cdh5.7权限测试示例

本文旨在展示CDH基于Kerberos身份认证和基于Sentry的权限控制功能的测试示例。

1. 准备测试数据

1
2
3
4
5
6
cat /tmp/events.csv
10.1 . 2.3 ,US,android,createNote
10.200 . 88.99 ,FR,windows,updateNote
10.1 . 2.3 ,US,android,updateNote
10.200 . 88.77 ,FR,ios,createNote
10.1 . 4.5 ,US,windows,updateTag


2. 创建用户
2.1. 创建系统用户
在集群所有节点创建系统用户并设置密码

1
2
3
4
5
6
useradd user1
passwd user1
useradd user2
passwd user2
useradd user3
passwd user3


2.2. 创建kerberos用户

1
2
3
kadmin.local -q  "addprinc user1"
kadmin.local -q  "addprinc user2"
kadmin.local -q  "addprinc user3"


3. 创建数据库和表
3.1. 创建数据库
admin为sentry的超级管理员,该用户配置权限时已设置

1
kinit admin

通过beeline连接 hiveserver2,运行下面命令创建hive库的超级管理员角色, 并将该角色赋予admin组,使admin有操作hive库的权力

1
2
3
4
beeline -u  "jdbc:hive2://vmw208:10000/;principal=hive/[email protected]"
create role admin_role;
GRANT ALL ON SERVER server1 TO ROLE admin_role;
GRANT ROLE admin_role TO GROUP admin;


创建两个测试数据库

1
2
create database db1;
create database db2;

3.2. 创建表
在两个测试数据库中各创建一张测试表,并导入测试数据

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
create table db1.table1 (
ip STRING, country STRING, client STRING, action STRING
) ROW FORMAT DELIMITED FIELDS TERMINATED BY  ',' ;
 
create table db2.table1 (
ip STRING, country STRING, client STRING, action STRING
) ROW FORMAT DELIMITED FIELDS TERMINATED BY  ',' ;
create table db2.table2 (
ip STRING, country STRING, client STRING, action STRING
) ROW FORMAT DELIMITED FIELDS TERMINATED BY  ',' ;
 
 
load data local inpath  '/home/iie/events.csv'  overwrite into table db1.table1;
load data local inpath  '/home/iie/events.csv'  overwrite into table db2.table1;
load data local inpath  '/home/iie/events.csv'  overwrite into table db2.table2;

  

4. 赋予用户权限
4.1. 给user1赋予db1的所有权限

1
2
3
create role user1_role;
GRANT ALL ON DATABASE db1 TO ROLE user1_role;
GRANT ROLE user1_role TO GROUP user1;


4.2. 给user2赋予db2的所有权限

1
2
3
create role user2_role;
GRANT ALL ON DATABASE db2 TO ROLE user2_role;
GRANT ROLE user2_role TO GROUP user2;


4.3. 给user3赋予db2.table1的所有权限

1
2
3
4
create role user3_role;
use db2;
GRANT select ON table table1 TO ROLE user3_role;
GRANT ROLE user3_role TO GROUP user3;

5. 测试用户权限
5.1. Hive测试
5.1.1. admin用户拥有整个hive库的权限

1
2
3
4
5
6
7
8
9
10
11
12
13
kinit admin
beeline -u  "jdbc:hive2://vmw208:10000/;principal=hive/[email protected]"
show databases;
5.1 . 2 .  user1用户只具有db1和 default 的权限
kinit user1
beeline -u  "jdbc:hive2://vmw208:10000/;principal=hive/[email protected]"
0 : jdbc:hive2: //vmw208:10000/> show databases;
+----------------+--+
| database_name |
+----------------+--+
| db1 |
default  |
+----------------+--+


5.1.3. user2用户只具有db2和default的权限

1
2
3
4
5
6
7
8
9
kinit user2
beeline -u  "jdbc:hive2://vmw208:10000/;principal=hive/[email protected]"
0 : jdbc:hive2: //vmw208:10000/> show databases;
+----------------+--+
| database_name |
+----------------+--+
| db2 |
default  |
+----------------+--+


