/usr/lib/python3/dist-packages/kafka/cluster.py is in python3-kafka 1.3.2-0ubuntu1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 | from __future__ import absolute_import
import collections
import copy
import logging
import threading
import time
from kafka.vendor import six
from . import errors as Errors
from .future import Future
from .structs import BrokerMetadata, PartitionMetadata, TopicPartition
log = logging.getLogger(__name__)
class ClusterMetadata(object):
DEFAULT_CONFIG = {
'retry_backoff_ms': 100,
'metadata_max_age_ms': 300000,
}
def __init__(self, **configs):
self._brokers = {} # node_id -> BrokerMetadata
self._partitions = {} # topic -> partition -> PartitionMetadata
self._broker_partitions = collections.defaultdict(set) # node_id -> {TopicPartition...}
self._groups = {} # group_name -> node_id
self._last_refresh_ms = 0
self._last_successful_refresh_ms = 0
self._need_update = False
self._future = None
self._listeners = set()
self._lock = threading.Lock()
self.need_all_topic_metadata = False
self.unauthorized_topics = set()
self.internal_topics = set()
self.controller = None
self.config = copy.copy(self.DEFAULT_CONFIG)
for key in self.config:
if key in configs:
self.config[key] = configs[key]
def brokers(self):
"""Get all BrokerMetadata
Returns:
set: {BrokerMetadata, ...}
"""
return set(self._brokers.values())
def broker_metadata(self, broker_id):
"""Get BrokerMetadata
Arguments:
broker_id (int): node_id for a broker to check
Returns:
BrokerMetadata or None if not found
"""
return self._brokers.get(broker_id)
def partitions_for_topic(self, topic):
"""Return set of all partitions for topic (whether available or not)
Arguments:
topic (str): topic to check for partitions
Returns:
set: {partition (int), ...}
"""
if topic not in self._partitions:
return None
return set(self._partitions[topic].keys())
def available_partitions_for_topic(self, topic):
"""Return set of partitions with known leaders
Arguments:
topic (str): topic to check for partitions
Returns:
set: {partition (int), ...}
"""
if topic not in self._partitions:
return None
return set([partition for partition, metadata
in six.iteritems(self._partitions[topic])
if metadata.leader != -1])
def leader_for_partition(self, partition):
"""Return node_id of leader, -1 unavailable, None if unknown."""
if partition.topic not in self._partitions:
return None
elif partition.partition not in self._partitions[partition.topic]:
return None
return self._partitions[partition.topic][partition.partition].leader
def partitions_for_broker(self, broker_id):
"""Return TopicPartitions for which the broker is a leader.
Arguments:
broker_id (int): node id for a broker
Returns:
set: {TopicPartition, ...}
"""
return self._broker_partitions.get(broker_id)
def coordinator_for_group(self, group):
"""Return node_id of group coordinator.
Arguments:
group (str): name of consumer group
Returns:
int: node_id for group coordinator
"""
return self._groups.get(group)
def ttl(self):
"""Milliseconds until metadata should be refreshed"""
now = time.time() * 1000
if self._need_update:
ttl = 0
else:
metadata_age = now - self._last_successful_refresh_ms
ttl = self.config['metadata_max_age_ms'] - metadata_age
retry_age = now - self._last_refresh_ms
next_retry = self.config['retry_backoff_ms'] - retry_age
return max(ttl, next_retry, 0)
def refresh_backoff(self):
"""Return milliseconds to wait before attempting to retry after failure"""
return self.config['retry_backoff_ms']
def request_update(self):
"""Flags metadata for update, return Future()
Actual update must be handled separately. This method will only
change the reported ttl()
Returns:
kafka.future.Future (value will be the cluster object after update)
"""
with self._lock:
self._need_update = True
if not self._future or self._future.is_done:
self._future = Future()
return self._future
def topics(self, exclude_internal_topics=True):
"""Get set of known topics.
Arguments:
exclude_internal_topics (bool): Whether records from internal topics
(such as offsets) should be exposed to the consumer. If set to
True the only way to receive records from an internal topic is
subscribing to it. Default True
Returns:
set: {topic (str), ...}
"""
topics = set(self._partitions.keys())
if exclude_internal_topics:
return topics - self.internal_topics
else:
return topics
def failed_update(self, exception):
"""Update cluster state given a failed MetadataRequest."""
f = None
with self._lock:
if self._future:
f = self._future
self._future = None
if f:
f.failure(exception)
self._last_refresh_ms = time.time() * 1000
def update_metadata(self, metadata):
"""Update cluster state given a MetadataResponse.
