# 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.
import json
import os
from ducktape.services.background_thread import BackgroundThreadService
from kafkatest.directory_layout.kafka_path import KafkaPathResolverMixin
from kafkatest.services.kafka import TopicPartition, consumer_group
from kafkatest.services.kafka.util import get_log4j_config_param, get_log4j_config_for_tools
from kafkatest.services.verifiable_client import VerifiableClientMixin
from kafkatest.version import get_version, DEV_BRANCH, V_2_3_0, V_2_3_1, V_3_7_0, V_4_0_0
class ConsumerState:
Started = 1
Dead = 2
Rebalancing = 3
Joined = 4
def _create_partition_from_dict(d):
topic = d["topic"]
partition = d["partition"]
return TopicPartition(topic, partition)
class ConsumerEventHandler(object):
def __init__(self, node, verify_offsets, idx, state=ConsumerState.Dead,
revoked_count=0, assigned_count=0, assignment=None,
position=None, committed=None, total_consumed=0,
shutdown_complete=False):
self.node = node
self.verify_offsets = verify_offsets
self.idx = idx
self.state = state
self.revoked_count = revoked_count
self.assigned_count = assigned_count
self.assignment = assignment if assignment is not None else []
self.position = position if position is not None else {}
self.committed = committed if committed is not None else {}
self.total_consumed = total_consumed
self.shutdown_complete = shutdown_complete
def handle_shutdown_complete(self, node=None, logger=None):
self.state = ConsumerState.Dead
self.assignment = []
self.position = {}
self.shutdown_complete = True
if node is not None and logger is not None:
logger.debug("Shut down %s" % node.account.hostname)
def handle_startup_complete(self, node, logger):
self.state = ConsumerState.Started
logger.debug("Started %s" % node.account.hostname)
def handle_offsets_committed(self, event, node, logger):
if event["success"]:
for offset_commit in event["offsets"]:
if offset_commit.get("error", "") != "":
logger.debug("%s: Offset commit failed for: %s" % (str(node.account), offset_commit))
continue
tp = _create_partition_from_dict(offset_commit)
offset = offset_commit["offset"]
assert tp in self.assignment, \
"Committed offsets for partition %s not assigned (current assignment: %s)" % \
(str(tp), str(self.assignment))
assert tp in self.position, "No previous position for %s: %s" % (str(tp), event)
assert self.position[tp] >= offset, \
"The committed offset %d was greater than the current position %d for partition %s" % \
(offset, self.position[tp], str(tp))
self.committed[tp] = offset
logger.debug("Offsets committed for %s" % node.account.hostname)
def handle_records_consumed(self, event, logger):
assert self.state == ConsumerState.Joined, \
"Consumed records should only be received when joined (current state: %s)" % str(self.state)
for record_batch in event["partitions"]:
tp = _create_partition_from_dict(record_batch)
min_offset = record_batch["minOffset"]
max_offset = record_batch["maxOffset"]
assert tp in self.assignment, \
"Consumed records for partition %s which is not assigned (current assignment: %s)" % \
(str(tp), str(self.assignment))
if tp not in self.position or self.position[tp] == min_offset:
self.position[tp] = max_offset + 1
else:
msg = "Consumed from an unexpected offset (%d, %d) for partition %s" % \
(self.position.get(tp), min_offset, str(tp))
if self.verify_offsets:
raise AssertionError(msg)
else:
if tp in self.position:
self.position[tp] = max_offset + 1
logger.warn(msg)
self.total_consumed += event["count"]
def handle_partitions_revoked(self, event, node, logger):
self.revoked_count += 1
self.state = ConsumerState.Rebalancing
self.position = {}
logger.debug("All partitions revoked from %s" % node.account.hostname)
def handle_partitions_assigned(self, event, node, logger):
self.assigned_count += 1
self.state = ConsumerState.Joined
assignment = []
for topic_partition in event["partitions"]:
tp = _create_partition_from_dict(topic_partition)
assignment.append(tp)
logger.debug("Partitions %s assigned to %s" % (assignment, node.account.hostname))
self.assignment = assignment
def handle_kill_process(self, clean_shutdown):
# if the shutdown was clean, then we expect the explicit
# shutdown event from the consumer
if not clean_shutdown:
self.handle_shutdown_complete()
def current_assignment(self):
return list(self.assignment)
def current_position(self, tp):
if tp in self.position:
return self.position[tp]
else:
return None
def last_commit(self, tp):
if tp in self.committed:
return self.committed[tp]
else:
