hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Prasad Mujumdar (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-5181) RetryingRawStore should not retry on logical failures (e.g. from commit)
Date Tue, 24 Sep 2013 23:58:03 GMT

    [ https://issues.apache.org/jira/browse/HIVE-5181?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13776955#comment-13776955
] 

Prasad Mujumdar commented on HIVE-5181:
---------------------------------------

[~sershe] yes, it was the real failures. I updated the patch y'day to allow retry on open,
but the test was still checking for the annotation on open which caused the failures. The
latest attachment .3.patch has that fixed. It runs fine locally, waiting for the next pre-commit
run should be clean.
                
> RetryingRawStore should not retry on logical failures (e.g. from commit)
> ------------------------------------------------------------------------
>
>                 Key: HIVE-5181
>                 URL: https://issues.apache.org/jira/browse/HIVE-5181
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Sergey Shelukhin
>            Assignee: Prasad Mujumdar
>            Priority: Minor
>         Attachments: HIVE-5181.1.patch, HIVE-5181.3.patch
>
>
> RetryingRawStore retries calls. Some method (e.g. drop_table_core in HiveMetaStore) explicitly
call openTransaction and commitTransaction on RawStore.
> When the commit call fails due to some real issue, it is retried, and instead of a real
cause for failure one gets some bogus exception about transaction open count.
> I doesn't make sense to retry logical errors, especially not from commitTransaction.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message