[ 
https://issues.apache.org/jira/browse/SOLR-938?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12659588#action_12659588
 ] 

Kay Kay commented on SOLR-938:
------------------------------

| I guess caching is pointless for a class which is used only twice for a full 
import

We have a much larger query , for which we do a full-import . We split the data 
by the primary key range and perform multiple smaller 'full-import's 
consolidated together to form a  full-import . Hence the necessity for 
retaining the instance for the event handlers. 

But I am ok with tracking that with a separate jira - if needed. When can we 
get this committed the current patch onto the trunk for now. 

> DataImportHandler: Add close hooks to the completion of a full-import process 
> ------------------------------------------------------------------------------
>
>                 Key: SOLR-938
>                 URL: https://issues.apache.org/jira/browse/SOLR-938
>             Project: Solr
>          Issue Type: New Feature
>          Components: contrib - DataImportHandler
>    Affects Versions: 1.4
>         Environment: JRE 6, Tomcat 6
>            Reporter: Kay Kay
>             Fix For: 1.4
>
>         Attachments: SOLR-938.patch, SOLR-938.patch, SOLR-938.patch
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> Adding a new feature that we need for continuation of a workflow based on a 
> full-import procedure. When   a full-import completes , the pre-defined hook 
> method is called at the end of the same. 
> Implementations that need to notified , need to implement the new interface - 
> DataImportCloseHook and register themselves in the data config file as 
> follows. 
> <dataConfig>
> .
> .
> .
> .
>    <closeHook type="com.myorg.fullimporter.closeHookImpl1" />
>    <closeHook type="com.myorg.fullimporter.closeHookImpl2" />
> </dataConfig>
> A single instance of them is created (as part of DataImporter) during 
> instantiation and the same insance is reused across multiple invocations. 
> Since the dataImporter is protected by a thread-lock - there is no thred-safe 
> guarantee for the implementations of the full-import close hook. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to