Files
impala/tests/custom_cluster/test_web_pages.py
Riza Suminto ff5cbadec4 IMPALA-12864: Deflake test_query_log_size_in_bytes.
test_query_log_size_in_bytes is flaky in exhaustive builds. The expected
QueryStateRecord is off by around 100KB per query, indicating that the
test query might be too complex and lead to variability in final query
profile that is being archived.

This patch simplifies the test query and relaxes the assertion. This
patch also adds QUERY_LOG_EST_TOTAL_BYTES metric (key name
"impala-server.query-log-est-total-bytes") that is validated as well in
test_query_log_size_in_bytes. query-state-record-test.cc is modified a
bit to resolve segmentation fault issue when building
unified-be-test-validated-executable target run_clang_tidy.sh.

Testing:
- Pass test_query_log_size_in_bytes in exhaustive build.

Change-Id: Ic9b543ffa931a01c11d58e45774ea46af78b3a25
Reviewed-on: http://gerrit.cloudera.org:8080/21100
Reviewed-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
Tested-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
2024-03-05 02:51:30 +00:00

439 lines
18 KiB
Python

# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
from __future__ import absolute_import, division, print_function
import json
import random
import re
import requests
import psutil
import pytest
import time
from tests.beeswax.impala_beeswax import ImpalaBeeswaxException
from tests.common.custom_cluster_test_suite import (
DEFAULT_CLUSTER_SIZE,
CustomClusterTestSuite)
from tests.shell.util import run_impala_shell_cmd
SMALL_QUERY_LOG_SIZE_IN_BYTES = 40 * 1024
class TestWebPage(CustomClusterTestSuite):
@classmethod
def get_workload(cls):
return 'functional-query'
@classmethod
def setup_class(cls):
if cls.exploration_strategy() != 'exhaustive':
pytest.skip('runs only in exhaustive')
super(TestWebPage, cls).setup_class()
@pytest.mark.execute_serially
@CustomClusterTestSuite.with_args(
impalad_args="--enable_extended_memory_metrics=true"
)
def test_varz_hidden_variables(self):
"""Tests that modified hidden variables show up in /varz"""
response = requests.get("http://localhost:25000/varz?json")
assert response.status_code == requests.codes.ok
varz_json = json.loads(response.text)
flag = [e for e in varz_json["flags"]
if e["name"] == "enable_extended_memory_metrics"]
assert len(flag) == 1
assert flag[0]["default"] == "false"
assert flag[0]["current"] == "true"
assert flag[0]["experimental"]
@pytest.mark.execute_serially
@CustomClusterTestSuite.with_args(
impalad_args="--webserver_max_post_length_bytes=100"
)
def test_max_post_length(self):
"""Tests that the maximum length of a POST request that will be accepted"""
too_big_post_content = "c" * 10000
# POST that exceeds the limit
response = requests.post("http://localhost:25000/", too_big_post_content)
assert response.status_code == requests.codes.request_entity_too_large
# POST within the limit
# This is on a URI that does not understand POST and treats it like a GET.
ok_post_content = "c" * 100
response = requests.post("http://localhost:25000/", ok_post_content)
assert response.status_code == requests.codes.ok
@pytest.mark.execute_serially
@CustomClusterTestSuite.with_args()
def test_webserver_interface(self):
addrs = psutil.net_if_addrs()
print("net_if_addrs returned: %s" % addrs)
ip_matcher = re.compile(r"\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}")
ip_addrs = []
for addr in addrs:
for snic in addrs[addr]:
if ip_matcher.match(snic.address):
ip_addrs.append(snic.address)
# There must be at least one available interface on the machine.
assert len(ip_addrs) > 0, addrs
ports = ["25000", "25010", "25020"]
# With default args, the webserver should be accessible over all interfaces for all
# daemons.
for ip in ip_addrs:
for port in ports:
response = requests.get("http://%s:%s/" % (ip, port))
assert response.status_code == requests.codes.ok, ip
# Pick a random interface and restart with the webserver on that interface.
interface = random.choice(ip_addrs)
self._start_impala_cluster(["--impalad_args=--webserver_interface=%s" % interface])
# Now the webserver should only be accessible over the choosen interface.
for ip in ip_addrs:
try:
response = requests.get("http://%s:25000/" % ip)
assert ip == interface
assert response.status_code == requests.codes.ok, ip
except requests.exceptions.ConnectionError:
assert ip != interface
@pytest.mark.execute_serially
@CustomClusterTestSuite.with_args(
impalad_args="--query_stmt_size=0"
)
def test_query_stmt_without_truncate(self):
"""Check if the full query string is displayed in the query list on the WebUI."""
