整体顺序:
关闭:gpaas——nginx——k8s——公共组件(zk、redis、mq、elk、文件预览)——PGmonitor——PG从——PG主——mdd——nfs
开机:nfs——nginx——公共组件(zk、redis、mq、elk、文件预览)———PGmonitor——PG主——PG从——k8s——mdd——gpaas
1、gpaas直接关机:shutdown
172.16.29.147
2、nginx检查nfs是否开机自动挂载,没有的话设置一下,直接关机:shutdown
172.16.29.142
172.16.29.141
3、k8s检查nfs是否开机自动挂载,没有的话设置一下,直接关机:shutdown
172.16.29.139
172.16.29.138
172.16.29.137
4、公共组件直接关机:shutdown
172.16.29.132
172.16.29.131
172.16.29.130
172.16.29.140
172.16.29.133
PG主从判断方法:
登录PGmonitor服务器
su - postgres
pg_autoctl show state read-write是主库,read-only是从库
5、PGmonitor关机:systemctl stop pg_mon,shutdown
172.16.29.134
6、PG从(假定是135,以查询为准):systemctl stop postgresql,shutdown
172.16.29.135
7、PG主(假定是136,以查询为准):systemctl stop postgresql,shutdown
172.16.29.136
8、MDD:查看vip在哪一台上,假定是143,先停从,后停主
172.16.29.143(主)
172.16.29.144(从)
172.16.29.144(从) 执行systemctl stop keepalived,systemctl disable keepalived
172.16.29.143(主) 执行systemctl stop keepalived
172.16.29.143(主) 执行systemctl stop mdd
172.16.29.144(从) 执行systemctl stop mdd
两台机器shutdown
9、nfs直接关机:shutdown
172.16.29.129
***************************************************************************************************************************
开机:nfs——nginx——公共组件(zk、redis、mq、elk、文件预览)———PGmonitor——PG主——PG从——k8s——mdd——gpaas
1、nfs开机,需要等nfs完全启动
172.16.29.129
2、nginx开机,检查nfs是否能自动挂载,没有挂载则手工挂载
172.16.29.142
172.16.29.141
3、公共组件直接开机,开机之后检查各服务(zk、redis、mq、elk、文件预览)状态是否正常,没问题之后继续再往下开机
172.16.29.132
172.16.29.131
172.16.29.130
172.16.29.140
172.16.29.133
zk集群:
netstat -nlp|grep 2181
/home/kingdee/common/zookeeper/zookeeper-3.5.9/bin/zkServer.sh status
redis节点和哨兵:
netstat -npl|egrep '6379|6380|6381|7505|7506|7507'
cd /home/kingdee/common/redis/redis-5.0.13/src
./redis-cli -h 172.16.29.132 -p 6379 -a Cosmic@1358 INFO|grep role
./redis-cli -h 172.16.29.131 -p 6379 -a Cosmic@1358 INFO|grep role
./redis-cli -h 172.16.29.130 -p 6379 -a Cosmic@1358 INFO|grep role
./redis-cli -h 172.16.29.132 -p 6380 -a Cosmic@1358 INFO|grep role
./redis-cli -h 172.16.29.131 -p 6380 -a Cosmic@1358 INFO|grep role
./redis-cli -h 172.16.29.130 -p 6380 -a Cosmic@1358 INFO|grep role
./redis-cli -h 172.16.29.132 -p 6381 -a Cosmic@1358 INFO|grep role
./redis-cli -h 172.16.29.131 -p 6381 -a Cosmic@1358 INFO|grep role
./redis-cli -h 172.16.29.130 -p 6381 -a Cosmic@1358 INFO|grep role
./redis-cli -h 172.16.29.132 -p 7505 -a Cosmic@1358 info Sentinel |grep master0
./redis-cli -h 172.16.29.131 -p 7505 -a Cosmic@1358 info Sentinel |grep master0
./redis-cli -h 172.16.29.130 -p 7505 -a Cosmic@1358 info Sentinel |grep master0
./redis-cli -h 172.16.29.132 -p 7506 -a Cosmic@1358 info Sentinel |grep master0
./redis-cli -h 172.16.29.131 -p 7506 -a Cosmic@1358 info Sentinel |grep master0
./redis-cli -h 172.16.29.130 -p 7506 -a Cosmic@1358 info Sentinel |grep master0
./redis-cli -h 172.16.29.132 -p 7507 -a Cosmic@1358 info Sentinel |grep master0
./redis-cli -h 172.16.29.131 -p 7507 -a Cosmic@1358 info Sentinel |grep master0
./redis-cli -h 172.16.29.130 -p 7507 -a Cosmic@1358 info Sentinel |grep master0
rabbitmq集群:
netstat -nlp|grep 5672
rabbitmqctl cluster_status
http://172.16.29.192:8090/mq
用户:admin或者cosmic
密码:Cosmic@1358
ELK:
netstat -nltp|grep 9092
netstat -nltp|grep 9200
netstat -nltp|grep 9600
4、PG-monitor开机
172.16.29.134
5、PG主(假定是136,以查询为准)开机,查询主库状态
172.16.29.136
PG主从判断方法:
登录PGmonitor服务器
su - postgres
pg_autoctl show state read-write是主库,read-only是从库,从库会显示异常
6、PG从(假定是135,以查询为准)开机,查询主从状态
172.16.29.135
PG主从判断方法:
登录PGmonitor服务器
su - postgres
pg_autoctl show state read-write是主库,read-only是从库,主从同步显示正常
7、k8s开机,检查nfs是否能自动挂载,没有挂载则手工挂载
172.16.29.139
172.16.29.138
172.16.29.137
检查k8s集群pod状态:kubectl get pod -A
8、MDD开机:假定主库是143,根据关机前的情况确认
先开:172.16.29.143(主),等待vip显示正常
后开:172.16.29.144(从)
登录web端查看是否正常
http://172.16.29.191:8080/bos-olap-webserver
http://172.16.29.143:8080/bos-olap-webserver
http://172.16.29.144:8080/bos-olap-webserver
admin/Cosmic@1358
9、gpaas开机
172.16.29.147
最后的所有服务器监控主机探针需要手动启动,这个不影响服务运行
推荐阅读
您的鼓励与嘉奖将成为创作者们前进的动力,如果觉得本文还不错,可以给予作者创作打赏哦!
请选择打赏金币数 *