.clusterfunc6 has no dependencies (level 0)",
|
|
"Table mydb.clusterfunc7 has no dependencies (level 0)",
|
|
"Table mydb.clusterfunc8 has no dependencies (level 0)",
|
|
"Table mydb.clusterfunc9 has no dependencies (level 0)",
|
|
"Table mydb.clusterfunc10 has no dependencies (level 0)",
|
|
]
|
|
for expect in expect_in_logs_1:
|
|
assert node1.contains_in_log(f"RestorerFromBackup: {expect}")
|
|
for expect in expect_in_logs_2:
|
|
assert node2.contains_in_log(f"RestorerFromBackup: {expect}")
|
|
for expect in expect_in_logs_3:
|
|
assert node3.contains_in_log(f"RestorerFromBackup: {expect}")
|
|
|
|
|
|
def test_get_error_from_other_host():
|
|
node1.query("CREATE TABLE tbl (`x` UInt8) ENGINE = MergeTree ORDER BY x")
|
|
node1.query("INSERT INTO tbl VALUES (3)")
|
|
|
|
backup_name = new_backup_name()
|
|
expected_error = "Got error from node2.*Table default.tbl was not found"
|
|
assert re.search(
|
|
expected_error,
|
|
node1.query_and_get_error(
|
|
f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name}"
|
|
),
|
|
)
|
|
|
|
|
|
@pytest.mark.parametrize("kill", [False, True])
|
|
def test_stop_other_host_during_backup(kill):
|
|
node1.query(
|
|
"CREATE TABLE tbl ON CLUSTER 'cluster' ("
|
|
"x UInt8"
|
|
") ENGINE=ReplicatedMergeTree('/clickhouse/tables/tbl/', '{replica}')"
|
|
"ORDER BY x"
|
|
)
|
|
|
|
node1.query("INSERT INTO tbl VALUES (3)")
|
|
node2.query("INSERT INTO tbl VALUES (5)")
|
|
|
|
backup_name = new_backup_name()
|
|
|
|
id = node1.query(
|
|
f"BACKUP TABLE tbl ON CLUSTER 'cluster' TO {backup_name} ASYNC"
|
|
).split("\t")[0]
|
|
|
|
# If kill=False the pending backup must be completed
|
|
# If kill=True the pending backup might be completed or failed
|
|
node2.stop_clickhouse(kill=kill)
|
|
|
|
assert_eq_with_retry(
|
|
node1,
|
|
f"SELECT status FROM system.backups WHERE id='{id}' AND status == 'CREATING_BACKUP'",
|
|
"",
|
|
retry_count=100,
|
|
)
|
|
|
|
status = node1.query(f"SELECT status FROM system.backups WHERE id='{id}'").strip()
|
|
|
|
if kill:
|
|
assert status in ["BACKUP_CREATED", "BACKUP_FAILED"]
|
|
else:
|
|
assert status == "BACKUP_CREATED"
|
|
|
|
node2.start_clickhouse()
|
|
|
|
if status == "BACKUP_CREATED":
|
|
node1.query("DROP TABLE tbl ON CLUSTER 'cluster' SYNC")
|
|
node1.query(f"RESTORE TABLE tbl ON CLUSTER 'cluster' FROM {backup_name}")
|
|
assert node1.query("SELECT * FROM tbl ORDER BY x") == TSV([3, 5])
|
|
elif status == "BACKUP_FAILED":
|
|
assert not os.path.exists(
|
|
os.path.join(get_path_to_backup(backup_name), ".backup")
|
|
)
|