当前位置：首页 > news >正文

Hadoop学习笔记（HDP）-Part.18 安装Flink

news 2026/2/8 3:51:01

目录
Part.01 关于HDP
Part.02 核心组件原理
Part.03 资源规划
Part.04 基础环境配置
Part.05 Yum源配置
Part.06 安装OracleJDK
Part.07 安装MySQL
Part.08 部署Ambari集群
Part.09 安装OpenLDAP
Part.10 创建集群
Part.11 安装Kerberos
Part.12 安装HDFS
Part.13 安装Ranger
Part.14 安装YARN+MR
Part.15 安装HIVE
Part.16 安装HBase
Part.17 安装Spark2
Part.18 安装Flink
Part.19 安装Kafka
Part.20 安装Flume

十八、安装Flink

1.配置Ambari的flink资源

(1)创建flink源

下载链接为
https://repo.huaweicloud.com/apache/flink/flink-1.9.3/flink-1.9.3-bin-scala_2.12.tgz
https://repo.maven.apache.org/maven2/org/apache/flink/flink-shaded-hadoop-2-uber/2.6.5-7.0/flink-shaded-hadoop-2-uber-2.6.5-7.0.jar
http://www.java2s.com/Code/JarDownload/javax.ws/javax.ws.rs-api-2.0.jar.zip
上传到hdp01上，并复制到/var/www/html下

mkdir /var/www/html/flink
cp /opt/flink-1.9.3-bin-scala_2.12.tgz /var/www/html/flink/
cp /opt/flink-shaded-hadoop-2-uber-2.6.5-7.0.jar /var/www/html/flink/
cp /opt/javax.ws.rs-api-2.0.jar /var/www/html/flink/

(2)下载ambari-flink-service服务

在外网服务器上

git clone https://ghproxy.com/https://gitee.com/liujingwen-git/ambari-flink-service-master.git /root/FLINK

注：github.com无法直接下载，可以使用gproxy.com进行代理加速，拼接形成加速URL；或者将github.com在本地强制解析为140.82.114.4
将FLINK文件夹上传到hdp01的/var/lib/ambari-server/resources/stacks/HDP/3.1/services/上，对应的版本可在集群的服务器中查看

hdp-select status hadoop-client | sed 's/hadoop-client - \([0-9]\.[0-9]\).*/\1/'

(3)文件说明

从github上下载的ambari-flink-service文件较多，有些可以删除以及修改
在这里插入图片描述

(4)修改metainfo.xml文件

文件位置：
/var/lib/ambari-server/resources/stacks/HDP/3.1/services/FLINK/metainfo.xml
修改版本

            <name>FLINK</name><displayName>Flink</displayName><comment>Apache Flink is a streaming dataflow engine that provides data distribution, communication, and fault tolerance for distributed computations over data streams.</comment><version>1.9.3</version>

删除FLINK_MASTER，仅保留安装FLINK_CLIENT，在Flink on YARN模式下，master与ResourceManager合设，无需单独安装

            <components><component><name>FLINK_CLIENT</name><displayName>FlinkCLIENT</displayName><category>CLIENT</category><cardinality>1+</cardinality><commandScript><script>scripts/flink_client.py</script><scriptType>PYTHON</scriptType><timeout>10000</timeout></commandScript></component></components>

(5)修改flink-ambari-config.xml文件

文件位置：
/var/lib/ambari-server/resources/stacks/HDP/3.1/services/FLINK/configuration/flink-ambari-config.xml
修改安装路径

  <property><name>flink_install_dir</name><value>/usr/hdp/3.1.5.0-152/flink</value><description>Location to install Flink</description></property>

修改安装包下载地址

  <property><name>flink_download_url</name><value>http://hdp01.hdp.com/flink/flink-1.9.3-bin-scala_2.12.tgz</value><description>Snapshot download location. Downloaded when setup_prebuilt is true</description></property><property><name>flink_hadoop_shaded_jar</name><value>http://hdp01.hdp.com/flink/flink-shaded-hadoop-2-uber-2.6.5-7.0.jar</value><description>Flink shaded hadoop jar download location. Downloaded when setup_prebuilt is true</description></property>
</configuration>

(6)修改flink-env.xml文件

文件位置：
/var/lib/ambari-server/resources/stacks/HDP/3.1/services/FLINK/configuration/flink-env.xml
修改JAVA环境变量

env.java.home: /usr/local/jdk1.8.0_351/jre/

(7)修改flink_client.py文件

文件位置：
/var/lib/ambari-server/resources/stacks/HDP/3.1/services/FLINK/package/scripts/flink_client.py
该文件的作用调用相关的环境变量，来实现整个安装过程
内容如下：

#!/usr/bin/env ptyhonfrom resource_management import *class FlinkClient(Script):def install(self, env):print 'Install the Flink Client'import paramsenv.set_params(params)self.configure(env)# create flink log dirDirectory(params.flink_log_dir,owner=params.flink_user,group=params.flink_group,create_parents=True,mode=0775)Execute(format("rm -rf {flink_install_dir}/log/flink"))Execute(format("ln -s {flink_log_dir} {flink_install_dir}/log/flink"))def configure(self, env):import paramsenv.set_params(params)# write out flink-conf.yamlproperties_content = InlineTemplate(params.flink_yaml_content)File(format(params.flink_install_dir + "/conf/flink-conf.yaml"), content=properties_content)def status(self, env):raise ClientComponentHasNoStatus()if __name__ == "__main__":FlinkClient().execute()

(8)修改parameter.py文件

文件位置：
/var/lib/ambari-server/resources/stacks/HDP/3.1/services/FLINK/package/scripts/params.py
该文件定义了部分环境变量
内容如下：