# The input query is a select + 450 'x ' long.
query_select = "x " * 450
query = 'select "{0}"'.format(query_select)
# In the site there is an extra \ before the " so we need that in the expected
# response too.
expected = 'select \\"{0}\\"'.format(query_select)
self.execute_query(query)
response = requests.get("http://localhost:25000/queries?json")
response_json = response.text
assert expected in response_json, "No matching statement found in the queries site."
assert '"resource_pool": "default-pool"' in response_json
@pytest.mark.execute_serially
@CustomClusterTestSuite.with_args(
impalad_args="--query_stmt_size=10"
)
def test_query_stmt_with_custom_length(self):
"""Check if the partial query with the correct length is displayed in the query list
on the WebUI."""
# The input query is a select + 450 'x ' long.
query = 'select "{0}"'.format("x " * 450)
# Searching for the custom, 10 chars long response. In the site there is an extra \
# before the " so we need that in the expected response too.
expected = 'select \\"x ...'
self.execute_query(query)
response = requests.get("http://localhost:25000/queries?json")
response_json = response.text
assert expected in response_json, "No matching statement found in the queries site."
assert '"resource_pool": "default-pool"' in response_json
@pytest.mark.execute_serially
@CustomClusterTestSuite.with_args(
cluster_size=1,
impalad_args="--query_log_size_in_bytes=" + str(SMALL_QUERY_LOG_SIZE_IN_BYTES)
)
def test_query_log_size_in_bytes(self):
"""Check if query list is limited by query_log_size_in_bytes flag."""
# This simple test query will produce ~8520 bytes of QueryStateRecord.
query = "select version()"
num_queries = 10
for i in range(0, num_queries):
self.execute_query_expect_success(self.client, query)
# Retrieve and verify the total size metrics.
metric_key = "impala-server.query-log-est-total-bytes"
metric_value = self.cluster.impalads[0].service.get_metric_value(metric_key)
assert metric_value > 0
assert metric_value <= SMALL_QUERY_LOG_SIZE_IN_BYTES
# Verify that the query page only contains a subset of the test queries.
queries_response = requests.get("http://localhost:25000/queries?json")
queries_json = json.loads(queries_response.text)
assert len(queries_json["completed_queries"]) > 0
assert len(queries_json["completed_queries"]) < num_queries
# Checks if 'messages' exists/does not exist in 'result_stderr' based on the value of
# 'should_exist'
def _validate_shell_messages(self, result_stderr, messages, should_exist=True):
for msg in messages:
if should_exist:
assert msg in result_stderr, result_stderr
else:
assert msg not in result_stderr, result_stderr
@pytest.mark.execute_serially
@CustomClusterTestSuite.with_args(
impalad_args="--ping_expose_webserver_url=false"
)
def test_webserver_url_not_exposed(self, vector):
if vector.get_value('table_format').file_format != 'text':
pytest.skip('runs only for text table_format')
# If webserver url is not exposed, debug web urls shouldn't be printed out.
shell_messages = ["Query submitted at: ", "(Coordinator: ",
"Query state can be monitored at: "]
query_shell_arg = '--query=select * from functional.alltypes'
# hs2
results = run_impala_shell_cmd(vector, [query_shell_arg])
self._validate_shell_messages(results.stderr, shell_messages, should_exist=False)
# beeswax
results = run_impala_shell_cmd(vector, ['--protocol=beeswax', query_shell_arg])
self._validate_shell_messages(results.stderr, shell_messages, should_exist=False)
# Even though webserver url is not exposed, it is still accessible.
page = requests.get('http://localhost:25000')
assert page.status_code == requests.codes.ok
@pytest.mark.execute_serially
@CustomClusterTestSuite.with_args(
impalad_args="--logtostderr=true --redirect_stdout_stderr=false",
statestored_args="--logtostderr=true --redirect_stdout_stderr=false",
catalogd_args="--logtostderr=true --redirect_stdout_stderr=false"
)
def test_webserver_hide_logs_link(self, vector):
"""Validate that there is no /logs link when we use --logtostderr=true """
ports = ["25000", "25010", "25020"]
for port in ports:
# Get the webui home page as json.
response = requests.get("http://localhost:%s?json" % port)
assert response.status_code == requests.codes.ok
home = json.loads(response.text)
# Get the items in the navbar.
navbar = home["__common__"]['navbar']
found_links = [link_item['link'] for link_item in navbar]
# The links that we expect to see in the navbar.
expected_coordinator_links = [
"/",
"/admission",
"/backends",
"/catalog",
"/hadoop-varz",
"/jmx",
"/log_level",
"/memz",
"/metrics",
"/profile_docs",
"/queries",
"/rpcz",
"/sessions",
"/threadz",
"/varz",
]
expected_statestore_links = [
"/",
"/log_level",
"/memz",
"/metrics",
"/profile_docs",
"/rpcz",
"/subscribers",
"/threadz",
"/topics",
"/varz",
]
expected_catalog_links = [
"/",
"/catalog",
"/events",
"/hadoop-varz",
"/jmx",
"/log_level",
"/memz",
"/metrics",
"/operations",
"/profile_docs",
"/rpcz",
"/threadz",
"/varz",
]
msg = "bad links from webui port %s" % port
if port == "25000":
assert found_links == expected_coordinator_links, msg
elif port == "25010":
assert found_links == expected_statestore_links, msg
elif port == "25020":
assert found_links == expected_catalog_links, msg
@pytest.mark.execute_serially
@CustomClusterTestSuite.with_args(
impalad_args="--disable_content_security_policy_header=true",
statestored_args="--disable_content_security_policy_header=true",
catalogd_args="--disable_content_security_policy_header=true"
)
def test_cdp_header_disabled(self):
"""Test that if servers are started with the flag
--disable_content_security_policy_header=true then the emission of the CDP header is
disabled."""
ports = ["25000", "25010", "25020"] # Respectively the impalad, statestore, catalog.
for port in ports:
response = requests.get("http://localhost:%s" % port)
assert 'Content-Security-Policy' not in response.headers, \
"CSP header present despite being disabled (port %s)" % port
@staticmethod
def _get_inflight_catalog_operations():
response = requests.get("http://localhost:25020/operations?json")
assert response.status_code == requests.codes.ok
operations = json.loads(response.text)
assert "inflight_catalog_operations" in operations
return operations["inflight_catalog_operations"]
@staticmethod
def _get_finished_catalog_operations():
response = requests.get("http://localhost:25020/operations?json")
assert response.status_code == requests.codes.ok
operations = json.loads(response.text)
assert "finished_catalog_operations" in operations
return operations["finished_catalog_operations"]
@CustomClusterTestSuite.with_args(catalogd_args="--catalog_operation_log_size=2")
def test_catalog_operations_limit(self, unique_database):
tbl = unique_database + ".tbl"
self.execute_query("create table {0}_1 (id int)".format(tbl))
self.execute_query("create table {0}_2 (id int)".format(tbl))
self.execute_query("create table {0}_3 (id int)".format(tbl))
self.execute_query("drop table {0}_1".format(tbl))
finished_operations = self._get_finished_catalog_operations()
# Verify only 2 operations are shown
assert len(finished_operations) == 2
op = finished_operations[0]
assert op["status"] == "FINISHED"
assert op["catalog_op_name"] == "DROP_TABLE"
assert op["target_name"] == tbl + "_1"
op = finished_operations[1]
assert op["status"] == "FINISHED"
assert op["catalog_op_name"] == "CREATE_TABLE"
assert op["target_name"] == tbl + "_3"
@CustomClusterTestSuite.with_args(catalogd_args="--catalog_operation_log_size=-1")
def test_catalog_operations_negative_limit(self, unique_database):
# Test negative limit on catalog_operation_log_size. The limit is converted to be
# Integer.MAX_VALUE. Run hundreds of commands and see whether they are all in the
# operation log.
tbl = unique_database + ".tbl"
self.execute_query("create table {0} (id int)".format(tbl))
num = 500
for i in range(num):
self.execute_query("invalidate metadata " + tbl)
finished_operations = self._get_finished_catalog_operations()
