MySQL高可用工具—orchestrator discovery 实例代码浅析

搬砖DBA 2021-04-21

1324

对于MySQL高可用方案，当前比较火的一个技术产品是orchestrator，本文不打算介绍它的整体功能及架构，相关内容可以在https://github.com/openark/orchestrator上查看手册。本文会详细介绍orchestrator discovery功能的大体流程，帮助大家理解orchestrator是怎么对一个实例、集群进行信息收集和维护的。

database_instance表

CREATE TABLE `database_instance` (
  `hostname` varchar(128) NOT NULL,
  `port` smallint(5) unsigned NOT NULL,
  `last_checked` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP,
  `last_attempted_check` timestamp NOT NULL DEFAULT '1971-01-01 00:00:00',
  `last_check_partial_success` tinyint(3) unsigned NOT NULL,
  `last_seen` timestamp NULL DEFAULT NULL,
  `uptime` int(10) unsigned NOT NULL,
  `server_id` int(10) unsigned NOT NULL,
  `server_uuid` varchar(64) NOT NULL,
  `version` varchar(128) NOT NULL,
  `binlog_server` tinyint(3) unsigned NOT NULL,
  `read_only` tinyint(3) unsigned NOT NULL,
  `binlog_format` varchar(16) NOT NULL,
  `log_bin` tinyint(3) unsigned NOT NULL,
  `log_slave_updates` tinyint(3) unsigned NOT NULL,
  `binary_log_file` varchar(128) NOT NULL,
  `binary_log_pos` bigint(20) unsigned NOT NULL,
  `master_host` varchar(128) NOT NULL,
  `master_port` smallint(5) unsigned NOT NULL,
  `slave_sql_running` tinyint(3) unsigned NOT NULL,
  `slave_io_running` tinyint(3) unsigned NOT NULL,
  `replication_sql_thread_state` tinyint(4) NOT NULL DEFAULT '0',
  `replication_io_thread_state` tinyint(4) NOT NULL DEFAULT '0',
  `has_replication_filters` tinyint(3) unsigned NOT NULL,
  `oracle_gtid` tinyint(3) unsigned NOT NULL,
  `master_uuid` varchar(64) NOT NULL,
  `ancestry_uuid` text NOT NULL,
  `executed_gtid_set` text NOT NULL,
  `gtid_purged` text NOT NULL,
  `gtid_errant` text NOT NULL,
  `supports_oracle_gtid` tinyint(3) unsigned NOT NULL,
  `mariadb_gtid` tinyint(3) unsigned NOT NULL,
  `pseudo_gtid` tinyint(3) unsigned NOT NULL,
  `master_log_file` varchar(128) NOT NULL,
  `read_master_log_pos` bigint(20) unsigned NOT NULL,
  `relay_master_log_file` varchar(128) NOT NULL,
  `exec_master_log_pos` bigint(20) unsigned NOT NULL,
  `relay_log_file` varchar(128) NOT NULL,
  `relay_log_pos` bigint(20) unsigned NOT NULL,
  `last_sql_error` text NOT NULL,
  `last_io_error` text NOT NULL,
  `seconds_behind_master` bigint(20) unsigned DEFAULT NULL,
  `slave_lag_seconds` bigint(20) unsigned DEFAULT NULL,
  `sql_delay` int(10) unsigned NOT NULL,
  `allow_tls` tinyint(3) unsigned NOT NULL,
  `num_slave_hosts` int(10) unsigned NOT NULL,
  `slave_hosts` text NOT NULL,
  `cluster_name` varchar(128) NOT NULL,
  `suggested_cluster_alias` varchar(128) NOT NULL,
  `data_center` varchar(32) NOT NULL,
  `physical_environment` varchar(32) NOT NULL,
  `instance_alias` varchar(128) NOT NULL,
  `semi_sync_enforced` tinyint(3) unsigned NOT NULL,
  `replication_depth` tinyint(3) unsigned NOT NULL,
  `is_co_master` tinyint(3) unsigned NOT NULL,
  `replication_credentials_available` tinyint(3) unsigned NOT NULL,
  `has_replication_credentials` tinyint(3) unsigned NOT NULL,
  `version_comment` varchar(128) NOT NULL DEFAULT '',
  `major_version` varchar(16) NOT NULL,
  `binlog_row_image` varchar(16) NOT NULL,
  `last_discovery_latency` bigint(20) NOT NULL,
  `semi_sync_master_enabled` tinyint(3) unsigned NOT NULL,
  `semi_sync_replica_enabled` tinyint(3) unsigned NOT NULL,
  `gtid_mode` varchar(32) NOT NULL,
  PRIMARY KEY (`hostname`,`port`),
  KEY `cluster_name_idx_database_instance` (`cluster_name`),
  KEY `last_checked_idx_database_instance` (`last_checked`),
  KEY `last_seen_idx_database_instance` (`last_seen`),
  KEY `master_host_port_idx_database_instance` (`master_host`,`master_port`),
  KEY `suggested_cluster_alias_idx_database_instance` (`suggested_cluster_alias`)
) ENGINE=InnoDB DEFAULT CHARSET=ascii