5.1.4. user3用户只具有db2.table1和default的权限

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
kinit user2
beeline -u  "jdbc:hive2://vmw208:10000/;principal=hive/[email protected]"
0 : jdbc:hive2: //vmw208:10000/> show databases;
+----------------+--+
| database_name |
+----------------+--+
| db2 |
default  |
+----------------+--+
0 : jdbc:hive2: //node0a17:10000/> use db2;
0 : jdbc:hive2: //node0a17:10000/> show tables;
INFO : OK
+-----------+--+
| tab_name |
+-----------+--+
| table1 |
+-----------+--+


5.2. Hdfs测试
配置hdfs acl与sentry同步后,hdfs权限与sentry监控的目录(/user/hive/warehouse)的权限同步
5.2.1. 切换到hive用户,查看hive库文件的权限
设置hdfs acl与sentry同步后,sentry监控的hive库的权限改动会同步到对应的hdfs文件权限

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
[root @vmw208  home]# kinit hive
[root @vmw208  home]# hdfs dfs -getfacl -R /user/hive/warehouse/
# file: /user/hive/warehouse
# owner: hive
# group: hive
user::rwx
user:hive:rwx
group::---
group:hive:rwx
mask::rwx
other::--x
 
# file: /user/hive/warehouse/db1.db
# owner: hive
# group: hive
user::rwx
user:hive:rwx
group:user1:rwx
group::---
group:hive:rwx
mask::rwx
other::--x
 
# file: /user/hive/warehouse/db1.db/table1
# owner: hive
# group: hive
user::rwx
user:hive:rwx
group:user1:rwx
group::---
group:hive:rwx
mask::rwx
other::--x
 
# file: /user/hive/warehouse/db1.db/table1/events.csv
# owner: hive
# group: hive
user::rwx
user:hive:rwx
group:user1:rwx
group::---
group:hive:rwx
mask::rwx
other::--x
 
# file: /user/hive/warehouse/db2.db
# owner: hive
# group: hive
user::rwx
user:hive:rwx
group:user2:rwx
group::---
group:hive:rwx
mask::rwx
other::--x
 
# file: /user/hive/warehouse/db2.db/table1
# owner: hive
# group: hive
user::rwx
user:hive:rwx
group:user2:rwx
group::---
group:hive:rwx
mask::rwx
other::--x
 
# file: /user/hive/warehouse/db2.db/table1/events.csv
# owner: hive
# group: hive
user::rwx
user:hive:rwx
group:user2:rwx
group::---
group:hive:rwx
mask::rwx
other::--x

  

5.2.2. 切换到user1用户,查看hdfs文件

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
[root @vmw208  home]# kinit user1
Password  for  user1 @HADOOP .COM:
[root @vmw208  home]# hdfs dfs -ls /user/hive/warehouse/db2.db
ls: Permission denied: user=user1, access=READ_EXECUTE, inode= "/user/hive/warehouse/db2.db" :hive:hive:drwxrwx—x
[root @vmw208  home]# hdfs dfs -cat /user/hive/warehouse/db2.db/table1/events.csv
cat: Permission denied: user=user1, access=READ, inode= "/user/hive/warehouse/db2.db/table1/events.csv" :hive:hive:-rwxrwx--x
 
[root @vmw208  home]# hdfs dfs -ls /user/hive/warehouse/db1.db
Found  1  items
drwxrwx--x+ - hive hive  0  2016 - 09 - 29  16 : 54  /user/hive/warehouse/db1.db/table1
[root @vmw208  home]# hdfs dfs -cat /user/hive/warehouse/db1.db/table1/events.csv
10.1 . 2.3 ,US,android,createNote
10.200 . 88.99 ,FR,windows,updateNote
10.1 . 2.3 ,US,android,updateNote
10.200 . 88.77 ,FR,ios,createNote
10.1 . 4.5 ,US,windows,updateTag

  

