You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@activemq.apache.org by "Alexander Fisher (JIRA)" <ji...@apache.org> on 2009/09/25 20:06:51 UTC
[jira] Reopened: (AMQ-2414) JDBC datastore locking does not work
when using MYSQL cluster
[ https://issues.apache.org/activemq/browse/AMQ-2414?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Fisher reopened AMQ-2414:
-----------------------------------
Two problems when testing with ndbcluser (mysql cluster).
1) Only the first table is created. The other CREATE TABLES fail mysteriously with Table 'ACTIVEMQ_ACKS' already exists SQLState: 42S01 Vendor code: 1050.
2) Assuming manually created tables, console output on slave is filled with SQL exceptions and stacktraces due to 'lock timeout exceeded' exceptions. These are quite normal and should be handled quietly.
> JDBC datastore locking does not work when using MYSQL cluster
> -------------------------------------------------------------
>
> Key: AMQ-2414
> URL: https://issues.apache.org/activemq/browse/AMQ-2414
> Project: ActiveMQ
> Issue Type: Bug
> Components: Message Store
> Affects Versions: 5.3.0
> Reporter: Alexander Fisher
> Fix For: 5.3.0
>
> Attachments: mysql-cluster-locking-r818849.patch
>
>
> JDBC Master/Slave doesn't work properly for mysql cluster when the master and slave brokers are using different MYSQL frontends as would be required in a HA setup. The code currently locks the database using a TABLE lock which are not cluster wide in mysql.
> http://dev.mysql.com/doc/refman/5.1/en/mysql-cluster-limitations-multiple-nodes.html
> The default code path uses a SELECT FOR UPDATE to get a lock. This is supported in mysql cluster.
> The attached patch works for me.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.