Mailing-List: contact notifications-help@accumulo.apache.org; run by ezmlm
Precedence: bulk
Reply-To: jira@apache.org
Date: Wed, 17 Aug 2016 17:51:20 +0000 (UTC)
From: "Michael Miller (JIRA)" <jira@apache.org>
To: notifications@accumulo.apache.org
Message-ID: <JIRA.12631517.1360356467000.333995.1471456280638@Atlassian.JIRA>
In-Reply-To: <JIRA.12631517.1360356467000@Atlassian.JIRA>
References: <JIRA.12631517.1360356467000@Atlassian.JIRA> <JIRA.12631517.1360356467432@arcas>
Subject: [jira] [Assigned] (ACCUMULO-1055) Configurable maximum size in
 which to perform merging minor compactions
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Wed, 17 Aug 2016 17:51:22 -0000


     [ https://issues.apache.org/jira/browse/ACCUMULO-1055?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Michael Miller reassigned ACCUMULO-1055:
----------------------------------------

    Assignee: Michael Miller  (was: Keith Turner)

> Configurable maximum size in which to perform merging minor compactions
> -----------------------------------------------------------------------
>
>                 Key: ACCUMULO-1055
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-1055
>             Project: Accumulo
>          Issue Type: Improvement
>          Components: tserver
>            Reporter: Josh Elser
>            Assignee: Michael Miller
>
> When performing a merging minor compaction, the smallest file for a tablet is chosen, and then data to be minor compacted is merged with said small file.
> When the smallest file isn't actually small, this can cause extremely large/long minor compactions. I don't want to completely "disable" merging minor compactions, for as I continue to run ingest, I'd like to take advantage of merging the small files together via minc.
> An upper bound on the file size to merge seems the most straightforward means to control this.


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