0032-如何在CDH启用Kerberos的情况下安装及使用Sentry(二)
温馨提示：要看高清无码套图，请使用手机打开并单击图片放大查看。
5.Sentry列权限管理
1.在集群所有节点添加fayson_r用户
[root@ip-172-31-6-148 cdh-shell-bak]# useradd fayson_r
[root@ip-172-31-6-148 cdh-shell-bak]# id fayson_r
uid=504(fayson_r) gid=504(fayson_r) groups=504(fayson_r)
[root@ip-172-31-6-148 cdh-shell-bak]#
2.创建Kerberos用户fayson_r
[root@ip-172-31-6-148 ~]# kadmin.local
Authenticating as principal hive/[email protected] with password.
kadmin.local: addprinc [email protected]
WARNING: no policy specified for [email protected]; defaulting to no policy
Enter password for principal "[email protected]":
Re-enter password for principal "[email protected]":
Principal "[email protected]" created.
kadmin.local:
3.使用hive用户登录Kerberos
使用beeline连接HiveServer2，创建columnread角色并授权test表s1列的读权限，将columnread角色授权给fayson_r用户组
[root@ip-172-31-6-148 ~]# klist
Ticket cache: FILE:/tmp/krb5cc_0
Default principal: hive/[email protected]
Valid starting Expires Service principal
09/07/17 15:27:58 09/08/17 15:27:58 krbtgt/[email protected]
renew until 09/12/17 15:27:58
[root@ip-172-31-6-148 ~]# beeline
Beeline version 1.1.0-cdh5.12.1 by Apache Hive
beeline> create role columnread;
No current connection
beeline> !connect jdbc:hive2://localhost:10000/;principal=hive/[email protected]
...
0: jdbc:hive2://localhost:10000/> create role columnread;
...
INFO : OK
No rows affected (0.183 seconds)
0: jdbc:hive2://localhost:10000/> grant select(s1) on table test to role columnread;
...
INFO : OK
No rows affected (0.105 seconds)
0: jdbc:hive2://localhost:10000/> grant role columnread to group fayson_r;
...
INFO : OK
No rows affected (0.105 seconds)
0: jdbc:hive2://localhost:10000/>
4.fayson_r用户测试
使用fayson_r登录Kerberos，通过beeline连接HiveServer2
[root@ip-172-31-6-148 ~]# kdestroy
[root@ip-172-31-6-148 ~]# kinit fayson_r
Password for [email protected]:
[root@ip-172-31-6-148 ~]# klist
Ticket cache: FILE:/tmp/krb5cc_0
Default principal: [email protected]
Valid starting Expires Service principal
09/08/17 03:16:47 09/09/17 03:16:47 krbtgt/[email protected]
renew until 09/15/17 03:16:47
[root@ip-172-31-6-148 ~]# beeline
Beeline version 1.1.0-cdh5.12.1 by Apache Hive
beeline> !connect jdbc:hive2://localhost:10000/;principal=hive/[email protected]
scan complete in 2ms
Connecting to jdbc:hive2://localhost:10000/;principal=hive/[email protected]
Connected to: Apache Hive (version 1.1.0-cdh5.12.1)
Driver: Hive JDBC (version 1.1.0-cdh5.12.1)
Transaction isolation: TRANSACTION_REPEATABLE_READ
0: jdbc:hive2://localhost:10000/> show databases;
...
INFO : OK
+----------------+--+
| database_name |
+----------------+--+
| default |
+----------------+--+
1 row selected (0.336 seconds)
0: jdbc:hive2://localhost:10000/> show tables;
...
INFO : OK
+-----------+--+
| tab_name |
+-----------+--+
| test |
+-----------+--+
1 row selected (0.202 seconds)
0: jdbc:hive2://localhost:10000/> select * from test;
Error: Error while compiling statement: FAILED: SemanticException No valid privileges
User fayson_r does not have privileges for QUERY
The required privileges: Server=server1->Db=default->Table=test->Column=s2->action=select; (state=42000,code=40000)
0: jdbc:hive2://localhost:10000/> select s1 from test;
...
INFO : OK
+---------+--+
| s1 |
+---------+--+
| a |
| 1 |
| 111 |
| a |
| 1 |
| 2 |
| testaa |
| 1 |
| 2 |
| 3 |
| 222 |
+---------+--+
11 rows selected (0.433 seconds)
0: jdbc:hive2://localhost:10000/> select count(*) from test;
Error: Error while compiling statement: FAILED: SemanticException No valid privileges
User fayson_r does not have privileges for QUERY
The required privileges: Server=server1->Db=default->Table=test->action=select; (state=42000,code=40000)
0: jdbc:hive2://localhost:10000/> select count(s1) from test;
...
