Files
MaxScale/include/maxscale/monitor.h
Markus Mäkelä a196420c2d Move monitor script processing and launching into the core
This removes parts of the nearly identical code from all monitors.

The removal of monitor type specific event checking is done based on the
assumption that only the monitor that is monitoring the server can be the
cause for a state change. This removes the need to actually check that the
state change is relevant for each monitor and allows the event handling to
be moved into the core.
2017-01-11 14:21:57 +02:00

332 lines
12 KiB
C

#pragma once
/*
* Copyright (c) 2016 MariaDB Corporation Ab
*
* Use of this software is governed by the Business Source License included
* in the LICENSE.TXT file and at www.mariadb.com/bsl.
*
* Change Date: 2019-07-01
*
* On the date above, in accordance with the Business Source License, use
* of this software will be governed by version 2 or later of the General
* Public License.
*/
/**
* @file monitor.h The interface to the monitor module
*
* @verbatim
* Revision History
*
* Date Who Description
* 07/07/13 Mark Riddoch Initial implementation
* 25/07/13 Mark Riddoch Addition of diagnotics
* 23/05/14 Mark Riddoch Addition of routine to find monitors by name
* 23/05/14 Massimiliano Pinto Addition of defaultId and setInterval
* 23/06/14 Massimiliano Pinto Addition of replicationHeartbeat
* 28/08/14 Massimiliano Pinto Addition of detectStaleMaster
* 30/10/14 Massimiliano Pinto Addition of disableMasterFailback
* 07/11/14 Massimiliano Pinto Addition of setNetworkTimeout
* 19/02/15 Mark Riddoch Addition of monitorGetList
* 19/11/15 Martin Brampton Automation of event and name declaration, absorption
* of what was formerly monitor_common.h
*
* @endverbatim
*/
#include <maxscale/cdefs.h>
#include <mysql.h>
#include <maxscale/server.h>
#include <maxscale/dcb.h>
#include <maxscale/log_manager.h>
#include <maxscale/resultset.h>
#include <maxscale/config.h>
#include <maxscale/externcmd.h>
#include <maxscale/secrets.h>
#include <maxscale/modinfo.h>
MXS_BEGIN_DECLS
/**
* The "Module Object" for a monitor module.
*
* The monitor modules are designed to monitor the backend databases that the gateway
* connects to and provide information regarding the status of the databases that
* is used in the routing decisions.
*
* startMonitor is called to start the monitoring process, it is called on the main
* thread of the gateway and is responsible for creating a thread for the monitor
* itself to run on. This should use the entry points defined in the thread.h
* header file rather than make direct calls to the operating system thrading libraries.
* The return from startMonitor is a void * handle that will be passed to all other monitor
* API calls.
*
* stopMonitor is responsible for shuting down and destroying a monitor, it is called
* with the void * handle that was returned by startMonitor.
*/
struct monitor;
typedef struct monitor MONITOR;
typedef struct
{
void *(*startMonitor)(MONITOR *monitor, const CONFIG_PARAMETER *params);
void (*stopMonitor)(MONITOR *monitor);
void (*diagnostics)(DCB *, const MONITOR *);
} MONITOR_OBJECT;
/**
* The monitor API version number. Any change to the monitor module API
* must change these versions usign the rules defined in modinfo.h
*/
#define MONITOR_VERSION {3, 0, 0}
/** Monitor's poll frequency */
#define MON_BASE_INTERVAL_MS 100
/**
* Monitor state bit mask values
*/
typedef enum
{
MONITOR_STATE_ALLOC = 0x00,
MONITOR_STATE_RUNNING = 0x01,
MONITOR_STATE_STOPPING = 0x02,
MONITOR_STATE_STOPPED = 0x04,
MONITOR_STATE_FREED = 0x08
} monitor_state_t;
/**
* Monitor network timeout types
*/
typedef enum
{
MONITOR_CONNECT_TIMEOUT = 0,
MONITOR_READ_TIMEOUT = 1,
MONITOR_WRITE_TIMEOUT = 2
} monitor_timeouts_t;
/*
* Results of attempt at database connection for monitoring
*/
typedef enum
{
MONITOR_CONN_OK,
MONITOR_CONN_REFUSED,
MONITOR_CONN_TIMEOUT
} connect_result_t;
#define MON_ARG_MAX 8192
#define DEFAULT_CONNECT_TIMEOUT 3
#define DEFAULT_READ_TIMEOUT 1
#define DEFAULT_WRITE_TIMEOUT 2
#define MONITOR_RUNNING 1
#define MONITOR_STOPPING 2
#define MONITOR_STOPPED 3
#define MONITOR_INTERVAL 10000 // in milliseconds
#define MONITOR_DEFAULT_ID 1UL // unsigned long value
#define MAX_MONITOR_USER_LEN 512
#define MAX_MONITOR_PASSWORD_LEN 512
/** Monitor events */
typedef enum
{
UNDEFINED_EVENT = 0,
MASTER_DOWN_EVENT = (1 << 0), /**< master_down */
MASTER_UP_EVENT = (1 << 1), /**< master_up */
SLAVE_DOWN_EVENT = (1 << 2), /**< slave_down */
SLAVE_UP_EVENT = (1 << 3), /**< slave_up */
SERVER_DOWN_EVENT = (1 << 4), /**< server_down */
SERVER_UP_EVENT = (1 << 5), /**< server_up */
SYNCED_DOWN_EVENT = (1 << 6), /**< synced_down */
SYNCED_UP_EVENT = (1 << 7), /**< synced_up */
DONOR_DOWN_EVENT = (1 << 8), /**< donor_down */
DONOR_UP_EVENT = (1 << 9), /**< donor_up */
NDB_DOWN_EVENT = (1 << 10), /**< ndb_down */
NDB_UP_EVENT = (1 << 11), /**< ndb_up */
LOST_MASTER_EVENT = (1 << 12), /**< lost_master */
LOST_SLAVE_EVENT = (1 << 13), /**< lost_slave */
LOST_SYNCED_EVENT = (1 << 14), /**< lost_synced */
LOST_DONOR_EVENT = (1 << 15), /**< lost_donor */
LOST_NDB_EVENT = (1 << 16), /**< lost_ndb */
NEW_MASTER_EVENT = (1 << 17), /**< new_master */
NEW_SLAVE_EVENT = (1 << 18), /**< new_slave */
NEW_SYNCED_EVENT = (1 << 19), /**< new_synced */
NEW_DONOR_EVENT = (1 << 20), /**< new_donor */
NEW_NDB_EVENT = (1 << 21), /**< new_ndb */
} monitor_event_t;
static const MXS_ENUM_VALUE monitor_event_enum_values[] =
{
{"master_down", MASTER_DOWN_EVENT},
{"master_up ", MASTER_UP_EVENT},
{"slave_down ", SLAVE_DOWN_EVENT},
{"slave_up ", SLAVE_UP_EVENT},
{"server_down ", SERVER_DOWN_EVENT},
{"server_up ", SERVER_UP_EVENT},
{"synced_down ", SYNCED_DOWN_EVENT},
{"synced_up ", SYNCED_UP_EVENT},
{"donor_down ", DONOR_DOWN_EVENT},
{"donor_up ", DONOR_UP_EVENT},
{"ndb_down ", NDB_DOWN_EVENT},
{"ndb_up ", NDB_UP_EVENT},
{"lost_master ", LOST_MASTER_EVENT},
{"lost_slave ", LOST_SLAVE_EVENT},
{"lost_synced ", LOST_SYNCED_EVENT},
{"lost_donor ", LOST_DONOR_EVENT},
{"lost_ndb ", LOST_NDB_EVENT},
{"new_master ", NEW_MASTER_EVENT},
{"new_slave ", NEW_SLAVE_EVENT},
{"new_synced ", NEW_SYNCED_EVENT},
{"new_donor ", NEW_DONOR_EVENT},
{"new_ndb ", NEW_NDB_EVENT},
{NULL}
};
/** Default value for the `events` parameter */
static const char MONITOR_EVENT_DEFAULT_VALUE[] = "master_down,master_up,slave_down,"
"slave_up,server_down,server_up,synced_down,synced_up,donor_down,donor_up,"
"ndb_down,ndb_up,lost_master,lost_slave,lost_synced,lost_donor,lost_ndb,"
"new_master,new_slave,new_synced,new_donor,new_ndb";
/**
* The linked list of servers that are being monitored by the monitor module.
*/
typedef struct monitor_servers
{
SERVER *server; /**< The server being monitored */
MYSQL *con; /**< The MySQL connection */
bool log_version_err;
int mon_err_count;
unsigned int mon_prev_status;
unsigned int pending_status; /**< Pending Status flag bitmap */
struct monitor_servers *next; /**< The next server in the list */
} MONITOR_SERVERS;
/**
* Representation of the running monitor.
