ruby-kafka 0.5.0 → 0.5.1.beta1
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.circleci/config.yml +33 -0
- data/CHANGELOG.md +8 -0
- data/README.md +29 -7
- data/docker-compose.yml +39 -0
- data/lib/kafka.rb +14 -0
- data/lib/kafka/broker.rb +48 -19
- data/lib/kafka/broker_pool.rb +3 -1
- data/lib/kafka/client.rb +65 -7
- data/lib/kafka/cluster.rb +68 -4
- data/lib/kafka/connection.rb +8 -18
- data/lib/kafka/connection_builder.rb +2 -1
- data/lib/kafka/consumer.rb +16 -4
- data/lib/kafka/fetch_operation.rb +3 -1
- data/lib/kafka/protocol.rb +8 -0
- data/lib/kafka/protocol/api_versions_request.rb +19 -0
- data/lib/kafka/protocol/api_versions_response.rb +47 -0
- data/lib/kafka/protocol/create_topics_request.rb +40 -0
- data/lib/kafka/protocol/create_topics_response.rb +24 -0
- data/lib/kafka/protocol/decoder.rb +9 -0
- data/lib/kafka/protocol/encoder.rb +8 -0
- data/lib/kafka/protocol/fetch_request.rb +4 -2
- data/lib/kafka/protocol/message.rb +13 -1
- data/lib/kafka/protocol/message_set.rb +7 -1
- data/lib/kafka/protocol/metadata_response.rb +10 -2
- data/lib/kafka/protocol/sasl_handshake_request.rb +1 -1
- data/lib/kafka/protocol/topic_metadata_request.rb +4 -0
- data/lib/kafka/round_robin_assignment_strategy.rb +5 -1
- data/lib/kafka/sasl/gssapi.rb +74 -0
- data/lib/kafka/sasl/plain.rb +37 -0
- data/lib/kafka/sasl/scram.rb +175 -0
- data/lib/kafka/sasl_authenticator.rb +31 -39
- data/lib/kafka/version.rb +1 -1
- metadata +14 -8
- data/circle.yml +0 -23
- data/lib/kafka/sasl_gssapi_authenticator.rb +0 -77
- data/lib/kafka/sasl_plain_authenticator.rb +0 -37
data/lib/kafka/cluster.rb
CHANGED
@@ -1,5 +1,5 @@
|
|
1
|
-
require "set"
|
2
1
|
require "kafka/broker_pool"
|
2
|
+
require "set"
|
3
3
|
|
4
4
|
module Kafka
|
5
5
|
|
@@ -49,6 +49,21 @@ module Kafka
|
|
49
49
|
end
|
50
50
|
end
|
51
51
|
|
52
|
+
def api_info(api_key)
|
53
|
+
apis.find {|api| api.api_key == api_key }
|
54
|
+
end
|
55
|
+
|
56
|
+
def apis
|
57
|
+
@apis ||=
|
58
|
+
begin
|
59
|
+
response = random_broker.api_versions
|
60
|
+
|
61
|
+
Protocol.handle_error(response.error_code)
