Mailing-List: contact issues-help@hive.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@hive.apache.org
Date: Wed, 4 May 2016 08:04:12 +0000 (UTC)
From: "Hari Sankar Sivarama Subramaniyan (JIRA)" <jira@apache.org>
To: issues@hive.apache.org
Message-ID: <JIRA.12962396.1461618738000.106203.1462349052837@Atlassian.JIRA>
In-Reply-To: <JIRA.12962396.1461618738000@Atlassian.JIRA>
References: <JIRA.12962396.1461618738000@Atlassian.JIRA> <JIRA.12962396.1461618738763@arcas>
Subject: [jira] [Commented] (HIVE-13608) We should provide better error
 message while constraints with duplicate names are created
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Wed, 04 May 2016 08:04:14 -0000


    [ https://issues.apache.org/jira/browse/HIVE-13608?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15270289#comment-15270289 ] 

Hari Sankar Sivarama Subramaniyan commented on HIVE-13608:
----------------------------------------------------------

[~ashutoshc] I looked at this option. I am a +0 for this idea of parsing the error message to see if it matches some string and prevent the retry altogether.

{code}
else if ((t instanceof MetaException) && t.getMessage().matches(
            "(?s).*(JDO[a-zA-Z]*|TProtocol|TTransport)Exception.*")) {
          caughtException = (MetaException)t;
        } 
{code} 
This can make the man-in-the-middle attack scenario with the current system worse. I realize why the above code was added in the first place to tackle JDO*/T* exceptions which are unrecoverable otherwise, but still this whole thing is hacky and adding more parsing messages here will worsen the existing code.

For practical purpose, the change you mentioned should be ok and I can upload the patch, but as I said, this is something that needs to be revisited to increase the security of the product.

Thanks
Hari

> We should provide better error message while constraints with duplicate names are created
> -----------------------------------------------------------------------------------------
>
>                 Key: HIVE-13608
>                 URL: https://issues.apache.org/jira/browse/HIVE-13608
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Hari Sankar Sivarama Subramaniyan
>            Assignee: Hari Sankar Sivarama Subramaniyan
>         Attachments: HIVE-13608.1.patch
>
>
> {code}
> PREHOOK: query: create table t1(x int, constraint pk1 primary key (x) disable novalidate)
> PREHOOK: type: CREATETABLE
> PREHOOK: Output: database:default
> PREHOOK: Output: default@t1
> POSTHOOK: query: create table t1(x int, constraint pk1 primary key (x) disable novalidate)
> POSTHOOK: type: CREATETABLE
> POSTHOOK: Output: database:default
> POSTHOOK: Output: default@t1
> PREHOOK: query: create table t2(x int, constraint pk1 primary key (x) disable novalidate)
> PREHOOK: type: CREATETABLE
> PREHOOK: Output: database:default
> PREHOOK: Output: default@t2
> FAILED: Execution Error, return code 1 from org.apache.hadoop.hive.ql.exec.DDLTask. MetaException(message:For direct MetaStore DB connections, we don't support retries at the client level.)
> {code}
> In the above case, it seems like useful error message is lost. It looks like a  generic problem with metastore server/client exception handling and message propagation. Seems like exception parsing logic of RetryingMetaStoreClient::invoke() needs to be updated.


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)