[jira] [Commented] (AMBARI-17788) Refactor spooler code in OutputHDFSFile to be reusable for OutputS3File

["Hemanth Yamijala (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/AMBARI-17788?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15384149#comment-15384149 ] 

Hemanth Yamijala commented on AMBARI-17788:
-------------------------------------------

I am planning to move out a Spooler class that would:

* Spool events to a configured local file
* Take an interface that would help specify whether to rollover or not (so that we could plugin different conditions if required)
* Take another interface to that would specify what to do upon rollover.

This would allow us to share similar logic for the spool file management, error handling scenarios etc and not need to fix things at multiple places.

> Refactor spooler code in OutputHDFSFile to be reusable for OutputS3File
> -----------------------------------------------------------------------
>
>                 Key: AMBARI-17788
>                 URL: https://issues.apache.org/jira/browse/AMBARI-17788
>             Project: Ambari
>          Issue Type: Technical task
>          Components: ambari-logsearch
>            Reporter: Hemanth Yamijala
>
> AMBARI-17785 proposed to enhance logsearch capabilities to periodically upload logs to S3, by using a spooling mechanism. [~mdharmesh] pointed out that such a facility already exists in the HDFS file copy handling. Looking at the code, there is a lot of similarity that can be reused in both cases. This task is to extract the common mechanisms and parameterize them so that it will make it easier to reuse and also help have similar improvements to spooling logic if any.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)