5.2.3. 切换到user2用户,查看hdfs文件

1
2
3
4
5
6
7
8
9
10
[root @vmw208  home]# kinit user2
Password  for  user2 @HADOOP .COM:
[root @vmw208  home]# hdfs dfs -cat /user/hive/warehouse/db1.db/table1/events.csv
cat: Permission denied: user=user2, access=READ, inode= "/user/hive/warehouse/db1.db/table1/events.csv" :hive:hive:-rwxrwx--x
[root @vmw208  home]# hdfs dfs -cat /user/hive/warehouse/db2.db/table1/events.csv
10.1 . 2.3 ,US,android,createNote
10.200 . 88.99 ,FR,windows,updateNote
10.1 . 2.3 ,US,android,updateNote
10.200 . 88.77 ,FR,ios,createNote
10.1 . 4.5 ,US,windows,updateTag  

5.3. Spark测试
5.3.1. Spark读hive表数据并打印到控制台
(1) 切换到user1用户测试

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
[root @vmw209  xdf]# kinit user1
Password  for  user1 @HADOOP .COM:
[root @vmw209  xdf]# spark-submit -- class  iie.hadoop.permission.QueryTable --master local /home/xdf/spark.jar db2 table1
……
Exception in thread  "main"  org.apache.hadoop.security.AccessControlException: Permission denied: user=user1, access=READ_EXECUTE, inode= "/user/hive/warehouse/db2.db/table1" :hive:hive:drwxrwx—x
[root @vmw209  xdf]# spark-submit -- class  iie.hadoop.permission.QueryTable --master local /home/xdf/spark.jar db1 table1
……
+------------+-------+-------+----------+
| ip|country| client| action|
+------------+-------+-------+----------+
10.1 . 2.3 | US|android|createNote|
| 10.200 . 88.99 | FR|windows|updateNote|
10.1 . 2.3 | US|android|updateNote|
| 10.200 . 88.77 | FR| ios|createNote|
10.1 . 4.5 | US|windows| updateTag|
+------------+-------+-------+----------+

  

(2) 切换到user2用户测试

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
[root @vmw209  xdf]# kinit user2
Password  for  user2 @HADOOP .COM:
[root @vmw209  xdf]# spark-submit -- class  iie.hadoop.permission.QueryTable --master local /home/xdf/spark.jar db1 table1
……
Exception in thread  "main"  org.apache.hadoop.security.AccessControlException: Permission denied: user=user2, access=READ_EXECUTE, inode= "/user/hive/warehouse/db1.db/table1" :hive:hive:drwxrwx—x
[root @vmw209  xdf]# spark-submit -- class  iie.hadoop.permission.QueryTable --master local /home/xdf/spark.jar db2 table1
……
+------------+-------+-------+----------+
| ip|country| client| action|
+------------+-------+-------+----------+
10.1 . 2.3 | US|android|createNote|
| 10.200 . 88.99 | FR|windows|updateNote|
10.1 . 2.3 | US|android|updateNote|
| 10.200 . 88.77 | FR| ios|createNote|
10.1 . 4.5 | US|windows| updateTag|
+------------+-------+-------+----------+

 

5.3.2. Spark读文件数据写入hive表中
调用工具程序spark.jar读本地文件/home/xdf/events.csv数据写到db2.table2
切换到user2用户测试

1
2
3
4
5
6
7
kinit user2
beeline -u  "jdbc:hive2://vmw208:10000/;principal=hive/[email protected]"
use db2;
create table table2 (
ip STRING, country STRING, client STRING, action STRING
) ROW FORMAT DELIMITED FIELDS TERMINATED BY  ',' ;
[root @vmw209  xdf]# spark-submit -- class  iie.hadoop.permission.HCatWriterTest --master local /home/xdf/spark.jar /home/xdf/events.csv db2 table2

成功!
写到db1.table1报错,没有权限!

