A.9. Continuent Tungsten 4.0.0 GA (17 Apr 2015)

Continuent Tungsten 4.0 is a major release which is designed to provide integration between Continuent Tungsten 4.0 and Tungsten Replicator 4.0. Providing MySQL clustering support, while providing replication for MySQL, Oracle, and out to datawarehouses such as HP Vertica, Amazon Redshift and Hadoop.

For more information on replicating data out of a cluster, see Section 3.4, “Replicating Data Out of an Existing Cluster”.

Known Issue

The following issues may affect the operation of Continuent Tungsten and should be taken into account when deploying or updating to this release.

  • When using read-only connectors, and making use of explicit transactions (i.e. with autocommit disabled), queries may be routed to the master, rather than a slave.

  • For security purposes you should ensure that you secure the following areas of your deployment:

  • Under certain circumstances, the rsync process can randomly fail during the installation/ deployment process when using the staging method of deployment. The error code returned by rsync may be 12 or 23.

    The error is transient and non-specific and deployment should be retried.

    Issues: CONT-1343

Improvements, new features and functionality

  • Installation and Deployment

    • tpm now correctly checks the functionality of the 'echo' protocol to validate 'echo'.

      Issues: CONT-90

    • Force a new directory under /opt/continuent/releases during tpm update if components are being added/removed.

      Issues: CONT-155

    • tpm configuration setting repl-thl-log-retention - tpm should check for a valid unit

      Issues: CONT-177

  • Tungsten Connector

    • Useless reverse DNS call at connection time can drastically affect performance.

      Issues: CONT-86

    • Add min and max to Connector statistics.

      Issues: CONT-107

  • Tungsten Manager

    • Start managers serially per-dataservice rather than globally. This prevents a race-condition.

      Issues: CONT-27

    • Add manager status command to cctrl.

      Issues: CONT-168

Bug Fixes

  • Installation and Deployment

    • Installing an RPM package can fail if the mysql user doesn't exist.

      Issues: CONT-43

    • Update tpm to force the replication timezone to GMT.

      Issues: CONT-85

  • Command-line Tools

    • tungsten_set_position previously did not work within SOR deployments.

      Issues: CONT-24

    • The dsctl set command does not work properly for events with multiple fragments.

      Issues: CONT-194

  • Tungsten Connector

    • The MySQL Connector/J prerequisite has now been removed from all installations.

      Issues: CONT-48

    • The Connector could raise a Null Pointer Exception after upgrading from Continuent Tungsten 2.0.5.

      Issues: CONT-196

  • Tungsten Manager

    • Connector should not allow a data source role change without intermediary offline.

      Issues: CONT-23

    • Isolated relay site does not resume replication correctly.

      Issues: CONT-26

    • Java library call InetAddress.isReachable() can produce false positives

      Issues: CONT-53

    • Switch should rollback upon connector un-ability to apply the change

      Issues: CONT-105

    • Threshold for checking for manager memory leaks too low.

      Issues: CONT-161

    • The 'last man standing' logic within the manager fails to identify the correct host.

      Issues: CONT-163

    • Manager not setting datasource to failed - isolation via ifdown.

      Issues: CONT-164

    • Attempting concurrent operations in cctrl generates an exception.

      Issues: CONT-165

    • Manager hangs on jmx call if the interface that it uses is put 'ifdown'.

      Issues: CONT-166

    • Non-isolated nodes see the isolated one as online

      Issues: CONT-169

    • Rule that checks for node liveness is firing too frequently.

      Issues: CONT-170

    • Using the recover command does a gratuitous change of policy.

      Issues: CONT-171

    • After isolation is removed from the master site, it is not recovered to online.

      Issues: CONT-173

    • Monitor is attempting to query for non-existent table on witness host.

      Issues: CONT-174

    • Composite failover does not succeed - replication handshake failure.

      Issues: CONT-175

    • In a composite dataservice, failover does not happen if the master site is isolated from the relay.

      Issues: CONT-188

    • Composite master does not return to online after a failover within the physical master service.

      Issues: CONT-403