[
https://issues.apache.org/jira/browse/HIVE-16719?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16017655#comment-16017655
]
Zsolt Fekete commented on HIVE-16719:
-------------------------------------
Note that during testing my refactor and modification I found that it is likely
that updating TABLE_PARAMS, SD_PARAMS, SERDE_PARAMS did not work previously,
because these calls in ObjectStore: {{mSerde.getParameters().put(serdeProp,
newSchemaLoc);}} does not have any affect. One should create a new map and call
{{mSerde.setParameters(..);}} with that.
So I also fixed this issue.
> HiveMetaTool fails when the data does not fit in memory
> -------------------------------------------------------
>
> Key: HIVE-16719
> URL: https://issues.apache.org/jira/browse/HIVE-16719
> Project: Hive
> Issue Type: Bug
> Components: Metastore
> Reporter: Zsolt Fekete
> Assignee: Zsolt Fekete
> Attachments: HIVE-16719.1.patch
>
>
> Currently HiveMetaTool reads full tables (as DataNucleus entities) into
> memory by calling PersistenceManager's retrieveAll().
> See these methods of ObjectStore: updateMDatabaseURI, updateTblPropURI,
> updateMStorageDescriptorTblPropURI, updateMStorageDescriptorTblURI,
> updateSerdeURI.
> This might cause failure when the affected tables (SDS, DBS, TABLE_PARAMS,
> SD_PARAMS, SERDES, SERDE_PARAMS) are too big.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
