The additions into the server.h header used C++ language which caused C
programs to fail to compile. Moved the implementation of the EMAverage
class into the private Server class in the server.hh header and exposed it
via functions in the server.h header. Also temporarily moved
almost_equal_server_scores into the public server.hh as there is no
service.hh header.
See script directory for method. The script to run in the top level
MaxScale directory is called maxscale-uncrustify.sh, which uses
another script, list-src, from the same directory (so you need to set
your PATH). The uncrustify version was 0.66.
Changes that allow slow or new servers to quickly apply samples towards the
server average. The most important changes are to not ignore the first N samples,
and apply an average to the server as soon as there is one available.
The new ResponseStat::make_valid() will use filter samples to add an average,
if no averages have yet been added, even if the number of filter samples is less
than the filter limit.
The math becomes simpler when the weight is inverted, i.e. a simple multiplication
to get the (inverse) score. Inverse weights are normalized to the range [0..1] where a lower
number is a higher weight,
The enum select_criteria_t is used to provide a std::function that takes the backends
as vector (rather than the prior pairwise compares) and returns the best backend.
This commit refactors slave selection. The compare is still pair-wise but isolated into a small run_comparison() function. The function get_slave_candidate() is used when new connections are created, which I both moved and modified (had to move due to scoping), so diff is off.
The slave selection for routing: get_slave_backend() now has the filtering logic from old get_slave_backend() and compare_backends(), the latter of which is removed.
Backend functions mostly take shared_ptr<SRWBackend> in various forms (as is, const ref, in a container). Ideally the shared_ptr would be used only to where it is really needed, and either naked ptrs or references to RWBackend would be used. This refactor does not address that issue, but compounds it by using even deeper shared_ptr structures. Fixing that in a future commit.
The main piece of code, slave selection (backend_cmp_response_time), uses the available
method of pair-wise comparison of slaves. This will be changed to selection using all
available slaves, along with removal of hard coded values.
This is to support calculating the average from a session, and the slave selection criteria to be able to route based on averages. This commit, like the next one, have TODOs which you should feel free to comment on. Undecided things.
The configuration updating in readwritesplit was the inspiration for the
mxs::rworker_local type. Due to this, taking it into use simply means that
the type changes from Config to mxs::rworker_local<Config>.
By storing the data gathere by readwritesplit inside the session, the
protocol will be aware of the state of the LOAD DATA LOCAL INFILE
execution. This prevents misinterpretation of the data which previously
led to closed connections, effectively rendering LOAD DATA LOCAL INFILE
unusable.
This change is a temporary solution to a problem that needs to be solved
at the protocol level. The changes required to implement this are too big
to add into a bug fix release.
- The ones that were not used were removed.
- The ones that were used were moved close to the actual type.
In most cases some values were missing and if the definition is
close to the type there is a remote chance that they will stay
in sync. If detached, they surely will not.
The configuration doesn't need to be contained in shared pointer as each
session holds its own version of it. This removes most of the overhead in
configuration reloading. The only thing that's left is any overhead added
by the use of thread-local storage.
By using the worker local data mechanism, data can be efficiently cached
on the local worker. This avoids all synchronization on reads and only
requires synchronization on a configuration update.
As an additional observation, the testing of std::mutex and SPINLOCK shows
that std::mutex far outperforms the MaxScale SPINLOCK even on
non-conflicting workloads.
The causal read queries were performed also when the target server was the
master. The extra functionality of the causal reads is only needed on
slaves.
Adjusted the test case to require GTID replication.
The configuration used the wrong parameter name. The test also did not
explicitly enable tracking of the last_gtid variable which caused it to
fail if it wasn't already on.
When the query queue does not contain a complete packet
(i.e. modutil_get_next_MySQL_packet return NULL), an informative dump of
how many bytes and what is stored is logged.
With the removal of the old session command implementation, the code that
used it can be removed or replaced with newer constructs. As a result, the
backend protocol no longer does any session command processing.
The three buffer types, GWBUF_TYPE_SESCMD_RESPONSE,
GWBUF_TYPE_RESPONSE_END and GWBUF_TYPE_SESCMD as well as their related
macros are no longer used and can be removed.
The protocol could leak memory in rare cases where several commands were
queued at the same time. Readwritesplit also didn't free the memory it
acquired via qc_get_table_names.
Relaced router_options with configuration parameters in the createInstance
router entry point. The same needs to be done for the filter API as barely
any filters use the feature.
Some routers (binlogrouter) still support router_options but using it is
deprecated. This had to be done as their use wasn't deprecated in 2.2.
The transaction migration in the case of a changed master never worked as
transaction replay would only be triggered when the master fails. To cover
this case, the transaction replay just needs to be started when the need
for a transaction migration is detected.
To help diagnose the behavior, the Trx class no longer logs a message when
a transaction is closed. This is now done by readwritesplit which has more
knowledge of the context in which the transaction is closed.
Moved transaction statistics calculations into a member function and
placed all target type specific processing into their respective
functions.
Also inverted the connection keepalive check to also cover hinted queries.
The parameter accepts both counts and percentages which requires special
handling in the router. This needs to be done when the configuration is
updated.
If a transaction is replayed, queued commands must not be processed. The
exception to this rule is when pending session commands are executed
before the first statement in the replayed transaction is executed.
If transaction replaying was enabled and a result was returned in more
than one call to clientReply, a NULL value would be added to the statement
which in turn would trigger a debug assertion.
Similarly any following statements in the transaction would be executed
regardless of whether the result was complete.
Renamed the statement execution function to better describe what it does.
Extended the basic functional test case to cover this.
Added a new router API entry point that allows configuration changes after
the instance has been created. This makes alterations to most service
parameters at runtime possible.
An option to reconfiguration would have been the creation of a new
service and the eventual destruction of the old one. This would be a more
complicated and costly method but from an architectural point of view it
is interesting.
The actual implementation of the configuration change is left to the
router. Currently, only readwritesplit performs reconfiguration as
implementing it with versioned configurations is very easy.
Versioned configurations can be considered an adequate first step but it
is not an optimal solution as it causes a bottleneck in the reference
counting of the shared configuration. Thread-specific configuration
definitions would make for a more efficient solution but the
implementation is more complex.
By using a shared pointer instead of a plain object, we can replace the
router configuration without it affecting existing sessions. This is a
change that is required to enable runtime reconfiguration of
readwritesplit.
When a valid target was not found, no error message was logged by the
router. This would cause the "Routing the query failed. Session will be
closed." message to be logged with no explanation as to why the routing
failed.
In addition to the above-mentioned case, no message would be logged if the
target for a COM_STMT_FETCH was not in use.
