Mailing-List: contact common-issues-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: common-issues@hadoop.apache.org
Date: Fri, 19 Aug 2011 03:08:29 +0000 (UTC)
From: "Aaron Morton (JIRA)" <jira@apache.org>
To: common-issues@hadoop.apache.org
Message-ID: 
 <2008763719.51750.1313723309929.JavaMail.tomcat@hel.zones.apache.org>
Subject: [jira] [Created] (HADOOP-7556) TaskLogAppender does not check if
 closed before flushing
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit

TaskLogAppender does not check if closed before flushing
--------------------------------------------------------

                 Key: HADOOP-7556
                 URL: https://issues.apache.org/jira/browse/HADOOP-7556
             Project: Hadoop Common
          Issue Type: Bug
    Affects Versions: 0.20.203.0
            Reporter: Aaron Morton
            Priority: Minor


For background see http://groups.google.com/group/brisk-users/browse_thread/thread/3a18f4679673bea8
http://mail-archives.apache.org/mod_mbox/hadoop-common-user/201108.mbox/%3C4E370C97-1541-4FDA-8456-1067DDDC4D77@thelastpickle.com%3E

Cassandra is using a log4j PropertyConfigurator which is closing all existing appenders. After a task has completed the TaskLogAppender.flush() is called and it tries to flush without checking if the writer has been closed. I'll upload a patch to check if the writer is closed, and log and silently fail if it is. 

The real problem is the log4j config collision, we're looking to different log4j LoggerRepositories for that. 


--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira