这个文档是 Cloudera Distribution of Apache Kafka 1.3.x. 其他版本的文档在Cloudera Documentation.
Using Kafka with Flume
在CDH 5.2.0 及更高的版本中, Flume 包含一个Kafka source and sink。使用它们可以让数据从Kafka流入Hadoop或者从任何Flume source 流入Kafka。
重要提示:不能配置一个Kafka source发送数据到 a Kafka sink.如果这么做, the Kafka source sets the topic in the event header, overriding the sink configuration and creating an infinite loop, sending messages back and forth between the source and sink. If you need to use both a source and a sink, use an interceptor to modify the event header and set a different topic.
Kafka Source
使用Kafka source 让数据从Kafka topics 流入 Hadoop. The Kafka source 可以与任何Flume sink合并, 这样很容易把数据从 Kafka 写到 HDFS, HBase, 以及Solr.
下面的 Flume 配置示例，是使用 Kafka source 发送数据到 HDFS sink:
1tier1.sources = source1 2 tier1.channels = channel1 3 tier1.sinks = sink1 4 5 tier1.sources.source1.type = org.apache.flume.source.kafka.KafkaSource 6 tier1.sources.source1.zookeeperConnect = zk01.example.com:2181 7 tier1.sources.source1.topic = weblogs 8 tier1.sources.source1.groupId = flume 9 tier1.sources.source1.channels = channel1 10 tier1.sources.source1.interceptors = i1 11 tier1.sources.source1.interceptors.i1.type = timestamp 12 tier1.sources.source1.kafka.consumer.timeout.ms = 100 13 14 tier1.channels.channel1.type = memory 15 tier1.channels.channel1.capacity = 10000 16 tier1.channels.channel1.transactionCapacity = 1000 17 18 tier1.sinks.sink1.type = hdfs 19 tier1.sinks.sink1.hdfs.path = /tmp/kafka/%{topic}/%y-%m-%d 20 tier1.sinks.sink1.hdfs.rollInterval = 5 21 tier1.sinks.sink1.hdfs.rollSize = 0 22 tier1.sinks.sink1.hdfs.rollCount = 0 23 tier1.sinks.sink1.hdfs.fileType = DataStream 24 tier1.sinks.sink1.channel = channel1
为了更高的吞吐量, 可以配置多个Kafka sources读取一个 topic.如果所有sources配置一个相同的groupID, 并且topic 有多个分区, 设置每一个source 从不同的分区读取数据,就可以改善效率.
下面的列表描述Kafka source 支持的参数; 必须的参数使用粗体列出 .
Table 1. Kafka Source Properties
Property Name
Default Value
Description
type
必须设置为org.apache.flume.source.kafka.KafkaSource.
zookeeperConnect
The URI of the ZooKeeper server or quorum used by Kafka. This can be a single node (for example, zk01.example.com:2181) or a comma-separated list of nodes in a ZooKeeper quorum (for example, zk01.example.com:2181,zk02.example.com:2181, zk03.example.com:2181).
topic
source 读取消息的Kafka topic。 Flume 每个source只支持一个 topic.。
groupID
flume
The unique identifier of the Kafka consumer group. Set the same groupID in all sources to indicate that they belong to the same consumer group.
batchSize
1000
向channel写入消息的最多条数
batchDurationMillis
1000
向channel书写的最大时间 (毫秒) 。
其他Kafka consumer 支持的属性
通过Kafka source配置Kafka consumer。可以使用任何consumer 支持的属性。 Prepend the consumer property name with the prefix kafka. (for example, kafka.fetch.min.bytes). See the Kafka documentation for the full list of Kafka consumer properties.
调优
Kafka source 重写了两个Kafka consumer 的属性:
- auto.commit.enable 设置为 false by the source, and every batch is committed. 为了改善性能, 设置为 true 改为使用 kafka.auto.commit.enable。 这个可能会丢失数据 if the source goes down before committing.
- consumer.timeout.ms设置为 10, so when Flume polls Kafka for new data, it waits no more than 10 ms for the data to be available. Setting this to a higher value can reduce CPU utilization due to less frequent polling, but introduces latency in writing batches to the channel.
Kafka Sink
使用Kafka sink 从一个 Flume source发送数据到 Kafka . You can use the Kafka sink in addition to Flume sinks such as HBase or HDFS.
The following Flume configuration example uses a Kafka sink with an exec source:
1tier1.sources = source1 2 tier1.channels = channel1 3 tier1.sinks = sink1 4 5 tier1.sources.source1.type = exec 6 tier1.sources.source1.command = /usr/bin/vmstat 1 7 tier1.sources.source1.channels = channel1 8 9 tier1.channels.channel1.type = memory 10 tier1.channels.channel1.capacity = 10000 11 tier1.channels.channel1.transactionCapacity = 1000 12 13 tier1.sinks.sink1.type = org.apache.flume.sink.kafka.KafkaSink 14 tier1.sinks.sink1.topic = sink1 15 tier1.sinks.sink1.brokerList = kafka01.example.com:9092,kafka02.example.com:9092 16 tier1.sinks.sink1.channel = channel1 17 tier1.sinks.sink1.batchSize = 20
The following table describes parameters the Kafka sink supports; required properties are listed in bold.
