Session commands that will not return a response can be completed
immediately. This requires some special code in the readwritesplit Backend
class implementation as well as a small addition to the Backend class
itself.
Since not all commands expect a response from the server, the queued query
routing function needs some adjustment. The routing of queued queries
should be attempted until a command which expects a response is found or
the queue is empty.
By properly handling these types of session commands, the router can
enable the execution of COM_STMT_CLOSE and COM_STMT_RESET on all
servers. This will prevent resource leakages in the server and allow
proper handling of COM_STMT type command.
If the internal ID is stored in the buffer when it is moving inside the
readwritesplit router, the RWBackend can manage the execution of all
commands with a statement ID by replacing the stored ID with the correct
value.
When a COM_STMT_EXECUTE or a COM_STMT_SEND_LONG_DATA command is executed,
the query type of the prepared statement is used. This allows read-only
prepared statements to be load balanced across slaves.
Mapping the handles returned to the client to a session command ID allows
the mapping of client handle to the backend specific handle. Currently,
the mapping is used for diagnostic output only.
The mxs_mysql_extract_ps_response function extracts the binary protocol
prepared statement components and stores them in a common structure.
The mxs_mysql_get_command extracts the command byte from a GWBUF object
containing a complete MySQL packet.
The default sql mode must now be provided explicitly when the query
classifier is setup. This is in preparation for "sql_mode" becoming
a global configuration parameter of MaxScale.
The current sql mode will be tracked in that variable. Since one
thread will handle multiple sessions, we need to track the sql mode
of a session, so that the sql mode of the query classifier can be
set whenever we handle a request of a particular session.
As the sql-mode will be managed by the client protocol, the data
can be private to it as well.
EXPLAIN statements are no longer parsed completely as doing so makes
it hard to modify the grammar for the needs or Oracle SQL.
Consequently, for an EXPLAIN statement you now bascially only get the
type and the operation (the newly added QUERY_OP_EXPLAIN and QUERY_OP_SHOW).
The other information is not interesting and is related to
information_schema and similar tables.
Several filters use a match-or-ignore logic with two regular
expressions when filtering queries. This commit adds a convenience
function for this task. Also adds a convenience function for reading
several regular expression parameters at once, compiling them and
saving the code while checking for errors.
Also, use the new functions in QLA and CCR filters.
Changed the query operation enum to contain implicit enum values instead
of providing. The operation was never used as a bitmask so it is pointless
to declare them as such.
Added the EXECUTE type to the enum and used it in qc_sqlite and
qc_mysqlembedded.
Providing helper functions for the commonly used parts of the server makes
code easier to read. It also removes any possibility for formatting
problems by moving the URI and name string handling inside the Backend
class.
Added helper functions that check various server states. This makes the
readwritesplit code easier to read as the function names convey the
intention better than the macro invokations.
As the session commands are always appended to the end of the list, the
name should reflect that action. For this reason, the function was renamed
to append_session_command.
Readwritesplit supports replacement of slave servers by storing all
executed session commands in a list. To make the copying of this list a
bit cleaner, an overload for a list of session commands was added. This
will allow relatively smooth addition of server replacement to all router
modules that use the Backend class.
Making the function allows higher level checking to be done by the derived
class.
The readwritesplit does some of the reply bookkeeping for session commands
in the function. This makes their execution less prone for errors as the
states are always updated correctly whenever a session command is
executed.
The class now allows simpler construction of session commands by
overloading the add_session_command with a version that accepts a const
reference to a shared pointer. This removes the need to copy the
references to the source buffer by calling gwbuf_clone.
Exposed the first session command as a const reference to allow
interaction with it. Currently, it is planned to be used to get the
session command position of each backend.
This is the first step to taking the Backend class into use. It is now
used in rwsplit_select_backends.cc and readwritesplit.hh. The module is
not yet functional and doesn't even compile.
Added some helper functions to the Backend class to get easier access to
the server referenced by the SERVER_REF and to check the state of the
backend.
The states are now internal to the Backend class. This simplifies the use
of the class by moving the burder of state tracking to the class
itself.
Refactored the way the schemarouter uses the Backend class.
Also fixed a memory leak in the schemarouter when `ignore_databases_regex`
was used..
Using the same implementation of the Backend and SessionCommand classes in
both schemarouter and readwritesplit will prevent duplication of code.
This commit only splits the generic parts of the class to a Backend class
which the schemarouter then extends. The session commands for both routers
are similar so they require no special handling.
The default database can now be manipulated with a set of functions
exposed by the maxscale/protocol/mysql.h header. This removes the need to
handle the structures themselves in the modules and is a step towards
moving the dcb->data contents inside the session.
That allows the version to be updated and read atomically. If
major/minor/patch are stored as separate variables, you can get an
inconsistent set. Now it may be out of date by the time it is used,
but it will never be internally inconsistent.
The behaviour of the query classifier needs to change depending
on the version of the servers of a service. With this function
"some" version of the service can be obtained.
The behaviour of the query classifier needs to be different
depending on the actual version of the server. There is already
a human readable string, but for programmatic use it needs to
be in a format that can easily be parsed.
With this change, it is no longer possible that the server version
is deallocated at the very moment it is read. There is still a race,
but it's mostly harmless.
The function now returns the compiled code or null or error. Also,
instead of capcount, it writes the ovector size. The ovector pointer
can be null. Removed the other config_get_compiled_regex()-function
and replaced it with the refactored function since they are almost
identical.
The module commands can now produce JSON formatted output which is passed
to the caller. The output should conform to the JSON API as closely as
possible.
Currently, the REST API wraps all JSON produced by module commands inside
a meta-object of the following type:
{
"meta": <output of module command>
}
This allows the output to be JSON API conformant without modifying the
modules and allows incremental updates to code.
The regex strings are compiled automatically. During file parsing,
the string is compiled once using default settings to check that
the pattern is valid. Once a module asks for the compiled pcre2_code,
the pattern is compiled again with module given settings.
The regex string in the config file should be enclosed within '/ ... /'
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.
