The ultimate diff is v1.5.6...v1.5.7
Merge pull request #245 from outbrain/remove-oauth removed oauth code; not ready yet
Merge pull request #217 from outbrain/merge-downstream-gh Merge downstream gh
- initial support for semi-sync - thanks Vitess!
- vagrant improvements - thanks Square!
- fuzzy cluster names accepted
- various fixes, thanks community!
- change of license: MIT
Long time no uploaded release. Lots of goodies here.
force-master-takeovercommand. Currently only available via CLI, for planned switchovers
- Binary log indexing. This makes Pseudo-GTID based recoveries significantly faster and usually non-dependent on binary log file size
- Improved and potentially we can all forget about internal db deployments bookkeeping.
- Introducing promotion-rules: still a way to go.
-c register-candidatemay specify
--promotion-rule=(prefer | neutral | must_not)
- Autoread replication credentials where possible via
which-heuristic-cluster-pool-instancescommand added, where dealing with pools
- Golang 1.5 package vendoring
- fqdn no longer requirement in CLI commands. As long as the hostname is uniquely identifiable you can use a shortname (substring) of the fqdn
community contributions: vagrant, documentation, logic, suggestions, issues -- thank you!
minor css change
aggregated instances can be drag-n-dropped in "smart mode". pattern is sent via URI. It is possible that pattern exceeds max URI length; this is not checked at this time.
"relocate" uses further heuristics for binlog servers, simplifying another scenario web: - modal dialog does not support "begin/end-maintenance" any more. Instead, it now supports "begin/end-downtime", with implied owner, with duration
Fixed resolve.go error where hostname_resolve would load into cache even on "none" configuration