kafka-connect-hive是基于kafka-connect平台实现的hive数据读取和写入插件，主要由source、sink两部分组成，source部分完成hive表数据的读取任务，kafka-connect将这些数据写入到其他数据存储层中，比如hive到ES数据的流入。sink部分完成向hive表写数据的任务，kafka-connect将第三方数据源（如MySQL）里的数据读取并写入到hive表中。
在这里我使用的是Landoop公司开发的kafka-connect-hive插件，项目文档地址Hive Sink，接下来看看如何使用该插件的sink部分。
环境准备
- Apache Kafka 2.11-2.1.0
- Confluent-5.1.0
- Apache Hadoop 2.6.3
- Apache Hive 1.2.1
- Java 1.8
功能
- 支持
KCQL路由查询，允许将kafka主题中的所有字段或部分字段写入hive表中 - 支持根据某一字段动态分区
- 支持全量和增量同步数据，不支持部分更新
开始使用
启动依赖
1、启动kafka：
cd kafka_2.11-2.1.0
bin/kafka-server-start.sh config/server.properties &
2、启动schema-registry：
cd confluent-5.1.0
bin/schema-registry-start etc/schema-registry/schema-registry.properties &
schema-registry组件提供了kafka topic的schema管理功能，保存了schema的各个演变版本，帮助我们解决新旧数据schema兼容问题。这里我们使用apache avro库来序列化kafka的key和value，因此需要依赖schema-registry组件，schema-registry使用默认的配置。
3、启动kafka-connect：
修改confluent-5.1.0/etc/schema-registry目录下connect-avro-distributed.properties文件的配置，修改后内容如下：
