Both the replication lag and the message printing state are saved in SERVER,
although the values are mostly used by readwritesplit. A log message is printed
both when a server goes over the limit and when it comes back below.
Because of concurrency issues, a message may be printed multiple times before
different threads detect the new message state.
Documentation updated to explain the change.
There is a race condition between the addition of the DCB into epoll and
the execution of the event that initiates the protocol pointer for the DCB
and sends the handshake to the client. If a hangup event would occur
before the handshake would be sent, it would be possible that the DCB
would get freed before the code that sends the handshake is executed.
By picking the worker who owns the DCB before the DCB is placed into the
owner's epoll instance, we make sure no events arrive on the DCB while the
control is transferred from the accepting worker to the owning
worker.
When poll_add_dcb was called for a DCB that once was polling system but
was subsequently removed, the DCB would appear twice in the worker's list
of DCBs. This caused a hang when the DCB was the last one in the worker's
list and dcb_foreach_local would be called.
To prevent the aforementioned problem, the DCBs are now added and removed
directly to and from the workers instead of indirectly via poll_add_dcb
and poll_remove_dcb.
If the connection to the master is lost, knowing what type of an error
caused the call to handleError helps deduce what was the real reason for
it. Logging the idle time of the connection helps detect when the
wait_timeout of a connection is exceeded.
The code used a rather questionable method for parsing SQL statements
instead of using the query classifier for detecting transaction start and
stop events.
The process of appending to a std::string always includes a size check in
case the internal storage needs to expand. Given that we know a
canonicalized version of a query string is never larger than the original
string and that we pre-allocate enough memory to cope with the worst-case
scenario, the extra logic in std::string::push_back is unnecessary and an
extra cost. Writing directly into the string avoids this cost and improves
the performance.
Switched from default character type functions to ones that use lookup
tables. Eliminated the internal state and replaced with in-place iteration
of the query. Added code to allow single-lookup detection of normal
characters.
The log message should now contain enough information to more easily debug
any configuration issues that result from updating the main config file
after runtime changes have been done.
Also fixed the theoretical race condition in detection of persisted
configuration files by only checking for their existence once.
By storing the server statistics object in side the session, the lookup
involved in getting a worker-local value is avoided. Since the lookup is
done multiple times for a single query, it is beneficial to store it in
the session.
As the worker-local value is never deleted, it is safe to store a
reference to it in the session. It is also never updated concurrently so
no atomic operations are necessary.
The code now only checks the need for a keepalive ping once every
keepalive interval. Reduced the number of mxs_clock calls to one so that
all servers use the same value.
All global parameters are now handled by the runtime configuration
modification code. The parameters that are trivial to update can now be
updated at runtime. All other global parameters cause a new error message
to be returned stating that the parameter in question cannot be modified
at runtime.
Also updated the list of modifiable parameters in MaxCtrl. This list
should not be stored in MaxCtrl and should be created by MaxScale at
runtime.
The prefix was always added even when the original version would've been
acceptable. For example, a version string of 5.5.40 would get converted to
5.5.5-5.5.40 which is quite confusing for older client applications.
Due to MDEV-15556 and MDEV-15840 recursive CTEs can't be reliably used
with older 10.2 versions. To prevent problems, only use the query that
extracts composite roles with newer versions.
Since we know the worst-case size of a canonical statement is the size of
the query string, we can reduce the number of memory allocations to one in
the get_canonical function.
The information stored for each prepared statement would not be cleared
until the end of the session. This is a problem if the sessions last for a
very long time as the stored information is unused once a COM_STMT_CLOSE
has been received.
In addition to this, the session command response maps were not cleared
correctly if all backends had processed all session commands.
When a response to a prepared statement was processed, the number of EOF
packets was used to see whether the response was complete. This code used
a function that does not work with the special packet returned by a PS
preparation that is similar to an OK packet.
The correct method is to count the total number of packets in the
response.
The deleter for std::unique_ptr<GWBUF> was not included in that file which
caused it to be deleted with the default deleter. The same should apply to
std::unique_ptr<json_t> as well.
Under heavy load some of the basic network operations could fail which led
to some of the allocated memory to leak.
Also the backend protocol never freed the current protocol command if it
was not completed. This would happen if a user executed a session command
as the first command but backend authentication would fail.
The authentication code did not initialize one of the buffers used to
calculate the password hashes. This resulted in the use of uninitialized
memory when the user provided no password.
By resetting the replay state the transaction replay can start again on a
new server. This allows the replay process work when a master server is
shutting down.
By delaying the replay for a second, we give the monitor a small chance to
adap to master failures. It'll also prevent rapid re-querying if multiple
transaction replays are supported.
A transaction that just completed will go through the start_trx_replay
function as from the client protocol's point of view the transaction is
still open. The debug assertion did not take this into account and would
fail if a successful commit was the last thing done on master that failed.
Also fixed the formatting.
If a Galera cluster drops down to a single node, the last node would not
be considered valid. During the failure of the second to last node, the
master would also temporarily lose the master status.
The behavior was changed to always keep the cluster UUID until the cluster
size drops down to zero. This guarantees that the same cluster is used as
long as possible.
When a server is stopping, it'll send an error to the client before
terminating the TCP connection. The code in readwritesplit would detect
this error and create a hangup event on the DCB. This would cause it to
appear as if the TCP connection was broken and the router would
immediately try to reconnect to the same server.
By ignoring the error and allowing the connection to die on its own, we
avoid immediately reconnecting and retrying any transactions on the
stopping server. This increases the chances that the monitor will see it
first and assign the server states correctly before the transaction replay
is attempted.
