[jira] [Commented] (FLUME-2207) HDFS file suffix style index suffix in ElasticSearchSink

["Satoshi Iijima (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/FLUME-2207?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14007574#comment-14007574 ] 

Satoshi Iijima commented on FLUME-2207:
---------------------------------------

This feature has already been implemented by TimeBasedIndexNameBuilder which is the default indexNameBuilder. But it does not work because of a small bug. I have attached a patch.

> HDFS file suffix style index suffix in ElasticSearchSink
> --------------------------------------------------------
>
>                 Key: FLUME-2207
>                 URL: https://issues.apache.org/jira/browse/FLUME-2207
>             Project: Flume
>          Issue Type: Improvement
>          Components: Sinks+Sources
>    Affects Versions: v1.5.0
>            Reporter: Dib Ghosh
>             Fix For: v1.6.0
>
>         Attachments: FLUME-2207.patch
>
>
> Currently ElasticSearchSink indexName is appended with "YYYY-MM-DD" to generate actual ElasticSearch index name and allow index rolling on a daily basis. This works well for simpler use cases but is restrictive imho. User should be able to add indexSuffix as per their own use case and also keep the indexSuffix empty if requirement demands.
> agent.sinks.essink.indexName = flume becomes flume-2013-10-10.
> It will be really useful to have HDFS Sink filepath style  indexSuffix naming i.e. "YYYY" will roll the indexes yearly, "YYYY-MM" will create indices monthly.



--
This message was sent by Atlassian JIRA
(v6.2#6252)