这两天一直在配置mysql数据库的多主一从的配置。其实现的思路是利用一个数据库启动两个实例（通过mysqld_multi），而这两个实例指向同一个数据库文件目录。但是配置过程中需要注意很多地方。
首先，由于启用了innodb，所以在启动数据服务的时候，服务进程会对innodb数据文件ibdata1加锁，导致第二个实例一致报错，在报了N多个 错误之后，放弃读取数据文件。所以在第二个实例启动时无法读取innodb的表数据，导致这种方案无法同步两个都包含innodb表的镜像。尝试采用指定 数据文件的方式（innodb_data_file_path=ibdata2:98M:autoextend，ibdata2是把ibdata1复制获 得的，注意大小应该和现有的数据文件大小一致，为大于等于数据文件大小的最小整数），启动多个实例，仍然报错，只是错误信息变成了innodb的log文 件无法读取，因此继续设置innodb_log_group_home_dir=test-innodb-log，指定日志文件的存储位置（启动时，提示 无法找到日志文件，自动建立新的日志文件ib_logfile1，ib_logfile0）。启动成功（./bin/mysqld_multi –config-file=./var/my.cnf –mysqld=mysqld_safe start 1,2）。
另外不能重复的还有如下选项：
port = 5507 socket = var/mysql.sock2 pid-file= var/test-db2a.pid
在配置同步镜像时，还有很多需要注意，由于是存在两个同步，很多配置不能采用默认设置，否则两个同步进程会读取同一个文件，互相干扰。有如下几个配置：
主服务器信息文件：master-info-file = test-master.info
relay日志文件：relay-log = test-relay-bin
relay日志索引文件：relay-log-index = test-relay-bin.index
relay日志信息文件：relay-log-info-file=test-relay-log.info
完整的配置文件如下：
[mysqld_multi] mysqld = /home/work/local/test/mysql/bin/mysqld_safe mysqladmin = /home/work/local/test/mysql/bin/mysqladmin # Here follows entries for some specific programs # server 1 [mysqld1] port = 5506 socket = /home/work/local/test/mysql/var/mysql.sock1 skip-locking pid-file= /home/work/local/test/mysql/var/test-db1a.pid datadir = /home/work/local/test/mysql/var log=/home/work/local/test/mysql/var/test-db1.log user = mysql log-slow-queries=/home/work/local/test/mysql/var/slowquery1.log long_query_time = 2 key_buffer = 256M max_allowed_packet = 1M read_buffer_size = 2M myisam_sort_buffer_size = 64M thread_cache = 32 query_cache_size = 32M thread_concurrency = 2 max_connections=500 server-id = 4 master-host = localhost master-user = replicuser master-password = 123456 master-port = 7890 #report-host = test-db3 master-connect-retry = 30 #replicate-rewrite-db=from_name->to_name replicate-do-table = test.usertest master-info-file = test-master.info relay-log = relay1-relay-bin relay-log-index = relay1-relay-bin.index relay-log-info-file=relay1-relay-log.info innodb_data_file_path=ibdata1:98M:autoextend log-slave-updates # # binary logging - not required for slaves, but recommended log-bin # The MySQL server 2[mysqld2] port = 5507 socket = /home/work/local/test/mysql/var/mysql.sock2 skip-locking pid-file= /home/work/local/test/mysql/var/test-db2a.pid datadir = /home/work/local/test/mysql/var log=/home/work/local/test/mysql/var/test-db2.log user = root log-slow-queries=/home/work/local/test/mysql/var/slowquery2.log long_query_time = 2 key_buffer = 256M max_allowed_packet = 1M table_cache = 512 sort_buffer_size = 2M read_buffer_size = 2M myisam_sort_buffer_size = 64M thread_cache = 32 query_cache_size = 32M thread_concurrency = 2 max_connections=500 server-id = 5 master-host = localhost master-user = replicuser master-password = 123456 master-port = 3336 master-connect-retry=60 replicate-do-table = test2.user2 master-info-file = relay2-master.info relay-log = relay2-relay-bin relay-log-index = relay2-relay-bin.index innodb_data_file_path=ibdata2:98M:autoextend innodb_log_group_home_dir=relayt2-innodb-log/ relay-log-info-file=relay2-relay-log.info log-slave-updates # # binary logging - not required for slaves, but recommended log-bin [mysqldump] quick max_allowed_packet = 16M