版本说明
Flink和kafka的版本号有一定的匹配关系，操作成功的版本：
- Flink1.17.1
- kafka_2.12-3.3.1
添加kafka连接器依赖
将flink-sql-connector-kafka-1.17.1.jar上传到flink的lib目录下
下载flink-sql-connector-kafka连接器jar包
https://mvnrepository.com/artifact/org.apache.flink/flink-connector-kafka/1.17.1
上传到flink的lib目录下
[hadoop@node2 ~]$ cp flink-connector-kafka-1.17.1.jar $FLINK_HOME/lib分发flink-connector-kafka-1.17.1.jar
xsync $FLINK_HOME/lib/flink-connector-kafka-1.17.1.jar启动yarn-session
[hadoop@node2 ~]$ myhadoop.sh start
[hadoop@node2 ~]$ yarn-session.sh -d启动kafka集群
[hadoop@node2 ~]$ zk.sh start
[hadoop@node2 ~]$ kf.sh start创建kafka主题
查看主题
[hadoop@node2 ~]$ kafka-topics.sh --bootstrap-server node2:9092,node3:9092,node4:9092 --list
如果没有ws1,则创建
[hadoop@node2 ~]$ kafka-topics.sh --bootstrap-server node2:9092,node3:9092,node4:9092 --create --replication-factor 1 --partitions 1 --topic ws1
普通Kafka表
'connector' = 'kafka'
进入Flink SQL客户端
[hadoop@node2 ~]$ sql-client.sh embedded -s yarn-session
...
省略若干日志输出
...
Flink SQL> 创建Kafka的映射表
sql
CREATE TABLE t1(
`event_time` TIMESTAMP(3) METADATA FROM 'timestamp',
--列名和元数据名一致可以省略 FROM 'xxxx', VIRTUAL表示只读
`partition` BIGINT METADATA VIRTUAL,
`offset` BIGINT METADATA VIRTUAL,
id int,
ts bigint ,
vc int )
WITH (
'connector' = 'kafka',
'properties.bootstrap.servers' = 'node2:9092,node3:9092,node4:9094',
'properties.group.id' = 'test',
-- 'earliest-offset', 'latest-offset', 'group-offsets', 'timestamp' and 'specific-offsets'
'scan.startup.mode' = 'earliest-offset',
-- fixed为flink实现的分区器，一个并行度只写往kafka一个分区
'sink.partitioner' = 'fixed',
'topic' = 'ws1',
'format' = 'json'
);可以往kafka读数据，也可以往kafka写数据。
插入数据到Kafka表
如果没有source表，先创建source表，如果source表存在则不需要再创建。
sql
CREATE TABLE source (
id INT,
ts BIGINT,
vc INT
) WITH (
'connector' = 'datagen',
'rows-per-second'='1',
'fields.id.kind'='random',
'fields.id.min'='1',
'fields.id.max'='10',
'fields.ts.kind'='sequence',
'fields.ts.start'='1',
'fields.ts.end'='1000000',
'fields.vc.kind'='random',
'fields.vc.min'='1',
'fields.vc.max'='100'
);把source表插入t1表
sql
insert into t1(id,ts,vc) select * from source;如果报错
[ERROR] Could not execute SQL statement. Reason:
java.lang.ClassNotFoundException: org.apache.kafka.common.serialization.ByteArraySerializer依然同样错误，还不行，把kafka libs目录下的kafka-clients-3.3.1.jar，把jar包发到Flink的lib目录，同时也注意重启sql-client、yarn-session也要重启（重要）
cp $KAFKA_HOME/libs/kafka-clients-3.3.1.jar $FLINK_HOME/lib查看是否复制成功
$ ls $FLINK_HOME/lib重启sql-client重新操作，成功如下：
Flink SQL> CREATE TABLE t1(
> `event_time` TIMESTAMP(3) METADATA FROM 'timestamp',
> --列名和元数据名一致可以省略 FROM 'xxxx', VIRTUAL表示只读
> `partition` BIGINT METADATA VIRTUAL,
> `offset` BIGINT METADATA VIRTUAL,
> id int,
> ts bigint ,
> vc int )
> WITH (
> 'connector' = 'kafka',
> 'properties.bootstrap.servers' = 'node2:9092,node3:9092,node4:9094',
> 'properties.group.id' = 'test',
> -- 'earliest-offset', 'latest-offset', 'group-offsets', 'timestamp' and 'specific-offsets'
> 'scan.startup.mode' = 'earliest-offset',
> -- fixed为flink实现的分区器，一个并��度只写往kafka一个分区
> 'sink.partitioner' = 'fixed',
> 'topic' = 'ws1',
> 'format' = 'json'
> );
[INFO] Execute statement succeed.
