[
https://issues.apache.org/jira/browse/HIVE-6137?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13878139#comment-13878139
]
Hari Sankar Sivarama Subramaniyan commented on HIVE-6137:
---------------------------------------------------------
[~thejas] No. It can happen because of broken connectivity issues as well. I
think the point here is that we need not expose the thrift exception to the
end-user while running from a CLI. For debugging purpose, we should log the
thrift exception. The end-user should just know that Hive encountered an
exception while trying to locate the data file. Probably, I should improve the
error message by mentioning that "Error while trying to access the data file :
" based on TTransportException instead of throwing a FileNotFound exception.
> Hive should report that the file/path doesn’t exist when it doesn’t (it now
> reports SocketTimeoutException)
> -----------------------------------------------------------------------------------------------------------
>
> Key: HIVE-6137
> URL: https://issues.apache.org/jira/browse/HIVE-6137
> Project: Hive
> Issue Type: Bug
> Reporter: Hari Sankar Sivarama Subramaniyan
> Assignee: Hari Sankar Sivarama Subramaniyan
> Attachments: HIVE-6137.1.patch
>
>
> Hive should report that the file/path doesn’t exist when it doesn’t (it now
> reports SocketTimeoutException):
> Execute a Hive DDL query with a reference to a non-existent blob (such as
> CREATE EXTERNAL TABLE...) and check Hive logs (stderr):
> FAILED: Error in metadata: org.apache.thrift.transport.TTransportException:
> java.net.SocketTimeoutException: Read timed out
> FAILED: Execution Error, return code 1 from
> org.apache.hadoop.hive.ql.exec.DDLTask
> This error message is not intuitive. If a file doesn't exist, Hive should
> report FileNotFoundException
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