这个表是orchestrator的一个非常重要的表，用来保存探测实例的信息，为后续判断实例是否存活提供基数据。

探测机制

func ContinuousDiscovery() {
  //省略若干行代码
  go handleDiscoveryRequests()
  //省略若干行代码
}

ContinuousDiscovery()函数是主要功能入口函数，它实现了异步发现、实例信息收集、探测实例失败、执行拓扑恢复等功能。handleDiscoveryRequests()实现并发执行discover的功能函数。

// handleDiscoveryRequests iterates the discoveryQueue channel and calls upon
// instance discovery per entry.
func handleDiscoveryRequests() {
  discoveryQueue = discovery.CreateOrReturnQueue("DEFAULT")
  // create a pool of discovery workers
  for i := uint(0); i < config.Config.DiscoveryMaxConcurrency; i++ {
    go func() {
      for {
        instanceKey := discoveryQueue.Consume()
        // Possibly this used to be the elected node, but has
        // been demoted, while still the queue is full.
        if !IsLeaderOrActive() {
          log.Debugf("Node apparently demoted. Skipping discovery of %+v. "+
            "Remaining queue size: %+v", instanceKey, discoveryQueue.QueueLen())
          discoveryQueue.Release(instanceKey)
          continue
        }
        DiscoverInstance(instanceKey)
        discoveryQueue.Release(instanceKey)
      }
    }()
  }
}

会创建一个discoveryQueue进行discover，开启DiscoveryMaxConcurrency（默认300）个并发执行。

instanceKey := discoveryQueue.Consume()从队列中取一个instanceKey，

type InstanceKey struct {
  Hostname string
  Port     int
}

instanceKey是InstanceKey类型的结构体对象，包括Hostname、Port成员。取得instanceKey后会传递给DiscoverInstance(instanceKey)进行处理。

// DiscoverInstance will attempt to discover (poll) an instance (unless
// it is already up to date) and will also ensure that its master and
// replicas (if any) are also checked.
func DiscoverInstance(instanceKey inst.InstanceKey) {
  //省略若干行代码
  // First we've ever heard of this instance. Continue investigation:
  instance, err = inst.ReadTopologyInstanceBufferable(&instanceKey, config.Config.BufferInstanceWrites, latency)


  // Investigate replicas:
  for _, replicaKey := range instance.SlaveHosts.GetInstanceKeys() {
    replicaKey := replicaKey // not needed? no concurrency here?
    if replicaKey.IsValid() {
      discoveryQueue.Push(replicaKey)
    }
  }
  // Investigate master:
  if instance.MasterKey.IsValid() {
    discoveryQueue.Push(instance.MasterKey)
  }
}

分析DiscoverInstance中重要的代码，ReeadTopologyInstanceBufferable会返回Instance类型数据，它是一个实例的具体描述。详细成员字段内容请参考https://github.com/openark/orchestrator/blob/master/go/inst/instance.go Instance结构体。

type Instance struct {
  Key                          InstanceKey
  InstanceAlias                string
  Uptime                       uint
  ServerID                     uint
  ServerUUID                   string
  Version                      string
  //省略若干行
}

