第4.1.3章 flume写入数据到hbase中

骑猪看日落 2021-12-14 06:09 592阅读 0赞

**1、环境准备**  
**1.1 flume组件开发环境**  
flume组件依赖的jar如下：

<dependency>
    			<groupId>commons-lang</groupId>
    			<artifactId>commons-lang</artifactId>
    			<version>2.5</version>
    		</dependency>
    
    		
    		<dependency>
                <groupId>org.apache.hbase</groupId>
                <artifactId>hbase-client</artifactId>
                <exclusions>
                    <exclusion>
                        <groupId>io.netty</groupId>
                        <artifactId>netty</artifactId>
                    </exclusion>
                </exclusions>
            </dependency>
    		
    		<dependency>
    		    <groupId>org.apache.flume.flume-ng-sinks</groupId>
    		    <artifactId>flume-ng-hbase-sink</artifactId>
    		    <version>${version.flume}</version>
    		</dependency>
    		
    		<dependency>
    			<groupId>jdk.tools</groupId>
    			<artifactId>jdk.tools</artifactId>
    			<version>${version.java}</version>
    			<scope>system</scope>
    			<systemPath>${JAVA_HOME}/lib/tools.jar</systemPath>
    		</dependency>
    
            
            <dependency>
                <groupId>org.slf4j</groupId>
                <artifactId>slf4j-api</artifactId>
            </dependency>

**1.2 自定义hbase序列化**  
下面定义了接口常量，虽然有些反模式，但较为容易理解。注意：java通过thrift框架生成文件，需要去除\_\_isset\_bitfield字段

import com.google.gson.Gson;
    import com.google.gson.GsonBuilder;
    
    public interface DcmFlumeBaseConstant { 
    
    	String SEPARATE = "$";
    	
    	Gson GSON = new GsonBuilder().create();
    
    	Charset UTF_8 = Charset.forName("UTF-8");
    	Charset ISO_8859_1 = Charset.forName("ISO-8859-1");
    	
    	String __isset_bitfield = "__isset_bitfield";
    
    	interface Fields { 
    		String ID = "id";
    	}
    }

《flume构建高可用、可扩展的海量日志采集系统》中描述Flume和HBase交互环境中，只关心Put和Increment，也就是下面的`PutRequest`和`AtomicIncrementRequest`  
Flume有两类HBase Sink，一类是Hbase sink和Async Hbase sink，Hbase sink采取的是逐个向hbase集群发送事件，而Async Hbase sink则采用非阻塞的且使用多线程写数据到Hbase。显然，两种方式Async Hbase sink的安全性较Hbase sink低，而性能要较Hbase sink高。

import org.apache.commons.lang.StringUtils;
    import org.apache.flume.Context;
    import org.apache.flume.Event;
    import org.apache.flume.conf.ComponentConfiguration;
    import org.apache.flume.sink.hbase.AsyncHbaseEventSerializer;
    import org.apache.hadoop.hbase.util.Bytes;
    import org.hbase.async.AtomicIncrementRequest;
    import org.hbase.async.PutRequest;
    import org.slf4j.Logger;
    import org.slf4j.LoggerFactory;
    
    import java.util.ArrayList;
    import java.util.HashMap;
    import java.util.List;
    import java.util.Map;
    
    /** * json数据hbase序列化，row_key默认使用ID * * @author dzm */
    public class HbaseSerializer implements AsyncHbaseEventSerializer { 
    
        private static final Logger logger = LoggerFactory.getLogger(HbaseSerializer.class);
        public static final ArrayList<PutRequest> EMPTY_ARRAYLIST = new ArrayList<>();
    	// hbase中的表
        private byte[] table;
        // 列簇
        private byte[] cf;
        private Event currentEvent;
    
        @Override
        public void initialize(byte[] table, byte[] cf) { 
            this.table = table;
            this.cf = cf;
        }
    
