Flink学习笔记——读写HBase

1.如果是csa(Cloudera Streaming Analytics)版本的高版本HBase
2.如果是低版本的HBase
- 1.一种是使用Table的put API
- 2.另一种是使用BufferedMutator

1.如果是csa(Cloudera Streaming Analytics)版本的高版本HBase

可以参考Cloudera官方例子，通过引入官方提供的flink-hbase来实现

<dependency>
    <groupId>org.apache.flink</groupId>
    <artifactId>flink-hbase_2.11</artifactId>
    <version>1.9.0-csa1.0.0.0</version>
</dependency>

要求flink最低版本1.9.0，hbase最低版本2.1.0-cdh6.3.0，然后就可以使用HBaseSinkFunction来写Hbase

1	`https://docs.cloudera.com/csa/1.2.0/datastream-connectors/topics/csa-hbase-configuration.html`

回到顶部

2.如果是低版本的HBase

则需要自行使用Java API来写HBase，比如我使用hbase版本为1.2.0-cdh5.1.6.2，可以参考：HBase读写的几种方式（三）flink篇和 Flink 消费kafka数据写入hbase

写HBase的时候注意开启Hbase regionserver的hbase.regionserver.port端口防火墙，默认为60020

Java API写Hbase的方式主要有2种：

1.一种是使用Table的put API

可以参考

1	`https://github.com/phillip2019/flink-parent/blob/master/flink-connectors/flink-connector-hbase/src/main/java/com/aikosolar/bigdata/flink/connectors/hbase/SimpleHBaseTableSink.java`

使用Flink来写Hbase的时候，需要继承RichSinkFunction，然后重写invoke方法，在invoke方法中调用HBase的put API

需要注意的是调用HBase Table的put API较为低效，即使是使用了List<Put>，也会消耗较多的资源，建议使用第二种的BufferedMutator

import org.apache.commons.io.IOUtils;
import org.apache.commons.lang3.StringUtils;
import org.apache.flink.configuration.Configuration;
import org.apache.flink.streaming.api.functions.sink.RichSinkFunction;
import org.apache.flink.util.Preconditions;
import org.apache.hadoop.hbase.HBaseConfiguration;
import org.apache.hadoop.hbase.TableName;
import org.apache.hadoop.hbase.client.*;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.kafka.clients.consumer.ConsumerRecord;
 
public class SimpleHBaseTableSink extends RichSinkFunction<ConsumerRecord<String, String>> {
 
    private final String tableName;
    private final HBaseWriterConfig writerConfig;
 
    private transient TableName tName;
    private transient Connection connection;
 
    public SimpleHBaseTableSink(HBaseWriterConfig writerConfig, String tableName) throws Exception {
        Preconditions.checkNotNull(writerConfig);
        Preconditions.checkNotNull(tableName);
 
        this.writerConfig = writerConfig;
        this.tableName = tableName;
    }
 
    @Override
    public void open(Configuration parameters) throws Exception {
        this.tName = TableName.valueOf(tableName);
 
        org.apache.hadoop.conf.Configuration conf = HBaseConfiguration.create();
        this.writerConfig.getHbaseConfig().forEach(conf::set);
        this.connection = ConnectionFactory.createConnection(conf);
    }
 
    @Override
    public void invoke(ConsumerRecord<String, String> record, Context context) throws Exception {
        Table table = null;
        try {
            table = connection.getTable(this.tName);
            String rowKey = String.valueOf(record.offset());
            String value = record.value();
            Put put = new Put(Bytes.toBytes(rowKey));
            put.addColumn(Bytes.toBytes("test_family"), Bytes.toBytes("test_col"), Bytes.toBytes(value));
 
            if (StringUtils.isNotBlank(writerConfig.getDurability())) {
                try {
                    Durability d = Durability.valueOf(writerConfig.getDurability());
                    put.setDurability(d);
                } catch (Exception e) {
                    e.printStackTrace();
                }
            }
            table.put(put);
        } finally {
            IOUtils.closeQuietly(table);
        }
    }
 
    @Override
    public void close() throws Exception {
        if (this.connection != null) {
            this.connection.close();
            this.connection = null;
        }
    }
}

HBaseWriterConfig

import lombok.AccessLevel;
import lombok.AllArgsConstructor;
import lombok.Getter;
import lombok.NoArgsConstructor;
import org.apache.commons.collections.MapUtils;
 
import java.io.Serializable;
import java.util.HashMap;
import java.util.Map;
 