return None
# This needs to be used for cooperative protocol.
class IncrementalAssignmentConsumerEventHandler(ConsumerEventHandler):
def __init__(self, node, verify_offsets, idx, **kwargs):
super().__init__(node, verify_offsets, idx, **kwargs)
def handle_partitions_revoked(self, event, node, logger):
self.revoked_count += 1
self.state = ConsumerState.Rebalancing
self.position = {}
revoked = []
for topic_partition in event["partitions"]:
tp = _create_partition_from_dict(topic_partition)
assert tp in self.assignment, \
"Topic partition %s cannot be revoked from %s as it was not previously assigned to that consumer" % \
(tp, node.account.hostname)
self.assignment.remove(tp)
revoked.append(tp)
logger.debug("Partitions %s revoked from %s" % (revoked, node.account.hostname))
def handle_partitions_assigned(self, event, node, logger):
self.assigned_count += 1
self.state = ConsumerState.Joined
assignment = []
for topic_partition in event["partitions"]:
tp = _create_partition_from_dict(topic_partition)
assignment.append(tp)
logger.debug("Partitions %s assigned to %s" % (assignment, node.account.hostname))
self.assignment.extend(assignment)
# This needs to be used for consumer protocol.
class ConsumerProtocolConsumerEventHandler(IncrementalAssignmentConsumerEventHandler):
def __init__(self, node, verify_offsets, idx, **kwargs):
super().__init__(node, verify_offsets, idx, **kwargs)
def handle_partitions_revoked(self, event, node, logger):
# The handler state is not transitioned to Rebalancing as the records can only be
# consumed in Joined state (see ConsumerEventHandler.handle_records_consumed).
# The consumer with consumer protocol should still be able to consume messages during rebalance.
self.revoked_count += 1
self.position = {}
revoked = []
for topic_partition in event["partitions"]:
tp = _create_partition_from_dict(topic_partition)
# tp existing in self.assignment is not guaranteed in the new consumer
# if it shuts down when revoking partitions for reconciliation.
if tp in self.assignment:
self.assignment.remove(tp)
revoked.append(tp)
logger.debug("Partitions %s revoked from %s" % (revoked, node.account.hostname))
class VerifiableConsumer(KafkaPathResolverMixin, VerifiableClientMixin, BackgroundThreadService):
"""This service wraps org.apache.kafka.tools.VerifiableConsumer for use in
system testing.
NOTE: this class should be treated as a PUBLIC API. Downstream users use
this service both directly and through class extension, so care must be
taken to ensure compatibility.
