The monitor permission checks didn't use the standard connection timeouts but
used hard-coded values. The config.c tried to connect to the embedded library
but since it is not used by it anymore, it always fails.
The monitor timeouts can now be larger than the monitor interval. This will
allow the combination of low monitoring intervals and large network timeouts.
If a network experiences some periodic lag, it is desirable to allow large
timeout values.
Shutting down monitors is not supposed to be done ofter so there is no true
benefit from keeping the connections open. With the refactoring of the monitor
interface, this can be done in a centralized place.
LOGIF and skygw_write_log removed from server/core/*.c and
replaced with calls to MXS_(ERROR|WARNING|NOTICE|INFO|DEBUG).
This is a mechanism change, no updating of the actual message
has been performed.
Currently this causes a very small performance hit, since the
check whether the priority is enabled or not is performed in
the function that is called and not before the function is called.
Once all LOGIFs and skygw_write_logs have been replaced, the
behaviour will be altered back to what it was.
The log manager variables lm_enabled_log_files_bitmask, log_ses_count
and tls_log_info that earlier were declared separately in every
c-file are now declared in the log_manager.h header.
Common variables, like databases, timeouts and interval, and functionality was moved to the MONITOR type.
This reduces the redundant functionality of the monitor API's functions like registerServer and setInterval.
void session_enable_log(SESSION* ses, logfile_id_t id)
and
void session_disable_log(SESSION* ses, logfile_id_t id)
Which switch specific log type on/off if the log type in question is not generally enabled.
Each thread carries a thread-specific struct log_info_t which includes members for current session id and bitfield for enabled log types for the current session. That information is checked before actual log write functions are called.
Each file where session-specific logging is used, must include the following exports:
/** Defined in log_manager.cc */
extern int lm_enabled_logfiles_bitmask;
extern size_t log_ses_count[];
extern __thread log_info_t tls_log_info;
Each monitor loops 10 times/second (sleep 100ms) and perform monitoring checks only when monitor's interval is spent. Monitors notice faster if the shutdown flag is set and thus overall shutdown is faster.
hint.c:added missing header
Changed interval from unsigned long to size_t which is guaranteed to be of same size also in windows (if possible).
Fix for broken replication has been added to mysql_monitor.
Both Slave_IO and Slave_SQL threads must be running in order to assign
the SERVER_SLAVE status but If only Slave_IO is running let’s assign
the master_id to current server and continue building the replication
tree; if no slaves at all the master will be still available.
The “detect_stale_master” option has been added, its default is 0.
If set to 1 the monitor will keep the last detected master even if the
replication setup is completely not working, i.e. both Slave_IO and
Slave_SQL threads are not running: this applies only to the server that
was master before.
After monitor or MaxScale are restarted and the replication is still
stopped or not configured there will be no master because it’s not
possible to compute the replication topology tree.