1
Exception in thread  "main"  org.apache.hive.hcatalog.common.HCatException :  2004  : HCatOutputFormat not initialized, setOutput has to be called. Cause : org.apache.hadoop.security.AccessControlException: Permission denied: user=user2, access=WRITE, inode= "/user/hive/warehouse/db1.db/table1" :hive:hive:drwxrwx--x

上面只是测试环境,因为kinit + 密码的方式有时效限制,不适合在生产环境运行,幸好spark提供了相关的参数:

1
2
3
4
spark-submit
……
--principal # 用户对应的kerberos principle
--keytab # 对应用户principle生成的密钥文件


spark的权限管理通过对hdfs/hive的文件目录设置权限来管理,不同的用户拥有不同的权限,用户在提交spark任务时,指定对应用户的kerberos principle和keytab来实现权限管理。任务提交命令如下:

1
spark-submit -- class  iie.hadoop.permission.QueryTable --master yarn-cluster --principal=user1 @HADOOP .COM --keytab=/home/user1/user1.keytab /home/user1/spark.jar db1 table1

其中--principal 和--keytab与用户一一对应

注意:spark-submit只有在yarn-cluster模式下,--principal 和--keytab才有效


5.4. Kafka测试
5.4.1. 认证
用户kafka为kafka权限控制的超级管理员

1
[root @node10  iie]#kinit -kt /home/iie/kafka.keytab kafka


5.4.2. 创建topic
创建topic1和topic2

1
2
[root @node10  iie]#kafka-topics --zookeeper node11: 2181 /kafka --create --topic topic1 --partitions  2  --replication-factor  1
[root @node10  iie]#kafka-topics --zookeeper node11: 2181 /kafka --create --topic topic2 --partitions  2  --replication-factor  1

5.4.3. 赋权
给user1用户附topic1的读写权限

1
2
[root @node10  iie]#kafka-acls --authorizer-properties zookeeper.connect=node11: 2181 /kafka --add --allow-principal User:user1 --allow-host node10  --producer --topic topic1 --group console-consumer- 9175
[root @node10  iie]#kafka-acls --authorizer-properties zookeeper.connect=node11: 2181 /kafka --add --allow-principal User:user1 --allow-host node10  --consumer --topic topic1 --group console-consumer- 9175

给user2用户附topic2的读写权限

1
2
[root @node10  iie]#kafka-acls --authorizer-properties zookeeper.connect=node11: 2181 /kafka --add --allow-principal User:user2 --allow-host node10  --producer --topic topic2 --group console-consumer- 9175
[root @node10  iie]#kafka-acls --authorizer-properties zookeeper.connect=node11: 2181 /kafka --add --allow-principal User:user2 --allow-host node10  --consumer --topic topic2 --group console-consumer- 9175

5.4.4. 查看权限

1
2
3
4
5
6
7
[root @node10  iie]#kafka-acls --authorizer-properties zookeeper.connect=node11: 2181 /kafka --list
Current ACLs  for  resource `Topic:topic1`:
User:user1 has Allow permission  for  operations: Write from hosts: node10
User:user1 has Allow permission  for  operations: Read from hosts: node10
Current ACLs  for  resource `Topic:topic2`:
User:user2 has Allow permission  for  operations: Read from hosts: node10
User:user2 has Allow permission  for  operations: Write from hosts: node10


5.4.5. 创建生产消费配置文件
创建consumer.properties

1
2
3
4
5
cat /etc/kafka/conf/consumer.properties
security.protocol=SASL_PLAINTEXT
sasl.mechanism=GSSAPI
sasl.kerberos.service.name=kafka
group.id=console-consumer- 9175

创建producer.properties

1
2
3
4
cat /etc/kafka/conf/producer.properties
security.protocol=SASL_PLAINTEXT
sasl.mechanism=GSSAPI
sasl.kerberos.service.name=kafka

5.4.6. 生产数据
命令行生产数据

1
2
3
4
[root @node10  iie]#kinit user1
[root @node10  iie]#kafka-console-producer --broker-list node12: 9092  --topic topic1 --producer.config /etc/kafka/conf/producer.properties
123123
123123

5.4.7. 消费数据
命令行消费数据

1
2
3
4
[root @node10  iie]#kinit user1
[root @node10  iie]#kafka-console-consumer --bootstrap-server node12: 9092  --topic topic1 -- new -consumer --from-beginning --consumer.config /etc/kafka/conf/consumer.properties
123123
123123

