设计背景

flink的用户想要将数据sink到StarRocks当中,但是flink官方只提供了flink-connector-jdbc, 不足以满足导入性能要求,为此我们新增了一个flink-connector-starrocks,内部实现是通过缓存并批量由stream load导入。

使用方式

点击下载插件

源码地址

com.starrocks.table.connector.flink.StarRocksDynamicTableSinkFactory加入到:src/main/resources/META-INF/services/org.apache.flink.table.factories.Factory

将以下内容加入pom.xml:

<dependency>
    <groupId>com.starrocks</groupId>
    <artifactId>flink-connector-starrocks</artifactId>
    <!-- for flink-1.11, flink-1.12 -->
    <version>1.1.1_flink-1.11</version>
    <!-- for flink-1.13 -->
    <version>1.1.1_flink-1.13</version>
</dependency>

使用方式如下:

// -------- sink with raw json string stream --------
fromElements(new String[]{
    "{\"score\": \"99\", \"name\": \"stephen\"}",
    "{\"score\": \"100\", \"name\": \"lebron\"}"
}).addSink(
    StarRocksSink.sink(
        // the sink options
        StarRocksSinkOptions.builder()
            .withProperty("jdbc-url", "jdbc:mysql://fe1_ip:query_port,fe2_ip:query_port,fe3_ip:query_port?xxxxx")
            .withProperty("load-url", "fe1_ip:http_port;fe2_ip:http_port;fe3_ip:http_port")
            .withProperty("username", "xxx")
            .withProperty("password", "xxx")
            .withProperty("table-name", "xxx")
            .withProperty("database-name", "xxx")
            .withProperty("sink.properties.format", "json")
            .withProperty("sink.properties.strip_outer_array", "true")
            .build()
    )
);


// -------- sink with stream transformation --------
class RowData {
    public int score;
    public String name;
    public RowData(int score, String name) {
        ......
    }
}
fromElements(
    new RowData[]{
        new RowData(99, "stephen"),
        new RowData(100, "lebron")
    }
).addSink(
    StarRocksSink.sink(
        // the table structure
        TableSchema.builder()
            .field("score", DataTypes.INT())
            .field("name", DataTypes.VARCHAR(20))
            .build(),
        // the sink options
        StarRocksSinkOptions.builder()
            .withProperty("jdbc-url", "jdbc:mysql://fe1_ip:query_port,fe2_ip:query_port,fe3_ip:query_port?xxxxx")
            .withProperty("load-url", "fe1_ip:http_port;fe2_ip:http_port;fe3_ip:http_port")
            .withProperty("username", "xxx")
            .withProperty("password", "xxx")
            .withProperty("table-name", "xxx")
            .withProperty("database-name", "xxx")
            .withProperty("sink.properties.column_separator", "\\x01")
            .withProperty("sink.properties.row_delimiter", "\\x02")
            .build(),
        // set the slots with streamRowData
        (slots, streamRowData) -> {
            slots[0] = streamRowData.score;
            slots[1] = streamRowData.name;
        }
    )
);

或者:

// create a table with `structure` and `properties`
// Needed: Add `com.starrocks.connector.flink.table.StarRocksDynamicTableSinkFactory` to: `src/main/resources/META-INF/services/org.apache.flink.table.factories.Factory`
tEnv.executeSql(
    "CREATE TABLE USER_RESULT(" +
        "name VARCHAR," +
        "score BIGINT" +
    ") WITH ( " +
        "'connector' = 'starrocks'," +
        "'jdbc-url'='jdbc:mysql://fe1_ip:query_port,fe2_ip:query_port,fe3_ip:query_port?xxxxx'," +
        "'load-url'='fe1_ip:http_port;fe2_ip:http_port;fe3_ip:http_port'," +
        "'database-name' = 'xxx'," +
        "'table-name' = 'xxx'," +
        "'username' = 'xxx'," +
        "'password' = 'xxx'," +
        "'sink.buffer-flush.max-rows' = '1000000'," +
        "'sink.buffer-flush.max-bytes' = '300000000'," +
        "'sink.buffer-flush.interval-ms' = '5000'," +
        "'sink.properties.column_separator' = '\\x01'," +
        "'sink.properties.row_delimiter' = '\\x02'," +
        "'sink.max-retries' = '3'" +
        "'sink.properties.*' = 'xxx'" + // stream load properties like `'sink.properties.columns' = 'k1, v1'`
    ")"
);

其中Sink选项如下:

OptionRequiredDefaultTypeDescription
connectorYESNONEStringstarrocks
jdbc-urlYESNONEStringthis will be used to execute queries in starrocks.
load-urlYESNONEStringfe_ip:http_port;fe_ip:http_port separated with ';', which would be used to do the batch sinking.
database-nameYESNONEStringstarrocks database name
table-nameYESNONEStringstarrocks table name
usernameYESNONEStringstarrocks connecting username
passwordYESNONEStringstarrocks connecting password
sink.semanticNOat-least-onceStringat-least-once or exactly-once(flush at checkpoint only and options like sink.buffer-flush.* won't work either).
sink.buffer-flush.max-bytesNO94371840(90M)Stringthe max batching size of the serialized data, range: [64MB, 10GB].
sink.buffer-flush.max-rowsNO500000Stringthe max batching rows, range: [64,000, 5000,000].
sink.buffer-flush.interval-msNO300000Stringthe flushing time interval, range: [1000ms, 3600000ms].
sink.max-retriesNO1Stringmax retry times of the stream load request, range: [0, 10].
sink.connect.timeout-msNO1000StringTimeout in millisecond for connecting to the load-url, range: [100, 60000].
sink.properties.*NONONEStringthe stream load properties like 'sink.properties.columns' = 'k1, k2, k3'.

注意事项

  • 支持exactly-once的数据sink保证,需要外部系统的 two phase commit 机制。由于 StarRocks 无此机制,我们需要依赖flink的checkpoint-interval在每次checkpoint时保存批数据以及其label,在checkpoint完成后的第一次invoke中阻塞flush所有缓存在state当中的数据,以此达到精准一次。但如果StarRocks挂掉了,会导致用户的flink sink stream 算子长时间阻塞,并引起flink的监控报警或强制kill。

  • 默认使用csv格式进行导入,用户可以通过指定'sink.properties.row_delimiter' = '\\x02'(此参数自 StarRocks-1.15.0 开始支持)与'sink.properties.column_separator' = '\\x01'来自定义行分隔符与列分隔符。

  • 如果遇到导入停止的 情况,请尝试增加flink任务的内存。

  • 如果代码运行正常且能接收到数据,但是写入不成功时请确认当前机器能访问BE的http_port端口,这里指能ping通集群show backends显示的ip:port。举个例子:如果一台机器有外网和内网ip,且FE/BE的http_port均可通过外网ip:port访问,集群里绑定的ip为内网ip,任务里loadurl写的FE外网ip:http_port,FE会将写入任务转发给BE内网ip:port,这时如果Client机器ping不通BE的内网ip就会写入失败。

基本原理

通过Flink-cdc和StarRocks-migrate-tools(简称smt)可以实现MySQL数据的秒级同步。

MySQL同步

如图所示,Smt可以根据MySQL和StarRocks的集群信息和表结构自动生成source table和sink table的建表语句。
通过Flink-cdc-connector消费MySQL的binlog,然后通过Flink-connector-starrocks写入StarRocks。

使用说明

  1. 下载 Flink, 推荐使用1.13,最低支持版本1.11。

  2. 下载 Flink CDC connector,请注意下载对应Flink版本的Flink-MySQL-CDC。

  3. 下载 Flink StarRocks connector,请注意1.13版本和1.11/1.12版本使用不同的connector.

  4. 解压 flink-sql-connector-mysql-cdc-xxx.jar, flink-connector-starrocks-xxx.jarflink-xxx/lib/

  5. 下载 smt.tar.gz

  6. 解压并修改配置文件 Db 需要修改成MySQL的连接信息。
    be_num 需要配置成StarRocks集群的节点数(这个能帮助更合理的设置bucket数量)。
    [table-rule.1] 是匹配规则,可以根据正则表达式匹配数据库和表名生成建表的SQL,也可以配置多个规则。
    flink.starrocks.* 是StarRocks的集群配置信息,参考Flink.

    [db]
    host = 192.168.1.1
    port = 3306
    user = root
    password =  
    
    [other]
    # number of backends in StarRocks
    be_num = 3
    # `decimal_v3` is supported since StarRocks-1.18.1
    use_decimal_v3 = false
    # file to save the converted DDL SQL
    output_dir = ./result
[table-rule.1]
# pattern to match databases for setting properties
database = ^console_19321.*$
# pattern to match tables for setting properties
table = ^.*$