"""
PERSISTENT_ROOT = "/mnt/verifiable_consumer"
STDOUT_CAPTURE = os.path.join(PERSISTENT_ROOT, "verifiable_consumer.stdout")
STDERR_CAPTURE = os.path.join(PERSISTENT_ROOT, "verifiable_consumer.stderr")
LOG_DIR = os.path.join(PERSISTENT_ROOT, "logs")
LOG_FILE = os.path.join(LOG_DIR, "verifiable_consumer.log")
CONFIG_FILE = os.path.join(PERSISTENT_ROOT, "verifiable_consumer.properties")
logs = {
"verifiable_consumer_stdout": {
"path": STDOUT_CAPTURE,
"collect_default": False},
"verifiable_consumer_stderr": {
"path": STDERR_CAPTURE,
"collect_default": False},
"verifiable_consumer_log": {
"path": LOG_FILE,
"collect_default": True}
}
def __init__(self, context, num_nodes, kafka, topic, group_id,
static_membership=False, max_messages=-1, session_timeout_sec=0, enable_autocommit=False,
assignment_strategy=None, group_protocol=None, group_remote_assignor=None,
version=DEV_BRANCH, stop_timeout_sec=30, log_level="INFO", jaas_override_variables=None,
on_record_consumed=None, reset_policy="earliest", verify_offsets=True, prop_file=""):
"""
:param jaas_override_variables: A dict of variables to be used in the jaas.conf template file
"""
super(VerifiableConsumer, self).__init__(context, num_nodes)
self.log_level = log_level
self.kafka = kafka
self.topic = topic
self.group_protocol = group_protocol
self.group_remote_assignor = group_remote_assignor
self.group_id = group_id
self.reset_policy = reset_policy
self.static_membership = static_membership
self.max_messages = max_messages
self.session_timeout_sec = session_timeout_sec
self.enable_autocommit = enable_autocommit
self.assignment_strategy = assignment_strategy
self.prop_file = prop_file
self.stop_timeout_sec = stop_timeout_sec
self.on_record_consumed = on_record_consumed
self.verify_offsets = verify_offsets
self.event_handlers = {}
self.global_position = {}
self.global_committed = {}
self.jaas_override_variables = jaas_override_variables or {}
for node in self.nodes:
node.version = version
def java_class_name(self):
return "VerifiableConsumer"
def create_event_handler(self, idx, node):
def create_handler_helper(handler_class, node, idx, existing_handler=None):
if existing_handler is not None:
return handler_class(node, self.verify_offsets, idx,
state=existing_handler.state,
revoked_count=existing_handler.revoked_count,
assigned_count=existing_handler.assigned_count,
assignment=existing_handler.assignment,
position=existing_handler.position,
committed=existing_handler.committed,
total_consumed=existing_handler.total_consumed,
shutdown_complete=existing_handler.shutdown_complete)
else:
return handler_class(node, self.verify_offsets, idx)
existing_handler = self.event_handlers[node] if node in self.event_handlers else None
if self.is_consumer_group_protocol_enabled():
return create_handler_helper(ConsumerProtocolConsumerEventHandler, node, idx, existing_handler)
elif self.is_eager():
return create_handler_helper(ConsumerEventHandler, node, idx, existing_handler)
else:
return create_handler_helper(IncrementalAssignmentConsumerEventHandler, node, idx, existing_handler)
def _worker(self, idx, node):
with self.lock:
self.event_handlers[node] = self.create_event_handler(idx, node)
handler = self.event_handlers[node]
handler.shutdown_complete = False
node.account.ssh("mkdir -p %s" % VerifiableConsumer.PERSISTENT_ROOT, allow_fail=False)
# Create and upload log properties
log_config = self.render(get_log4j_config_for_tools(node), log_file=VerifiableConsumer.LOG_FILE)
node.account.create_file(get_log4j_config_for_tools(node), log_config)
# Create and upload config file
self.security_config = self.kafka.security_config.client_config(self.prop_file, node,
self.jaas_override_variables)
self.security_config.setup_node(node)
self.prop_file += str(self.security_config)
self.logger.info("verifiable_consumer.properties:")
self.logger.info(self.prop_file)
node.account.create_file(VerifiableConsumer.CONFIG_FILE, self.prop_file)
self.security_config.setup_node(node)
# apply group.instance.id to the node for static membership validation
node.group_instance_id = None
if self.static_membership:
assert node.version >= V_2_3_0, \
"Version %s does not support static membership (must be 2.3 or higher)" % str(node.version)
node.group_instance_id = self.group_id + "-instance-" + str(idx)
cmd = self.start_cmd(node)
self.logger.debug("VerifiableConsumer %d command: %s" % (idx, cmd))
for line in node.account.ssh_capture(cmd):
event = self.try_parse_json(node, line.strip())
if event is not None:
with self.lock:
name = event["name"]
if name == "shutdown_complete":
handler.handle_shutdown_complete(node, self.logger)
elif name == "startup_complete":
handler.handle_startup_complete(node, self.logger)
elif name == "offsets_committed":
handler.handle_offsets_committed(event, node, self.logger)
self._update_global_committed(event)
elif name == "records_consumed":
handler.handle_records_consumed(event, self.logger)
self._update_global_position(event, node)
elif name == "record_data" and self.on_record_consumed:
self.on_record_consumed(event, node)
elif name == "partitions_revoked":
handler.handle_partitions_revoked(event, node, self.logger)
elif name == "partitions_assigned":
handler.handle_partitions_assigned(event, node, self.logger)
else:
self.logger.debug("%s: ignoring unknown event: %s" % (str(node.account), event))
def is_eager(self):
return self.group_protocol == consumer_group.classic_group_protocol and self.assignment_strategy != "org.apache.kafka.clients.consumer.CooperativeStickyAssignor"
def _update_global_position(self, consumed_event, node):
for consumed_partition in consumed_event["partitions"]:
tp = _create_partition_from_dict(consumed_partition)
if tp in self.global_committed:
# verify that the position never gets behind the current commit.
if self.global_committed[tp] > consumed_partition["minOffset"]:
msg = "Consumed position %d is behind the current committed offset %d for partition %s" % \
(consumed_partition["minOffset"], self.global_committed[tp], str(tp))
if self.verify_offsets:
raise AssertionError(msg)
else:
self.logger.warn(msg)
# the consumer cannot generally guarantee that the position increases monotonically
# without gaps in the face of hard failures, so we only log a warning when this happens
if tp in self.global_position and self.global_position[tp] != consumed_partition["minOffset"]:
self.logger.warn("%s: Expected next consumed offset of %d for partition %s, but instead saw %d" %
(str(node.account), self.global_position[tp], str(tp), consumed_partition["minOffset"]))
self.global_position[tp] = consumed_partition["maxOffset"] + 1
def _update_global_committed(self, commit_event):
if commit_event["success"]:
for offset_commit in commit_event["offsets"]:
tp = _create_partition_from_dict(offset_commit)
offset = offset_commit["offset"]
assert self.global_position[tp] >= offset, \
"Committed offset %d for partition %s is ahead of the current position %d" % \
(offset, str(tp), self.global_position[tp])
self.global_committed[tp] = offset
def start_cmd(self, node):
cmd = ""
cmd += "export LOG_DIR=%s;" % VerifiableConsumer.LOG_DIR
cmd += " export KAFKA_OPTS=%s;" % self.security_config.kafka_opts
cmd += " export KAFKA_LOG4J_OPTS=\"%s%s\"; " % (get_log4j_config_param(node), get_log4j_config_for_tools(node))
cmd += self.impl.exec_cmd(node)
if self.on_record_consumed:
cmd += " --verbose"
if node.group_instance_id:
cmd += " --group-instance-id %s" % node.group_instance_id
elif node.version == V_2_3_0 or node.version == V_2_3_1:
# In 2.3, --group-instance-id was required, but would be left empty
# if `None` is passed as the argument value
cmd += " --group-instance-id None"
# 3.7.0 includes support for KIP-848 which introduced a new implementation of the consumer group protocol.
# The two implementations use slightly different configuration, hence these arguments are conditional.
#
# See the Java class/method VerifiableConsumer.createFromArgs() for how the command line arguments are
# parsed and used as configuration in the runner.
if node.version >= V_3_7_0 and self.is_consumer_group_protocol_enabled():
cmd += " --group-protocol %s" % self.group_protocol
if self.group_remote_assignor:
cmd += " --group-remote-assignor %s" % self.group_remote_assignor
else:
# Either we're an older consumer version or we're using the old consumer group protocol.
if self.assignment_strategy:
cmd += " --assignment-strategy %s" % self.assignment_strategy
if self.enable_autocommit:
cmd += " --enable-autocommit "
if node.version < V_4_0_0:
cmd += " --broker-list %s" % self.kafka.bootstrap_servers(self.security_config.security_protocol)
else:
cmd += " --bootstrap-server %s" % self.kafka.bootstrap_servers(self.security_config.security_protocol)
cmd += " --reset-policy %s --group-id %s --topic %s" % \
(self.reset_policy, self.group_id, self.topic)
if self.session_timeout_sec > 0:
cmd += " --session-timeout %s" % self.session_timeout_sec*1000
if self.max_messages > 0:
cmd += " --max-messages %s" % str(self.max_messages)
version = get_version(node)
if version.supports_command_config():
cmd += " --command-config %s" % VerifiableConsumer.CONFIG_FILE
else:
cmd += " --consumer.config %s" % VerifiableConsumer.CONFIG_FILE
cmd += " 2>> %s | tee -a %s &" % (VerifiableConsumer.STDOUT_CAPTURE, VerifiableConsumer.STDOUT_CAPTURE)
return cmd
def pids(self, node):
return self.impl.pids(node)
def try_parse_json(self, node, string):
"""Try to parse a string as json. Return None if not parseable."""
try:
return json.loads(string)
except ValueError:
self.logger.debug("%s: Could not parse as json: %s" % (str(node.account), str(string)))
return None
def stop_all(self):
for node in self.nodes:
self.stop_node(node)
def kill_node(self, node, clean_shutdown=True, allow_fail=False):
sig = self.impl.kill_signal(clean_shutdown)
for pid in self.pids(node):
node.account.signal(pid, sig, allow_fail)
with self.lock:
self.event_handlers[node].handle_kill_process(clean_shutdown)
def stop_node(self, node, clean_shutdown=True):
self.kill_node(node, clean_shutdown=clean_shutdown)
stopped = self.wait_node(node, timeout_sec=self.stop_timeout_sec)
assert stopped, "Node %s: did not stop within the specified timeout of %s seconds" % \
(str(node.account), str(self.stop_timeout_sec))
def clean_node(self, node):
self.kill_node(node, clean_shutdown=False)
node.account.ssh("rm -rf " + self.PERSISTENT_ROOT, allow_fail=False)
self.security_config.clean_node(node)
def current_assignment(self):
with self.lock:
return { handler.node: handler.current_assignment() for handler in self.event_handlers.values() }
def current_position(self, tp):
with self.lock:
if tp in self.global_position:
return self.global_position[tp]
else:
return None
def owner(self, tp):
with self.lock:
for handler in self.event_handlers.values():
if tp in handler.current_assignment():
return handler.node
return None
def last_commit(self, tp):
with self.lock:
if tp in self.global_committed:
return self.global_committed[tp]
else:
return None
def total_consumed(self):
with self.lock:
return sum(handler.total_consumed for handler in self.event_handlers.values())
def num_rebalances(self):
with self.lock:
return max(handler.assigned_count for handler in self.event_handlers.values())
def num_revokes_for_alive(self, keep_alive=1):
with self.lock:
return max(handler.revoked_count for handler in self.event_handlers.values()
if handler.idx <= keep_alive)
def joined_nodes(self):
with self.lock:
return [handler.node for handler in self.event_handlers.values()
if handler.state == ConsumerState.Joined]
def rebalancing_nodes(self):
with self.lock:
return [handler.node for handler in self.event_handlers.values()
if handler.state == ConsumerState.Rebalancing]
def dead_nodes(self):
with self.lock:
return [handler.node for handler in self.event_handlers.values()
if handler.state == ConsumerState.Dead]
def alive_nodes(self):
with self.lock:
return [handler.node for handler in self.event_handlers.values()
if handler.state != ConsumerState.Dead]
def shutdown_complete_nodes(self):
with self.lock:
return [handler.node for handler in self.event_handlers.values()
if handler.shutdown_complete]
def is_consumer_group_protocol_enabled(self):
return self.group_protocol and self.group_protocol.lower() == consumer_group.consumer_group_protocol