获得一个instance对象后，会把这个instance下的slaves和master依次Push到队列中，以方便对集群相关成员进行discover，这也是为什么我们把一个集群的单个实例加入orchestrator后，它很快就能发现其他实例信息的原因。

ReadTopologyInstanceBufferable这个函数的内容比较多，我们分块介绍些。详细内容请参考：

https://github.com/openark/orchestrator/blob/master/go/inst/instance_dao.go

func ReadTopologyInstanceBufferable(instanceKey *InstanceKey, bufferWrites bool, latency *stopwatch.NamedStopwatch) (*Instance, error) {
    //省略若干行
    //初始化instance
    instance := NewInstance()
    
    
    //把实例的Uptime保存到instance.Uptime字段
    var dummy string
    // show global status works just as well with 5.6 & 5.7 (5.7 moves variables to performance_schema)
    err = db.QueryRow("show global status like 'Uptime'").Scan(&dummy, &instance.Uptime) 
    
    
    //把实例的hostname、server_id、version等信息存入instance对象
    var mysqlHostname, mysqlReportHost string
    err = db.QueryRow("select @@global.hostname, ifnull(@@global.report_host, ''), @@global.server_id, @@global.version, @@global.version_comment, @@global.read_only, @@global.binlog_format, @@global.log_bin, @@global.log_slave_updates").Scan(&mysqlHostname, &mysqlReportHost, &instance.ServerID, &instance.Version, &instance.VersionComment, &instance.ReadOnly, &instance.Binlog_format, &instance.LogBinEnabled, &instance.LogSlaveUpdatesEnabled)
    if err != nil {
      goto Cleanup
    }
    
    
    //如果开启了LogBinEnabled 则把logfile、logpos、Executed_Gtid_Set信息存入instance实例
    go func() {
        defer waitGroup.Done()
            err = sqlutils.QueryRowsMap(db, "show master status", func(m sqlutils.RowMap) error {
            var err error
            instance.SelfBinlogCoordinates.LogFile = m.GetString("File")
            instance.SelfBinlogCoordinates.LogPos = m.GetInt64("Position")
            instance.ExecutedGtidSet = m.GetStringD("Executed_Gtid_Set", "")
            return err
         })
     }()
     
     
     //如果实例开启了半同步，则把半同步信息保存到instance
     waitGroup.Add(1)
     go func() {
         defer waitGroup.Done()
         err = sqlutils.QueryRowsMap(db, "show global status like 'rpl_semi_sync_%_status'", funsqlutils.RowMap) error {
             if m.GetString("Variable_name") == "Rpl_semi_sync_master_status" {
             instance.SemiSyncMasterEnabled = (m.GetString("Value") == "ON")
              } else if m.GetString("Variable_name") == "Rpl_semi_sync_slave_status" {
              instance.SemiSyncReplicaEnabled = (m.GetString("Value") == "ON")
            }
          return nil
        })
     }()
      
      
      //读取实例中gtid_mode、gtid_purged等信息存入instance对象。
      go func() {
        defer waitGroup.Done()
        var masterInfoRepositoryOnTable bool
        // Stuff only supported on Oracle MySQL >= 5.6
        // ...
        // @@gtid_mode only available in Orcale MySQL >= 5.6
        // Previous version just issued this query brute-force, but I don't like errors being issued where they shouldn't.
        _ = db.QueryRow("select @@global.gtid_mode, @@global.server_uuid, @@global.gtid_purged, @@global.master_info_repository = 'TABLE', @@global.binlog_row_image").Scan(&instance.GTIDMode, &instance.ServerUUID, &instance.GtidPurged, &masterInfoRepositoryOnTable, &instance.BinlogRowImage)
        if instance.GTIDMode != "" && instance.GTIDMode != "OFF" {
          instance.SupportsOracleGTID = true
        }
        if config.Config.ReplicationCredentialsQuery != "" {
          instance.ReplicationCredentialsAvailable = true
        } else if masterInfoRepositoryOnTable {
          _ = db.QueryRow("select count(*) > 0 and MAX(User_name) != '' from mysql.slave_master_info").Scan(&instance.ReplicationCredentialsAvailable)
        }
      }()
     
