A.3. Tungsten Clustering 5.3.3 GA (20 September 2018)

Version End of Life. 7 June 2019

This is a bugfix release.

Improvements, new features and functionality

  • Installation and Deployment

    • tpm now outputs a note and recommendation for performing backups of your cluster when installation has been completed.

      Issues: CT-730

  • Command-line Tools

    • The tungsten_prep_upgrade command has been updated to support an explicit host definition for the MySQL host in place of defaulting to the localhost (127.0.0.1). Use the --host option.

      Issues: CT-656

    • A new Nagios compatible check script has been added to the release, check_tungsten_policy, which returns the currently active policy mode.

      Issues: CT-675

      For more information, see Section 8.5, “The check_tungsten_policy Command”.

  • Tungsten Connector

    • When receiving an error within MySQLPacket, the Connector now prints out the full content of the underlying error message.

      Issues: CT-636

  • Tungsten Manager

    • The router gateway which provides communication between the manager and connector could shutdown even when quorum was available in a two-node cluster.

      Issues: CT-676

Bug Fixes

  • Installation and Deployment

    • tpm would fail during installation if the current directory was not writable by the current user.

      Issues: CT-564

    • When performing a tpm update in a cluster with an active witness, the host with the witness will not be restarted correctly resulting in the witness being down on that host.

      Issues: CT-596

  • Command-line Tools

    • Using tpm diag, the command would ignore options on the command-line, including --net-ssh-option.

      Issues: CT-610

    • Using tpm connector at the command-line would fail if the core MySQL configuration file (i.e. /etc/my.cnf) did not exist.

      Issues: CT-641

  • Tungsten Connector

    • The connector would fail to set reusable network addresses during configuration which could delay or slow startup until the address/port become available again.

      Issues: CT-694

    • When operating in bridge mode, the connector would fail to check whether the driver was in enabled/disabled mode, which could cause upgrades to fail as part of a graceful shutdown/update operation.

      Issues: CT-696

    • Multiple connectors within a cluster could all connect to the same manager within a given service, increasing the load on the single manager.

      Issues: CT-717

  • Tungsten Manager

    • When using the connector, the connector --cluster-status --json command would output header and footer information in place of bare JSON which would then cause JSON parsing to fail.

      Issues: CT-685

    • A memory leak within the manager, particularly in multimaster deployments, could cause the Java VM to consume more and more CPU cycles and then restart.

      Issues: CT-673, CT-691

    • During a relay failover within a composite or multi-site multi-master deployment, if the communications had also failed between sites when the failover occured the manager would be unable to determine the correct master of the remote site.

      Issues: CT-703

    • A memory leak was identified in the router manager component that manages the communicating between the manager and the connector.

      Issues: CT-715

    • In a deployment, single cluster or composite multimaster where there is either the potential for high-latency across sites, or high latency within a site due to high loads on the connectors, the manager could mis-identify this high latency as a failure. This would trigger a quorum validation. These would be reported as network hangs, even though the result of the quorum check would be valid.

      To address this, the processing of router notifications processed by the connector and all other operations have been separated. This reduces the change of a heartbeat gap between hosts and therefore the connectors are available to the managers even under high loads or latency.

      Issues: CT-725

Tungsten Clustering 5.3.3 Includes the following changes made in Tungsten Replicator 5.3.3

Release 5.3.3 is a bug fix release.

Improvements, new features and functionality

  • Core Replicator

    • The output from thl list now includes the name of the file for the correspnding THL event. For example:

      SEQ# = 0 / FRAG# = 0 (last frag)
      - FILE = thl.data.0000000001
      - TIME = 2018-08-29 12:40:57.0
      - EPOCH# = 0
      - EVENTID = mysql-bin.000050:0000000000000508;-1
      - SOURCEID = demo-c11
      - METADATA = [mysql_server_id=5;dbms_type=mysql;tz_aware=true;is_metadata=true;service=alpha;shard=tungsten_alpha;heartbeat=MASTER_ONLINE]
      - TYPE = com.continuent.tungsten.replicator.event.ReplDBMSEvent
      - OPTIONS = [foreign_key_checks = 1, unique_checks = 1, time_zone = '+00:00', ##charset = US-ASCII]

      Issues: CT-550

Bug Fixes

  • Command-line Tools

    • Using tpm diag, the command would ignore options on the command-line, including --net-ssh-option.

      Issues: CT-610

    • When running tpm diag, the operation would fail if the /etc/mysql directory does not exist.

      Issues: CT-724

  • Core Replicator

    • The LOAD DATA INFILE would fail to be executed and replicated properly.

      Issues: CT-10, CT-652

    • The trepsvc.log displayed information without highlighting the individual services reporting the entries making it difficult to identify individual log entries.

      Issues: CT-659