|
62
|
+
|
63
|
+
response.apis
|
64
|
+
end
|
65
|
+
end
|
66
|
+
|
52
67
|
# Clears the list of target topics.
|
53
68
|
#
|
54
69
|
# @see #add_target_topics
|
@@ -130,6 +145,41 @@ module Kafka
|
|
130
145
|
raise
|
131
146
|
end
|
132
147
|
|
148
|
+
def create_topic(name, num_partitions: 1, replication_factor: 1, timeout: 30)
|
149
|
+
options = {
|
150
|
+
topics: {
|
151
|
+
name => {
|
152
|
+
num_partitions: num_partitions,
|
153
|
+
replication_factor: replication_factor,
|
154
|
+
}
|
155
|
+
},
|
156
|
+
timeout: timeout,
|
157
|
+
}
|
158
|
+
|
159
|
+
broker = controller_broker
|
160
|
+
|
161
|
+
@logger.info "Creating topic `#{name}` using controller broker #{broker}"
|
162
|
+
|
163
|
+
response = broker.create_topics(**options)
|
164
|
+
|
165
|
+
response.errors.each do |topic, error_code|
|
166
|
+
Protocol.handle_error(error_code)
|
167
|
+
end
|
168
|
+
|
169
|
+
begin
|
170
|
+
partitions_for(name).each do |info|
|
171
|
+
Protocol.handle_error(info.partition_error_code)
|
172
|
+
end
|
173
|
+
rescue Kafka::LeaderNotAvailable
|
174
|
+
@logger.warn "Leader not yet available for `#{name}`, waiting 1s..."
|
175
|
+
sleep 1
|
176
|
+
|
177
|
+
retry
|
178
|
+
end
|
179
|
+
|
180
|
+
@logger.info "Topic `#{name}` was created"
|
181
|
+
end
|
182
|
+
|
133
183
|
def resolve_offsets(topic, partitions, offset)
|
134
184
|
add_target_topics([topic])
|
135
185
|
refresh_metadata_if_necessary!
|
@@ -178,6 +228,7 @@ module Kafka
|
|
178
228
|
end
|
179
229
|
|
180
230
|
def topics
|
231
|
+
refresh_metadata_if_necessary!
|
181
232
|
cluster_info.topics.map(&:topic_name)
|
182
233
|
end
|
183
234
|
|
@@ -213,11 +264,15 @@ module Kafka
|
|
213
264
|
broker = @broker_pool.connect(node.hostname, node.port)
|
214
265
|
cluster_info = broker.fetch_metadata(topics: @target_topics)
|
215
266
|
|
216
|
-
|
267
|
+
if cluster_info.brokers.empty?
|
268
|
+
@logger.error "No brokers in cluster"
|
269
|
+
else
|
270
|
+
@logger.info "Discovered cluster metadata; nodes: #{cluster_info.brokers.join(', ')}"
|
217
271
|
|
218
|
-
|
272
|
+
@stale = false
|
219
273
|
|
220
|
-
|
274
|
+
return cluster_info
|
275
|
+
end
|
221
276
|
rescue Error => e
|
222
277
|
@logger.error "Failed to fetch metadata from #{node}: #{e}"
|
223
278
|
errors << [node, e]
|
@@ -231,10 +286,19 @@ module Kafka
|
|
231
286
|
raise ConnectionError, "Could not connect to any of the seed brokers:\n#{error_description}"
|
232
287
|
end
|
233
288
|
|
289
|
+
def random_broker
|
290
|
+
node_id = cluster_info.brokers.sample.node_id
|
291
|
+
connect_to_broker(node_id)
|
292
|
+
end
|
293
|
+
|
234
294
|
def connect_to_broker(broker_id)
|
235
295
|
info = cluster_info.find_broker(broker_id)
|
236
296
|
|
237
297
|
@broker_pool.connect(info.host, info.port, node_id: info.node_id)
|
238
298
|
end
|
299
|
+
|
300
|
+
def controller_broker
|
301
|
+
connect_to_broker(cluster_info.controller_id)
