You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hive.apache.org by "Rajesh Balamohan (JIRA)" <ji...@apache.org> on 2017/02/14 08:13:41 UTC

[jira] [Commented] (HIVE-15909) HiveMetaStoreChecker::checkTable can be expensive in ObjectStores

    [ https://issues.apache.org/jira/browse/HIVE-15909?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15865334#comment-15865334 ] 

Rajesh Balamohan commented on HIVE-15909:
-----------------------------------------

Instead of multi-threading, one option could be to check if all partitions retrieved from metastore have same parent. If so, {{listStatus/listFiles}} can be used to reduce the number of lookups. HADOOP-13208 has optimized implementations for S3.

> HiveMetaStoreChecker::checkTable can be expensive in ObjectStores
> -----------------------------------------------------------------
>
>                 Key: HIVE-15909
>                 URL: https://issues.apache.org/jira/browse/HIVE-15909
>             Project: Hive
>          Issue Type: Bug
>          Components: Metastore
>            Reporter: Rajesh Balamohan
>            Priority: Minor
>
> With objectstores like s3, HiveMetaStoreChecker::checkTable can be expensive with partitioned dataset.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)