*/
struct monitor
{
char *name; /**< The name of the monitor module */
char user[MAX_MONITOR_USER_LEN]; /*< Monitor username */
char password[MAX_MONITOR_PASSWORD_LEN]; /*< Monitor password */
SPINLOCK lock;
CONFIG_PARAMETER* parameters; /*< configuration parameters */
MONITOR_SERVERS* databases; /*< List of databases the monitor monitors */
monitor_state_t state; /**< The state of the monitor */
int connect_timeout; /**< Connect timeout in seconds for mysql_real_connect */
int read_timeout; /**< Timeout in seconds to read from the server.
* There are retries and the total effective timeout
* value is three times the option value.
*/
int write_timeout; /**< Timeout in seconds for each attempt to write to the server.
* There are retries and the total effective timeout value is
* two times the option value.
*/
MONITOR_OBJECT *module; /**< The "monitor object" */
char *module_name; /**< Name of the monitor module */
void *handle; /**< Handle returned from startMonitor */
size_t interval; /**< The monitor interval */
bool created_online; /**< Whether this monitor was created at runtime */
volatile bool server_pending_changes;
/**< Are there any pending changes to a server?
* If yes, the next monitor loop starts early. */
struct monitor *next; /**< Next monitor in the linked list */
};
extern MONITOR *monitor_alloc(char *, char *);
extern void monitor_free(MONITOR *);
extern MONITOR *monitor_find(const char *);
extern bool monitorAddServer(MONITOR *mon, SERVER *server);
extern void monitorRemoveServer(MONITOR *mon, SERVER *server);
extern void monitorAddUser(MONITOR *, char *, char *);
extern void monitorAddParameters(MONITOR *monitor, CONFIG_PARAMETER *params);
extern bool monitorRemoveParameter(MONITOR *monitor, const char *key);
extern void monitorStop(MONITOR *);
extern void monitorStart(MONITOR *, void*);
extern void monitorStopAll();
extern void monitorStartAll();
extern void monitorShowAll(DCB *);
extern void monitorShow(DCB *, MONITOR *);
extern void monitorList(DCB *);
extern void monitorSetInterval (MONITOR *, unsigned long);
extern bool monitorSetNetworkTimeout(MONITOR *, int, int);
extern RESULTSET *monitorGetList();
extern bool check_monitor_permissions(MONITOR* monitor, const char* query);
monitor_event_t mon_get_event_type(MONITOR_SERVERS* node);
const char* mon_get_event_name(MONITOR_SERVERS* node);
void monitor_clear_pending_status(MONITOR_SERVERS *ptr, int bit);
void monitor_set_pending_status(MONITOR_SERVERS *ptr, int bit);
bool mon_status_changed(MONITOR_SERVERS* mon_srv);
bool mon_print_fail_status(MONITOR_SERVERS* mon_srv);
connect_result_t mon_connect_to_db(MONITOR* mon, MONITOR_SERVERS *database);
void mon_log_connect_error(MONITOR_SERVERS* database, connect_result_t rval);
void mon_log_state_change(MONITOR_SERVERS *ptr);
void lock_monitor_servers(MONITOR *monitor);
void release_monitor_servers(MONITOR *monitor);
void servers_status_pending_to_current(MONITOR *monitor);
void servers_status_current_to_pending(MONITOR *monitor);
/**
* @brief Handle state change events
*
* This function should be called by all monitors at the end of each monitoring
* cycle. This will log state changes and execute any scripts that should be executed.
*
* @param monitor Monitor object
* @param script Script to execute or NULL for no script
* @param events Enabled events
*/
void mon_process_state_changes(MONITOR *monitor, const char *script, uint64_t events);
/**
* @brief Hangup connections to failed servers
*
* Injects hangup events for DCB that are connected to servers that are down.
*
* @param monitor Monitor object
*/
void mon_hangup_failed_servers(MONITOR *monitor);
/**
* @brief Serialize the servers of a monitor to a file
*
* This partially converts @c monitor into an INI format file. Only the servers
* of the monitor are serialized. This allows the monitor to keep monitoring
* the servers that were added at runtime even after a restart.
*
* NOTE: This does not persist the complete monitor configuration and requires
* that an existing monitor configuration is in the main configuration file.
* Changes to monitor parameters are not persisted.
*
* @param monitor Monitor to serialize
* @return False if the serialization of the monitor fails, true if it was successful
*/
bool monitor_serialize_servers(const MONITOR *monitor);
/**
* @brief Serialize a monitor to a file
*
* This converts the static configuration of the monitor into an INI format file.
*
* @param monitor Monitor to serialize
* @return True if serialization was successful
*/
bool monitor_serialize(const MONITOR *monitor);
/**
* Check if a server is being monitored and return the monitor.
* @param server Server that is queried
* @return The monitor watching this server, or NULL if not monitored
*/
MONITOR* monitor_server_in_use(const SERVER *server);
MXS_END_DECLS