|
302
|
+
end
|
239
303
|
end
|
240
304
|
end
|
data/lib/kafka/connection.rb
CHANGED
@@ -48,7 +48,7 @@ module Kafka
|
|
48
48
|
# broker. Default is 10 seconds.
|
49
49
|
#
|
50
50
|
# @return [Connection] a new connection.
|
51
|
-
def initialize(host:, port:, client_id:, logger:, instrumenter:,
|
51
|
+
def initialize(host:, port:, client_id:, logger:, instrumenter:, connect_timeout: nil, socket_timeout: nil, ssl_context: nil)
|
52
52
|
@host, @port, @client_id = host, port, client_id
|
53
53
|
@logger = logger
|
54
54
|
@instrumenter = instrumenter
|
@@ -56,11 +56,6 @@ module Kafka
|
|
56
56
|
@connect_timeout = connect_timeout || CONNECT_TIMEOUT
|
57
57
|
@socket_timeout = socket_timeout || SOCKET_TIMEOUT
|
58
58
|
@ssl_context = ssl_context
|
59
|
-
@sasl_authenticator = sasl_authenticator
|
60
|
-
end
|
61
|
-
|
62
|
-
def address_match?(host, port)
|
63
|
-
@host == host && @port == port
|
64
59
|
end
|
65
60
|
|
66
61
|
def to_s
|
@@ -75,8 +70,6 @@ module Kafka
|
|
75
70
|
@logger.debug "Closing socket to #{to_s}"
|
76
71
|
|
77
72
|
@socket.close if @socket
|
78
|
-
|
79
|
-
@socket = nil
|
80
73
|
end
|
81
74
|
|
82
75
|
# Sends a request over the connection.
|
@@ -86,20 +79,25 @@ module Kafka
|
|
86
79
|
#
|
87
80
|
# @return [Object] the response.
|
88
81
|
def send_request(request)
|
82
|
+
api_name = Protocol.api_name(request.api_key)
|
83
|
+
|
89
84
|
# Default notification payload.
|
90
85
|
notification = {
|
91
86
|
broker_host: @host,
|
92
|
-
api:
|
87
|
+
api: api_name,
|
93
88
|
request_size: 0,
|
94
89
|
response_size: 0,
|
95
90
|
}
|
96
91
|
|
97
92
|
@instrumenter.instrument("request.connection", notification) do
|
98
93
|
open unless open?
|
99
|
-
|
94
|
+
|
95
|
+
raise IdleConnection if idle?
|
100
96
|
|
101
97
|
@correlation_id += 1
|
102
98
|
|
99
|
+
@logger.debug "Sending #{api_name} API request #{@correlation_id} to #{to_s}"
|
100
|
+
|
103
101
|
write_request(request, notification)
|
104
102
|
|
105
103
|
response_class = request.response_class
|
@@ -133,7 +131,6 @@ module Kafka
|
|
133
131
|
@correlation_id = 0
|
134
132
|
|
135
133
|
@last_request = nil
|
136
|
-
@sasl_authenticator.authenticate!(self)
|
137
134
|
rescue Errno::ETIMEDOUT => e
|
138
135
|
@logger.error "Timed out while trying to connect to #{self}: #{e}"
|
139
136
|
raise ConnectionError, e
|
@@ -142,11 +139,6 @@ module Kafka
|
|
142
139
|
raise ConnectionError, e
|
143
140
|
end
|
144
141
|
|
145
|
-
def reopen
|
146
|
-
close
|
147
|
-
open
|
148
|
-
end
|
149
|
-
|
150
142
|
def idle?
|
151
143
|
@last_request && @last_request < Time.now - IDLE_TIMEOUT
|
152
144
|
end
|
@@ -157,8 +149,6 @@ module Kafka
|
|
157
149
|
#
|
158
150
|
# @return [nil]
|
159
151
|
def write_request(request, notification)
|
160
|
-
@logger.debug "Sending request #{@correlation_id} to #{to_s}"