用户对topic没有权限时报错

1
2
3
4
5
[root @node10  iie]# kinit user2
Password  for  user2 @HADOOP .COM:
[root @node10  iie]# kafka-console-consumer --bootstrap-server node12: 9092  --topic topic1 -- new -consumer --from-beginning --consumer.config /etc/kafka/conf/consumer.properties
[ 2016 - 10 - 12  15 : 38 : 01 , 599 ] ERROR Unknown error when running consumer:  (kafka.tools.ConsoleConsumer$)
org.apache.kafka.common.errors.TopicAuthorizationException: Not authorized to access topics: [topic1]


5.4.8. 移除权限
登陆管理员用户移除权限

1
[root @node10  iie]#kinit -kt /home/iie/kafka.keytab kafka

删除user1对topic1的消费权限

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
[root @node10  iie]# kafka-acls --authorizer-properties zookeeper.connect=node11: 2181 /kafka --remove --allow-principal User:user1 --allow-host node10 --consumer --topic topic1 --group console-consumer- 92175
Are you sure you want to remove ACLs:
     User:user1 has Allow permission  for  operations: Read from hosts: node10
     User:user1 has Allow permission  for  operations: Describe from hosts: node10
  from resource `Topic:topic1`? (y/n)
y
Are you sure you want to remove ACLs:
     User:user1 has Allow permission  for  operations: Read from hosts: node10
  from resource `Group:console-consumer- 92175 `? (y/n)
y
Current ACLs  for  resource `Topic:topic1`:
     User:Aluser1 has Allow permission  for  operations: Read from hosts: node10
     User:Aluser1 has Allow permission  for  operations: Describe from hosts: node10
     User:user1 has Allow permission  for  operations: Write from hosts: node10
 
Current ACLs  for  resource `Group:console-consumer- 92175 `:

  

测试user1消费topic1报错,说明权限已经移除

1
2
3
4
5
6
7
8
9
10
11
12
[root @node10  iie]# kinit user1
Password  for  user1 @HADOOP .COM:
[root @node10  iie]# kafka-console-consumer --bootstrap-server node12: 9092  --topic topic1 -- new -consumer --from-beginning --consumer.config /etc/kafka/conf/consumer.properties
[ 2016 - 10 - 12  15 : 45 : 11 , 572 ] WARN The configuration sasl.mechanism = GSSAPI was supplied but isn't a known config. (org.apache.kafka.clients.consumer.ConsumerConfig)
[ 2016 - 10 - 12  15 : 45 : 11 , 914 ] WARN Not authorized to read from topic topic1. (org.apache.kafka.clients.consumer.internals.Fetcher)
[ 2016 - 10 - 12  15 : 45 : 11 , 916 ] ERROR Error processing message, terminating consumer process:  (kafka.tools.ConsoleConsumer$)
org.apache.kafka.common.errors.TopicAuthorizationException: Not authorized to access topics: [topic1]
[ 2016 - 10 - 12  15 : 45 : 11 , 920 ] WARN Not authorized to read from topic topic1. (org.apache.kafka.clients.consumer.internals.Fetcher)
[ 2016 - 10 - 12  15 : 45 : 11 , 921 ] ERROR Not authorized to commit to topics [topic1]  for  group console-consumer- 9175  (org.apache.kafka.clients.consumer.internals.ConsumerCoordinator)
[ 2016 - 10 - 12  15 : 45 : 11 , 922 ] WARN Auto offset commit failed  for  group console-consumer- 9175 : Not authorized to access topics: [topic1] (org.apache.kafka.clients.consumer.internals.ConsumerCoordinator)
[ 2016 - 10 - 12  15 : 45 : 11 , 927 ] WARN TGT renewal thread has been interrupted and will exit. (org.apache.kafka.common.security.kerberos.Login)
Processed a total of  0  messages

  

转载于:https://www.cnblogs.com/royfans/p/7299490.html

你可能感兴趣的:(cdh5.7权限测试示例)