        @Override
        public void configure(Context context) { 
    
        }
    
        @Override
        public void configure(ComponentConfiguration conf) { 
    
        }
    
        @SuppressWarnings("unchecked")
        @Override
        public List<PutRequest> getActions() { 
            String eventStr = new String(currentEvent.getBody(), DcmFlumeBaseConstant.UTF_8);
            if (logger.isDebugEnabled()) { 
                logger.debug("event body: " + eventStr);
            }
            if (StringUtils.isNotEmpty(eventStr)) { 
                // 这里看到数据取自eventStr
                Map<String, Object> dataMap = DcmFlumeBaseConstant.GSON.fromJson(eventStr, HashMap.class);
                try { 
                    filter(dataMap);
                } catch (Exception e) { 
                    return EMPTY_ARRAYLIST;
                }
                if (dataMap == null || dataMap.size() == 0) { 
                    return EMPTY_ARRAYLIST;
                }
                List<PutRequest> rows = new ArrayList<PutRequest>();
                // hbase中写入行记录
                List<byte[]> qualifiers = new ArrayList<byte[]>();
                List<byte[]> values = new ArrayList<byte[]>();
                for (Map.Entry<String, Object> entry : dataMap.entrySet()) { 
                    // __isset_bitfield为thrift字段，不需要存储
                    if (DcmFlumeBaseConstant.__isset_bitfield.equals(entry.getKey())) { 
                        continue;
                    }
    
                    if (entry.getValue() != null) { 
                        // 为空的数据，不用写入到hbase中
                        if (entry.getValue() instanceof java.util.List) { 
                            // list类型在detail的sink中处理
                        } else if (StringUtils.isNotEmpty(entry.getValue().toString())) { 
                            qualifiers.add(Bytes.toBytes(entry.getKey()));
                            values.add(Bytes.toBytes(entry.getValue().toString()));
                        }
                    }
                }
                byte[][] qualifiersB = new byte[qualifiers.size()][];
                qualifiers.toArray(qualifiersB);
                byte[][] valuesB = new byte[values.size()][];
                values.toArray(valuesB);
                // getRowKey(dataMap)获取rowkey
                PutRequest put = new PutRequest(table, getRowKey(dataMap), cf, qualifiersB, valuesB);
                rows.add(put);
                return rows;
            }
            return EMPTY_ARRAYLIST;
        }
    
        /** * Hbase的 row key=id * * @param dataMap * @return */
        protected byte[] getRowKey(Map<String, Object> dataMap) { 
            String rowKey = dataMap.get(DcmFlumeBaseConstant.Fields.ID).toString();
            dataMap.remove("DcmFlumeBaseConstant.Fields.ID");
            return Bytes.toBytes(rowKey);
        }
    
        @Override
        public void setEvent(Event event) { 
            this.currentEvent = event;
        }
    
        @Override
        public void cleanUp() { 
            table = null;
            cf = null;
            currentEvent = null;
        }
    
        /** * hbase计数器，不需要 */
        @Override
        public List<AtomicIncrementRequest> getIncrements() { 
            return new ArrayList<AtomicIncrementRequest>();
        }
    
        /** * 是否需要保存 * * @param dataMap */
        public void filter(Map<String, Object> dataMap) { 
    
        }

如果想更改一下rowkey，设为自己的

private static String DATE_FORMAT = "yyyyMMddHHmmss";
    
    	/** * 对传入的数据进行过滤 * * @param dataMap */
    	@Override
    	public void filter(Map<String, Object> dataMap) { 
    		if (!dataMap.containsKey(("id"))) { 
                dataMap = null;
    		}
            String userId = dataMap.get("id").toString();
            String eventType = dataMap.get("eventType").toString();
            String time = DateUtil.formatDate(DateUtil.parseDate(dataMap.get("time")), DATE_FORMAT);
    		dataMap.put("userId", userId);
                dataMap.put("id", userId + DcmFlumeBaseConstant.SEPARATE + eventType + DcmFlumeBaseConstant.SEPARATE + time);
    	}

