Flink CDC 2.2 版本新增了 OceanBase CE，PolarDB-X，SqlServer，TiDB 四种数据源接入。其中新增 OceanBase CDC，SqlServer CDC，TiDB CDC 三个连接器，而 PolarDB-X 的支持则是通过对 MySQL CDC 连接器进行兼容适配实现。
发布原文链接：https://mp.weixin.qq.com/s/oLezzfC6LFENwO--BJpelQ
最近在测tidb原生的ticdc（PingCap提供）导入数据到kafka，再使用flink StarRocks cdc导入到StarRocks，刚好可以使用flink新发布的flink tidb cdc来对接，实现从tidb数据同步到StarRocks。
1.tidb环境部署
- 下载tidb部署安装包tidb-community-server-v4.0.9-linux-amd64.tar.gz，下载地址
https://download.pingcap.org/tidb-community-server-v4.0.9-linux-amd64.tar.gz
- 安装环境
tar -xvf tidb-community-server-v4.0.9-linux-amd64.tar.gz cd tidb-community-server-v4.0.9-linux-amd64 sh local_install.sh
- 使用tidb playground模式启动tidb单节点测试环境
sr@cs03:~/.tiup/bin$pwd /home/disk1/sr/.tiup/bin sr@cs03:~/.tiup/bin$./tiup playground Starting component `playground`: /home/disk1/sr/.tiup/components/playground/v1.3.2/tiup-playground Use the latest stable version: v4.0.9 Specify version manually: tiup playground <version> The stable version: tiup playground v4.0.0 The nightly version: tiup playground nightly Playground Bootstrapping... Start pd instance Start tikv instance Start tidb instance Waiting for tidb instances ready 127.0.0.1:4000 ... Done Start tiflash instance Waiting for tiflash instances ready 127.0.0.1:3930 ... Done CLUSTER START SUCCESSFULLY, Enjoy it ^-^ To connect TiDB: mysql --host 127.0.0.1 --port 4000 -u root To view the dashboard: http://127.0.0.1:2379/dashboard To view the Prometheus: http://127.0.0.1:37449 To view the Grafana: http://127.0.0.1:3000
2.Flink tidb cdc connector下载
下载flink tidb cdc connector，放在flink lib下：
wget https://repo1.maven.org/maven2/com/ververica/flink-sql-connector-tidb-cdc/2.2.0/flink-sql-connector-tidb-cdc-2.2.0.jar
sr@cs03:~/app/flink-1.13.3/lib$pwd /home/disk1/sr/app/flink-1.13.3/lib sr@cs03:~/app/flink-1.13.3/lib$ll total 353940 -rw-rw-r-- 1 sr sr 9404460 Mar 19 17:29 flink-connector-starrocks-1.2.1_flink-1.13_2.12.jar -rw-r--r-- 1 sr sr 92313 Oct 13 2021 flink-csv-1.13.3.jar -rw-r--r-- 1 sr sr 115418686 Oct 13 2021 flink-dist_2.11-1.13.3.jar -rw-r--r-- 1 sr sr 19583 Nov 5 21:27 flink-format-changelog-json-1.4.0.jar -rw-r--r-- 1 sr sr 148127 Oct 13 2021 flink-json-1.13.3.jar -rw-rw-r-- 1 sr sr 41368997 Feb 12 2020 flink-shaded-hadoop-2-uber-2.7.5-10.0.jar.bak -rwxrwxr-x 1 sr sr 7709740 Jun 8 2021 flink-shaded-zookeeper-3.4.14.jar -rw-rw-r-- 1 sr sr 3674114 Oct 12 2021 flink-sql-connector-kafka_2.12-1.13.3.jar -rw-rw-r-- 1 sr sr 19648014 Dec 16 23:36 flink-sql-connector-mysql-cdc-2.1.1.jar -rw-rw-r-- 1 sr sr 85355380 Apr 19 17:03 flink-sql-connector-tidb-cdc-2.2.0.jar -rw-r--r-- 1 sr sr 36453353 Oct 13 2021 flink-table_2.11-1.13.3.jar -rw-r--r-- 1 sr sr 41061738 Oct 13 2021 flink-table-blink_2.11-1.13.3.jar -rwxrwxr-x 1 sr sr 67114 Mar 31 2021 log4j-1.2-api-2.12.1.jar -rwxrwxr-x 1 sr sr 276771 Mar 31 2021 log4j-api-2.12.1.jar -rwxrwxr-x 1 sr sr 1674433 Mar 31 2021 log4j-core-2.12.1.jar -rwxrwxr-x 1 sr sr 23518 Mar 31 2021 log4j-slf4j-impl-2.12.1.jar
3.Yarn session模式启动flink环境
sr@cs03:~/app/flink-1.13.3/bin$./yarn-session.sh SLF4J: Class path contains multiple SLF4J bindings. SLF4J: Found binding in [jar:file:/home/disk1/sr/app/flink-1.13.3/lib/log4j-slf4j-impl-2.12.1.jar!/org/slf4j/impl/StaticLoggerBinder.class] ...... ...... ...... 2022-04-19 17:23:37,134 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - The configured TaskManager memory is 2728 MB. YARN will allocate 3072 MB to make up an integer multiple of its minimum allocation memory (1024 MB, configured via 'yarn.scheduler.minimum-allocation-mb'). The extra 344 MB may not be used by Flink. 