############################################
### flink sink configurations
### DO NOT set `connector`, `table-name`, `database-name`, they are auto-generated
############################################
flink.starrocks.jdbc-url=jdbc:mysql://192.168.1.1:9030
flink.starrocks.load-url= 192.168.1.1:8030
flink.starrocks.username=root
flink.starrocks.password=
flink.starrocks.sink.properties.column_separator=\x01
flink.starrocks.sink.properties.row_delimiter=\x02
flink.starrocks.sink.buffer-flush.interval-ms=15000
```
  1. 执行starrocks-migrate-tool,所有建表语句都生成在result目录下

    $./starrocks-migrate-tool
    $ls result
    flink-create.1.sql    smt.tar.gz              starrocks-create.all.sql
    flink-create.all.sql  starrocks-create.1.sql
  2. 生成StarRocks的表结构

    Mysql -hxx.xx.xx.x -P9030 -uroot -p < starrocks-create.1.sql
  3. 生成Flink table并开始同步

    bin/sql-client.sh -f flink-create.1.sql

    这个执行以后同步任务会持续执行

    如果是Flink 1.13之前的版本可能无法直接执行脚本,需要逐行提交 注意 记得打开MySQL binlog

  4. 观察任务状况

    bin/flink list 

    如果有任务请查看log日志,或者调整conf中的系统配置中内存和slot。

注意事项

  1. 如果有多组规则,需要给每一组规则匹配database,table和 flink-connector的配置

    [table-rule.1]
    # pattern to match databases for setting properties
    database = ^console_19321.*$
    # pattern to match tables for setting properties
    table = ^.*$
    
    ############################################
    ### flink sink configurations
    ### DO NOT set `connector`, `table-name`, `database-name`, they are auto-generated
    ############################################
    flink.starrocks.jdbc-url=jdbc:mysql://192.168.1.1:9030
    flink.starrocks.load-url= 192.168.1.1:8030
    flink.starrocks.username=root
    flink.starrocks.password=
    flink.starrocks.sink.properties.column_separator=\x01
    flink.starrocks.sink.properties.row_delimiter=\x02
    flink.starrocks.sink.buffer-flush.interval-ms=15000
    
    [table-rule.2]
    # pattern to match databases for setting properties
    database = ^database2.*$
    # pattern to match tables for setting properties
    table = ^.*$
    
    ############################################
    ### flink sink configurations
    ### DO NOT set `connector`, `table-name`, `database-name`, they are auto-generated
    ############################################
    flink.starrocks.jdbc-url=jdbc:mysql://192.168.1.1:9030
    flink.starrocks.load-url= 192.168.1.1:8030
    flink.starrocks.username=root
    flink.starrocks.password=
    # 如果导入数据不方便选出合适的分隔符可以考虑使用Json格式,但是会有一定的性能损失
    flink.starrocks.sink.properties.format=json
    flink.starrocks.sink.buffer-flush.interval-ms=5000
    ~~~
    
  2. Flink.starrocks.sink 的参数可以参考上文,比如可以给不同的规则配置不同的导入频率等参数。

  3. 针对分库分表的大表可以单独配置一个规则,比如:有两个数据库 edu_db_1,edu_db_2,每个数据库下面分别有course_1,course_2 两张表,并且所有表的数据结构都是相同的,通过如下配置把他们导入StarRocks的一张表中进行分析。

    [table-rule.3]
    # pattern to match databases for setting properties
    database = ^edu_db_[0-9]*$
    # pattern to match tables for setting properties
    table = ^course_[0-9]*$
    
    ############################################
    ### flink sink configurations
    ### DO NOT set `connector`, `table-name`, `database-name`, they are auto-generated
    ############################################
    flink.starrocks.jdbc-url=jdbc:mysql://192.168.1.1:9030
    flink.starrocks.load-url= 192.168.1.1:8030
    flink.starrocks.username=root
    flink.starrocks.password=
    flink.starrocks.sink.properties.column_separator=\x01
    flink.starrocks.sink.properties.row_delimiter=\x02
    flink.starrocks.sink.buffer-flush.interval-ms=5000

    这样会自动生成一个多对一的导入关系,在StarRocks默认生成的表名是 course__auto_shard,也可以自行在生成的配置文件中修改。

  4. 如果在sql-client中命令行执行建表和同步任务,需要做对'\'字符进行转义

    'sink.properties.column_separator' = '\\x01'
    'sink.properties.row_delimiter' = '\\x02'  
  5. 如何开启MySQL binlog
    修改/etc/my.cnf

    #开启binlog日志
    log-bin=/var/lib/mysql/mysql-bin
    
    #log_bin=ON
    ##binlog日志的基本文件名
    #log_bin_basename=/var/lib/mysql/mysql-bin
    ##binlog文件的索引文件,管理所有binlog文件
    #log_bin_index=/var/lib/mysql/mysql-bin.index
    #配置serverid
    server-id=1
    binlog_format = row

    重启mysqld,然后可以通过 SHOW VARIABLES LIKE 'log_bin'; 确认是否已经打开。