|
161
|
-
|
162
152
|
message = Kafka::Protocol::RequestMessage.new(
|
163
153
|
api_key: request.api_key,
|
164
154
|
api_version: request.respond_to?(:api_version) ? request.api_version : 0,
|
data/lib/kafka/consumer.rb
CHANGED
@@ -179,6 +179,8 @@ module Kafka
|
|
179
179
|
# @param min_bytes [Integer] the minimum number of bytes to read before
|
180
180
|
# returning messages from each broker; if `max_wait_time` is reached, this
|
181
181
|
# is ignored.
|
182
|
+
# @param max_bytes [Integer] the maximum number of bytes to read before
|
183
|
+
# returning messages from each broker.
|
182
184
|
# @param max_wait_time [Integer, Float] the maximum duration of time to wait before
|
183
185
|
# returning messages from each broker, in seconds.
|
184
186
|
# @param automatically_mark_as_processed [Boolean] whether to automatically
|
@@ -190,10 +192,11 @@ module Kafka
|
|
190
192
|
# The original exception will be returned by calling `#cause` on the
|
191
193
|
# {Kafka::ProcessingError} instance.
|
192
194
|
# @return [nil]
|
193
|
-
def each_message(min_bytes: 1, max_wait_time: 1, automatically_mark_as_processed: true)
|
195
|
+
def each_message(min_bytes: 1, max_bytes: 10485760, max_wait_time: 1, automatically_mark_as_processed: true)
|
194
196
|
consumer_loop do
|
195
197
|
batches = fetch_batches(
|
196
198
|
min_bytes: min_bytes,
|
199
|
+
max_bytes: max_bytes,
|
197
200
|
max_wait_time: max_wait_time,
|
198
201
|
automatically_mark_as_processed: automatically_mark_as_processed
|
199
202
|
)
|
@@ -253,6 +256,8 @@ module Kafka
|
|
253
256
|
# @param min_bytes [Integer] the minimum number of bytes to read before
|
254
257
|
# returning messages from each broker; if `max_wait_time` is reached, this
|
255
258
|
# is ignored.
|
259
|
+
# @param max_bytes [Integer] the maximum number of bytes to read before
|
260
|
+
# returning messages from each broker.
|
256
261
|
# @param max_wait_time [Integer, Float] the maximum duration of time to wait before
|
257
262
|
# returning messages from each broker, in seconds.
|
258
263
|
# @param automatically_mark_as_processed [Boolean] whether to automatically
|
@@ -261,10 +266,11 @@ module Kafka
|
|
261
266
|
# messages can be committed to Kafka.
|
262
267
|
# @yieldparam batch [Kafka::FetchedBatch] a message batch fetched from Kafka.
|
263
268
|
# @return [nil]
|
264
|
-
def each_batch(min_bytes: 1, max_wait_time: 1, automatically_mark_as_processed: true)
|
269
|
+
def each_batch(min_bytes: 1, max_bytes: 10485760, max_wait_time: 1, automatically_mark_as_processed: true)
|
265
270
|
consumer_loop do
|
266
271
|
batches = fetch_batches(
|
267
272
|
min_bytes: min_bytes,
|
273
|
+
max_bytes: max_bytes,
|
268
274
|
max_wait_time: max_wait_time,
|
269
275
|
automatically_mark_as_processed: automatically_mark_as_processed
|
270
276
|
)
|
@@ -349,6 +355,9 @@ module Kafka
|
|
349
355
|
yield
|
350
356
|
rescue HeartbeatError, OffsetCommitError
|
351
357
|
join_group
|
358
|
+
rescue RebalanceInProgress
|
359
|
+
@logger.warn "Group rebalance in progress, re-joining..."
|
360
|
+
join_group
|
352
361
|
rescue FetchError, NotLeaderForPartition, UnknownTopicOrPartition
|
353
362
|
@cluster.mark_as_stale!