     
     //执行show slave status信息，把相关内容保存到instance实例 
    err = sqlutils.QueryRowsMap(db, "show slave status", func(m sqlutils.RowMap) error {
      instance.HasReplicationCredentials = (m.GetString("Master_User") != "")
      instance.Slave_IO_Running = (m.GetString("Slave_IO_Running") == "Yes")
      if isMaxScale110 {
        // Covering buggy MaxScale 1.1.0
        instance.Slave_IO_Running = instance.Slave_IO_Running && (m.GetString("Slave_IO_State") == "Binlog Dump")
      }
      instance.Slave_SQL_Running = (m.GetString("Slave_SQL_Running") == "Yes")
      instance.ReadBinlogCoordinates.LogFile = m.GetString("Master_Log_File")
      instance.ReadBinlogCoordinates.LogPos = m.GetInt64("Read_Master_Log_Pos")
      instance.ExecBinlogCoordinates.LogFile = m.GetString("Relay_Master_Log_File")
      instance.ExecBinlogCoordinates.LogPos = m.GetInt64("Exec_Master_Log_Pos")
      instance.IsDetached, _ = instance.ExecBinlogCoordinates.ExtractDetachedCoordinates()
      instance.RelaylogCoordinates.LogFile = m.GetString("Relay_Log_File")
      instance.RelaylogCoordinates.LogPos = m.GetInt64("Relay_Log_Pos")
      instance.RelaylogCoordinates.Type = RelayLog
      instance.LastSQLError = emptyQuotesRegexp.ReplaceAllString(strconv.QuoteToASCII(m.GetString("Last_SQL_Error")), "")
      instance.LastIOError = emptyQuotesRegexp.ReplaceAllString(strconv.QuoteToASCII(m.GetString("Last_IO_Error")), "")
      instance.SQLDelay = m.GetUintD("SQL_Delay", 0)
      instance.UsingOracleGTID = (m.GetIntD("Auto_Position", 0) == 1)
      instance.UsingMariaDBGTID = (m.GetStringD("Using_Gtid", "No") != "No")
      instance.MasterUUID = m.GetStringD("Master_UUID", "No")
      instance.HasReplicationFilters = ((m.GetStringD("Replicate_Do_DB", "") != "") || (m.GetStringD("Replicate_Ignore_DB", "") != "") || (m.GetStringD("Replicate_Do_Table", "") != "") || (m.GetStringD("Replicate_Ignore_Table", "") != "") || (m.GetStringD("Replicate_Wild_Do_Table", "") != "") || (m.GetStringD("Replicate_Wild_Ignore_Table", "") != ""))
      masterHostname := m.GetString("Master_Host")
      if isMaxScale110 {
        // Buggy buggy maxscale 1.1.0. Reported Master_Host can be corrupted.
        // Therefore we (currently) take @@hostname (which is masquarading as master host anyhow)
        masterHostname = maxScaleMasterHostname
      }
      masterKey, err := NewResolveInstanceKey(masterHostname, m.GetInt("Master_Port"))
      if err != nil {
        logReadTopologyInstanceError(instanceKey, "NewResolveInstanceKey", err)
      }
      masterKey.Hostname, resolveErr = ResolveHostname(masterKey.Hostname)
      if resolveErr != nil {
        logReadTopologyInstanceError(instanceKey, fmt.Sprintf("ResolveHostname(%q)", masterKey.Hostname), resolveErr)
      }
      instance.MasterKey = *masterKey
      instance.IsDetachedMaster = instance.MasterKey.IsDetached()
      instance.SecondsBehindMaster = m.GetNullInt64("Seconds_Behind_Master")
      if instance.SecondsBehindMaster.Valid && instance.SecondsBehindMaster.Int64 < 0 {
        log.Warningf("Host: %+v, instance.SecondsBehindMaster < 0 [%+v], correcting to 0", instanceKey, instance.SecondsBehindMaster.Int64)
        instance.SecondsBehindMaster.Int64 = 0
      }
      // And until told otherwise:
      instance.SlaveLagSeconds = instance.SecondsBehindMaster
      instance.AllowTLS = (m.GetString("Master_SSL_Allowed") == "Yes")
      // Not breaking the flow even on error
      slaveStatusFound = true
      return nil
    })    
 