**1.3 hadoop、hbase环境变量**  
flume写入hadoop、hbase需要一些jar，因为jar太多，最简单的办法是通过环境变量来设置，  
hadoop、hbase自身是不用启动的

# flume
    export FLUME_HOME=/application/flume
    export PATH=$PATH:$FLUME_HOME/bin
    # hadoop hbase
    export HADOOP_HOME=/application/hadoop
    export PATH=$PATH:$HADOOP_HOME/bin:$HADOOP_HOME/sbin
    
    export HBASE_HOME=/application/hbase
    export PATH=$PATH:$HBASE_HOME/bin
    export HBASE_LIBRARY_PATH=$HBASE_HOME/lib/native/Linux-amd64-64

![3][]  
**2 接口实现**  
**2.1 flume配置文件**  
参考[Flume1.7.0入门：安装、部署、及flume的案例][Flume1.7.0_flume]，`Source`是采集数据,`sink`是从Channel收集数据，运行在一个独立线程,sink组件是用于把数据发送到目的地的组件，Channel:连接 sources 和 sinks ，这个有点像一个队列,source组件把数据收集来以后，临时存放在channel中，即channel组件在agent中是专门用来存放临时数据的——对采集到的数据进行简单的缓存，可以存放在memory、jdbc、file等等。  
下面的配置信息，可以解读为：`35001`端口采集avro对象，采集后的数据先保存在kafka中，接着由sink来处理，序列化采用我自定义的`HbaseSerializer`,将数据写入到hbase中。  
那么`assemble`是做什么的呢？它用于定义`source`和`sink`的绑定关系，两者之间通过`channel`进行关联

# 上传flume-xx-conf.properties到$FLUME_HOME/conf目录下
    [root@bwsc73 conf]# cat dcm-order-flume-conf.properties 
    # read from kafka and write to hbase
    dcm-order-agent.sources = dcm-order-source
    dcm-order-agent.channels = dcm-order-channel
    dcm-order-agent.sinks = dcm-order-sink
    
    # source
    dcm-order-agent.sources.dcm-order-source.type=avro
    dcm-order-agent.sources.dcm-order-source.bind=0.0.0.0
    dcm-order-agent.sources.dcm-order-source.port=35001
    
    # channel
    dcm-order-agent.channels.dcm-order-channel.flumeBatchSize = 100
    dcm-order-agent.channels.dcm-order-channel.type = org.apache.flume.channel.kafka.KafkaChannel
    dcm-order-agent.channels.dcm-order-channel.kafka.bootstrap.servers = bwsc68:9092,bwsc68:9092,bwsc70:9092
    dcm-order-agent.channels.dcm-order-channel.kafka.topic = dcm_order
    dcm-order-agent.channels.dcm-order-channel.kafka.consumer.group.id = dcm_order_flume_channel
    #dcm-order-agent.channels.dcm-order-channel.kafka.consumer.auto.offset.reset = latest
    
    # sink
    dcm-order-agent.sinks.dcm-order-sink.type = asynchbase
    dcm-order-agent.sinks.dcm-order-sink.table = dcm_order
    dcm-order-agent.sinks.dcm-order-sink.columnFamily = i
    dcm-order-agent.sinks.dcm-order-sink.zookeeperQuorum = bwsc65:2181,bwsc66:2181,bwsc67:2181
    dcm-order-agent.sinks.dcm-order-sink.serializer = com.dzm.dcm.flume.sink.HbaseSerializer
    
    
    # assemble
    dcm-order-agent.sources.dcm-order-source.channels = dcm-order-channel
    dcm-order-agent.sinks.dcm-order-sink.channel = dcm-order-channel

**3 flume组件部署**  
**3.1 文件上传在哪里**

mkdir -p $FLUME_HOME/plugins.d/your_project/lib
    上传到jar文件至$FLUME_HOME/plugins.d/your_project/lib目录

![1][]  
![2][]  
**3.2 启动**

# 在$FLUME_HOME目录输入以下命令启动(注：必须在flume根目录启动)
    nohup bin/flume-ng agent -n your-order-agent -c conf -f conf/your-order-flume-conf.properties > your-order-flume.out 2>&1 &

