The MySQL monitor replication_heartbeat table now uses the default storage
engine of the database when creating the table. Most of the time the
default is InnoDB which makes the table crash-safe.
The `monitoruser` and `monitorpw` parameters were mislabeled as `monuser`
and `monpw`. To allow backwards compatibility, the `monuser` and `monpw`
still work as aliases for the correct commands.
The type and name parsing functions could move outside of allocated memory
as they didn't check for the terminating null character. Also fixed the
printf format string used when the list of used tables is being created.
Fixed CDC testing connector to abort on error and added some extra output
to the cdc_datatypes test.
The tests should no longer expect client session creation to fail if the
branch session creation is not successful.
The main session is kept alive so that a failure in the branch session
does not cause a failure of the main session. The main session can be
considered more important as it is what the client will use. If the branch
session fails, the failure will be logged so no information is lost.
Also added a missing configuration file and removed some extra-verbose
output in a test script.
The tee filter iterates over the available ports and finds the first
available network listener and tries to connect. This allows the use of
UNIX domain sockets with services that are targeted by the tee filter.
The tee filter now uses the local client class to clone the queries. This
imposes some restrictions on how the filter can be used but also makes
MaxScale as a whole more robust.
The local client class handles a network connection to a local service. It
is meant to be used with the tee filter so that the internal cloning of
DCBs can be removed.
The local client class connects to a port on the local host. This requires
that the user connecting to the tee filter has adequate grants to connect
locally to a MaxScale service.
After authentication for the local client is complete, the client will
pipe all queries to the service while ignoring any responses it
generates. This could be expanded so that a special handler would be given
as an argument. The handler would then handle the responses to the
queries.
Refactored the backend authentication functions so that they can be
exposed to the tee filter. This allows the tee filter to use the same
functions as the protocol modules use without having to reimplement them
inside the tee filter.
The module command self links now point to an endpoint that executes the
module command. Depending on the type of the module command, either a GET
or a POST request must be made.
Files are saved in GTID repo at creation time: this allows to show
files without transactions via SHOW [FULL] BINARY LOGS
The FULL keywords add domain_id and server id to the output as file
prefix: 0/10122/mysql-bin.000080
A set of the core MaxScale parameters can now be altered at runtime. This
set consists of the authentication timeouts and the admin interface
authentication. Other parameters either can't be modified due to internal
limitations or aren't sensible to modify at runtime.
Standalone events in transaction (no COMMIT event) are now handled:
the GTID is saved to gtid_maps storage.
Transaction detection assumes there is only one query_event after GTID
event with MARIADB_FL_STANDALONE flag set.
Saving GTID components into gtid_maps storage will allow to create a
hiearchical binlog cache dir.
Empty GTID for master registration can be specified with SET
@@global.gtid_slave_pos = ‘’
The template class wraps a HashMap such that only a few operations
are allowed. Usage requires specializing a RegistryTraits class
template for each entry type.
The schema generator program needs to add the real_type and length fields
if the data types define them.
Also fixed a bug where the real_type and length fields were checked for
generated fields.
When a cloned DCB is created, the service pointer is not copied and it
needs to be manually set in the newSession entry point. This most likely
due to the fact that the cloned DCB always has a different service and it
is not possible to deduce it.
Another option would be to pass the target service as a parameter but the
whole DCB cloning process could use with a rewrite so any modifications
beyond the required minimum are wasteful.
If a connection has not been fully established (i.e. authentication has
been completed) then it should not be considered as a connection pool
candidate.
The avro schema allows custom properties to be defined for the schema
fields. The avrorouter stored extra information about the table into the
schema for later use.
Currently, this information is only generated by the avrorouter
itself. Further improvements to the schema generator scripts need to be
done.
When the binlog has been read, it needs to be treated as if the
transaction or row limit has been hit. This will cause all tables to be
flushed to disk before the files are indexed.
When a MariaDB 10.0 DATETIME field with a custom length was defined, the
field offsets weren't calculated properly.
As there is no metadata for pre-10.1 DATETIME types with decimal
precision, the metadata (i.e. decimal count) needs to be gathered from the
CREATE TABLE statement. This information is then used to calculate the
correct field length when the value is decoded.
This change does not fix the incorrect interpretation of the old DATETIME
value. The converted values are still garbled due to the fact that the
value needs to be shifted out of the decimal format before it can be
properly converted.
MariaDB GTID Master registration:
creating missing binlog files (with 4 byes) between current one and the
filename coming from ROTATE_EVENT.
blr_slave_binlog_dump() is also checking possible empty files.
The avro schema allows custom properties to be defined for the schema
fields. The avrorouter stored extra information about the table into the
schema for later use.
Currently, this information is only generated by the avrorouter
itself. Further improvements to the schema generator scripts need to be
done.
When the binlog has been read, it needs to be treated as if the
transaction or row limit has been hit. This will cause all tables to be
flushed to disk before the files are indexed.
