一、主备双集群部署
架构：主集群：一主一备；灾备集群：一主一备
1. 安装集群 1
主节点：192.168.0.141 备节点：192.168.0.145
xml 文件参考：
2. 安装集群 2
主节点：192.168.0.142 备节点：192.168.0.143
xml 文件参考：
3. 在主集群创建容灾用户
gsql -d postgres -p PORT -c "create user DR_USERNAME with replication password 'DR_PASSWORD';"
4. 将集群 1 启动为主集群
gs_sdr -t start -m primary -X XMLFILE [-U DR_USERNAME [-W DR_PASSWORD]] [--time-out=SECS]
5. 将集群 2 启动为备集群
gs_sdr -t start -m disaster_standby -U USER -W PASSWORD -X XMLFILE [-U DR_USERNAME [-W DR_PASSWORD]] [--time-out=SECS]
6. 查询
gs_sdr -t query
主集群：
备集群：
以上：
-U 容灾用户名，-W 容灾用户密码，为可选参数，如果未携带需要在搭建过程中交互输入；
–time-out 超时时间，为可选参数，默认为 1200 秒。
二、搭建容灾关系前后参数对比
1. 方案
搭建前后将流复制相关参数导入文件，使用 diff 对比
2. 步骤
gsql -p 26000 -d postgres -c "select name,reset_val from pg_settings where category ='Write-Ahead Log / Settings'">settings1.log
gsql -p 26000 -d postgres -c "select name,reset_val from pg_settings where category ='Write-Ahead Log / Checkpoints'" >>settings1.log
gsql -p 26000 -d postgres -c "select name,reset_val from pg_settings where category ='Replication / Sending Servers'" >>settings1.log
gsql -p 26000 -d postgres -c "select name,reset_val from pg_settings where category ='Resource usage / Recovery'" >>settings1.log
gsql -p 26000 -d postgres -c "select name,reset_val from pg_settings where category ='Replication / Master Server'" >>settings1.log
gsql -p 26000 -d postgres -c "select name,reset_val from pg_settings where category ='Replication / Standby Servers'" >>settings1.log
- 对比结果
前
< replconninfo1| localhost=192.168.0.141 localport=26001 localheartbeatport=26005 localservice=26004 remotehost=192.168.0.145 remoteport=26001 remoteheartbeatport=26005 remoteservice=26004
< replconninfo2|
< replconninfo3|
后
> replconninfo1| localhost=192.168.0.141 localport=26001 localheartbeatport=26005 localservice=26004 remotehost=192.168.0.145 remoteport=26001 remoteheartbeatport=26005 remoteservice=26004 iscascade=true iscrossregion=false
> replconninfo2| localhost=192.168.0.141 localport=26001 localheartbeatport=26005 localservice=26004 remotehost=192.168.0.142 remoteport=26001 remoteheartbeatport=26005 remoteservice=26004 iscascade=false iscrossregion=true
> replconninfo3| localhost=192.168.0.141 localport=26001 localheartbeatport=26005 localservice=26004 remotehost=192.168.0.143 remoteport=26001 remoteheartbeatport=26005 remoteservice=26004 iscascade=false iscrossregion=true
前
< wal_keep_segments | 16
后
> wal_keep_segments | 128
注：主集群搭建后 wal_keep_segments=128,备集群搭建后 wal_keep_segments=16
前
< most_available_sync | off
后
> most_available_sync | on
后（新增）
opgs1com/opgs2com
stream_cluster_run_mode = 'cluster_primary'
opgs3com/opgs4com
stream_cluster_run_mode = 'cluster_standby'