    Cleanup:
      waitGroup.Wait()
      //省略若干行
      if instanceFound {
        instance.LastDiscoveryLatency = time.Since(readingStartTime)
        instance.IsLastCheckValid = true
        instance.IsRecentlyChecked = true
        instance.IsUpToDate = true
        latency.Start("backend")
        if bufferWrites {
          enqueueInstanceWrite(instance, instanceFound, err)
        } else {
          WriteInstance(instance, instanceFound, err)
        }
        WriteLongRunningProcesses(&instance.Key, longRunningProcesses)
        lastAttemptedCheckTimer.Stop()
        latency.Stop("backend")
        return instance, nil
      }
    
      // Something is wrong, could be network-wise. Record that we
      // tried to check the instance. last_attempted_check is also
      // updated on success by writeInstance.
      latency.Start("backend")
      _ = UpdateInstanceLastChecked(&instance.Key, partialSuccess)
      latency.Stop("backend")
      return nil, err
      //完整内容请参考https://github.com/openark/orchestrator/blob/master/go/inst/instance_dao.go      
}

将instance对象信息保存到后端database_instance表。

*************************** 1. row ***************************
                         hostname: 192.168.1.1
                             port: 3306
                     last_checked: 2021-01-24 12:11:41
             last_attempted_check: 2021-01-24 12:11:41
       last_check_partial_success: 1
                        last_seen: 2021-01-24 12:11:41
                           uptime: 42599071
                        server_id: 48301
                      server_uuid: 16db875e-e197-11e8-9c8f-b4969111dff0
                          version: 5.6.32-78.0-log
                    binlog_server: 0
                        read_only: 0
                    binlog_format: ROW
                          log_bin: 1
                log_slave_updates: 1
                  binary_log_file: mysql-bin.000075
                   binary_log_pos: 583135957
                      master_host:
                      master_port: 0
                slave_sql_running: 0
                 slave_io_running: 0
     replication_sql_thread_state: -1
      replication_io_thread_state: -1
          has_replication_filters: 0
                      oracle_gtid: 0
                      master_uuid:
                    ancestry_uuid: 16db875e-e197-11e8-9c8f-b4969111dff0
                executed_gtid_set: 16db875e-e197-11e8-9c8f-b4969111dff0:1-86855534
                      gtid_purged: 16db875e-e197-11e8-9c8f-b4969111dff0:1-83262861
                      gtid_errant:
             supports_oracle_gtid: 1
                     mariadb_gtid: 0
                      pseudo_gtid: 0
                  master_log_file:
              read_master_log_pos: 0
            relay_master_log_file:
              exec_master_log_pos: 0
                   relay_log_file:
                    relay_log_pos: 0
                   last_sql_error:
                    last_io_error:
            seconds_behind_master: NULL
                slave_lag_seconds: NULL
                        sql_delay: 0
                        allow_tls: 0
                  num_slave_hosts: 2
                      slave_hosts: [{"Hostname":"192.168.1.2","Port":3306},{"Hostname":"192.168.1.3","Port":3306}]
                     cluster_name: 192.168.1.1:3306
          suggested_cluster_alias: mysql_test
                      data_center: RW
             physical_environment:
                   instance_alias: bjlf-mysql-001.bjlf.com
               semi_sync_enforced: 0
                replication_depth: 0
                     is_co_master: 0
replication_credentials_available: 0
      has_replication_credentials: 0
                  version_comment: Percona Server (GPL), Release 78.0, Revision 8a8e016
                    major_version: 5.6
                 binlog_row_image: FULL
           last_discovery_latency: 10706003
         semi_sync_master_enabled: 0
        semi_sync_replica_enabled: 0
                        gtid_mode: ON

last_checked、last_attempted_check、last_check_partial_success、last_seen是判断实例健康状态的几个重要字段。

last_check_partial_success：从上文中可以看到ReadTopologyInstanceBufferable函数会连接实例获得很多的实例信息，如果函数连接到实例之前异常，那么这个字段为0，如果能连接到实例能获得部分或全部信息这个字段值为1。

last_attempted_check：ReadTopologyInstanceBufferable函数开始执行的时候会对instanceKey合法性进行判断，不合法情况下会更新last_attempted_check时间戳为当前时间，返回err信息为：ReadTopologyInstance will not act on invalid instance key。