# Verify all operations are in the history. There are one DROP_DATABASE, one
# CREATE_DATABASE, one CREATE_TABLE and 'num' INVALIDATEs in the list.
assert len(finished_operations) == 3 + num
for i in range(num):
op = finished_operations[i]
assert op["status"] == "FINISHED"
assert op["catalog_op_name"] == "INVALIDATE_METADATA"
assert op["target_name"] == tbl
op = finished_operations[-3]
assert op["status"] == "FINISHED"
assert op["catalog_op_name"] == "CREATE_TABLE"
assert op["target_name"] == tbl
op = finished_operations[-2]
assert op["status"] == "FINISHED"
assert op["catalog_op_name"] == "CREATE_DATABASE"
assert op["target_name"] == unique_database
op = finished_operations[-1]
assert op["status"] == "FINISHED"
assert op["catalog_op_name"] == "DROP_DATABASE"
assert op["target_name"] == unique_database
@CustomClusterTestSuite.with_args(
impalad_args="--catalog_client_rpc_timeout_ms=100 "
"--catalog_client_rpc_retry_interval_ms=10 "
"--catalog_client_connection_num_retries=2")
def test_catalog_operations_with_rpc_retry(self):
"""Test that catalog RPC retries are all shown in the /operations page"""
# Run a DESCRIBE to ensure the table is loaded. So the first RPC attempt will
# time out in its real work. This triggers a PrioritizeLoad RPC which usually
# finishes in 40ms. So 100ms for catalog RPC timeout is enough.
self.execute_query("describe functional.alltypes")
try:
# This runs around 600ms with the debug action so the catalog RPC will timeout.
self.execute_query("refresh functional.alltypes", {
"debug_action": "catalogd_refresh_hdfs_listing_delay:SLEEP@100"
})
except ImpalaBeeswaxException as e:
assert "RPC recv timed out" in str(e)
# In impalad side, the query fails by the above error. However, in catalogd side,
# the RPCs are still running. Check the in-flight operations.
inflight_operations = self._get_inflight_catalog_operations()
assert len(inflight_operations) == 2
for op in inflight_operations:
assert op["status"] == "STARTED"
assert op["catalog_op_name"] == "REFRESH"
assert op["target_name"] == "functional.alltypes"
assert inflight_operations[0]["query_id"] == inflight_operations[1]["query_id"]
assert inflight_operations[0]["thread_id"] != inflight_operations[1]["thread_id"]
# Wait until the catalog operations finish
while len(self._get_inflight_catalog_operations()) != 0:
time.sleep(1)
# Verify both RPC attempts are shown as finished operations.
finished_operations = self._get_finished_catalog_operations()
assert len(finished_operations) == 2
for op in finished_operations:
assert op["status"] == "FINISHED"
assert op["catalog_op_name"] == "REFRESH"
assert op["target_name"] == "functional.alltypes"
assert finished_operations[0]["query_id"] == finished_operations[1]["query_id"]
assert finished_operations[0]["thread_id"] != finished_operations[1]["thread_id"]
def _verify_topic_size_metrics(self):
# Calculate the total topic metrics from the /topics page
response = requests.get("http://localhost:25010/topics?json")
assert response.status_code == requests.codes.ok
topics_json = json.loads(response.text)
total_key_size = 0
total_value_size = 0
total_topic_size = 0
for topic in topics_json["topics"]:
total_key_size += topic["key_size_bytes"]
total_value_size += topic["value_size_bytes"]
total_topic_size += topic["total_size_bytes"]
# Retrieve and verify the total topic metrics from the /metrics page
response = requests.get("http://localhost:25010/metrics?json")
assert response.status_code == requests.codes.ok
metrics_json = json.loads(response.text)["metric_group"]["metrics"]
for metric in metrics_json:
if metric["name"] == "statestore.total-key-size-bytes":
assert metric["value"] == total_key_size
elif metric["name"] == "statestore.total-value-size-bytes":
assert metric["value"] == total_value_size
elif metric["name"] == "statestore.total-topic-size-bytes":
assert metric["value"] == total_topic_size
@CustomClusterTestSuite.with_args()
def test_transient_topic_size(self):
self._verify_topic_size_metrics()
# Kill an impalad and wait until it's removed
killed_impalad = self.cluster.impalads[2]
killed_impalad.kill()
# Before we kill an impalad, there are DEFAULT_CLUSTER_SIZE + 1 subscribers
# (DEFAULT_CLUSTER_SIZE impalads and 1 for catalogd). After we kill one impalad,
# there should be DEFAULT_CLUSTER_SIZE subscribers.
self.cluster.statestored.service.wait_for_live_subscribers(DEFAULT_CLUSTER_SIZE)
# Verify the topic metrics again
self._verify_topic_size_metrics()