Skip to content

Latest commit

 

History

History
30 lines (16 loc) · 1.65 KB

RELEASE_NOTES.md

File metadata and controls

30 lines (16 loc) · 1.65 KB
Raw

0.6.0 Release Notes

  • replication-manager now uses the POSIX flag syntax instead of the Go flag syntax. All flags now use the double dash (--) syntax.

  • Failover and switchover logic has been entirely rewritten. As a consequence, failover can now use the monitoring console without restarting it, and topologies are preserved across failovers.

  • Failed masters can be reintroduced automatically if the --autorejoin option is true. This option is enabled by default.

  • The monitoring console has a default

  • Failed servers are now displayed as Standalone Servers on the monitoring console.

  • Servers now correctly set read-only or read-write mode. In the monitoring console, Ctrl-R sets read-only mode interactively, and Ctrl-W sets read-write mode on the slaves.

  • replication-manager can now log its messages to a file with the --log option.

  • It is now possible to set a connection timeout with --timeout. It helps if some servers are unresponsive.

  • Automatic failover mode comes now with 3 tunables:

    • It is now possible to set the number of master failures with the --failcount option.

    • It is now possible to set the number of automatic failovers with the --failover-limit option. Default is 0 (unlimited). If you want to have similar behavior as MHA, use --failover-limit=1

    • It is now possible to set the interval between automatic failovers with the --failover-time-limit option. Default is 0 (no interval).

  • A bug with failover setting the read-only flag on the promoted master has been corrected.

  • Several major bugs have been corrected and replication-manager should not crash anymore in various conditions.