Arguments:
metadata (MetadataResponse): broker response to a metadata request
Returns: None
"""
# In the common case where we ask for a single topic and get back an
# error, we should fail the future
if len(metadata.topics) == 1 and metadata.topics[0][0] != 0:
error_code, topic = metadata.topics[0][:2]
error = Errors.for_code(error_code)(topic)
return self.failed_update(error)
if not metadata.brokers:
log.warning("No broker metadata found in MetadataResponse")
for broker in metadata.brokers:
if metadata.API_VERSION == 0:
node_id, host, port = broker
rack = None
else:
node_id, host, port, rack = broker
self._brokers.update({
node_id: BrokerMetadata(node_id, host, port, rack)
})
if metadata.API_VERSION == 0:
self.controller = None
else:
self.controller = self._brokers.get(metadata.controller_id)
_new_partitions = {}
_new_broker_partitions = collections.defaultdict(set)
_new_unauthorized_topics = set()
_new_internal_topics = set()
for topic_data in metadata.topics:
if metadata.API_VERSION == 0:
error_code, topic, partitions = topic_data
is_internal = False
else:
error_code, topic, is_internal, partitions = topic_data
if is_internal:
_new_internal_topics.add(topic)
error_type = Errors.for_code(error_code)
if error_type is Errors.NoError:
_new_partitions[topic] = {}
for p_error, partition, leader, replicas, isr in partitions:
_new_partitions[topic][partition] = PartitionMetadata(
topic=topic, partition=partition, leader=leader,
replicas=replicas, isr=isr, error=p_error)
if leader != -1:
_new_broker_partitions[leader].add(
TopicPartition(topic, partition))
elif error_type is Errors.LeaderNotAvailableError:
log.warning("Topic %s is not available during auto-create"
" initialization", topic)
elif error_type is Errors.UnknownTopicOrPartitionError:
log.error("Topic %s not found in cluster metadata", topic)
elif error_type is Errors.TopicAuthorizationFailedError:
log.error("Topic %s is not authorized for this client", topic)
_new_unauthorized_topics.add(topic)
elif error_type is Errors.InvalidTopicError:
log.error("'%s' is not a valid topic name", topic)
else:
log.error("Error fetching metadata for topic %s: %s",
topic, error_type)
with self._lock:
self._partitions = _new_partitions
self._broker_partitions = _new_broker_partitions
self.unauthorized_topics = _new_unauthorized_topics
self.internal_topics = _new_internal_topics
f = None
if self._future:
f = self._future
self._future = None
self._need_update = False
now = time.time() * 1000
self._last_refresh_ms = now
self._last_successful_refresh_ms = now
if f:
f.success(self)
log.debug("Updated cluster metadata to %s", self)
for listener in self._listeners:
listener(self)
def add_listener(self, listener):
"""Add a callback function to be called on each metadata update"""
self._listeners.add(listener)
def remove_listener(self, listener):
"""Remove a previously added listener callback"""
self._listeners.remove(listener)
def add_group_coordinator(self, group, response):
"""Update with metadata for a group coordinator
Arguments:
group (str): name of group from GroupCoordinatorRequest
response (GroupCoordinatorResponse): broker response
Returns:
bool: True if metadata is updated, False on error
"""
log.debug("Updating coordinator for %s: %s", group, response)
error_type = Errors.for_code(response.error_code)
if error_type is not Errors.NoError:
log.error("GroupCoordinatorResponse error: %s", error_type)
self._groups[group] = -1
return False
node_id = response.coordinator_id
coordinator = BrokerMetadata(
response.coordinator_id,
response.host,
response.port,
None)
# Assume that group coordinators are just brokers
# (this is true now, but could diverge in future)
if node_id not in self._brokers:
self._brokers[node_id] = coordinator
# If this happens, either brokers have moved without
# changing IDs, or our assumption above is wrong
else:
node = self._brokers[node_id]
if coordinator.host != node.host or coordinator.port != node.port:
log.error("GroupCoordinator metadata conflicts with existing"
" broker metadata. Coordinator: %s, Broker: %s",
coordinator, node)
self._groups[group] = node_id
return False
log.info("Group coordinator for %s is %s", group, coordinator)
self._groups[group] = node_id
return True
def with_partitions(self, partitions_to_add):
"""Returns a copy of cluster metadata with partitions added"""
new_metadata = ClusterMetadata(**self.config)
new_metadata._brokers = copy.deepcopy(self._brokers)
new_metadata._partitions = copy.deepcopy(self._partitions)
new_metadata._broker_partitions = copy.deepcopy(self._broker_partitions)
new_metadata._groups = copy.deepcopy(self._groups)
new_metadata.internal_topics = copy.deepcopy(self.internal_topics)
new_metadata.unauthorized_topics = copy.deepcopy(self.unauthorized_topics)
for partition in partitions_to_add:
new_metadata._partitions[partition.topic][partition.partition] = partition
if partition.leader is not None and partition.leader != -1:
new_metadata._broker_partitions[partition.leader].add(
TopicPartition(partition.topic, partition.partition))
return new_metadata
def __str__(self):
return 'ClusterMetadata(brokers: %d, topics: %d, groups: %d)' % \
(len(self._brokers), len(self._partitions), len(self._groups))
|