[ 
https://issues.apache.org/jira/browse/SENTRY-191?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14007595#comment-14007595
 ] 

SentryQA commented on SENTRY-191:
---------------------------------

Here are the results of testing the latest attachment
https://issues.apache.org/jira/secure/attachment/12646556/SENTRY-191.2.patch 
against master.

{color:red}Overall:{color} -1 due to an error

{color:red}ERROR:{color} failed to build with patch (exit code 1)

Console output: 
http://bigtop01.cloudera.org:8080/job/PreCommit-SENTRY-Build/81/console

This message is automatically generated.

> Sentry Policy Service should not require passing the RPC requestor's 
> user/group information
> -------------------------------------------------------------------------------------------
>
>                 Key: SENTRY-191
>                 URL: https://issues.apache.org/jira/browse/SENTRY-191
>             Project: Sentry
>          Issue Type: Bug
>    Affects Versions: 1.3.0
>            Reporter: Lenni Kuff
>            Assignee: Prasad Mujumdar
>            Priority: Blocker
>         Attachments: SENTRY-191.1.patch, SENTRY-191.2.patch
>
>
> Sentry Policy Service should not require passing the RPC requestor's 
> user/group information. Currently this is done to "authorize" whether a user 
> can execute a GRANT/REVOKE statement since only pre-selected set of admin 
> users run grant/revoke statements. This does not seem very secure and also 
> couples "authorization" with the storing of policy metadata.
> I propose that instead of this model, a default "admin" role be introduced. 
> On Sentry Service startup the the role be populated with set of valid admin 
> users as specified in the sentry-service.xml configuration file.
> When GRANT/REVOKE statements are run they should be treated the same as any 
> other SQL statement and authorized at the binding layer (if the give user 
> isn't part of the "admin" role then fail the request). 



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Reply via email to