从下面可以看到那些依赖的jar，停止flume的进程，通过`kill`命令即可

[root@bwsc73 conf]# ps -ef|grep flume-multi-hbase-sink
    root      6122  4792  0 11:53 pts/1    00:00:00 grep --color=auto flume-multi-hbase-sink
    root     20523     1  0 Jun12 ?        01:20:57 /usr/java/jdk1.8.0_151/bin/java -Xmx20m -cp /application/flume/lib/*:/application/flume/plugins.d/bw-msg-flume/lib/*:/application/flume/plugins.d/dcm-flume/lib/*:/application/flume/plugins.d/flume-invoice-bft/lib/*:/application/flume/plugins.d/flume-invoice-wp/lib/*:/application/flume/plugins.d/flume-invoice-wp/libext/*:/application/hadoop-2.6.4/etc/hadoop:/application/hadoop-2.6.4/share/hadoop/common/lib/*:/application/hadoop-2.6.4/share/hadoop/common/*:/application/hadoop-2.6.4/share/hadoop/hdfs:/application/hadoop-2.6.4/share/hadoop/hdfs/lib/*:/application/hadoop-2.6.4/share/hadoop/hdfs/*:/application/hadoop-2.6.4/share/hadoop/yarn/lib/*:/application/hadoop-2.6.4/share/hadoop/yarn/*:/application/hadoop-2.6.4/share/hadoop/mapreduce/lib/*:/application/hadoop-2.6.4/share/hadoop/mapreduce/*:/application/hadoop/contrib/capacity-scheduler/*.jar:/application/hbase/conf:/usr/java/jdk/lib/tools.jar:/application/hbase:/application/hbase/lib/activation-1.1.jar:/application/hbase/lib/antisamy-1.4.3.jar:/application/hbase/lib/aopalliance-1.0.jar:/application/hbase/lib/apacheds-i18n-2.0.0-M15.jar:/application/hbase/lib/apacheds-kerberos-codec-2.0.0-M15.jar:/application/hbase/lib/api-asn1-api-1.0.0-M20.jar:/application/hbase/lib/api-util-1.0.0-M20.jar:/application/hbase/lib/asm-3.1.jar:/application/hbase/lib/avro-1.7.4.jar:/application/hbase/lib/batik-css-1.7.jar:/application/hbase/lib/batik-ext-1.7.jar:/application/hbase/lib/batik-util-1.7.jar:/application/hbase/lib/bsh-core-2.0b4.jar:/application/hbase/lib/commons-beanutils-1.7.0.jar:/application/hbase/lib/commons-beanutils-core-1.7.0.jar:/application/hbase/lib/commons-cli-1.2.jar:/application/hbase/lib/commons-codec-1.9.jar:/application/hbase/lib/commons-collections-3.2.2.jar:/application/hbase/lib/commons-compress-1.4.1.jar:/application/hbase/lib/commons-configuration-1.6.jar:/application/hbase/lib/commons-daemon-1.0.13.jar:/application/hbase/lib/commons-digester-1.8.jar:/application/hbase/lib/commons-el-1.0.jar:/application/hbase/lib/commons-fileupload-1.2.jar:/application/hbase/lib/commons-httpclient-3.1.jar:/application/hbase/lib/commons-io-2.4.jar:/application/hbase/lib/commons-lang-2.6.jar:/application/hbase/lib/commons-logging-1.2.jar:/application/hbase/lib/commons-math-2.2.jar:/application/hbase/lib/commons-math3-3.1.1.jar:/application/hbase/lib/commons-net-3.1.jar:/application/hbase/lib/disruptor-3.3.0.jar:/application/hbase/lib/esapi-2.1.0.jar:/application/hbase/lib/findbugs-annotations-1.3.9-1.jar:/application/hbase/lib/guava-12.0.1.jar:/application/hbase/lib/guice-3.0.jar:/application/hbase/lib/guice-servlet-3.0.jar:/application/hbase/lib/hadoop-annotations-2.5.1.jar:/application/hbase/lib/hadoop-auth-2.5.1.jar:/application/hbase/lib/hadoop-client-2.5.1.jar:/application/hbase/lib/hadoop-common-2.5.1.jar:/application/hbase/lib/hadoop-hdfs-2.5.1.jar:/application/hbase/lib/hadoop-mapreduce-client-app-2.5.1.jar:/application/hbase/lib/hadoop-mapreduce-client-common-2.5.1.jar:/application/hbase/lib/hadoop-mapreduce-client-core-2.5.1.jar:/application/hbase/lib/hadoop-mapreduce-client-jobclient-2.5.1.jar:/application/hbase/lib/hadoop-mapreduce-client-shuffle-2.5.1.jar:/application/hbase/lib/hadoop-yarn-api-2.5.1.jar:/application/hbase/lib/hadoop-yarn-client-2.5.1.jar:/application/hbase/lib/hadoop-yarn-common-2.5.1.jar:/application/hbase/lib/hadoop-yarn-server-common-2.5.1.jar:/application/hbase/lib/hbase-annotations-1.1.4.jar:/application/hbase/lib/hbase-annotations-1.1.4-tests.jar:/application/hbase/lib/hbase