|
354
363
|
rescue LeaderNotAvailable => e
|
@@ -370,7 +379,7 @@ module Kafka
|
|
370
379
|
|
371
380
|
def make_final_offsets_commit!(attempts = 3)
|
372
381
|
@offset_manager.commit_offsets
|
373
|
-
rescue ConnectionError, EOFError
|
382
|
+
rescue ConnectionError, OffsetCommitError, EOFError
|
374
383
|
# It's important to make sure final offsets commit is done
|
375
384
|
# As otherwise messages that have been processed after last auto-commit
|
376
385
|
# will be processed again and that may be huge amount of messages
|
@@ -379,6 +388,8 @@ module Kafka
|
|
379
388
|
@logger.error "Retrying to make final offsets commit (#{attempts} attempts left)"
|
380
389
|
sleep(0.1)
|
381
390
|
make_final_offsets_commit!(attempts - 1)
|
391
|
+
rescue Kafka::Error => e
|
392
|
+
@logger.error "Encountered error while shutting down; #{e.class}: #{e.message}"
|
382
393
|
end
|
383
394
|
|
384
395
|
def join_group
|
@@ -400,7 +411,7 @@ module Kafka
|
|
400
411
|
end
|
401
412
|
end
|
402
413
|
|
403
|
-
def fetch_batches(min_bytes:, max_wait_time:, automatically_mark_as_processed:)
|
414
|
+
def fetch_batches(min_bytes:, max_bytes:, max_wait_time:, automatically_mark_as_processed:)
|
404
415
|
join_group unless @group.member?
|
405
416
|
|
406
417
|
subscribed_partitions = @group.subscribed_partitions
|
@@ -411,6 +422,7 @@ module Kafka
|
|
411
422
|
cluster: @cluster,
|
412
423
|
logger: @logger,
|
413
424
|
min_bytes: min_bytes,
|
425
|
+
max_bytes: max_bytes,
|
414
426
|
max_wait_time: max_wait_time,
|
415
427
|
)
|
416
428
|
|
@@ -18,10 +18,11 @@ module Kafka
|
|
18
18
|
# operation.execute
|
19
19
|
#
|
20
20
|
class FetchOperation
|
21
|
-
def initialize(cluster:, logger:, min_bytes: 1, max_wait_time: 5)
|
21
|
+
def initialize(cluster:, logger:, min_bytes: 1, max_bytes: 10485760, max_wait_time: 5)
