The read-write distribution in readwritesplit is now stored in a map
partitioned by the servers that the router has used. Currently, the
statistics for removed servers aren't dropped so some filtering still
needs to be added.
The statistic.hh header defines a set of functions that complement the
standard library numeric functions. They differ from the standard library
functions in that they take a container reference and a pointer-to-member
as parameters and calculate the statistic based on the pointed-to member.
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.
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.
By removing the constness of the type for the actual stored value, the
return values of the access operators are defined by the type itself. This
allows both const and non-const types to be stored in a rworker_local
object.
Given that worker.hh was public, it made sense to make routingworker.hh
public as well. This removes the need to include private headers in
modules and allows C++ constructs to be used in C++ code when previously
only the C API was available.
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.
No longer depends on monitor events as the other operations do not
either. The active_event/new_event detection was removed, as these
only protect against a rare situation. A similar feature which
protects all the cluster modifications will be implemented later.
The maxscale_is_shutting_down function is used to detect when MaxScale
should stop. This fixes a race condition in the code where the workers has
not yet been initialized but a termination signal has been received. It
also replaces the misuse of the service_should_stop variable with a proper
function.
qc_thread_init() must now explicitly be called in every thread
and not just in other threads but the one where qc_process_init()
is called.
This change was caused by QC_INIT_SELF initialization actually
being performed in query_classifier.cc. Before this change, there
actually was a leak in the routing worker running in the main
thread, the query classification cache was created twice.
Enabling it will give a performance improvement and it offloads work from
the CPU. This leads to better and more complete utilization of system
resources.
The master server of a monitor can now be extracted with the
mon_get_external_master function. It uses the REST API to find a master
server in the set of servers monitored by a monitor.
When the replication status from the external master is checked, the
pending status must be used. This makes sure that the SlaveStatusArray and
the server state are sync.
Also extended the message that was logged when the external master was
lost. By adding the network address there, it makes it easier to see where
the server was replicating from if only the log file is available.