另一个地方更新此字段的地方是ReadTopologyInstanceBufferable做了以下定义：

  lastAttemptedCheckTimer := time.AfterFunc(time.Second, func() {
    go UpdateInstanceLastAttemptedCheck(instanceKey)
  })

lastAttemptedCheckTimer会延迟执行，不过如果ReadTopologyInstanceBufferable快于1秒内执行并且能正常discover到实例信息的话，lastAttemptedCheckTimer.Stop()会暂停更新last_attempted_check字段。

ReadTopologyInstanceBufferable函数中label Cleanup 代码块定义了相关逻辑，如果能对一个实例正常discover的情况下会执行WriteInstance函数把获得的instance实例信息保存到后端database_instance表里面。保存的过程中会一次性把last_checked、last_attempted_check、last_seen更新成当前时间，把last_check_partial_success更新成1。

总结

orchestrator会间隔InstancePollSeconds（默认5s）连接实例上拉取实例状态，并将这些状态信息存入orchestrator的元数据库的orchestrator.database_instance表中。获得的状态信息有（NOT MaxScale）：

show global status like 'Uptime'
select @@global.hostname, ifnull(@@global.report_host, ''), @@global.server_id, @@global.version, @@global.version_comment, @@global.read_only, @@global.binlog_format, @@global.log_bin, @@global.log_slave_updates
show master status
show global status like 'rpl_semi_sync_%_status'
select @@global.gtid_mode, @@global.server_uuid, @@global.gtid_executed, @@global.gtid_purged, @@global.master_info_repository = 'TABLE', @@global.binlog_row_image
show slave status
select count(*) > 0 and MAX(User_name) != '' from mysql.slave_master_info
show slave hosts
select substring_index(host, ':', 1) as slave_hostname from information_schema.processlist where command IN ('Binlog Dump', 'Binlog Dump GTID')
SELECT SUBSTRING_INDEX(@@hostname, '.', 1)