-client-1.1.4.jar:/application/hbase/lib/hbase-common-1.1.4.jar:/application/hbase/lib/hbase-common-1.1.4-tests.jar:/application/hbase/lib/hbase-examples-1.1.4.jar:/application/hbase/lib/hbase-hadoop2-compat-1.1.4.jar:/application/hbase/lib/hbase-hadoop-compat-1.1.4.jar:/application/hbase/lib/hbase-it-1.1.4.jar:/application/hbase/lib/hbase-it-1.1.4-tests.jar:/application/hbase/lib/hbase-prefix-tree-1.1.4.jar:/application/hbase/lib/hbase-procedure-1.1.4.jar:/application/hbase/lib/hbase-protocol-1.1.4.jar:/application/hbase/lib/hbase-resource-bundle-1.1.4.jar:/application/hbase/lib/hbase-rest-1.1.4.jar:/application/hbase/lib/hbase-server-1.1.4.jar:/application/hbase/lib/hbase-server-1.1.4-tests.jar:/application/hbase/lib/hbase-shell-1.1.4.jar:/application/hbase/lib/hbase-thrift-1.1.4.jar:/application/hbase/lib/htrace-core-3.1.0-incubating.jar:/application/hbase/lib/httpclient-4.2.5.jar:/application/hbase/lib/httpcore-4.1.3.jar:/application/hbase/lib/jackson-core-asl-1.9.13.jar:/application/hbase/lib/jackson-jaxrs-1.9.13.jar:/application/hbase/lib/jackson-mapper-asl-1.9.13.jar:/application/hbase/lib/jackson-xc-1.9.13.jar:/application/hbase/lib/jamon-runtime-2.3.1.jar:/application/hbase/lib/jasper-compiler-5.5.23.jar:/application/hbase/lib/jasper-runtime-5.5.23.jar:/application/hbase/lib/javax.inject-1.jar:/application/hbase/lib/java-xmlbuilder-0.4.jar:/application/hbase/lib/jaxb-api-2.2.2.jar:/application/hbase/lib/jaxb-impl-2.2.3-1.jar:/application/hbase/lib/jcodings-1.0.8.jar:/application/hbase/lib/jersey-client-1.9.jar:/application/hbase/lib/jersey-core-1.9.jar:/application/hbase/lib/jersey-guice-1.9.jar:/application/hbase/lib/jersey-json-1.9.jar:/application/hbase/lib/jersey-server-1.9.jar:/application/hbase/lib/jets3t-0.9.0.jar:/application/hbase/lib/jettison-1.3.3.jar:/application/hbase/lib/jetty-6.1.26.jar:/application/hbase/lib/jetty-sslengine-6.1.26.jar:/application/hbase/lib/jetty-util-6.1.26.jar:/application/hbase/lib/joni-2.1.2.jar:/application/hbase/lib/jruby-complete-1.6.8.jar:/application/hbase/lib/jsch-0.1.42.jar:/application/hbase/lib/jsp-2.1-6.1.14.jar:/application/hbase/lib/jsp-api-2.1-6.1.14.jar:/application/hbase/lib/jsr305-1.3.9.jar:/application/hbase/lib/junit-4.12.jar:/application/hbase/lib/leveldbjni-all-1.8.jar:/application/hbase/lib/libthrift-0.9.0.jar:/application/hbase/lib/log4j-1.2.17.jar:/application/hbase/lib/metrics-core-2.2.0.jar:/application/hbase/lib/nekohtml-1.9.12.jar:/application/hbase/lib/netty-3.2.4.Final.jar:/application/hbase/lib/netty-all-4.0.23.Final.jar:/application/hbase/lib/paranamer-2.3.jar:/application/hbase/lib/protobuf-java-2.5.0.jar:/application/hbase/lib/servlet-api-2.5-6.1.14.jar:/application/hbase/lib/servlet-api-2.5.jar:/application/hbase/lib/slf4j-api-1.7.7.jar:/application/hbase/lib/slf4j-log4j12-1.7.5.jar:/application/hbase/lib/snappy-java-1.0.4.1.jar:/application/hbase/lib/spymemcached-2.11.6.jar:/application/hbase/lib/xalan-2.7.0.jar:/application/hbase/lib/xml-apis-1.3.03.jar:/application/hbase/lib/xml-apis-ext-1.3.04.jar:/application/hbase/lib/xmlenc-0.52.jar:/application/hbase/lib/xom-1.2.5.jar:/application/hbase/lib/xz-1.0.jar:/application/hbase/lib/zookeeper-3.4.6.jar:/application/hadoop-2.6.4/etc/hadoop:/application/hadoop-2.6.4/share/hadoop/common/lib/*:/application/hadoop-2.6.4/share/hadoop/common/*:/application/hadoop-2.6.4/share/hadoop/hdfs:/application/hadoop-2.6.4/share/hadoop/hdfs/lib/*:/application/hadoop-2.6.4/share/hadoop/hdfs/*:/application/hadoop-2.6.4/share/hadoop/yarn/lib/*:/application/hadoop-2.6.4/share/hadoop/yarn/*:/application/hadoop-2.6.4/share/hadoop/mapreduce/lib/*:/application/hadoop-2.6.4/share/hadoop/mapreduce/*:/application/hadoop/contrib/capacity-scheduler/*.jar:/application/hbase/conf:/lib/* -Dja

