Export - CSV (All fields)
Export - CSV (Current fields)
MXS-1524
Send FDE event to slaves with one buffer instead of twoMXS-1982
Removing all servers from a service creates invalid persisted configurations
Allow switchover with empty new master server
Save all SlaveStatus rows when querying SHOW ALL SLAVES STATUS
The Gtid-class should save all triplets
SHOW DATABASES only returns lowercase results
Provide clear documentation of Monitor operationMXS-1632
Maxscale Workload Statistics By Server By readwrite available via REST API
Changing masters does not close readwritesplit session immediately
MaxAdmin command history saves duplicates of a command
Check server version, type and monitorability only when (re)connecting
Move maxscale_shutdown into the core
Query hint to route queries to last used server
Intermittent client connection hang due to packet-timing-sensitive defect in multi-statement response handling
Fix ASAN issues in Cache test programs.
Failover/switchover does not disable/enable server eventsMXS-1971
performance test builders
Router reconfiguration
Master is taken from last row of SHOW ALL SLAVES STATUS
Create test that verifies MXS-1774
Make it easier to test with local MaxScale
Provide detailed information why a slave is not promoted to master
Turn server maintenance bit into separate variable
Refactor avrorouter
Support Galera cluster nodes as masters for Binlog Router
support of prepared statement (stmt) for schemarouter
Escape newlines in qlafilter output
Transaction replay
Delayed query retry
Retry failed writes
Master migration
Master reconnection
Compacted session command history
Create on-demand caching that can easily be enabled
Binlog Filter documentation
Aggressive read write split
Support PAM group mapping, like MariaDB Server does
Unknown global parameters aren't detected
Generalize worker mechanism and MessageQueue for other threads than routing threads.
Create more master re-connection tests
Resultset concatenation router
Failed rejoin can lead to a slave with read-only OFF
Can't change maxscale_id for replication_heartbeat
Existing ReadWriteSplit connection can't find master after the master changes
Compile all protocols as C++
Compile all monitors as C++
multi_loaddata command connection closed by mxs
Fix output of "[Total|Average] event queue length"
GTID values not updated if only slaves are present
Allow "terminating" filters to reply via normal response processing chain.