# Sample configuration for a distributed Kafka Connect worker that uses Avro serialization and
# integrates the the Schema Registry. This sample configuration assumes a local installation of
# Confluent Platform with all services running on their default ports.
# Bootstrap Kafka servers. If multiple servers are specified, they should be comma-separated.
bootstrap.servers=localhost:9092
# The group ID is a unique identifier for the set of workers that form a single Kafka Connect
# cluster
group.id=connect-cluster
# The converters specify the format of data in Kafka and how to translate it into Connect data.
# Every Connect user will need to configure these based on the format they want their data in
# when loaded from or stored into Kafka
key.converter=io.confluent.connect.avro.AvroConverter
key.converter.schema.registry.url=http://localhost:8081
value.converter=io.confluent.connect.avro.AvroConverter
value.converter.schema.registry.url=http://localhost:8081
# Internal Storage Topics.
#
# Kafka Connect distributed workers store the connector and task configurations, connector offsets,
# and connector statuses in three internal topics. These topics MUST be compacted.
# When the Kafka Connect distributed worker starts, it will check for these topics and attempt to create them
# as compacted topics if they don't yet exist, using the topic name, replication factor, and number of partitions
# as specified in these properties, and other topic-specific settings inherited from your brokers'
# auto-creation settings. If you need more control over these other topic-specific settings, you may want to
# manually create these topics before starting Kafka Connect distributed workers.
#
# The following properties set the names of these three internal topics for storing configs, offsets, and status.
config.storage.topic=connect-configs
offset.storage.topic=connect-offsets
status.storage.topic=connect-statuses
# The following properties set the replication factor for the three internal topics, defaulting to 3 for each
# and therefore requiring a minimum of 3 brokers in the cluster. Since we want the examples to run with
# only a single broker, we set the replication factor here to just 1. That's okay for the examples, but
# ALWAYS use a replication factor of AT LEAST 3 for production environments to reduce the risk of
# losing connector offsets, configurations, and status.
config.storage.replication.factor=1
offset.storage.replication.factor=1
status.storage.replication.factor=1
# The config storage topic must have a single partition, and this cannot be changed via properties.
# Offsets for all connectors and tasks are written quite frequently and therefore the offset topic
# should be highly partitioned; by default it is created with 25 partitions, but adjust accordingly
# with the number of connector tasks deployed to a distributed worker cluster. Kafka Connect records
# the status less frequently, and so by default the topic is created with 5 partitions.
#offset.storage.partitions=25
#status.storage.partitions=5
# The offsets, status, and configurations are written to the topics using converters specified through
# the following required properties. Most users will always want to use the JSON converter without schemas.
# Offset and config data is never visible outside of Connect in this format.
internal.key.converter=org.apache.kafka.connect.json.JsonConverter
internal.value.converter=org.apache.kafka.connect.json.JsonConverter
internal.key.converter.schemas.enable=false
internal.value.converter.schemas.enable=false
# Confluent Control Center Integration -- uncomment these lines to enable Kafka client interceptors
# that will report audit data that can be displayed and analyzed in Confluent Control Center
# producer.interceptor.classes=io.confluent.monitoring.clients.interceptor.MonitoringProducerInterceptor
# consumer.interceptor.classes=io.confluent.monitoring.clients.interceptor.MonitoringConsumerInterceptor
# These are provided to inform the user about the presence of the REST host and port configs
# Hostname & Port for the REST API to listen on. If this is set, it will bind to the interface used to listen to requests.
#rest.host.name=0.0.0.0
#rest.port=8083
# The Hostname & Port that will be given out to other workers to connect to i.e. URLs that are routable from other servers.
#rest.advertised.host.name=0.0.0.0
#rest.advertised.port=8083
# Set to a list of filesystem paths separated by commas (,) to enable class loading isolation for plugins
# (connectors, converters, transformations). The list should consist of top level directories that include
# any combination of:
# a) directories immediately containing jars with plugins and their dependencies
# b) uber-jars with plugins and their dependencies
# c) directories immediately containing the package directory structure of classes of plugins and their dependencies
# Examples:
# plugin.path=/usr/local/share/java,/usr/local/share/kafka/plugins,/opt/connectors,
# Replace the relative path below with an absolute path if you are planning to start Kafka Connect from within a
# directory other than the home directory of Confluent Platform.
plugin.path=/kafka/confluent-5.1.0/plugins/lib
这里需要设置plugin.path参数，该参数指定了kafka-connect插件包的保存地址，必须得设置。
下载kafka-connect-hive-1.2.1-2.1.0-all.tar.gz，解压后将kafka-connect-hive-1.2.1-2.1.0-all.jar放到plugin.path指定的目录下，然后执行如下命令启动kafka-connect：
cd confluent-5.1.0
bin/connect-distributed etc/schema-registry/connect-avro-distributed.properties
准备测试数据
1、在hive服务器上使用beeline执行如下命令：
# 创建hive_connect数据库
create database hive_connect;
# 创建cities_orc表
use hive_connect;
create table cities_orc (city string, state string, population int, country string) stored as orc;
2、使用postman添加kafka-connect-hive sink的配置到kafka-connect：
URL：localhost:8083/connectors/
请求类型：POST
请求体如下：
{
"name": "hive-sink-example",
"config": {
"name": "hive-sink-example",
"connector.class": "com.landoop.streamreactor.connect.hive.sink.hiveSinkConnector",
"tasks.max": 1,
"topics": "hive_sink_orc",
"connect.hive.kcql": "insert into cities_orc select * from hive_sink_orc AUTOCREATE PARTITIONBY state STOREAS ORC WITH_FLUSH_INTERVAL = 10 WITH_PARTITIONING = DYNAMIC",
"connect.hive.database.name": "hive_connect",
"connect.hive.hive.metastore": "thrift",
"connect.hive.hive.metastore.uris": "thrift://quickstart.cloudera:9083",
"connect.hive.fs.defaultFS": "hdfs://quickstart.cloudera:9001",
"connect.hive.error.policy": "NOOP",
"connect.progress.enabled": true
}
}
开始测试，查看结果
启动kafka producer，写入测试数据，scala测试代码如下：
class AvroTest {
/**
* 测试kafka使用avro方式生产数据
* 参考 https://docs.confluent.io/current/schema-registry/docs/serializer-formatter.html
*/
@Test
def testProducer: Unit = {
// 设置kafka broker地址、序列化方式、schema-registry组件的地址
val props = new Properties()
props.put(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG, "localhost:9092")
props.put(ProducerConfig.KEY_SERIALIZER_CLASS_CONFIG, classOf[io.confluent.kafka.serializers.KafkaAvroSerializer])
props.put(ProducerConfig.VALUE_SERIALIZER_CLASS_CONFIG, classOf[io.confluent.kafka.serializers.KafkaAvroSerializer])
props.put("citySchema.registry.url", "http://localhost:8081")
// 设置schema
val citySchema = "{"type":"record","name":"myrecord","fields":[{"name":"city","type":"string"},{"name":"state","type":"string"},{"name":"population","type":"int"},{"name":"country","type":"string"}]}"
val parser = new Schema.Parser()
val schema = parser.parse(citySchema)
// 构造测试数据
val avroRecord1 = new GenericData.Record(schema)
avroRecord1.put("city", "Philadelphia")
avroRecord1.put("state", "PA")
avroRecord1.put("population", 1568000)
avroRecord1.put("country", "USA")
val avroRecord2 = new GenericData.Record(schema)
avroRecord2.put("city", "Chicago")
avroRecord2.put("state", "IL")
avroRecord2.put("population", 2705000)
avroRecord2.put("country", "USA")
val avroRecord3 = new GenericData.Record(schema)
avroRecord3.put("city", "New York")
avroRecord3.put("state", "NY")
avroRecord3.put("population", 8538000)
avroRecord3.put("country", "USA")
// 生产数据
val producer = new KafkaProducer[String, GenericData.Record](props)
try {
val recordList = List(avroRecord1, avroRecord2, avroRecord3)
val key = "key1"
for (elem <- recordList) {
val record = new ProducerRecord("hive_sink_orc", key, elem)
for (i <- 0 to 100) {
val ack = producer.send(record).get()
println(s"${ack.toString} written to partition ${ack.partition.toString}")
}
}
} catch {
case e: Throwable => e.printStackTrace()
} finally {
// When you're finished producing records, you can flush the producer to ensure it has all been written to Kafka and
// then close the producer to free its resources.