Table 2. Kafka Sink Properties
Property Name
Default Value
Description
type
必须设置为： org.apache.flume.sink.kafka.KafkaSink.
brokerList
The brokers the Kafka sink uses to discover topic partitions, formatted as a comma-separated list of hostname:port entries. You do not need to specify the entire list of brokers, but Cloudera recommends that you specify at least two for high availability.
topic
default-flume-topic
The Kafka topic to which messages are published by default. If the event header contains a topic field, the event is published to the designated topic, overriding the configured topic.
batchSize
100
The number of messages to process in a single batch. Specifying a larger batchSize can improve throughput and increase latency.
requiredAcks
1
The number of replicas that must acknowledge a message before it is written successfully. Possible values are 0 (do not wait for an acknowledgement), 1 (wait for the leader to acknowledge only), and -1 (wait for all replicas to acknowledge). To avoid potential loss of data in case of a leader failure, set this to -1.
其他Kafka producer所支持的属性
Used to configure the Kafka producer used by the Kafka sink. You can use any producer properties supported by Kafka. Prepend the producer property name with the prefix kafka. (for example, kafka.compression.codec). See the Kafka documentation for the full list of Kafka producer properties.
Kafka sink 使用 topic 以及 key properties from the FlumeEvent headers to determine where to send events in Kafka. If the header contains the topic property, that event is sent to the designated topic, overriding the configured topic. If the header contains the key property, that key is used to partition events within the topic. Events with the same key are sent to the same partition. If the key parameter is not specified, events are distributed randomly to partitions. Use these properties to control the topics and partitions to which events are sent through the Flume source or interceptor.
Kafka Channel
CDH 5.3 以及更高的版本包含一个Kafka channel to Flume in addition to the existing memory and file channels. 可以使用Kafka channel:
- To write to Hadoop directly from Kafka without using a source.不使用source，从Kafka直接向hadoop中写数据。
- To write to Kafka directly from Flume sources without additional buffering.不使用额外的缓冲区直接从Flume source向Kafka写数据。
- As a reliable and highly available channel for any source/sink combination.可以与任何source/sink结合。
如下的 Flume 配置使用了一个Kafka channel 以及一个exec source 和 hdfs sink:
1tier1.sources = source1 2tier1.channels = channel1 3tier1.sinks = sink1 4 5tier1.sources.source1.type = exec 6tier1.sources.source1.command = /usr/bin/vmstat 1 7tier1.sources.source1.channels = channel1 8 9tier1.channels.channel1.type = org.apache.flume.channel.kafka.KafkaChannel 10tier1.channels.channel1.capacity = 10000 11tier1.channels.channel1.transactionCapacity = 1000 12tier1.channels.channel1.brokerList = kafka02.example.com:9092,kafka03.example.com:9092 13tier1.channels.channel1.topic = channel2 14tier1.channels.channel1.zookeeperConnect = zk01.example.com:2181 15tier1.channels.channel1.parseAsFlumeEvent = true 16 17tier1.sinks.sink1.type = hdfs 18tier1.sinks.sink1.hdfs.path = /tmp/kafka/channel 19tier1.sinks.sink1.hdfs.rollInterval = 5 20tier1.sinks.sink1.hdfs.rollSize = 0 21tier1.sinks.sink1.hdfs.rollCount = 0 22tier1.sinks.sink1.hdfs.fileType = DataStream 23tier1.sinks.sink1.channel = channel1
下面的列表描述了Kafka channel 所支持的参数; 粗体为必要参数.
Table 3. Kafka Channel Properties
Property Name
Default Value
Description
type
必须设置为：org.apache.flume.channel.kafka.KafkaChannel.
brokerList
The brokers the Kafka channel uses to discover topic partitions, formatted as a comma-separated list of hostname:port entries. You do not need to specify the entire list of brokers, but Cloudera recommends that you specify at least two for high availability.
zookeeperConnect
The URI of the ZooKeeper server or quorum used by Kafka. This can be a single node (for example, zk01.example.com:2181) or a comma-separated list of nodes in a ZooKeeper quorum (for example, zk01.example.com:2181,zk02.example.com:2181, zk03.example.com:2181).
topic
flume-channel
The Kafka topic the channel will use.
groupID
flume
The unique identifier of the Kafka consumer group the channel uses to register with Kafka.
parseAsFlumeEvent
true
Set to true if a Flume source is writing to the channel and expects AvroDataums with the FlumeEvent schema (org.apache.flume.source.avro.AvroFlumeEvent) in the channel. Set to false if other producers are writing to the topic that the channel is using.
readSmallestOffset
false
If true, reads all data in the topic. If false, reads only data written after the channel has started. Only used when parseAsFlumeEvent is false.
kafka.consumer.timeout.ms
100
当向sink写数据时轮询的间隔时间.
其他Kafka producer所支持的属性
Used to configure the Kafka producer. You can use any producer properties supported by Kafka. Prepend the producer property name with the prefix kafka. (for example, kafka.compression.codec). See the Kafka documentation for the full list of Kafka producer properties.