获得到实例信息后，通过下面语句将状态值存入到orchestrator的元数据库中：

INSERT INTO database_instance
                (hostname, port, last_checked, last_attempted_check, last_check_partial_success, uptime, server_id, server_uuid, version, major_version, version_comment, binlog_server, read_only, binlog_format, binlog_row_image, log_bin, log_slave_updates, binary_log_file, binary_log_pos, master_host, master_port, slave_sql_running, slave_io_running, replication_sql_thread_state, replication_io_thread_state, has_replication_filters, supports_oracle_gtid, oracle_gtid, master_uuid, ancestry_uuid, executed_gtid_set, gtid_mode, gtid_purged, gtid_errant, mariadb_gtid, pseudo_gtid, master_log_file, read_master_log_pos, relay_master_log_file, exec_master_log_pos, relay_log_file, relay_log_pos, last_sql_error, last_io_error, seconds_behind_master, slave_lag_seconds, sql_delay, num_slave_hosts, slave_hosts, cluster_name, suggested_cluster_alias, data_center, region, physical_environment, replication_depth, is_co_master, replication_credentials_available, has_replication_credentials, allow_tls, semi_sync_enforced, semi_sync_master_enabled, semi_sync_replica_enabled, instance_alias, last_discovery_latency, last_seen)
        VALUES
                ('192.168.1.1', 3306, NOW(), NOW(), 1, 42599071, 48301, '16db875e-e197-11e8-9c8f-b4969111dff0', '5.6.32-78.0-log', '5.6', 'Percona Server (GPL), Release 78.0, Revision 8a8e016', 0, 1, 'ROW', 'FULL', 1, 1, 'mysql-bin.000075', 583135957, '', 0, 1, 1, 1, 1, 0, 1, 1, '16db875e-e197-11e8-9c8f-b4969111dff0', '16db875e-e197-11e8-9c8f-b4969111dff0:1-86855534', '16db875e-e197-11e8-9c8f-b4969111dff0:1-83262861', 'ON', '', '', 0, 0, '', 0, '', , '', 137056344, '', '', 0, 0, 0, 1, '[{\"Hostname\":\"192.168.1.2\",\"Port\":3306}]', '192.168.1.1:3306', 'bjlf-mysql-001.bjlf.com', '', '', '', 1, 1, 1, 1, 0, 0, 0, 0, '', 10706003, NOW())
        ON DUPLICATE KEY UPDATE
                hostname=VALUES(hostname), port=VALUES(port), last_checked=VALUES(last_checked), last_attempted_check=VALUES(last_attempted_check), last_check_partial_success=VALUES(last_check_partial_success), uptime=VALUES(uptime), server_id=VALUES(server_id), server_uuid=VALUES(server_uuid), version=VALUES(version), major_version=VALUES(major_version), version_comment=VALUES(version_comment), binlog_server=VALUES(binlog_server), read_only=VALUES(read_only), binlog_format=VALUES(binlog_format), binlog_row_image=VALUES(binlog_row_image), log_bin=VALUES(log_bin), log_slave_updates=VALUES(log_slave_updates), binary_log_file=VALUES(binary_log_file), binary_log_pos=VALUES(binary_log_pos), master_host=VALUES(master_host), master_port=VALUES(master_port), slave_sql_running=VALUES(slave_sql_running), slave_io_running=VALUES(slave_io_running), replication_sql_thread_state=VALUES(replication_sql_thread_state), replication_io_thread_state=VALUES(replication_io_thread_state), has_replication_filters=VALUES(has_replication_filters), supports_oracle_gtid=VALUES(supports_oracle_gtid), oracle_gtid=VALUES(oracle_gtid), master_uuid=VALUES(master_uuid), ancestry_uuid=VALUES(ancestry_uuid), executed_gtid_set=VALUES(executed_gtid_set), gtid_mode=VALUES(gtid_mode), gtid_purged=VALUES(gtid_purged), gtid_errant=VALUES(gtid_errant), mariadb_gtid=VALUES(mariadb_gtid), pseudo_gtid=VALUES(pseudo_gtid), master_log_file=VALUES(master_log_file), read_master_log_pos=VALUES(read_master_log_pos), relay_master_log_file=VALUES(relay_master_log_file), exec_master_log_pos=VALUES(exec_master_log_pos), relay_log_file=VALUES(relay_log_file), relay_log_pos=VALUES(relay_log_pos), last_sql_error=VALUES(last_sql_error), last_io_error=VALUES(last_io_error), seconds_behind_master=VALUES(seconds_behind_master), slave_lag_seconds=VALUES(slave_lag_seconds), sql_delay=VALUES(sql_delay), num_slave_hosts=VALUES(num_slave_hosts), slave_hosts=VALUES(slave_hosts), cluster_name=VALUES(cluster_name), suggested_cluster_alias=VALUES(suggested_cluster_alias), data_center=VALUES(data_center), region=VALUES(region), physical_environment=VALUES(physical_environment), replication_depth=VALUES(replication_depth), is_co_master=VALUES(is_co_master), replication_credentials_available=VALUES(replication_credentials_available), has_replication_credentials=VALUES(has_replication_credentials), allow_tls=VALUES(allow_tls), semi_sync_enforced=VALUES(semi_sync_enforced), semi_sync_master_enabled=VALUES(semi_sync_master_enabled), semi_sync_replica_enabled=VALUES(semi_sync_replica_enabled), instance_alias=VALUES(instance_alias), last_discovery_latency=VALUES(last_discovery_latency), last_seen=VALUES(last_seen)

后期会针对orchestrator探测实例失败的流程进行分析，敬请期待。

mysql

文章转载自搬砖DBA，如果涉嫌侵权，请发送邮件至：contact@modb.pro进行举报，并提供相关证据，一经查实，墨天轮将立刻删除相关内容。

MySQL高可用工具—orchestrator discovery 实例代码浅析

探测机制

总结

评论