杀掉全部flume进程  
`ps -ef | grep flume | grep -v grep | awk '{print $2}' | xargs kill -9`  
**3.3 调试**

#查看日志
    tail -fn 100 logs/flume.log
    # 如果需要检查报文是否接收成功，修改flume/conf/log4j.properties，加入配置，配置需要debug模式的包名全路径后重新启动
    log4j.lodcgger.com.your_package.your_path= DEBUG

**3.4 haproxy配置**  
`35001`端口在properties文件中  
![1][1 1]

frontend dcm_order_flume_front
        bind *:35101
        mode tcp
        log global
        option tcplog
        timeout client 3600s
        backlog 4096
        maxconn 1000000
        default_backend dcm_order_flume_back
    
    backend dcm_order_flume_back
        mode tcp
        option log-health-checks
        option redispatch
        option tcplog
        balance roundrobin
        timeout connect 1s
        timeout queue 5s
        timeout server 3600s
        balance roundrobin
        server f1 bwhs180:35001 check inter 2000 rise 3 fall 3 weight 1
        server f2 bwhs181:35001 check inter 2000 rise 3 fall 3 weight 1
        server f3 bwhs183:35001 check inter 2000 rise 3 fall 3 weight 1

**3.5 flume进程是否活着在**  
根据端口来确定进程是否还活着，如果挂掉了，那么就重启一下

hdfsAgent=`lsof -i:36001 | awk 'NR==2{print $2}'`
    echo $hdfsAgent
    
    if [ "$hdfsAgent" =  "" ]; then
            echo "hdfs-agent is restart!"
            nohup bin/flume-ng agent -n hdfs-agent -c conf -f conf/hdfs-flume-conf.properties &
    else
            echo "hdfs-agent is alive!"
    fi

