All RPATH setting have to be set before generation of any binary.
Defining of CMAKE_INSTALL_RPATH is moved to separate file which is called from main CMakeLists.txt in the beginning, just after install_layout.cmake
Now considers other routing hints if first one fails. The order is inverted compared
to e.g. namedserver filter settings because of how routing hints are stored. If all hints
are unsuccessful, route to any slave.
Lack of Docker can cause build failure
To avoid it if Docker is not available
test is not started, simply returns 0
It is hard to make Docker installation reliable especially for very old or very new distros
Basic tests can be executed with 15 machines Master/slave backend. Tests have label BIG_REPL_BACKEND Default template modified to support big backend. Tests temporaraly labeled as UNSTABLE to prevent their execution nightly
For big test maxscale.cnf is automatically generated for any number of nodes
If the DCB was closed before the handshake for the LocalCliet connection
was received, the gw_decode_mysql_server_handshake would use the closed
DCB to log the connection ID. Clearing out the pointer prevents it.
The DCB callbacks shouldn't be used to send more events as they cause the
callback to be called recursively. The recursive calls caused rows to be
sent before the schemas for the rows were sent. Queuing the events via the
worker mechanism prevents this.
Reduced the default cache size from 40% to 15%. Most cases don't benefit
from that much memory and the defaults have caused problems in live
environments.
If a query spans more than a single packet, it will never be successfully
classified due to the fact that the complete SQL is never available to the
query classifier. For this reason, it is pointless to cache them.
If a monitor was created with an argument that was not a key-value type, a
promise would be rejected outside of the main maxctrl function. Added a
test case that covers this and fixed a few other test coverage problems
that were present.
The STL regex implementations have proven to be unreliable on older
systems and replacing the regex with hand-written code for version
extraction is less prone to break.
The test seems to fail from time to time due to the monitor not picking up
the master state fast enough. To prevent this, the test needs to wait for
at least one monitor interval.
If the monitor setting "replication_master_ssl" is set to on, any CHANGE MASTER TO-command
will have MASTER_SSL=1. If set to off or unset, MASTER_SSL is left unchanged to match existing
behaviour.
Basic tests can be executed with 15 machines Master/slave backend. Tests have label BIG_REPL_BACKEND Default template modified to support big backend. Tests temporaraly labeled as UNSTABLE to prevent their execution nightly
