KingbaseES数据库部署及双机热备配置
Dataguard双机热备部署
注:配置期间使用数据库的属主用户进行操作
1.1 数据一致性准备
(1)修改数据库的配置文件参数:(2阶段初始化的数据库实例)
创建在归档日志存放目录(需要根据机器的存储空间进行创建,空间大于100GB):
mkdir /datas_local/archive vim /DBDATA/data/kingbase.conf 打开归档: log_archive_start = on log_archive_dest = ‘/datas_local/archive‘ --目录根据操作系统服务器存储设置
备机只读设置:
hot_standby=on wal_level=‘hot_standby‘
使用kingbase用户操作
1、将从机指定数据目录下的(/DBDATA)data目录删除; 2、将主机指定数据目录下的(/DBDATA)data目录打包压缩:tar –czvf data.tbz data 3、将主机指定数据目录下的(/home/kingbase/KingbaseES)的data.tbz复制到从机指定路径下(/DBDATA)。 主机IP:10.150.73.182 从机IP:10.150.73.187 4、 在从机上指定目录(/DBDATA)解压数据文件data.tbz,生成本地数据目录:tar -xzvf data.tbz; (2) 增加ifconfig文件的SUID和SGID,使普通用户运行ifconfig时暂时具备root权限。 chmod u+s /sbin/ifconfig chmod g+s /sbin/ifconfig
1.2 Dataguard配置文件准备
1、先配置主机,再配置从机;
将/home/kingbase/KingbaseES/bin/template/standby.conf分别拷贝到主备数据库安装目录的config下,并进行修改; A机外网IP:10.150.73.182 A机私网IP:10.10.10.1 B机外网IP:10.150.73.187 B机私网IP:10.10.10.2 浮动ip:10.150.73.189
主机OS配置:
Dataguard和kingbase运行在普通用户模式下,普通用户没有设置和删除IP的权限,IP浮动过程中,需要设置和删除IP,因此,如果需要浮动IP功能,需要增加普通用户设置和删除IP的权限。方法示例:
增加ifconfig文件的SUID和SGID,使普通用户运行ifconfig时暂时具备root权限,使用root用户执行:
chmod u+s /sbin/ifconfig chmod g+s /sbin/ifconfig
注:IP地址随环境准备配置
主机
[standby] managed = false is_debug = false log_timestamp = true local_server_port = 54322 remote_server_ip = 10.10.10.2 remote_server_port = 54322 log_dir = ‘/home/kingbase/KingbaseES/log‘ standby_log = "standby-%w.log" server_mode = ‘primary‘ kingbase_fault_num = 0 enable_float_ip = true floatip_readonly_addcmd = ‘/sbin/ifconfig eth0:0 10.150.73.188/24‘ floatip_readonly_delcmd = ‘/sbin/ifconfig eth0:0 down‘ floatip_readwrite_addcmd = ‘/sbin/ifconfig eth0:0 10.150.73.189/24‘ floatip_readwrite_delcmd = ‘/sbin/ifconfig eth0:0 down‘ [kingbase] managed = false lifecheck_interval = 2 kingbase_log = "kingbase-%w.log" bin_dir = ‘/home/kingbase/KingbaseES/bin‘ data_dir = ‘/DBDATA/data‘ log_archive_dest = ‘/datas_local/archive‘ primary_conninfo = ‘host=10.10.10.2 port=54321‘ parallel_recover_threads = 1 local_kingbase_cmdport = 54323 protection_mode = maxprotection --同步模式(maxperformance异步模式) send_taillog = true retrytimes = 0 [network] managed = true startlevel = 99 stoplevel = 0 checkalive = ‘touch /DBDATA/kingbase/kingbase.txt && ping -c 1 -w 1 10.150.73.190 1>/dev/null 2>/dev/null‘ checkinterval = 5 checkfails = 3 hotstandby = false startcmd = "" stopcmd = "" needswitch = true retrytimes = 0 retrycmd = "" [klogminer] managed = false startlevel = 99 stoplevel = 0 checkalive = "kdb_logminerd status" checkinterval = 3 checkfails = 1 hotstandby = false startcmd = "kdb_logminerd start $MODE$" stopcmd = "kdb_logminerd stop $PID$" needswitch = true retrytimes = 0 retrycmd = "kdb_logminerd restart $FAILEDTIMES$"