另外重启flume，可以参考

hdfsAgent=`ps -ef | grep flume | grep hdfs-agent | awk '{print $2}'`
    echo $hdfsAgent
    
    if [ "$hdfsAgent" =  "" ]; then
            echo "hdfs-agent is restart!"
            nohup bin/flume-ng agent -n hdfs-agent -c conf -f conf/hdfs-flume-conf.properties &
    else
            echo "hdfs-agent is alive!"
    fi

**4 问题集**  
**4.1 java.lang.OutOfMemoryError: Java heap space**  
参考[flume系列之Java heap space大小设置][flume_Java heap space]，在`conf/flume-env.sh`中增加配置

export JAVA_OPTS="-Xms512m -Xmx2000m -Dcom.sun.management.jmxremote"

即使增加了这个配置，还是会发生[kafka相关操作及问题汇总][kafka]，总是会出现`org.apache.kafka.common.errors.NotLeaderForPartitionException`的问题，前期研究[第4.1.2章 flume的拓扑结构][4.1.2_ flume]，就想到了，flume中`source`节点与`sink`节点分开，这样就可以降低flume资源消耗的风险.

23 Jul 2019 13:58:41,136 ERROR [kafka-producer-network-thread | producer-1] (org.apache.kafka.common.utils.KafkaThread$1.uncaughtException:30)  - Uncaught exception in kafka-producer-network-thread | producer-1: 
    java.lang.OutOfMemoryError: Java heap space
            at java.util.HashMap.newNode(HashMap.java:1747)
            at java.util.HashMap.putVal(HashMap.java:631)
            at java.util.HashMap.put(HashMap.java:612)
            at org.apache.kafka.common.Cluster.<init>(Cluster.java:48)
            at org.apache.kafka.common.requests.MetadataResponse.<init>(MetadataResponse.java:176)
            at org.apache.kafka.clients.NetworkClient$DefaultMetadataUpdater.handleResponse(NetworkClient.java:578)
            at org.apache.kafka.clients.NetworkClient$DefaultMetadataUpdater.maybeHandleCompletedReceive(NetworkClient.java:565)
            at org.apache.kafka.clients.NetworkClient.handleCompletedReceives(NetworkClient.java:441)
            at org.apache.kafka.clients.NetworkClient.poll(NetworkClient.java:265)
            at org.apache.kafka.clients.producer.internals.Sender.run(Sender.java:216)
            at org.apache.kafka.clients.producer.internals.Sender.run(Sender.java:128)
            at java.lang.Thread.run(Thread.java:748)
    23 Jul 2019 13:58:41,136 ERROR [Thread-1] (org.apache.thrift.server.TThreadedSelectorServer$SelectorThread.run:544)  - run() exiting due to uncaught error
    java.lang.OutOfMemoryError: Java heap space
            at java.nio.HeapByteBuffer.<init>(HeapByteBuffer.java:57)
            at java.nio.ByteBuffer.allocate(ByteBuffer.java:335)
            at org.apache.thrift.server.AbstractNonblockingServer$FrameBuffer.read(AbstractNonblockingServer.java:338)
            at org.apache.thrift.server.AbstractNonblockingServer$AbstractSelectThread.handleRead(AbstractNonblockingServer.java:202)
            at org.apache.thrift.server.TThreadedSelectorServer$SelectorThread.select(TThreadedSelectorServer.java:576)
            at org.apache.thrift.server.TThreadedSelectorServer$SelectorThread.run(TThreadedSelectorServer.java:536)

