After the download, unzip kafka, there are three production profile to the topic under kafka / config, consumer-related.
- server.properties-- server configuration
- Production side configuration producer.properties--
- consumer.properties-- consumer side configuration
This introduces the configuration parameters of the server.
server.properties
# Number specified kafka cluster global broker, each broker's number can not be repeated. broker. ID = 0 Address list specifies broker #Listeners listen for client connections, i.e. the connection to the client's broker address list entry, # configured to format Protocol: // hostname: Port, a plurality of intermediate addresses separated by commas, # protocol which represents the protocol type, Kafka currently supported protocols include PLAINTEXT, SSL, SASL_SSL etc. If security authentication is not turned on, you can use simple PLAINTEXT; # hostname represents the host name, host name is best not to be empty, here is localhost; #port behalf of the service port, producer or consumer will establish a connection on this port, this is 9092 . listeners = PLAINTEXT: // localhost: 9092 # advertised.listeners mainly used IaaS (Infrastructure as a Service) environment, # if multiple network cards comprising a public network and private network card NIC, listeners configuration parameters for the private network IP address binding broker using interprocess communication, # advertised.listeners configuration parameter bindings for external public IP clients to use, advertised.listeners = PLAINTEXT: //localhost: 9092 number of threads #broker processing network requests, i.e. the number of threads of the received message, generally you do not need to modify. # Receiving thread will received message into memory, and then written to disk from memory. num.network.threads = 3 the number of threads to use when writing to disk # messages from memory. # Of threads to handle the number of disk IO. num.io.threads = . 8 # send socket buffer size. socket.send.buffer.bytes = 102400 # to accept the socket buffer size. socket.receive.buffer.bytes = 102400 # socket buffer size requested. socket.request.max.bytes = 104857600 #kafka operation log storage path log.dirs = / usr / local / Kafka / log / Kafka #topic current slice number in the broker num.partitions = . 1 # is the default log seven days time, overtime, then it will be clean up, clean up this thing you need to have some threads to do. # This is used to set the number of threads to restore and clean up the data under the data. num.recovery.threads.per.data. dir = 1 #-themed number of copies offsets.topic.replication.factor = 1 # affairs theme replication factor, set higher in order to ensure availability. transaction.state.log.replication.factor = 1 # affairs topics covered by the min.insync.replicas configuration. transaction.state.log.min.isr = 1 # thread receives the message will be received into memory, and then from memory to disk, when the message is written to the disk from memory, there is a time limit (time threshold) and a quantity (threshold number) # set here is the threshold number, the number of the message reaches a threshold, triggers the flush to disk. The next parameter is the time threshold. log.flush.interval.messages = 10000 Time # of the message buffer, reaching a threshold, triggers the flush message from memory to disk, in milliseconds. log.flush.interval.ms = 1000 # Maximum time to keep log files, the unit hours, the default seven days (168 hours), overtime will be deleted. log.retention.hours = 168 #topic partition segment is a bunch of files stored # This parameter controls the log file size of each segment, in bytes, which is the default for the 1G. log.segment.bytes = 1073741824 time periodically checks the file size of # milliseconds. I.e. # periodically check segment file has not reached. 1G (a parameter) log.retention.check.interval.ms = 300000 #zookeeper cluster address, may be a plurality, the plurality of addresses separated by commas, #broker need use zookeeper save meta data. zookeeper.connect = localhost: 2181 #zookeeper connection timeout zookeeper.connection.timeout.ms = 6000 # prior to performing the first re-balancing, group coordinator will wait for more consumers to join the group time. # Possibility of a longer delay time means rebalancing of the smaller, but the waiting time begins to increase. group.initial.rebalance.delay.ms = 0