Flink SQL> CREATE TABLE source (
> id INT,
> ts BIGINT,
> vc INT
> ) WITH (
> 'connector' = 'datagen',
> 'rows-per-second'='1',
> 'fields.id.kind'='random',
> 'fields.id.min'='1',
> 'fields.id.max'='10',
> 'fields.ts.kind'='sequence',
> 'fields.ts.start'='1',
> 'fields.ts.end'='1000000',
> 'fields.vc.kind'='random',
> 'fields.vc.min'='1',
> 'fields.vc.max'='100'
> );
[INFO] Execute statement succeed.
Flink SQL> insert into t1(id,ts,vc) select * from source;2024-06-14 10:45:30,125 WARN org.apache.flink.yarn.configuration.YarnLogConfigUtil [] - The configuration directory ('/home/hadoop/soft/flink-1.17.1/conf') already contains a LOG4J config file.If you want to use logback, then please delete or rename the log configuration file.
2024-06-14 10:45:30,673 INFO org.apache.hadoop.yarn.client.RMProxy [] - Connecting to ResourceManager at node3/192.168.193.143:8032
2024-06-14 10:45:31,027 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - No path for the flink jar passed. Using the location of class org.apache.flink.yarn.YarnClusterDescriptor to locate the jar
2024-06-14 10:45:31,227 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - Found Web Interface node3:41749 of application 'application_1718331886020_0001'.
insert into t1(id,ts,vc) select * from source;
[INFO] Submitting SQL update statement to the cluster...
[INFO] SQL update statement has been successfully submitted to the cluster:
Job ID: b1765f969c3ae637bd4c8100efbb0c4e
查询Kafka表
sql
select * from t1;报错
[ERROR] Could not execute SQL statement. Reason:
java.lang.ClassNotFoundException: org.apache.kafka.clients.consumer.ConsumerRecord
重启yarn session，重新操作，成功如下：
Flink SQL> CREATE TABLE t1(
> `event_time` TIMESTAMP(3) METADATA FROM 'timestamp',
> --列名和元数据名一致可以省略 FROM 'xxxx', VIRTUAL表示只读
> `partition` BIGINT METADATA VIRTUAL,
> `offset` BIGINT METADATA VIRTUAL,
> id int,
> ts bigint ,
> vc int )
> WITH (
> 'connector' = 'kafka',
> 'properties.bootstrap.servers' = 'node2:9092,node3:9092,node4:9094',
> 'properties.group.id' = 'test',
> -- 'earliest-offset', 'latest-offset', 'group-offsets', 'timestamp' and 'specific-offsets'
> 'scan.startup.mode' = 'earliest-offset',
> -- fixed为flink实现的分区器，一个并??度只写往kafka一个分区
> 'sink.partitioner' = 'fixed',
> 'topic' = 'ws1',
> 'format' = 'json'
> );
[INFO] Execute statement succeed.
Flink SQL> CREATE TABLE source (
> id INT,
> ts BIGINT,
> vc INT
> ) WITH (
> 'connector' = 'datagen',
> 'rows-per-second'='1',
> 'fields.id.kind'='random',
> 'fields.id.min'='1',
> 'fields.id.max'='10',
> 'fields.ts.kind'='sequence',
> 'fields.ts.start'='1',
> 'fields.ts.end'='1000000',
> 'fields.vc.kind'='random',
> 'fields.vc.min'='1',
> 'fields.vc.max'='100'
> );
[INFO] Execute statement succeed.
