ClickHouse/tests/integration/test_keeper_restore_from_snapshot/test.py

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

132 lines
4.1 KiB
Python
Raw Normal View History

2021-03-04 13:40:43 +00:00
#!/usr/bin/env python3
import pytest
from helpers.cluster import ClickHouseCluster
2022-09-06 10:58:14 +00:00
import helpers.keeper_utils as keeper_utils
2021-03-04 13:40:43 +00:00
import random
import string
import os
import time
cluster = ClickHouseCluster(__file__)
node1 = cluster.add_instance(
"node1", main_configs=["configs/enable_keeper1.xml"], stay_alive=True
)
node2 = cluster.add_instance(
"node2", main_configs=["configs/enable_keeper2.xml"], stay_alive=True
)
node3 = cluster.add_instance(
"node3", main_configs=["configs/enable_keeper3.xml"], stay_alive=True
)
2021-03-04 13:40:43 +00:00
from kazoo.client import KazooClient, KazooState
2021-03-04 13:40:43 +00:00
@pytest.fixture(scope="module")
def started_cluster():
try:
cluster.start()
yield cluster
finally:
cluster.shutdown()
2021-03-04 13:40:43 +00:00
def get_fake_zk(nodename, timeout=30.0):
_fake_zk_instance = KazooClient(
hosts=cluster.get_instance_ip(nodename) + ":9181", timeout=timeout
)
2021-03-04 13:40:43 +00:00
_fake_zk_instance.start()
return _fake_zk_instance
2021-03-04 13:40:43 +00:00
def stop_zk(zk):
try:
if zk:
zk.stop()
zk.close()
except:
pass
def test_recover_from_snapshot(started_cluster):
try:
node1_zk = node2_zk = node3_zk = None
node1_zk = get_fake_zk("node1")
node2_zk = get_fake_zk("node2")
node3_zk = get_fake_zk("node3")
node1_zk.create("/test_snapshot_multinode_recover", "somedata".encode())
node2_zk.sync("/test_snapshot_multinode_recover")
node3_zk.sync("/test_snapshot_multinode_recover")
assert node1_zk.get("/test_snapshot_multinode_recover")[0] == b"somedata"
assert node2_zk.get("/test_snapshot_multinode_recover")[0] == b"somedata"
assert node3_zk.get("/test_snapshot_multinode_recover")[0] == b"somedata"
node3.stop_clickhouse(kill=True)
2021-03-06 07:10:51 +00:00
# at least we will have 2 snapshots
2021-03-06 14:14:38 +00:00
for i in range(435):
node1_zk.create(
"/test_snapshot_multinode_recover" + str(i),
("somedata" + str(i)).encode(),
)
2021-03-04 13:40:43 +00:00
2021-03-06 14:14:38 +00:00
for i in range(435):
2021-03-04 13:40:43 +00:00
if i % 10 == 0:
node1_zk.delete("/test_snapshot_multinode_recover" + str(i))
finally:
for zk in [node1_zk, node2_zk, node3_zk]:
stop_zk(zk)
# stale node should recover from leader's snapshot
2021-03-09 07:19:14 +00:00
# with some sanitizers can start longer than 5 seconds
node3.start_clickhouse(20)
2022-09-06 10:58:14 +00:00
keeper_utils.wait_until_connected(cluster, node3)
print("Restarted")
2021-03-04 13:40:43 +00:00
try:
node1_zk = node2_zk = node3_zk = None
node1_zk = get_fake_zk("node1")
node2_zk = get_fake_zk("node2")
node3_zk = get_fake_zk("node3")
node1_zk.sync("/test_snapshot_multinode_recover")
node2_zk.sync("/test_snapshot_multinode_recover")
node3_zk.sync("/test_snapshot_multinode_recover")
assert node1_zk.get("/test_snapshot_multinode_recover")[0] == b"somedata"
assert node2_zk.get("/test_snapshot_multinode_recover")[0] == b"somedata"
assert node3_zk.get("/test_snapshot_multinode_recover")[0] == b"somedata"
2021-03-06 14:14:38 +00:00
for i in range(435):
2021-03-04 13:40:43 +00:00
if i % 10 != 0:
assert (
node1_zk.get("/test_snapshot_multinode_recover" + str(i))[0]
== ("somedata" + str(i)).encode()
)
assert (
node2_zk.get("/test_snapshot_multinode_recover" + str(i))[0]
== ("somedata" + str(i)).encode()
)
assert (
node3_zk.get("/test_snapshot_multinode_recover" + str(i))[0]
== ("somedata" + str(i)).encode()
)
2021-03-04 13:40:43 +00:00
else:
assert (
node1_zk.exists("/test_snapshot_multinode_recover" + str(i)) is None
)
assert (
node2_zk.exists("/test_snapshot_multinode_recover" + str(i)) is None
)
assert (
node3_zk.exists("/test_snapshot_multinode_recover" + str(i)) is None
)
2021-03-04 13:40:43 +00:00
finally:
for zk in [node1_zk, node2_zk, node3_zk]:
stop_zk(zk)