Mailing-List: contact dev-help@activemq.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@activemq.apache.org
Message-ID: <18404747.1201493845879.JavaMail.jira@brutus>
Date: Sun, 27 Jan 2008 20:17:25 -0800 (PST)
From: "Shahram Javey (JIRA)" <jira@apache.org>
To: dev@activemq.apache.org
Subject: [jira] Updated: (AMQ-992) MySQL doesn't honor lock in JDBC Master
 Slave configuration?
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit


     [ https://issues.apache.org/activemq/browse/AMQ-992?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Shahram Javey updated AMQ-992:
------------------------------

    Attachment: patch.txt

The current lock statement with MyISAM tables has led to unstable database. After a few hours of use, MySQL database becomes progressively  slower and eventually no other process can get any database lock. The only option is to kill the slave broker and to restart the database. Assuming that changing the table from MyISAM to InnoDB will fix the problem, I've attached a patch to alter the storage engine after the tables have been created. Has anyone else observed this problem?

> MySQL doesn't honor lock in JDBC Master Slave configuration?
> ------------------------------------------------------------
>
>                 Key: AMQ-992
>                 URL: https://issues.apache.org/activemq/browse/AMQ-992
>             Project: ActiveMQ
>          Issue Type: Bug
>    Affects Versions: 4.1.0
>         Environment: RHEL 4
> MySQL 4.x, 5.x
> mysql-ab_jdbc_driver
>            Reporter: Steven Lotito
>             Fix For: 5.1.0
>
>         Attachments: mysql_obtain_lock.txt, patch.txt
>
>
> I have been attempting to get the new 4.1 JDBC Master Slave configuration working with MySQL.
> The log from the first broker to start up states:
> 2006-10-18 09:35:08,558 [main           ] INFO  DefaultDatabaseLocker          - Attempting to acquire the exclusive lock to become the Master broker
> 2006-10-18 09:35:08,559 [main           ] INFO  DefaultDatabaseLocker          - Becoming the master on dataSource: org.apache.commons.dbcp.BasicDataSource@136d9d8
> The 2nd broker to start up has an identical message and both brokers listen for connections.
> The 2nd broker should be waiting for the lock and NOT accepting connections, if I understand http://www.activemq.org/site/jdbc-master-slave.html correctly...
> Oracle exhibits the expected behavior:
> When running the exact same configuration (except using an Oracle datasource), the first broker has the same log message as above,  while the 2nd broker halts at the "Attempting to acquire the exclusive lock to become the Master broker" message until I fail the master.  Then it becomes the master.
> Is this a known issue?  I was able to replicate it using both MySql 4 and 5 (trying both the MySQL Connector/J 3.1 and MySQL Connector/J 5.0 drivers)

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.