Merge pull request #17486 from ClickHouse/fix_more_flaky_tests

Fix more flaky integration tests

(cherry picked from commit 1d179dac)

docker/test/integration/runner/compose/docker_compose_kerberized_kafka.yml

@@ -50,7 +50,7 @@ services:
         - label:disable
 
   kafka_kerberos:
-    image: yandex/clickhouse-kerberos-kdc:${DOCKER_KERBEROS_KDC_TAG}
+    image: yandex/clickhouse-kerberos-kdc:${DOCKER_KERBEROS_KDC_TAG:-latest}
     hostname: kafka_kerberos
     volumes:
         - ${KERBERIZED_KAFKA_DIR}/secrets:/tmp/keytab

docker/test/integration/runner/compose/docker_compose_mysql_golang_client.yml

 version: '2.3'
 services:
   golang1:
-    image: yandex/clickhouse-mysql-golang-client:${DOCKER_MYSQL_GOLANG_CLIENT_TAG}
+    image: yandex/clickhouse-mysql-golang-client:${DOCKER_MYSQL_GOLANG_CLIENT_TAG:-latest}
     # to keep container running
     command: sleep infinity

docker/test/integration/runner/compose/docker_compose_mysql_java_client.yml

 version: '2.3'
 services:
   java1:
-    image: yandex/clickhouse-mysql-java-client:${DOCKER_MYSQL_JAVA_CLIENT_TAG}
+    image: yandex/clickhouse-mysql-java-client:${DOCKER_MYSQL_JAVA_CLIENT_TAG:-latest}
     # to keep container running
     command: sleep infinity

docker/test/integration/runner/compose/docker_compose_mysql_js_client.yml

 version: '2.3'
 services:
   mysqljs1:
-    image: yandex/clickhouse-mysql-js-client:${DOCKER_MYSQL_JS_CLIENT_TAG}
+    image: yandex/clickhouse-mysql-js-client:${DOCKER_MYSQL_JS_CLIENT_TAG:-latest}
     # to keep container running
     command: sleep infinity

docker/test/integration/runner/compose/docker_compose_mysql_php_client.yml

 version: '2.3'
 services:
   php1:
-    image: yandex/clickhouse-mysql-php-client:${DOCKER_MYSQL_PHP_CLIENT_TAG}
+    image: yandex/clickhouse-mysql-php-client:${DOCKER_MYSQL_PHP_CLIENT_TAG:-latest}
     # to keep container running
     command: sleep infinity

docker/test/integration/runner/compose/docker_compose_postgesql_java_client.yml

 version: '2.2'
 services:
   java:
-    image: yandex/clickhouse-postgresql-java-client:${DOCKER_POSTGRESQL_JAVA_CLIENT_TAG}
+    image: yandex/clickhouse-postgresql-java-client:${DOCKER_POSTGRESQL_JAVA_CLIENT_TAG:-latest}
     # to keep container running
     command: sleep infinity

tests/integration/helpers/network.py

@@ -179,16 +179,19 @@ class _NetworkManager:
                 except docker.errors.NotFound:
                     pass
 
-            # for some reason docker api may hang if image doesn't exist, so we download it
-            # before running
-            for i in range(5):
-                try:
-                    subprocess.check_call("docker pull yandex/clickhouse-integration-helper", shell=True)
-                    break
-                except:
-                    time.sleep(i)
-            else:
-                raise Exception("Cannot pull yandex/clickhouse-integration-helper image")
+            image = subprocess.check_output("docker images -q yandex/clickhouse-integration-helper 2>/dev/null", shell=True)
+            if not image.strip():
+                print("No network image helper, will try download")
+                # for some reason docker api may hang if image doesn't exist, so we download it
+                # before running
+                for i in range(5):
+                    try:
+                        subprocess.check_call("docker pull yandex/clickhouse-integration-helper", shell=True)
+                        break
+                    except:
+                        time.sleep(i)
+                else:
+                    raise Exception("Cannot pull yandex/clickhouse-integration-helper image")
 
             self._container = self._docker_client.containers.run('yandex/clickhouse-integration-helper',
                                                                  auto_remove=True,

tests/integration/test_reload_zookeeper/test.py

@@ -57,6 +57,7 @@ def test_reload_zookeeper(start_cluster):
 </yandex >
 """
     node.replace_config("/etc/clickhouse-server/conf.d/zookeeper.xml", new_config)
+    node.query("SYSTEM RELOAD CONFIG")
     ## config reloads, but can still work
     assert_eq_with_retry(node, "SELECT COUNT() FROM test_table", '1000', retry_count=120, sleep_time=0.5)
 
