Mailing-List: contact notifications-help@accumulo.apache.org; run by ezmlm
Precedence: bulk
Reply-To: jira@apache.org
Date: Wed, 7 Aug 2013 14:12:48 +0000 (UTC)
From: "Josh Elser (JIRA)" <jira@apache.org>
To: notifications@accumulo.apache.org
Message-ID: <JIRA.12662385.1375884665036.16337.1375884768022@arcas>
In-Reply-To: <JIRA.12662385.1375884665036@arcas>
References: <JIRA.12662385.1375884665036@arcas>
Subject: [jira] [Created] (ACCUMULO-1645) Provide better information on
 unassigned tablets on the monitor
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit

Josh Elser created ACCUMULO-1645:
------------------------------------

             Summary: Provide better information on unassigned tablets on the monitor
                 Key: ACCUMULO-1645
                 URL: https://issues.apache.org/jira/browse/ACCUMULO-1645
             Project: Accumulo
          Issue Type: Improvement
          Components: monitor
    Affects Versions: 1.4.3
            Reporter: Josh Elser
             Fix For: 1.4.4


Take the situation when live ingest occurs on a system. During live ingest, many tablet servers die, due to unexpected issues (data-center issues, for example). When the tablets are reassigned to a new tabletserver, if all copies of the necessary WAL files for that tablet were contained on servers which are now dead, the tablet cannot be assigned until those files can be found.

This presents itself on the monitor as unassigned tablets, but doesn't denote any additional information about which table they belong to, the extent, or even that this is due to missing WAL files. It would be nice to definitively know which tablets are offline, the WAL files which are missing, and any additional information Accumulo might have. This would alleviate the administrator of the system from having to guess as to why said tablets are offline.

Due to the HDFS WAL support in 1.5, this is even less of a concern than it is in 1.4 due to the decreased likelihood of a WAL file going missing.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira