1.1. Tungsten Clustering 6.0.5 GA (20 March 2019)

Version End of Life. 20 March 2022

Release 6.0.5 contains both significant improvments as well as some needed bugfixes.

Improvements, new features and functionality

  • Command-line Tools

    • A new utility script has been added to the release, tungsten_reset_manager (in [Tungsten Clustering (for MySQL) 6.0 Manual]), which assists with the graceful reset of the manager's dynamic state files on disk.

      Issues: CT-850

      For more information, see The tungsten_reset_manager Command (in [Tungsten Clustering (for MySQL) 6.0 Manual]).

Bug Fixes

  • Installation and Deployment

    • Fixing the rpm-based post-install chown command so that symlinked directories get correct ownership.

      Issues: CT-767

    • The Tungsten Clustering RPM now preserves the original OS group memberships for the tungsten user.

      Issues: CT-867

  • Command-line Tools

    • Do not try to backup a witness server.

      Issues: CT-669

    • Include additional views of cctrl output in tpm diag (cctrl_status_simple_SVCNAME).

      Issues: CT-681

    • The MySQL MyISAM check seems to fail intermittantly with no way to bypass it so the check has been disabled completely.

      Issues: CT-756

    • Fixed an issue where the tpm (in [Tungsten Clustering (for MySQL) 6.0 Manual]) command would allocate inconsistent THL listener ports for the Composite Multimaster (CMM) topology.

      The new, correct behavior is for the main cluster replicator to always be allocated port 2112, and then relay sub-services are incremented per remote cluster.

      For example, in a 4-site CMM deployment, ports 2112 through 2115 would be allocated - 2112 for the main cluster and 2113, 2114 and 2115 for the remote site relays.

      Issues: CT-799

    • The tpm diag (in [Tungsten Clustering (for MySQL) 6.0 Manual]) command now collects cctrl status without a "WARNING: Unrecognized option 'multi'" error.

      Issues: CT-821

    • Remove any clear-text passwords gathered via tpm diag.

      Issues: CT-822

    • Fixed NullPointerException in cctrl 'ls -l' output when the dataserver is down.

      Issues: CT-826

  • Tungsten Connector

    • MySQL ping commands are now reconnected/retried upon "server gone away" error (Proxy mode ONLY).

      Issues: CT-863

  • Tungsten Manager

    • When a replicator is not running insert the Replicator.HOST to the ReplicationNotification. It was wrongly inserted into the Replicator.DATASERVERHOST. This fixes the get_replicator_roles script. Also substituted hard-coded strings for their constant values.

      Issues: CT-760

    • mysql_checker_query script was returning unexpected errors and creating false positives. Changed the script logic to use the timestampdiff function for better accuracy.

      Issues: CT-824

    • Change the Manager behavior so as to place the replicator online asynchronously to prevent cctrl from hanging if a slave replicator is put online while the master is offline. Now, if the master is offline the slave will go into the SYNCHRONIZING state. As the master comes online the slaves will come online as well.

      Issues: CT-825

Tungsten Clustering 6.0.5 Includes the following changes made in Tungsten Replicator 6.0.5

Release 6.0.5 is a bugfix release.

Improvements, new features and functionality

Bug Fixes

  • Command-line Tools

    • The --hosts (in [Tungsten Replicator 6.0 Manual]) option was not working with the diag sub-command of the tpm (in [Tungsten Replicator 6.0 Manual]) command on nodes installed using the INI method.

      The corrected behavior is as follows:

      • With Staging-method deployments, the tpm diag (in [Tungsten Replicator 6.0 Manual]) command continues to behave as before:

        • The tpm diag (in [Tungsten Replicator 6.0 Manual]) command alone will obtain diagnostics from all hosts in the cluster.

        • The tpm diag --hosts host1,host2,hostN command will obtain diagnostics from the specified host(s) only.

      • With INI-method deployments, the new behavior is as follows:

        • The tpm diag (in [Tungsten Replicator 6.0 Manual]) command alone will obtain diagnostics from the local host only.

        • The tpm diag --hosts host1,host2,hostN command will obtain diagnostics from the specified host(s) only.

          Warning

          Limitation: the host list MUST include the local hostname or the command will fail.

      Issues: CT-345

    • The trepctl (in [Tungsten Replicator 6.0 Manual]) command now properly handles the -all-services option for the reset sub-command.

      Issues: CT-762

    • The command tpm reverse --ini-format now outputs without the double-dashes and the trailing backslash

      Issues: CT-827

    • The command tpm diag (in [Tungsten Replicator 6.0 Manual]) was not collecting config dirs other than the localhost ones.

      Now the mysql, manager, cluster and connector config directories are properly gathered in the diag zip file.

      Issues: CT-860

    • The tpm (in [Tungsten Replicator 6.0 Manual]) command now properly handles network interface names containing colons and/or dots.

      Issues: CT-864

    • Fixed an issue where the tpm (in [Tungsten Replicator 6.0 Manual]) command could print warnings about nil verify_host_key.

      Issues: CT-873

  • Core Replicator

    • The postgres applier now respects the database name set by pgsql-dbname.

      Specifically, the tungsten-replicator/samples/conf/datasources/postgresql.tpl was updated to use the correct variable for the value.

      Issues: CT-704

    • Instead of searching for a master with appropriate role (i.e. matching the slave preferred role) until timeout is reached, the Replicator will now loop twice before accepting connection to any host, no matter what its role is.

      Issues: CT-712

    • The backup process fails with 0-byte store*.properties files or store*.properties files with invalid dates.

      Changed the process so that invalid backup properties files are skipped.

      Issues: CT-820

    • Fix the ability to enable parallel apply within a Composite Multimaster (CMM) topology.

      Now handling relay as slave to make the relay use the same code as a slave concerning its internal connections (disable binary logging of its internal SQL queries).

      Issues: CT-851