[HIVE-24342] isPathEncrypted should make sure resolved path also from HDFS

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.1.2
• Fix Version/s: 4.0.0-alpha-1
• Component/s: HiveServer2, Shims
• Labels: pull-request-available
• Target Version/s: 3.1.3
• Hadoop Flags: Reviewed
• Assignee: Uma Maheswara Rao G
• Reporter: Uma Maheswara Rao G
• Created: 31/Oct/20 21:25
• Updated: 08/Mar/24 08:41
• Resolved: 04/Nov/20 07:26

Description

Currently isPathEncrypted will make sure path is from hdfs by checking the path scheme is "hdfs"

In the case if mounted ViewFileSystem based files systems like ViewFSOverloadScheme or ViewHDFS (HDFS-15289) may need o check resolved path is really hdfs.

In ViewHDFS case, we can mount hdfs://ns1/test ---> o3fs://b.v.ozone1/test

When user calling queries with the path hdfs://ns1/test, isPathEncrypted will think the path is from hdfs only as its checking path scheme.

 

0: jdbc:hive2://umag-1.umag.root.xxx.site:218> select * from test30;
Error: Error while compiling statement: FAILED: SemanticException Unable to determine if hdfs://ns1/test is encrypted: java.lang.UnsupportedOperationException: This API:getEZForPath is specific to DFS. Can't run on other fs:o3fs://bucket.volume.ozone1 (state=42000,code=40000)
0: jdbc:hive2://umag-1.umag.root.xxx.site:218> cd Closing: 0: jdbc:hive2://umag-1.umag.root.xxx.site:2181,umag-2.umag.root.xxx.site:2181,umag-5.umag.root.xxx.site:2181/default;password=root;principal=hive/[e-mail];retries=5;serviceDiscoveryMode=zooKeeper;user=root;zooKeeperNamespace=hiveserver2

 

So, here we should use resolvePath to make sure the resolved path really in hdfs. If the resolved path is not from hdfs (in above case, it o3fs path), then it will return false.

After fixing this, the query is passing.:

 

0: jdbc:hive2://umag-1.umag.root.xxx.site:218> select * from test30;
INFO  : Compiling command(queryId=hive_20201031002253_1691548f-6fa8-4ea9-9cd4-87b70fe8f6bb): select * from test30
INFO  : No Stats for default@test30, Columns: item, user_id, state, order_id
INFO  : Semantic Analysis Completed (retrial = false)
INFO  : Created Hive schema: Schema(fieldSchemas:[FieldSchema(name:test30.order_id, type:bigint, comment:null), FieldSchema(name:test30.user_id, type:string, comment:null), FieldSchema(name:test30.item, type:string, comment:null), FieldSchema(name:test30.state, type:string, comment:null)], properties:null)
INFO  : Completed compiling command(queryId=hive_20201031002253_1691548f-6fa8-4ea9-9cd4-87b70fe8f6bb); Time taken: 4.47 seconds
INFO  : Executing command(queryId=hive_20201031002253_1691548f-6fa8-4ea9-9cd4-87b70fe8f6bb): select * from test30
INFO  : Completed executing command(queryId=hive_20201031002253_1691548f-6fa8-4ea9-9cd4-87b70fe8f6bb); Time taken: 0.09 seconds
INFO  : OK
+------------------+-----------------+--------------+---------------+
| test30.order_id  | test30.user_id  | test30.item  | test30.state  |
+------------------+-----------------+--------------+---------------+
| 1234             | u1              | iphone7      | CA            |
| 2345             | u1              | ipad         | CA            |
| 3456             | u2              | desktop      | NY            |
 
 
+------------------+-----------------+--------------+---------------+
11 rows selected (6.975 seconds)

 

Issue Links

• is related to: HDFS-15289 Allow viewfs mounts with HDFS/HCFS scheme and centralized mount table
• links to: GitHub Pull Request #1639
• links to: GitHub Pull Request #1639