// 调用flush方法确保所有数据都被写入到Kafka
producer.flush()
// 调用close方法释放资源
producer.close()
}
}
}
4、使用beeline查询hive数据：
use hive_connect;
select * from cities_orc;
输出部分结果如下：
+------------------+------------------------+---------------------+-------------------+--+
| cities_orc.city | cities_orc.population | cities_orc.country | cities_orc.state |
+------------------+------------------------+---------------------+-------------------+--+
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Chicago | 2705000 | USA | IL |
| Philadelphia | 1568000 | USA | PA |
| Philadelphia | 1568000 | USA | PA |
| Philadelphia | 1568000 | USA | PA |
| Philadelphia | 1568000 | USA | PA |
| Philadelphia | 1568000 | USA | PA |
| Philadelphia | 1568000 | USA | PA |
| Philadelphia | 1568000 | USA | PA |
| Philadelphia | 1568000 | USA | PA |
| Philadelphia | 1568000 | USA | PA |
| Philadelphia | 1568000 | USA | PA |
配置说明
KCQL配置
connect.hive.kcql中的配置项说明如下：
WITH_FLUSH_INTERVAL：long类型，表示文件提交的时间间隔，单位是毫秒WITH_FLUSH_SIZE：long类型，表示执行提交操作之前，已提交到HDFS的文件长度WITH_FLUSH_COUNT：long类型，表示执行提交操作之前，未提交到HDFS的记录数WITH_SCHEMA_EVOLUTION：string类型，默认值是MATCH，表示hive schema和kafka topic record的schema的兼容策略，hive connector会使用该策略来添加或移除字段WITH_TABLE_LOCATION：string类型，表示hive表在HDFS中的存储位置，如果不指定的话，将使用hive中默认的配置WITH_OVERWRITE：boolean类型，表示是否覆盖hive表中已存在的记录，使用该策略时，会先删除已有的表，再新建PARTITIONBY：List<String>类型，保存分区字段。指定后，将从指定的列中获取分区字段的值WITH_PARTITIONING：string类型，默认值是STRICT，表示分区创建方式。主要有DYNAMIC和STRICT两种方式。DYNAMIC方式将根据PARTITIONBY指定的分区字段创建分区，STRICT方式要求必须已经创建了所有分区AUTOCREATE：boolean类型，表示是否自动创建表
Kafka connect配置
Kafka connect的配置项说明如下：
name：string类型，表示connector的名称，在整个kafka-connect集群中唯一topics：string类型，表示保存数据的topic名称，必须与KCQL语句中的topic名称一致tasks.max：int类型，默认值为1，表示connector的任务数量connector.class：string类型，表示connector类的名称，值必须是com.landoop.streamreactor.connect.hive.sink.HiveSinkConnectorconnect.hive.kcql：string类型，表示kafka-connect查询语句connect.hive.database.name：string类型，表示hive数据库的名称connect.hive.hive.metastore：string类型，表示连接hive metastore所使用的网络协议connect.hive.hive.metastore.uris：string类型，表示hive metastore的连接地址connect.hive.fs.defaultFS：string类型，表示HDFS的地址