kerberos系列之flink认证配置

大数据安全系列的其它文章

https://www.cnblogs.com/bainianminguo/p/12548076.html-----------安装kerberos

https://www.cnblogs.com/bainianminguo/p/12548334.html-----------hadoop的kerberos认证

https://www.cnblogs.com/bainianminguo/p/12548175.html-----------zookeeper的kerberos认证

https://www.cnblogs.com/bainianminguo/p/12584732.html-----------hive的kerberos认证

https://www.cnblogs.com/bainianminguo/p/12584880.html-----------es的search-guard认证

https://www.cnblogs.com/bainianminguo/p/12639821.html-----------flink的kerberos认证

https://www.cnblogs.com/bainianminguo/p/12639887.html-----------spark的kerberos认证

今天的博客介绍大数据安全系列之flink的kerberos配置

一、flink安装

1、解压安装包

tar -zxvf flink-1.8.0-bin-scala_2.11.tgz -C /usr/local/

2、重命名安装目录

[root@cluster2-host1 local]# mv flink-1.8.0/ flink

3、修改环境变量文件

export FLINK_HOME=/usr/local/flink
export PATH=${PATH}:${FLINK_HOME}/bin

[root@cluster2-host1 data]# source /etc/profile

[root@cluster2-host1 data]# echo $FLINK_HOME

/usr/local/flink

4、修改flink的配置文件

[root@cluster2-host1 conf]# vim flink-conf.yaml 
[root@cluster2-host1 conf]# pwd
/usr/local/flink/conf

jobmanager.rpc.address: cluster2-host1

修改slaver文件

[root@cluster2-host1 conf]# vim slaves 
[root@cluster2-host1 conf]# pwd
/usr/local/flink/conf

cluster2-host2
cluster2-host3

修改master文件

[root@cluster2-host1 bin]# cat /usr/local/flink/conf/masters 
cluster2-host1

修改yarn-site.xml文件

<property>
  <name>yarn.nodemanager.vmem-pmem-ratio</name>
  <value>5</value>
</property>

5、创建flink用户

[root@cluster2-host3 hadoop]# useradd flink -g flink
[root@cluster2-host3 hadoop]# passwd flink 
Changing password for user flink.
New password: 
BAD PASSWORD: The password fails the dictionary check - it is based on a dictionary word
Retype new password:

6、修改flink安装目录的属主和属组

[root@cluster2-host3 hadoop]# chown -R flink:flink /usr/local/flink/

7、启动flink验证安装步骤

[root@cluster2-host1 bin]# ./start-cluster.sh 
Starting cluster.
[INFO] 1 instance(s) of standalonesession are already running on cluster2-host1.
Starting standalonesession daemon on host cluster2-host1.
Starting taskexecutor daemon on host cluster2-host2.
Starting taskexecutor daemon on host cluster2-host3.

检查进程

[root@cluster2-host1 bin]# jps
10400 Secur
30817 StandaloneSessionClusterEntrypoint
12661 ResourceManager
12805 NodeManager
4998 QuorumPeerMain
30935 Jps
2631 NameNode

登陆页面

http://10.87.18.34:8081/#/overview

关闭flink，上面的standalone的启动方法，下面启动flink-session模式

拷贝hadoop的依赖包到flink的lib目录

scp flink-shaded-hadoop2-uber-2.7.5-1.8.0.jar /usr/local/flink/lib/

启动flink-session模式

./yarn-session.sh -n 2 -s 6 -jm 1024 -tm 1024 -nm test -d

检查yanr的页面

二、配置flink的kerberos的配置

1、创建flink的kerberos认证主体文件

kadmin.local:  addprinc flink/cluster2-host1
kadmin.local:  addprinc flink/cluster2-host2
kadmin.local:  addprinc flink/cluster2-host3

kadmin.local:  ktadd -norandkey -k /etc/security/keytab/flink.keytab flink/cluster2-host1
kadmin.local:  ktadd -norandkey -k /etc/security/keytab/flink.keytab flink/cluster2-host2
kadmin.local:  ktadd -norandkey -k /etc/security/keytab/flink.keytab flink/cluster2-host3

2、拷贝keytab文件到其它节点

[root@cluster2-host1 bin]# scp /etc/security/keytab/flink.keytab root@cluster2-host2:/usr/local/flink/
flink.keytab                                                                                                                                                                                                                                100% 1580     1.5KB/s   00:00    
[root@cluster2-host1 bin]# scp /etc/security/keytab/flink.keytab root@cluster2-host3:/usr/local/flink/
flink.keytab

3、修改flink的配置文件

security.kerberos.login.use-ticket-cache: true
security.kerberos.login.keytab: /usr/local/flink/flink.keytab
security.kerberos.login.principal: flink/cluster2-host3
yarn.log-aggregation-enable: true

4、启动yarn-session，看到如下操作，则配置完成

flink@cluster2-host1 bin]$  ./yarn-session.sh -n 2 -s 6 -jm 1024 -tm 1024 -nm flink5 -d

2020-03-05 02:42:23,706 INFO  org.apache.hadoop.security.UserGroupInformation               - Login successful for user flink/cluster2-host1 using keytab file /usr/local/flink/flink.keytab

查看页面

检查进程

[root@cluster2-host1 sbin]# jps
6118 ResourceManager
15975 NameNode
22472 -- process information unavailable
6779 NodeManager
23483 YarnSessionClusterEntrypoint
24717 Master
9790 QuorumPeerMain
25534 Jps
20239 Secur

5、flink的kerberos的配置完成

来源：oschina

链接：https://my.oschina.net/u/4272135/blog/4302693

标签

Hadoop

yarn

dictionary