@Getter
@AllArgsConstructor(access = AccessLevel.PRIVATE)
@NoArgsConstructor(access = AccessLevel.PRIVATE)
public class HBaseWriterConfig implements Serializable {
    private int writeBufferSize = -1;
    private int asyncFlushSize = -1;
    private long asyncFlushInterval = -1;
    private boolean async;
    private String durability;
    private Map<String, String> hbaseConfig = new HashMap<>();
 
    public static class Builder {
        private int writeBufferSize = 5 * 1024 * 1024;
        private int asyncFlushSize = 5000;
        // 单位:秒
        private long asyncFlushInterval = 60;
        private boolean async;
        private String durability;
        private Map<String, String> config = new HashMap<>();
 
        public static synchronized Builder me() {
            return new Builder();
        }
 
        public Builder setDurability(String durability) {
            this.durability = durability;
            return this;
        }
 
        public Builder writeBufferSize(int writeBufferSize) {
            this.writeBufferSize = writeBufferSize;
            return this;
        }
 
        public Builder conf(String key, String value) {
            this.config.put(key, value);
            return this;
        }
 
        public Builder conf(Map<String, String> config) {
            if (MapUtils.isNotEmpty(config)) {
                for (Map.Entry<String, String> e : config.entrySet()) {
                    this.config.put(e.getKey(), e.getValue());
                }
            }
            return this;
        }
 
        public Builder aync() {
            this.async = true;
            return this;
        }
 
        public Builder flushInterval(long interval) {
            this.asyncFlushInterval = interval;
            return this;
        }
 
        public Builder flushSize(int size) {
            this.asyncFlushSize = size;
            return this;
        }
 
        public Builder async(boolean enable) {
            this.async = enable;
            return this;
        }
 
        public HBaseWriterConfig build() {
            return new HBaseWriterConfig(writeBufferSize, asyncFlushSize, asyncFlushInterval, async, durability, config);
        }
 
    }
}

addSink

dataStream.addSink(new SimpleHBaseTableSink(HBaseWriterConfig.Builder.me()
        .async(true)
        .build(),
        "test_table")
);

2.另一种是使用BufferedMutator

可以参考

1	`https://github.com/phillip2019/flink-parent/blob/master/flink-connectors/flink-connector-hbase/src/main/java/com/aikosolar/bigdata/flink/connectors/hbase/writter/HBaseWriterWithBuffer.java`

使用BufferedMutator进行异步批量插入的方式效率更高

import org.apache.flink.configuration.Configuration;
import org.apache.flink.streaming.api.functions.sink.RichSinkFunction;
import org.apache.flink.util.Preconditions;
import org.apache.hadoop.hbase.HBaseConfiguration;
import org.apache.hadoop.hbase.TableName;
import org.apache.hadoop.hbase.client.*;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.kafka.clients.consumer.ConsumerRecord;
 
public class SimpleHBaseTableSink extends RichSinkFunction<ConsumerRecord<String, String>> {
 
    private final String tableName;
    private final HBaseWriterConfig writerConfig;
 
    private transient TableName tName;
    private transient Connection connection;
 
    public SimpleHBaseTableSink(HBaseWriterConfig writerConfig, String tableName) throws Exception {
        Preconditions.checkNotNull(writerConfig);
        Preconditions.checkNotNull(tableName);
 
        this.writerConfig = writerConfig;
        this.tableName = tableName;
    }
 
    @Override
    public void open(Configuration parameters) throws Exception {
        this.tName = TableName.valueOf(tableName);
 
        org.apache.hadoop.conf.Configuration conf = HBaseConfiguration.create();
        this.writerConfig.getHbaseConfig().forEach(conf::set);
        this.connection = ConnectionFactory.createConnection(conf);
    }
 
    @Override
    public void invoke(ConsumerRecord<String, String> record, Context context) throws Exception {
        String rowKey = String.valueOf(record.offset());
        String value = record.value();
 
        // 设置缓存1m，当达到1m时数据会自动刷到hbase，替代了hTable.setWriteBufferSize(30 * 1024 * 1024)
        BufferedMutatorParams params = new BufferedMutatorParams(this.tName);
        params.writeBufferSize(1024 * 1024);
        // 创建一个批量异步与hbase通信的对象
        BufferedMutator mutator = connection.getBufferedMutator(params);
        Put put = new Put(Bytes.toBytes(rowKey));
        put.addColumn(Bytes.toBytes("test_family"), Bytes.toBytes("test_col"), Bytes.toBytes(value));
        // 向hbase插入数据,达到缓存会自动提交,这里也可以通过传入List<put>的方式批量插入
        mutator.mutate(put);
        // 不用每次put后就调用flush，最后调用就行，这个方法替代了旧api的hTable.setAutoFlush(false, false)
        mutator.flush();
        mutator.close();
    }
 
    @Override
    public void close() throws Exception {
        if (this.connection != null) {
            this.connection.close();
            this.connection = null;
        }
    }
}

一些Hbase sink需要的constants

1	`https://github.com/apache/hbase/blob/master/hbase-common/src/main/java/org/apache/hadoop/hbase/HConstants.java`

需要引入

<dependency>
    <groupId>org.apache.hbase</groupId>
    <artifactId>hbase-common</artifactId>
    <version>1.2.0-cdh5.15.1</version>
</dependency>

参考flume的sink实现

flume在1.8.0支持hbase的1.x版本，使用

1	`https://github.com/apache/flume/blob/trunk/flume-ng-sinks/flume-ng-hbase-sink/src/main/java/org/apache/flume/sink/hbase/HBaseSinkConfigurationConstants.java`

在1.9.0支持了habse的2.x版本，使用

1	`https://github.com/apache/flume/blob/trunk/flume-ng-sinks/flume-ng-hbase2-sink/src/main/java/org/apache/flume/sink/hbase2/HBase2SinkConfigurationConstants.java`