The connection creation is now internal to RWSplitSession. This makes the
code more readable by removing the need to pass parameters and allowing
easier reuse of existing functions. The various conditions require to
create connections are now also checked in only one place.
Readwritesplit now picks the best available master if no open master
connection is available. This is required if the server rank is to be
taken into account when master selection is done.
Given the following query:
PREPARE ps FROM 'PREPARE ps2 FROM \'SELECT 1\'';
The debug assertion is hit even though this is valid, albeit unsupported,
SQL. An optimization would be to ignore the query if the prepared
statement type is another prepared statement.
When a single GWBUF was split into two with gwbuf_split, the new GWBUF
would point to the start of the shared data and the old one to the
end. Data-wise, this is fine but as the parsing info for queries is stored
in the shared buffer it causes problems when multiple packets get read in
one network payload. The end result would be that only the first query in
the lot would get parsed and the rest would get the same classification as
the first one.
To properly fix this without the need to deep clone the buffer would
require a reorganization of the buffer mechanism in MaxScale.
This commit alone doesn't fix the queued query routing problems in
readwritesplit. The commit from 2.2 which fixes the ordering problems with
queued queries is also required for a fully functional queued query
mechanism.
If a routing of a queued query caused it to be put back on the query
queue, the order in which the queue was reorganized was wrong. The first
query would get appended as the last query which caused the order to be
reversed.
Th discarding of connections in maintenance mode must be done after any
results have been written to them. This prevents closing of the connection
before the actual result is returned.
The candidate selection code used default values that would cause reads
past buffers. The code could also dereference the end iterator which
causes undefined behavior.
The parameters are now written in the order they appear in the module
parameter definitions. Also enabled a previously disabled part in
server unit test.
This fixes some situations where MaxAdmin/MaxCtrl would block and wait
until a monitor operation or tick is complete. This also fixes a deadlock
caused by calling monitor diagnostics inside a monitor script.
Concurrency is enabled by adding one mutex per server object to protect
array-like fields from concurrent reading/writing.
Previously, runtime monitor modifications could directly alter monitor fields,
which could leave the text-form parameters and reality out-of-sync. Also,
the configure-function was not called for the entire monitor-object, only the
module-implementation.
Now, all modifications go through the overridden configure-function, which calls the
base-class function. As most configuration changes are given in text-form, this
removes the need for specific setters. The only exceptions are the server add/remove
operations, which must modify the text-form serverlist.
Was deprecated in 2.3. Similar features are in MariaDB-Monitor.
One system test was modified to use MariaDB-Monitor instead. Some parts
of the test are disabled for now to make it pass.
Queries in the query queue need to be explicitly parsed since they are
stored in a single buffer and thus share the query classification
information. In the next major version this should be changed into an
array of individual buffers instead of a shared buffer.
Always storing runtime configuration changes prevents problems when the
change causes another parameter to change. One example of this is
transaction_replay that implicitly enables other parameters.
If a DCB was closed and a hangup event was sent to it via
dcb_hangup_foreach shortly after it was closed, the DCB would still
receive it even if it was closed. To prevent this, events must only be
delivered to DCBs if they haven't been closed.
If a session command is executed when lazy_connect is enabled and no
connections have been created, a connection must be made. This makes sure
that the session isn't closed and that the client receives a response.
The lazy connection creation reduces the burden that short sessions place
on the backend servers. This also prevents the problems caused by early
disconnections that happen when only one server is used but multiple
connections are created. This does not solve the problem (MXS-619) but it
does mitigate it to acceptable levels.
This commit also adds a change to the weighting algorithm that prefers
existing connections over unopened ones. This helps avoid the
flip-flopping that happens when the absolute scores are very similar. The
hard-coded value might need to be tuned once testing is done.
The protocol should not track the session state as the parsing is quite
expensive with the current code. This change is a workaround that enables
the parsing only when required. A proper way to handle this would be to do
all the response processing in one place thus avoiding the duplication of
work.
Given the fact that there exist only three possible categories, the map
can be replaced with a static array that needs no memory
allocations. Making this array thread-local allows it to be reused which
places an upper limit on the number of memory allocations.