#!/usr/bin/env python
from resource_management import *
from resource_management.libraries.script.script import Script
import sys, os, glob
from resource_management.libraries.functions.version import format_stack_version
from resource_management.libraries.functions.default import default# server configurations
config = Script.get_config()# params from flink-ambari-config
flink_install_dir = config['configurations']['flink-ambari-config']['flink_install_dir']
flink_numcontainers = config['configurations']['flink-ambari-config']['flink_numcontainers']
flink_numberoftaskslots= config['configurations']['flink-ambari-config']['flink_numberoftaskslots']
flink_jobmanager_memory = config['configurations']['flink-ambari-config']['flink_jobmanager_memory']
flink_container_memory = config['configurations']['flink-ambari-config']['flink_container_memory']
setup_prebuilt = config['configurations']['flink-ambari-config']['setup_prebuilt']
flink_appname = config['configurations']['flink-ambari-config']['flink_appname']
flink_queue = config['configurations']['flink-ambari-config']['flink_queue']
flink_streaming = config['configurations']['flink-ambari-config']['flink_streaming']hadoop_conf_dir = config['configurations']['flink-ambari-config']['hadoop_conf_dir']
flink_download_url = config['configurations']['flink-ambari-config']['flink_download_url']
flink_hadoop_shaded_jar_url = config['configurations']['flink-ambari-config']['flink_hadoop_shaded_jar']
javax_ws_rs_api_jar = config['configurations']['flink-ambari-config']['javax_ws_rs_api_jar']conf_dir=''
bin_dir=''# params from flink-conf.yaml
flink_yaml_content = config['configurations']['flink-env']['content']
flink_user = config['configurations']['flink-env']['flink_user']
flink_group = config['configurations']['flink-env']['flink_group']
flink_log_dir = config['configurations']['flink-env']['flink_log_dir']
flink_log_file = os.path.join(flink_log_dir,'flink-setup.log')temp_file='/tmp/flink.tgz'

(9)添加用户及组

添加用户和组

groupadd flink
useradd -d /home/flink -g flink flink

(10)重启ambari服务

重启服务

ambari-server restart

在ambari中的Stack and Versions中可以看到flink的信息
在这里插入图片描述

2.安装

添加flink服务
在这里插入图片描述

在Custom flink-env中新增
Key：yarn.client.failover-proxy-provider
Value：org.apache.hadoop.yarn.client.ConfiguredRMFailoverProxyProvider

3.实验：wordcount

实验：wordcount
从尚硅谷下载实验用的程序代码，在idea中对以socket形式接收数据流的代码进行修改，从192.168.111.1的nc处接收数据流，然后对词频统计后输出到本地文件中
在这里插入图片描述

public static void main(String[] args) throws Exception {final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();env.setParallelism(1);DataStream<String> stream = env.socketTextStream("192.168.111.1", 1234);DataStream<Tuple2<String, Integer>> resultStream = stream.flatMap(new Tokenizer()).keyBy(0).sum(1);resultStream.print();resultStream.writeAsText("/tmp/wc_result.txt");env.execute("Flink Streaming Java API Skeleton");
}

在idea中的Build->Build Artifacts中选择build生成jar包，然后上传到hdp05上，使用一个租户进行kerberos认证后，提交flink任务。

cd /usr/hdp/3.1.5.0-152/flink/bin/
./flink run -m yarn-cluster /root/flink-tutorial-master.jar -c WordCountFromSocket

在这里插入图片描述
在192.168.111.1上启动nc

nc -l 1234

在这里插入图片描述
输入数据流后，中断nc进程，然后在hdp05上查看结果文件/tmp/wc_result.txt

在yarn上可以查看到对应的应用信息

4.常见报错

(1)安装时报错Unable to run the custom hook script

报错信息：Error: Error: Unable to run the custom hook script [‘/usr/bin/python’, ‘/var/lib/ambari-agent/cache/stack-hooks/before-ANY/scripts/hook.py’, ‘ANY’, ‘/var/lib/ambari-agent/data/command-1360.json’, ‘/var/lib/ambari-agent/cache/stack-hooks/before-ANY’, ‘/var/lib/ambari-agent/data/structured-out-1360.json’, ‘INFO’, ‘/var/lib/ambari-agent/tmp’, ‘PROTOCOL_TLSv1_2’, ‘’]
2023-03-26 22:23:31,445 - The repository with version 3.1.5.0-152 for this command has been marked as resolved. It will be used to report the version of the component which was installed
在这里插入图片描述
通过ambari添加自定义服务时，总是不能自动增加service账号

python configs.py -u admin -p lnyd@LNsy115 -n HDP315 -l hdp01 -t 8080 -a get -c cluster-env | grep -i ignore_groupsusers_create
python configs.py -u admin -p lnyd@LNsy115 -n HDP315 -l hdp01 -t 8080 -a set -c cluster-env -k ignore_groupsusers_create -v true

在这里插入图片描述

(2)安装后启动报错parent directory /usr/local/flink/conf doesn’t exist

安装后在启动时报错：resource_management.core.exceptions.Fail: Applying File[‘/usr/local/flink/conf/flink-conf.yaml’] failed, parent directory /usr/local/flink/conf doesn’t exist
在这里插入图片描述
不知什么原因没解压过去，需要手动解压到该目录

tar -zxvf /opt/flink-1.9.3-bin-scala_2.12.tgz -C /root/
mv /root/flink-1.9.3/* /usr/local/flink/

(3)提交任务报错java.lang.NoClassDefFoundError

报错信息：
java.lang.NoClassDefFoundError: com/sun/jersey/core/util/FeaturesAndProperties
在这里插入图片描述
maven会自动下载相关的依赖jar包，因此需要将project下的jersey依赖jar包拷贝至flink的lib目录下