@@ -86,6 +87,7 @@ def test_reload_zookeeper(start_cluster):
 </yandex>
 """
     node.replace_config("/etc/clickhouse-server/conf.d/zookeeper.xml", new_config)
+    node.query("SYSTEM RELOAD CONFIG")
 
     assert_eq_with_retry(node, "SELECT COUNT() FROM test_table", '1000', retry_count=120, sleep_time=0.5)
 

tests/integration/test_settings_constraints_distributed/test.py

@@ -17,19 +17,14 @@ distributed = cluster.add_instance('distributed', main_configs=["configs/config.
 def started_cluster():
     try:
         cluster.start()
-
         for node in [node1, node2]:
-            node.query("CREATE TABLE sometable (date Date, id UInt32, value Int32) ENGINE = MergeTree() ORDER BY id;")
-            node.query("INSERT INTO sometable VALUES (toDate('2010-01-10'), 1, 1)")
             node.query("CREATE USER shard")
             node.query("GRANT ALL ON *.* TO shard")
 
-        distributed.query(
-            "CREATE TABLE proxy (date Date, id UInt32, value Int32) ENGINE = Distributed(test_cluster, default, sometable, toUInt64(date));")
-        distributed.query(
-            "CREATE TABLE shard_settings (name String, value String) ENGINE = Distributed(test_cluster, system, settings);")
         distributed.query("CREATE ROLE admin")
         distributed.query("GRANT ALL ON *.* TO admin")
+        distributed.query(
+            "CREATE TABLE shard_settings (name String, value String) ENGINE = Distributed(test_cluster, system, settings);")
 
         yield cluster
 
@@ -46,6 +41,14 @@ def restart_distributed():
 
 
 def test_select_clamps_settings():
+    for node in [node1, node2]:
+        node.query("CREATE TABLE sometable_select (date Date, id UInt32, value Int32) ENGINE = MergeTree() ORDER BY id;")
+        node.query("INSERT INTO sometable_select VALUES (toDate('2010-01-10'), 1, 1)")
+
+    distributed.query(
+        "CREATE TABLE proxy_select (date Date, id UInt32, value Int32) ENGINE = Distributed(test_cluster, default, sometable_select, toUInt64(date));")
+
+
     distributed.query("CREATE USER normal DEFAULT ROLE admin SETTINGS max_memory_usage = 80000000")
     distributed.query("CREATE USER wasteful DEFAULT ROLE admin SETTINGS max_memory_usage = 2000000000")
     distributed.query("CREATE USER readonly DEFAULT ROLE admin SETTINGS readonly = 1")
@@ -53,7 +56,7 @@ def test_select_clamps_settings():
     node2.query("ALTER USER shard SETTINGS readonly = 1")
 
     # Check that shards doesn't throw exceptions on constraints violation
-    query = "SELECT COUNT() FROM proxy"
+    query = "SELECT COUNT() FROM proxy_select"
     assert distributed.query(query) == '2\n'
     assert distributed.query(query, user='normal') == '2\n'
     assert distributed.query(query, user='wasteful') == '2\n'
@@ -62,7 +65,7 @@ def test_select_clamps_settings():
     assert distributed.query(query, settings={"max_memory_usage": 40000000, "readonly": 2}) == '2\n'
     assert distributed.query(query, settings={"max_memory_usage": 3000000000, "readonly": 2}) == '2\n'
 
-    query = "SELECT COUNT() FROM remote('node{1,2}', 'default', 'sometable')"
+    query = "SELECT COUNT() FROM remote('node{1,2}', 'default', 'sometable_select')"
     assert distributed.query(query) == '2\n'
     assert distributed.query(query, user='normal') == '2\n'
     assert distributed.query(query, user='wasteful') == '2\n'
@@ -103,10 +106,17 @@ def test_select_clamps_settings():
 
 
 def test_insert_clamps_settings():
+    for node in [node1, node2]:
+        node.query("CREATE TABLE sometable_insert (date Date, id UInt32, value Int32) ENGINE = MergeTree() ORDER BY id;")
+        node.query("INSERT INTO sometable_insert VALUES (toDate('2010-01-10'), 1, 1)")
+
+    distributed.query(
+        "CREATE TABLE proxy_insert (date Date, id UInt32, value Int32) ENGINE = Distributed(test_cluster, default, sometable_insert, toUInt64(date));")
+
     node1.query("ALTER USER shard SETTINGS max_memory_usage = 50000000 MIN 11111111 MAX 99999999")
     node2.query("ALTER USER shard SETTINGS max_memory_usage = 50000000 MIN 11111111 MAX 99999999")
 
-    distributed.query("INSERT INTO proxy VALUES (toDate('2020-02-20'), 2, 2)")
-    distributed.query("INSERT INTO proxy VALUES (toDate('2020-02-21'), 2, 2)", settings={"max_memory_usage": 5000000})
-    distributed.query("SYSTEM FLUSH DISTRIBUTED proxy")
-    assert_eq_with_retry(distributed, "SELECT COUNT() FROM proxy", "4")
+    distributed.query("INSERT INTO proxy_insert VALUES (toDate('2020-02-20'), 2, 2)")
+    distributed.query("INSERT INTO proxy_insert VALUES (toDate('2020-02-21'), 2, 2)", settings={"max_memory_usage": 5000000})
+    distributed.query("SYSTEM FLUSH DISTRIBUTED proxy_insert")
+    assert_eq_with_retry(distributed, "SELECT COUNT() FROM proxy_insert", "4")