2022-04-19 17:23:37,134 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - Cluster specification: ClusterSpecification{masterMemoryMB=2600, taskManagerMemoryMB=2728, slotsPerTaskManager=2} 2022-04-19 17:23:39,039 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - Submitting application master application_1641365150688_0057 2022-04-19 17:23:39,069 INFO org.apache.hadoop.yarn.client.api.impl.YarnClientImpl [] - Submitted application application_1641365150688_0057 2022-04-19 17:23:39,070 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - Waiting for the cluster to be allocated 2022-04-19 17:23:39,071 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - Deploying cluster, current state ACCEPTED 2022-04-19 17:23:44,097 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - YARN application has been deployed successfully. 2022-04-19 17:23:44,159 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - Found Web Interface cs03:38081 of application 'application_1641365150688_0057'. JobManager Web Interface: http://cs03:38081
4.tidb环境建表，插入数据，更新等
mysql> create database gong; mysql> use gong; mysql> CREATE TABLE orders ( -> order_id INT, -> order_date TIMESTAMP(3), -> customer_name varchar(20), -> price DECIMAL(10, 5), -> product_id INT, -> order_status BOOLEAN, -> PRIMARY KEY(order_id) -> ) ; mysql> insert into orders values(2,now(),"gong",13.99,123,1); Query OK, 1 row affected (0.00 sec) mysql> insert into orders values(3,now(),"gong",13.99,123,1); Query OK, 1 row affected (0.00 sec) mysql> insert into orders values(4,now(),"gong",13.99,123,1); Query OK, 1 row affected (0.01 sec) ......
...... mysql> insert into orders values(12,now(),"gong",13.99,123,1); Query OK, 1 row affected (0.00 sec)
mysql> SELECT * FROM ORDERS;
+----------+-------------------------+---------------+----------+------------+--------------+
| order_id | order_date | customer_name | price | product_id | order_status |
+----------+-------------------------+---------------+----------+------------+--------------+
| 1 | 2022-04-19 17:12:33.000 | gong | 13.99000 | 123 | 1 |
| 2 | 2022-04-19 17:14:20.000 | gong | 13.99000 | 123 | 1 |
| 3 | 2022-04-19 17:14:25.000 | gong | 13.99000 | 123 | 1 |
| 4 | 2022-04-19 17:14:46.000 | gong | 13.99000 | 123 | 1 |
......
......
| 12 | 2022-04-19 17:20:37.000 | gong | 13.99000 | 123 | 1 |
+----------+-------------------------+---------------+----------+------------+--------------+
mysql> update orders set customer_name = "xxxxxx" where order_id = 12;
Query OK, 1 row affected (0.02 sec)
Rows matched: 1 Changed: 1 Warnings: 0
5.测试tidb数据到flink sql
开启flink sql：
cd $FLINK_HOME/bin ./sql_client.sh embedded
-- checkpoint every 3000 milliseconds Flink SQL> SET 'execution.checkpointing.interval' = '3s'; -- register a TiDB table 'orders' in Flink SQL Flink SQL> CREATE TABLE orders ( order_id INT, order_date TIMESTAMP(3), customer_name STRING, price DECIMAL(10, 5), product_id INT, order_status BOOLEAN, PRIMARY KEY(order_id) NOT ENFORCED ) WITH ( 'connector' = 'tidb-cdc', 'tikv.grpc.timeout_in_ms' = '20000', 'pd-addresses' = 'localhost:2379', 'database-name' = 'gong', 'table-name' = 'orders' ); -- read snapshot and binlogs from orders table Flink SQL> SELECT * FROM orders;
check生成的flink任务，确保running状态：
sr@cs03:~/app/flink-1.13.3/bin$./flink list SLF4J: Class path contains multiple SLF4J bindings ...... ...... 2022-04-19 18:16:02,146 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - Found Web Interface cs03:38081 of application 'application_1641365150688_0057'. Waiting for response... ------------------ Running/Restarting Jobs ------------------- 19.04.2022 17:25:27 : 4852d0bdb40aba6ae4041f1e36eadd1c : collect (RUNNING) --------------------------------------------------------------
上图中在flink sql cli中能查出数据说明数据flow已经与flink sql打通，在后面就是导入到StarRocks；
6.对接StarRocks，将数据写入到StarRocks
在StarRocks建表
