You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "Vaibhav Gumashta (JIRA)" <ji...@apache.org> on 2016/05/13 00:07:12 UTC
[jira] [Commented] (HIVE-13753) Make metastore client thread safe
in DbTxnManager
[ https://issues.apache.org/jira/browse/HIVE-13753?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15282262#comment-15282262 ]
Vaibhav Gumashta commented on HIVE-13753:
-----------------------------------------
[~wzheng] Is there a need to expose the underlying IMetaStoreClient object via SynchronizedMetaStoreClient?
> Make metastore client thread safe in DbTxnManager
> -------------------------------------------------
>
> Key: HIVE-13753
> URL: https://issues.apache.org/jira/browse/HIVE-13753
> Project: Hive
> Issue Type: Bug
> Components: Transactions
> Affects Versions: 1.3.0, 2.1.0
> Reporter: Wei Zheng
> Assignee: Wei Zheng
> Attachments: HIVE-13753.1.patch
>
>
> The fact that multiple threads sharing the same metastore client which is used for RPC to Thrift is not thread safe.
> Race condition can happen when one sees "out of sequence response" error message from Thrift server. That means the response from the Thrift server is for a different request (by a different thread).
> Solution will be to synchronize methods from the client side.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)