Flink SQL> insert into t1(id,ts,vc) select * from source;2024-06-14 11:22:17,971 WARN org.apache.flink.yarn.configuration.YarnLogConfigUtil [] - The configuration directory ('/home/hadoop/soft/flink-1.17.1/conf') already contains a LOG4J config file.If you want to use logback, then please delete or rename the log configuration file.
2024-06-14 11:22:18,422 INFO org.apache.hadoop.yarn.client.RMProxy [] - Connecting to ResourceManager at node3/192.168.193.143:8032
2024-06-14 11:22:18,895 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - No path for the flink jar passed. Using the location of class org.apache.flink.yarn.YarnClusterDescriptor to locate the jar
2024-06-14 11:22:19,052 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - Found Web Interface node4:38788 of application 'application_1718331886020_0002'.
insert into t1(id,ts,vc) select * from source;
[INFO] Submitting SQL update statement to the cluster...
[INFO] SQL update statement has been successfully submitted to the cluster:
Job ID: 84292f84d1fce4756ccd8ae294b6163a
Flink SQL> select * from t1;2024-06-14 11:23:38,338 WARN org.apache.flink.yarn.configuration.YarnLogConfigUtil [] - The configuration directory ('/home/hadoop/soft/flink-1.17.1/conf') already contains a LOG4J config file.If you want to use logback, then please delete or rename the log configuration file.
2024-06-14 11:23:38,606 INFO org.apache.hadoop.yarn.client.RMProxy [] - Connecting to ResourceManager at node3/192.168.193.143:8032
2024-06-14 11:23:38,617 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - No path for the flink jar passed. Using the location of class org.apache.flink.yarn.YarnClusterDescriptor to locate the jar
2024-06-14 11:23:38,649 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - Found Web Interface node4:38788 of application 'application_1718331886020_0002'.
select * from t1;
[INFO] Result retrieval cancelled.
Flink SQL>
upsert-kafka表
'connector' = 'upsert-kafka'
如果当前表存在更新操作，那么普通的kafka连接器将无法满足，此时可以使用Upsert Kafka连接器。
创建upsert-kafka的映射表(必须定义主键)
sql
CREATE TABLE t2(
id int ,
sumVC int ,
primary key (id) NOT ENFORCED
)
WITH (
'connector' = 'upsert-kafka',
'properties.bootstrap.servers' = 'node2:9092',
'topic' = 'ws2',
'key.format' = 'json',
'value.format' = 'json'
);如果没有kafka名为ws2的topic，将自动被创建。
插入upsert-kafka表
sql
insert into t2 select id,sum(vc) sumVC from source group by id;查询upsert-kafka表
upsert-kafka 无法从指定的偏移量读取，只会从主题的源读取。如此，才知道整个数据的更新过程。并且通过 -U，+U，+I 等符号来显示数据的变化过程。
设置显示模式
sql
SET sql-client.execution.result-mode=tableau;查询t2表数据
sql
select * from t2;如果发现没有输出数据，原因是之前的source表已经生成到end（1000000）就不再生成数据了。
进入Flink Web UI，cancel掉所有running job，重新操作成功如下：
删除表
Flink SQL> show tables;
+------------+
| table name |
+------------+
| source |
| t1 |
| t2 |
+------------+
3 rows in set
Flink SQL> drop table source;
Flink SQL> drop table t1;
Flink SQL> drop table t2;创建表
sql
CREATE TABLE source (
id INT,
ts BIGINT,
vc INT
) WITH (
'connector' = 'datagen',
'rows-per-second'='1',
'fields.id.kind'='random',
'fields.id.min'='1',
'fields.id.max'='10',
'fields.ts.kind'='sequence',
'fields.ts.start'='1',
'fields.ts.end'='1000000',
'fields.vc.kind'='random',
'fields.vc.min'='1',
'fields.vc.max'='100'
);
sql
CREATE TABLE t2(
id int ,
sumVC int ,
primary key (id) NOT ENFORCED
)
WITH (
'connector' = 'upsert-kafka',
'properties.bootstrap.servers' = 'node2:9092',
'topic' = 'ws2',
'key.format' = 'json',
'value.format' = 'json'
);设置显示模式
SET sql-client.execution.result-mode=tableau;查询表
select * from t2;
完成！enjoy it!