INFO : OK
+------+--+
| _c0 |
+------+--+
| 11 |
+------+--+
1 row selected (33.012 seconds)
0: jdbc:hive2://localhost:10000/>
5.浏览HDFS目录
[root@ip-172-31-6-148 ~]# klist
Ticket cache: FILE:/tmp/krb5cc_0
Default principal: [email protected]
Valid starting Expires Service principal
09/08/17 03:16:47 09/09/17 03:16:47 krbtgt/[email protected]
renew until 09/15/17 03:16:47
[root@ip-172-31-6-148 ~]# hadoop fs -ls /user/hive/warehouse
ls: Permission denied: user=fayson_r, access=READ_EXECUTE, inode="/user/hive/warehouse":hive:hive:drwxrwx--x
[root@ip-172-31-6-148 ~]# hadoop fs -ls /user/hive/warehouse/test
ls: Permission denied: user=fayson_r, access=READ_EXECUTE, inode="/user/hive/warehouse/test":hive:hive:drwxrwx--x
[root@ip-172-31-6-148 ~]#
6.使用admin用户登录Hue，创建fayson_r用户
使用fayson_r用户登录
不能对test表所有列进行查询
可以对test表的s1列进行查询
不能通过File Brwoser浏览test表的数据目录/user/hive/warehouse/test
测试总结：
fayson_r用户所属用户组为fayson_r，该组只拥有对test表s1列的读权限，因此在select和count的时候只能对s1列进行select和count，fayson_r用户无权限浏览/user/hive/warehouse下的所有目录；使用hue只能对test表s1列进行select和count操作，无权限浏览/user/hive/warehouse目录及目录下所有子目录。
注意：Sentry只支持SELECT的列授权，不能用于INSERT和ALL的列授权。
6.备注
在集群启用Sentry服务后，由于Sentry不支持Hive CLI权限管理，所以建议禁用Hive CLI。
- 如何限制用户使用Hive CLI操作
进入Hive服务，修改hadoop.proxyuser.hive.group配置，此配置会覆盖HDFS服务中hive代理用户组配置，默认值为空则继承HDFS服务中的hive代理用户配置
以上配置表示hue、hive、impala用户组可以使用Hive CLI，配置完成重启Hive及其相关服务。
注：如果配置为空则表示禁止所有用户组，需要注意配置为空会导致Hue不可用，hive用户不能通过beeline和Hive CLI访问Hive。
2.测试配置是否生效
l 使用hive用户登录Kerberos，然后使用Hive CLI操作
[root@ip-172-31-6-148 251-hive-HIVEMETASTORE]# kinit -kt hive.keytab hive/[email protected]
[root@ip-172-31-6-148 251-hive-HIVEMETASTORE]# klist
Ticket cache: FILE:/tmp/krb5cc_0
Default principal: hive/[email protected]
Valid starting Expires Service principal
09/07/17 13:33:21 09/08/17 13:33:21 krbtgt/[email protected]
renew until 09/12/17 13:33:21
[root@ip-172-31-6-148 251-hive-HIVEMETASTORE]#
登录Hive CLI，进行SQL操作
[root@ip-172-31-6-148 251-hive-HIVEMETASTORE]# hive
...
hive> show databases;
OK
default
Time taken: 1.881 seconds, Fetched: 1 row(s)
hive> show tables;
OK
test
test_hive_delimiter
test_table
Time taken: 0.034 seconds, Fetched: 3 row(s)
hive> select * from test;
OK
a b
1 2
111 222
a b
1 2
2 333
testaa testbbb
1 test
2 fayson
3 zhangsan
222 2323
Time taken: 0.477 seconds, Fetched: 11 row(s)
hive> select count(*) from test;
...
OK
11
Time taken: 31.143 seconds, Fetched: 1 row(s)
hive>
- 使用hue用户登录Kerberos进行测试
[root@ip-172-31-6-148 259-hue-HUE_SERVER]# kdestroy
[root@ip-172-31-6-148 259-hue-HUE_SERVER]# kinit -kt hue.keytab hue/[email protected]
[root@ip-172-31-6-148 259-hue-HUE_SERVER]# klist
Ticket cache: FILE:/tmp/krb5cc_0
Default principal: hue/[email protected]
Valid starting Expires Service principal
09/07/17 13:37:22 09/08/17 13:37:22 krbtgt/[email protected]
renew until 09/12/17 13:37:22
[root@ip-172-31-6-148 259-hue-HUE_SERVER]#
通过Hive CLI操作
[root@ip-172-31-6-148 259-hue-HUE_SERVER]# hive
...