**4.2 SinkRunner-PollingRunner-DefaultSinkProcessor**

26 Jul 2019 08:49:38,274 ERROR [SinkRunner-PollingRunner-DefaultSinkProcessor] (org.apache.kafka.clients.consumer.internals.ConsumerCoordinator$OffsetCommitResponseHandler.handle:550)  - Error UNKNOWN_MEMBER_ID occurred while committing offsets for group flume_scrapy_snapshot_channel
    26 Jul 2019 08:49:38,274 ERROR [SinkRunner-PollingRunner-DefaultSinkProcessor] (org.apache.flume.SinkRunner$PollingRunner.run:158)  - Unable to deliver event. Exception follows.
    org.apache.kafka.clients.consumer.CommitFailedException: Commit cannot be completed due to group rebalance
            at org.apache.kafka.clients.consumer.internals.ConsumerCoordinator$OffsetCommitResponseHandler.handle(ConsumerCoordinator.java:552)
            at org.apache.kafka.clients.consumer.internals.ConsumerCoordinator$OffsetCommitResponseHandler.handle(ConsumerCoordinator.java:493)
            at org.apache.kafka.clients.consumer.internals.AbstractCoordinator$CoordinatorResponseHandler.onSuccess(AbstractCoordinator.java:665)
            at org.apache.kafka.clients.consumer.internals.AbstractCoordinator$CoordinatorResponseHandler.onSuccess(AbstractCoordinator.java:644)
            at org.apache.kafka.clients.consumer.internals.RequestFuture$1.onSuccess(RequestFuture.java:167)
            at org.apache.kafka.clients.consumer.internals.RequestFuture.fireSuccess(RequestFuture.java:133)
            at org.apache.kafka.clients.consumer.internals.RequestFuture.complete(RequestFuture.java:107)
            at org.apache.kafka.clients.consumer.internals.ConsumerNetworkClient$RequestFutureCompletionHandler.onComplete(ConsumerNetworkClient.java:380)
            at org.apache.kafka.clients.NetworkClient.poll(NetworkClient.java:274)
            at org.apache.kafka.clients.consumer.internals.ConsumerNetworkClient.clientPoll(ConsumerNetworkClient.java:320)
            at org.apache.kafka.clients.consumer.internals.ConsumerNetworkClient.poll(ConsumerNetworkClient.java:213)
            at org.apache.kafka.clients.consumer.internals.ConsumerNetworkClient.poll(ConsumerNetworkClient.java:193)
            at org.apache.kafka.clients.consumer.internals.ConsumerNetworkClient.poll(ConsumerNetworkClient.java:163)
            at org.apache.kafka.clients.consumer.internals.ConsumerCoordinator.commitOffsetsSync(ConsumerCoordinator.java:358)
            at org.apache.kafka.clients.consumer.KafkaConsumer.commitSync(KafkaConsumer.java:968)
            at org.apache.flume.channel.kafka.KafkaChannel$ConsumerAndRecords.commitOffsets(KafkaChannel.java:684)
            at org.apache.flume.channel.kafka.KafkaChannel$KafkaTransaction.doCommit(KafkaChannel.java:567)
            at org.apache.flume.channel.BasicTransactionSemantics.commit(BasicTransactionSemantics.java:151)
            at com.bwjf.flume.hbase.flume.sink.MultiAsyncHBaseSink.process(MultiAsyncHBaseSink.java:288)
            at org.apache.flume.sink.DefaultSinkProcessor.process(DefaultSinkProcessor.java:67)
            at org.apache.flume.SinkRunner$PollingRunn

![1][1 2]

[3]: /images/20211213/aca00835bb1c431f8d6ab069ac19c296.png
[Flume1.7.0_flume]: https://www.cnblogs.com/netbloomy/p/6666683.html
[1]: /images/20211213/18537968610b425096918c2f344e3268.png
[2]: /images/20211213/6c894eeae070409b9ed7779d88b53311.png
[1 1]: /images/20211213/60c2454458c94a83b613056ac5d05a5a.png
[flume_Java heap space]: https://blog.csdn.net/welcome66/article/details/73225609
[kafka]: https://blog.csdn.net/warrah/article/details/78909798
[4.1.2_ flume]: https://blog.csdn.net/warrah/article/details/78780260
[1 2]: /images/20211213/7dfff3b75d0443aebe4b0a63108f40c8.png