CREATE TABLE `orders` ( `order_id` int(11) NOT NULL COMMENT "", `order_date` datetime NULL COMMENT "", `customer_name` varchar(20) NULL COMMENT "", `price` decimal64(10, 5) NULL COMMENT "", `product_id` int(11) NULL COMMENT "", `order_status` boolean NULL COMMENT "" ) ENGINE=OLAP PRIMARY KEY(`order_id`) COMMENT "OLAP" DISTRIBUTED BY HASH(`order_id`) BUCKETS 1 PROPERTIES ( "replication_num" = "3", "in_memory" = "false", "storage_format" = "DEFAULT" );
7.使用Flink sql将数据同步到StarRocks
先将flink StarRocks cdc 依赖包 flink-connector-starrocks-1.2.1_flink-1.13_2.12.jar下载好，放到flink lib目录下,重启flink服务；(下载链接 Flink StarRocks connector，请注意1.13版本和1.11/1.12版本使用不同的connector）
在flink sql cli端
CREATE TABLE IF NOT EXISTS `orders_sink` ( order_id INT, order_date TIMESTAMP(3), customer_name varchar(20), price DECIMAL(10, 5), product_id INT, order_status BOOLEAN, PRIMARY KEY(`order_id`) NOT ENFORCED ) with ( 'load-url' = 'cs01:9011', 'sink.buffer-flush.interval-ms' = '15000', 'sink.properties.row_delimiter' = '\x02', 'sink.properties.column_separator' = '\x01', 'connector' = 'starrocks', 'database-name' = 'gong', 'table-name' = 'orders', 'jdbc-url' = 'jdbc:mysql://cs01:9013', 'password' = '', 'username' = 'root' );
Flink SQL> show tables; +-------------+ | table name | +-------------+ | orders | | orders_sink | +-------------+
2 rows in set
将数据插入到StarRocks
Flink SQL> insert into orders_sink select * from orders; [INFO] Submitting SQL update statement to the cluster... 2022-04-19 18:37:14,581 INFO org.apache.hadoop.yarn.client.RMProxy [] - Connecting to ResourceManager at /172.26.194.184:38035 2022-04-19 18:37:14,582 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 2022-04-19 18:37:14,587 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - Found Web Interface cs03:38081 of application 'application_1641365150688_0057'. [INFO] SQL update statement has been successfully submitted to the cluster: Job ID: 33b03fb9d257b57c5243e721c19a6041
查看新flink sql生成的任务：
sr@cs03:~/app/flink-1.13.3/bin$./flink list SLF4J: Class path contains multiple SLF4J bindings. SLF4J: Found binding in [jar:file:/home/disk1/sr/app/flink-1.13.3/lib/log4j-slf4j-impl-2.12.1.jar!/org/slf4j/impl/StaticLoggerBinder.class] ...... ...... 2022-04-19 18:37:46,875 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 2022-04-19 18:37:46,953 INFO org.apache.flink.yarn.YarnClusterDescriptor [] - Found Web Interface cs03:38081 of application 'application_1641365150688_0057'. Waiting for response... ------------------ Running/Restarting Jobs ------------------- 19.04.2022 18:37:14 : 33b03fb9d257b57c5243e721c19a6041 : insert-into_default_catalog.default_database.orders_sink (RUNNING) -------------------------------------------------------------- No scheduled jobs.
在StarRocks验证变更数据是否都已经进去了
mysql> select current_version();
+-------------------+ | current_version() | +-------------------+ | 2.1.0 1864de0 | +-------------------+ 1 row in set (0.02 sec) mysql> select * from orders; +----------+---------------------+---------------+----------+------------+--------------+ | order_id | order_date | customer_name | price | product_id | order_status | +----------+---------------------+---------------+----------+------------+--------------+ | 1 | 2022-04-19 09:12:33 | gong | 13.99000 | 123 | 1 | | 2 | 2022-04-19 09:14:20 | gong | 13.99000 | 123 | 1 | | 3 | 2022-04-19 09:14:25 | gong | 13.99000 | 123 | 1 | | 4 | 2022-04-19 09:14:46 | gong | 13.99000 | 123 | 1 |
......
......
| 12 | 2022-04-19 09:20:37 | xxxxxx | 13.99000 | 123 | 1 | | 13 | 2022-04-19 09:26:56 | gong | 13.99000 | 123 | 1 | | 14 | 2022-04-19 09:33:21 | gong | 13.99000 | 123 | 1 | | 15 | 2022-04-19 10:08:03 | gong | 13.99000 | 123 | 1 | +----------+---------------------+---------------+----------+------------+--------------+ 15 rows in set (0.01 sec)
8.总结：
整个测试过程还比较顺滑，flink tidb cdc简化整个数据流链路，不依赖kafka，是tidb同步数据到starrocks做后续极速数据分析的一大利器。