hive> show databases;
OK
default
Time taken: 1.892 seconds, Fetched: 1 row(s)
hive> show tables;
OK
test
test_hive_delimiter
test_table
Time taken: 0.036 seconds, Fetched: 3 row(s)
hive> select * from test;
FAILED: SemanticException Unable to determine if hdfs://ip-172-31-6-148.fayson.com:8020/user/hive/warehouse/test is encrypted: org.apache.hadoop.security.AccessControlException: Permission denied: user=hue, access=READ, inode="/user/hive/warehouse/test":hive:hive:drwxrwx--x
at org.apache.hadoop.hdfs.server.namenode.DefaultAuthorizationProvider.checkAccessAcl(DefaultAuthorizationProvider.java:363)
at org.apache.hadoop.hdfs.server.namenode.DefaultAuthorizationProvider.check(DefaultAuthorizationProvider.java:256)
at org.apache.hadoop.hdfs.server.namenode.DefaultAuthorizationProvider.checkPermission(DefaultAuthorizationProvider.java:168)
at org.apache.sentry.hdfs.SentryAuthorizationProvider.checkPermission(SentryAuthorizationProvider.java:178)
at org.apache.hadoop.hdfs.server.namenode.FSPermissionChecker.checkPermission(FSPermissionChecker.java:152)
at org.apache.hadoop.hdfs.server.namenode.FSDirectory.checkPermission(FSDirectory.java:3530)
at org.apache.hadoop.hdfs.server.namenode.FSDirectory.checkPermission(FSDirectory.java:3513)
at org.apache.hadoop.hdfs.server.namenode.FSDirectory.checkPathAccess(FSDirectory.java:3484)
at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.checkPathAccess(FSNamesystem.java:6624)
at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.getEZForPath(FSNamesystem.java:9267)
at org.apache.hadoop.hdfs.server.namenode.NameNodeRpcServer.getEZForPath(NameNodeRpcServer.java:1637)
at org.apache.hadoop.hdfs.server.namenode.AuthorizationProviderProxyClientProtocol.getEZForPath(AuthorizationProviderProxyClientProtocol.java:928)
at org.apache.hadoop.hdfs.protocolPB.ClientNamenodeProtocolServerSideTranslatorPB.getEZForPath(ClientNamenodeProtocolServerSideTranslatorPB.java:1360)
at org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos$ClientNamenodeProtocol$2.callBlockingMethod(ClientNamenodeProtocolProtos.java)
at org.apache.hadoop.ipc.ProtobufRpcEngine$Server$ProtoBufRpcInvoker.call(ProtobufRpcEngine.java:617)
at org.apache.hadoop.ipc.RPC$Server.call(RPC.java:1073)
at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:2217)
at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:2213)
at java.security.AccessController.doPrivileged(Native Method)
at javax.security.auth.Subject.doAs(Subject.java:415)
at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1917)
at org.apache.hadoop.ipc.Server$Handler.run(Server.java:2211)
hive>
- 使用impala用户登录Kerberos测试
[root@ip-172-31-6-148 253-impala-STATESTORE]# kdestroy
[root@ip-172-31-6-148 253-impala-STATESTORE]# kinit -kt impala.keytab impala/[email protected]
[root@ip-172-31-6-148 253-impala-STATESTORE]# klist
Ticket cache: FILE:/tmp/krb5cc_0
Default principal: impala/[email protected]
Valid starting Expires Service principal
09/07/17 13:41:25 09/08/17 13:41:25 krbtgt/[email protected]
renew until 09/12/17 13:41:25
[root@ip-172-31-6-148 253-impala-STATESTORE]#
使用Hive CLI操作
[root@ip-172-31-6-148 253-impala-STATESTORE]# hive
...
hive> show databases;
OK
default
Time taken: 1.941 seconds, Fetched: 1 row(s)
hive> show tables;
OK
test
test_hive_delimiter
test_table
Time taken: 0.037 seconds, Fetched: 3 row(s)
hive> select * from test;
OK
a b
1 2
111 222
a b
1 2
2 333
testaa testbbb
1 test
2 fayson
3 zhangsan
222 2323
Time taken: 0.523 seconds, Fetched: 11 row(s)
hive>
- 使用fayson用户测试，此用户被禁止使用HiveCLI
使用fayson用户登录Kerberos测试
[root@ip-172-31-6-148 ~]# kdestroy
[root@ip-172-31-6-148 ~]# kinit fayson
Password for [email protected]:
[root@ip-172-31-6-148 ~]# klist
Ticket cache: FILE:/tmp/krb5cc_0
Default principal: [email protected]
Valid starting Expires Service principal
09/07/17 13:44:29 09/08/17 13:44:29 krbtgt/[email protected]
renew until 09/14/17 13:44:29
[root@ip-172-31-6-148 ~]#
使用Hive CLI操作
[root@ip-172-31-6-148 ~]# hive
...
hive> show databases;
FAILED: SemanticException org.apache.hadoop.hive.ql.metadata.HiveException: org.apache.hadoop.hive.ql.metadata.HiveException: org.apache.thrift.transport.TTransportException: java.net.SocketException: Connection reset
hive> show tables;
FAILED: SemanticException org.apache.hadoop.hive.ql.metadata.HiveException: org.apache.hadoop.hive.ql.metadata.HiveException: org.apache.thrift.transport.TTransportException: java.net.SocketException: Connection reset
hive>
测试总结：
通过Hive配置（hadoop.proxyuser.hive.groups）可以限制用户组使用HiveCLI访问Hive，未配置在内的用户组是不可以通过Hive CLI访问（如fayson用户）。在测试中发现impala和hive用户可以通过Hive CLI访问hive表并未受权限控制，而hue用户只能show databases和show tables不能select表。报HDFS访问权限异常，由于hive表的属主均为hive，hive和impala用户都属于hive组，hue用户不属于hive组导致。
注意：hadoop.proxyuser.hive.groups是针对用户组限制，如配置了hive用户组可以通过Hive CLI访问Hive，则属于hive组的所有用户均可以通过Hive CLI访问Hive表且不受Sentry权限控制。