|
22
22
|
@cluster = cluster
|
23
23
|
@logger = logger
|
24
24
|
@min_bytes = min_bytes
|
25
|
+
@max_bytes = max_bytes
|
25
26
|
@max_wait_time = max_wait_time
|
26
27
|
@topics = {}
|
27
28
|
end
|
@@ -66,6 +67,7 @@ module Kafka
|
|
66
67
|
options = {
|
67
68
|
max_wait_time: @max_wait_time * 1000, # Kafka expects ms, not secs
|
68
69
|
min_bytes: @min_bytes,
|
70
|
+
max_bytes: @max_bytes,
|
69
71
|
topics: topics,
|
70
72
|
}
|
71
73
|
|
data/lib/kafka/protocol.rb
CHANGED
@@ -24,6 +24,8 @@ module Kafka
|
|
24
24
|
LEAVE_GROUP_API = 13
|
25
25
|
SYNC_GROUP_API = 14
|
26
26
|
SASL_HANDSHAKE_API = 17
|
27
|
+
API_VERSIONS_API = 18
|
28
|
+
CREATE_TOPICS_API = 19
|
27
29
|
|
28
30
|
# A mapping from numeric API keys to symbolic API names.
|
29
31
|
APIS = {
|
@@ -39,6 +41,8 @@ module Kafka
|
|
39
41
|
LEAVE_GROUP_API => :leave_group,
|
40
42
|
SYNC_GROUP_API => :sync_group,
|
41
43
|
SASL_HANDSHAKE_API => :sasl_handshake,
|
44
|
+
API_VERSIONS_API => :api_versions,
|
45
|
+
CREATE_TOPICS_API => :create_topics,
|
42
46
|
}
|
43
47
|
|
44
48
|
# A mapping from numeric error codes to exception classes.
|
@@ -131,5 +135,9 @@ require "kafka/protocol/offset_fetch_request"
|
|
131
135
|
require "kafka/protocol/offset_fetch_response"
|
132
136
|
require "kafka/protocol/offset_commit_request"
|
133
137
|
require "kafka/protocol/offset_commit_response"
|
138
|
+
require "kafka/protocol/api_versions_request"
|
139
|
+
require "kafka/protocol/api_versions_response"
|
134
140
|
require "kafka/protocol/sasl_handshake_request"
|
135
141
|
require "kafka/protocol/sasl_handshake_response"
|
142
|
+
require "kafka/protocol/create_topics_request"
|
143
|
+
require "kafka/protocol/create_topics_response"
|
@@ -0,0 +1,47 @@
|
|
1
|
+
module Kafka
|
2
|
+
module Protocol
|
3
|
+
|
4
|
+
class ApiVersionsResponse
|
5
|
+
class ApiInfo
|
6
|
+
attr_reader :api_key, :min_version, :max_version
|
7
|
+
|
8
|
+
def initialize(api_key:, min_version:, max_version:)
|
9
|
+
@api_key, @min_version, @max_version = api_key, min_version, max_version
|
10
|
+
end
|
11
|
+
|
12
|
+
def api_name
|
13
|
+
Protocol.api_name(api_key)
|
14
|
+
end
|
15
|
+
|
16
|
+
def to_s
|
17
|
+
"#{api_name}=#{min_version}..#{max_version}"
|
18
|
+
end
|
19
|
+
|
20
|
+
def inspect
|
21
|
+
"#<Kafka api version #{to_s}>"
|
22
|
+
end
|
23
|
+
end
|
24
|
+
|
25
|
+
attr_reader :error_code, :apis
|
26
|
+
|
27
|
+
def initialize(error_code:, apis:)
|
28
|
+
@error_code = error_code
|
29
|
+
@apis = apis
|
30
|
+
end
|
31
|
+
|
32
|
+
def self.decode(decoder)
|
33
|
+
error_code = decoder.int16
|
34
|
+
|
35
|
+
apis = decoder.array do
|
36
|
+
ApiInfo.new(
|
37
|
+
api_key: decoder.int16,
|
38
|
+
min_version: decoder.int16,
|
39
|
+
max_version: decoder.int16,
|
40
|
+
)
|
41
|
+
end
|
42
|
+
|
43
|
+
new(error_code: error_code, apis: apis)
|
44
|
+
end
|
45
|
+
end
|
46
|
+
end
|
47
|
+
end
|
@@ -0,0 +1,40 @@
|
|
1
|
+
module Kafka
|
2
|
+
module Protocol
|
3
|
+
|
4
|
+
class CreateTopicsRequest
|
5
|
+
def initialize(topics:, timeout:)
|
6
|
+
@topics, @timeout = topics, timeout
|
7
|
+
end
|
8
|
+
|
9
|
+
def api_key
|
10
|
+
CREATE_TOPICS_API
|
11
|
+
end
|
12
|
+
|
13
|
+
def api_version
|
14
|
+
0
|
15
|
+
end
|
16
|
+
|
17
|
+
def response_class
|
18
|
+
Protocol::CreateTopicsResponse
|
19
|
+
end
|
20
|
+
|
21
|
+
def encode(encoder)
|
22
|
+
encoder.write_array(@topics) do |topic, config|
|
23
|
+
encoder.write_string(topic)
|
24
|
+
encoder.write_int32(config.fetch(:num_partitions))
|
25
|
+
encoder.write_int16(config.fetch(:replication_factor))
|
26
|
+
|
27
|
+
# Replica assignments. We don't care.
|
28
|
+
encoder.write_array([])
|
29
|
+
|
30
|
+
# Config entries. We don't care.
|
31
|
+
encoder.write_array([])
|
32
|
+
end
|
33
|
+
|
34
|
+
# Timeout is in ms.
|
35
|
+
encoder.write_int32(@timeout * 1000)
|
36
|
+
end
|
37
|
+
end
|
38
|
+
|
39
|
+
end
|
40
|
+
end
|