Chris Nauroth commented on HIVE-9736:

Hi [~mithun].  Thank you for uploading a new patch.

I was unable to apply patch v3 to the master branch.  Does it need to be 
rebased, or should I be working with a different branch?

There was one suggestion I made on Review Board that still isn't implemented.  
In {{Hadoop23Shims#checkFileAccess}}, we can combine the multiple {{actions}} 
by using {{FsAction#or}}, and then call {{accessMethod.invoke}} just once to do 
the check in a single RPC (per file).  Were you planning to make this change, 
or is there a reason you decided not to do it?

Aside from that, I can see all of my other feedback has been addressed.  Thanks 

> StorageBasedAuthProvider should batch namenode-calls where possible.
> --------------------------------------------------------------------
>                 Key: HIVE-9736
>                 URL: https://issues.apache.org/jira/browse/HIVE-9736
>             Project: Hive
>          Issue Type: Bug
>          Components: Metastore, Security
>            Reporter: Mithun Radhakrishnan
>            Assignee: Mithun Radhakrishnan
>         Attachments: HIVE-9736.1.patch, HIVE-9736.2.patch, HIVE-9736.3.patch
> Consider a table partitioned by 2 keys (dt, region). Say a dt partition could 
> have 10000 associated regions. Consider that the user does:
> {code:sql}
> ALTER TABLE my_table DROP PARTITION (dt='20150101');
> {code}
> As things stand now, {{StorageBasedAuthProvider}} will make individual 
> {{DistributedFileSystem.listStatus()}} calls for each partition-directory, 
> and authorize each one separately. It'd be faster to batch the calls, and 
> examine multiple FileStatus objects at once.

This message was sent by Atlassian JIRA

Reply via email to