备机
[standby] managed = false is_debug = false log_timestamp = true local_server_port = 54322 remote_server_ip = 10.10.10.1 remote_server_port = 54322 log_dir = ‘/home/kingbase/KingbaseES/log‘ standby_log = "standby-%w.log" server_mode = ‘standby‘ kingbase_fault_num = 0 enable_float_ip = true floatip_readonly_addcmd = ‘/sbin/ifconfig eth0:0 10.150.73.188/24‘ floatip_readonly_delcmd = ‘/sbin/ifconfig eth0:0 down‘ floatip_readwrite_addcmd = ‘/sbin/ifconfig eth0:0 10.150.73.189/24‘ floatip_readwrite_delcmd = ‘/sbin/ifconfig eth0:0 down‘ [kingbase] managed = false lifecheck_interval = 2 kingbase_log = "kingbase-%w.log" bin_dir = ‘/home/kingbase/KingbaseES/bin‘ data_dir = ‘/DBDATA/data‘ log_archive_dest = ‘/datas_local/archive‘ primary_conninfo = ‘host=10.10.10.1 port=54321‘ parallel_recover_threads = 1 local_kingbase_cmdport = 54323 protection_mode = maxprotection send_taillog = true retrytimes = 0 [network] managed = true startlevel = 99 stoplevel = 0 checkalive = ‘ touch /DBDATA/kingbase/kingbase.txt && ping -c 1 -w 1 10.150.73.190 1>/dev/null 2>/dev/null ‘ checkinterval = 5 checkfails = 3 hotstandby = false startcmd = "" stopcmd = "" needswitch = true retrytimes = 0 retrycmd = "" [klogminer] managed = false startlevel = 99 stoplevel = 0 checkalive = "kdb_logminerd status" checkinterval = 3 checkfails = 1 hotstandby = false startcmd = "kdb_logminerd start $MODE$" stopcmd = "kdb_logminerd stop $PID$" needswitch = true retrytimes = 0 retrycmd = "kdb_logminerd restart $FAILEDTIMES$"
Dataguard双机集群启动
上述配置工作完成之后,首次启动standby服务,需要先启动主机standby控制器,然后再启动从机standby控制器;
使用kingbase用户进入目录/home/kingbase/KingbaseES/bin设置双机管理用户名\密码:
./standby –UKLS –WKLS ——主备机都需要进行设置 使用root用户将/home/kingbase/KingbaseES/bin/standby7d文件拷贝到主备机的/etc/init.d文件夹下; 先启动主机: 使用root用户:/etc/init.d/standby7d start 启动备机:/etc/init.d/standby7d start istandby登陆: ./istandby –UKLS –WKLS 查看standby状态:get all status 停止控制器监听:stop listener 切换主从状态:swich over 如果配置正确,主从机网络正常。启动主从机standby后,主从机自动识别各自模式状态:主机为primary,从机为standby,通过客户端连接主机数据库,可以连接成功正常访问, 客户端连接从机数据库,不能正常访问,提示“处于standby模式的数据库系统不能接受连接”。
Dataguard观察器
观察器配置部署(主备机都需要配置并启动)
cd /home/kingbase/KingbaseES/config vim observer.conf [observer] server_ip_1 = ‘10.150.73.182‘ ----外网的IP地址 server_ip_2 = ‘10.150.73.187‘ server_port_1 = 54322 server_port_2 = 54322 server_timeout = 10 network_timeout = 10 log_directory = ‘/home/kingbase/KingbaseES/observer.log‘ query_interval = 1 TCP = 54324 启动观察器: ./observer &
本文出自 “岁伏” 博客,请务必保留此出处http://suifu.blog.51cto.com/9167728/1618572
原文地址:http://suifu.blog.51cto.com/9167728/1618572