炫“库”行动-人大金仓有奖征文-KingbaseES V8R3 读写分离集群在线扩容案例

短命女 2022-09-15 05:41 329阅读 0赞

【本文正在参与炫“库”行动-人大金仓有奖征文】

[CSDN![icon-default.png_t_L892][]https://marketing.csdn.net/p/98bd30353e7cb998b6070a89e8b91edb][CSDN_icon-default.png_t_L892_https_marketing.csdn.net_p_98bd30353e7cb998b6070a89e8b91edb]

**案例说明：**

> 此次测试案例用于kingbaseES V8R3读写分离集群在线扩容测试，主要分为三个步骤。
> 
> 1、 通过sys\_basebackup创建新备库。  
> 2、 将备库加入到Cluster nodes管理，可以用kingbase\_monitor.sh一键启停。  
> 3、 主备复制切换测试。

**操作系统和数据库版本：**

> 1）操作系统环境  
> \[kingbase@\#localhost ~\]$ cat /etc/centos-release  
> CentOS Linux release 7.6.1810 (Core)  
>    
> 2）数据库版本  
> TEST=\# select version();  
>                                                       version  
> \-------------------------------------------------------------------------------  
>  Kingbase V008R003C002B0100 on x86\_64-pc-linux-gnu, compiled by gcc (GCC) 4.1.2 20080704 (Red Hat 4.1.2-46), 64-bit  
> (1 row)

**前期准备：**

> 1、 原有节点，主库(192.168.4.127)、备库(192.168.4.129)，新增节点(192.168.4.159)。  
> 2、 在新增节点安装相同版本的kingbaseES数据库。  
> 3、 配置新增节点到原节点的ssh信任关系（root、kingbase）。  
> 4、 关闭新增节点防火墙和selinux并启动crond服务。

![watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16][]

**一、在线增加数据库复制节点**

**1.1 查看kingbase cluster节点进程**

===首先确定原主备库的Cluster和kingbaseES服务状态正常。===

>   
> 主库：
> 
> kingbase@srv127 bin\]$ ps -ef |grep kingbase  
> kingbase 15845     1  0 15:11 ?        00:00:00 /home/kingbase/cluster/ESHA/db/bin/kingbase -D /home/kingbase/cluster/ESHA/db/data  
> kingbase 15877 15845  0 15:11 ?        00:00:00 kingbase: logger process  
> kingbase 15879 15845  0 15:11 ?        00:00:00 kingbase: checkpointer process  
> kingbase 15880 15845  0 15:11 ?        00:00:00 kingbase: writer process  
> kingbase 15881 15845  0 15:11 ?        00:00:00 kingbase: wal writer process  
> kingbase 15882 15845  0 15:11 ?        00:00:00 kingbase: autovacuum launcher process  
> kingbase 15883 15845  0 15:11 ?        00:00:00 kingbase: archiver process  
> kingbase 15884 15845  0 15:11 ?        00:00:00 kingbase: stats collector process  
> kingbase 15885 15845  0 15:11 ?        00:00:00 kingbase: bgworker: syslogical supervisor  
> kingbase 16851 15845  0 15:11 ?        00:00:00 kingbase: wal sender process SYSTEM 192.168.4.127(52967) streaming 0/180000D0  
> root     17781     1  0 15:11 ?        00:00:00 ./kingbasecluster -n  
> root     17830 17781  0 15:11 ?        00:00:00 kingbasecluster: watchdog  
> root     17895 17781  0 15:11 ?        00:00:00 kingbasecluster: lifecheck  
> root     17897 17895  0 15:11 ?        00:00:00 kingbasecluster: heartbeat receiver  
> root     17898 17895  0 15:11 ?        00:00:00 kingbasecluster: heartbeat sender  
> root     17899 17781  0 15:11 ?        00:00:00 kingbasecluster: wait for connection request  
> root     17900 17781  0 15:11 ?        00:00:00 kingbasecluster: wait for connection request  
> root     17901 17781  0 15:11 ?        00:00:00 kingbasecluster: wait for connection request  
> root     17912 17781  0 15:11 ?        00:00:00 kingbasecluster: wait for connection request  
> root     17913 17781  0 15:11 ?        00:00:00 kingbasecluster: wait for connection request  
> root     17914 17781  0 15:11 ?        00:00:00 kingbasecluster: wait for connection request  
> root     17924 17781  0 15:11 ?        00:00:00 kingbasecluster: PCP: wait for connection request  
> root     17925 17781  0 15:11 ?        00:00:00 kingbasecluster: worker process  
> kingbase 20397 15845  0 15:12 ?        00:00:00 kingbase: SUPERMANAGER\_V8ADMIN TEST 192.168.4.127(53139) idle
> 
> 备库：
> 
> \[kingbase@srv129 bin\]$ ps -ef |grep kingbase  
> kingbase 30708     1  0 15:11 ?        00:00:00 /home/kingbase/cluster/ESHA/db/bin/kingbase -D /home/kingbase/cluster/ESHA/db/data  
> kingbase 30709 30708  0 15:11 ?        00:00:00 kingbase: logger process  
> kingbase 30710 30708  0 15:11 ?        00:00:00 kingbase: startup process   recovering 000000030000000000000018  
> kingbase 30714 30708  0 15:11 ?        00:00:00 kingbase: checkpointer process  
> kingbase 30715 30708  0 15:11 ?        00:00:00 kingbase: writer process  
> kingbase 30716 30708  0 15:11 ?        00:00:00 kingbase: stats collector process  
> kingbase 30995     1  0 10月18 ?      00:00:17 /home/kingbase/cluster/KHA6/KHA/kingbase/bin/kbha -A daemon -f /home/kingbase/cluster/KHA6/KHA/kingbase/bin/../etc/repmgr.conf  
> kingbase 31173 30708  0 15:11 ?        00:00:00 kingbase: wal receiver process   streaming 0/180000D0  
> root     31222     1  0 15:11 ?        00:00:00 ./kingbasecluster -n  
> root     31264 31222  0 15:11 ?        00:00:00 kingbasecluster: watchdog  
> root     31396 31222  0 15:11 ?        00:00:00 kingbasecluster: lifecheck  
> root     31398 31396  0 15:11 ?        00:00:00 kingbasecluster: heartbeat receiver  
> root     31399 31396  0 15:11 ?        00:00:00 kingbasecluster: heartbeat sender  
> root     31400 31222  0 15:11 ?        00:00:00 kingbasecluster: wait for connection request  
> root     31402 31222  0 15:11 ?        00:00:00 kingbasecluster: wait for connection request  
> ……  
> root     31421 31222  0 15:11 ?        00:00:00 kingbasecluster: PCP: wait for connection request  
> root     31422 31222  0 15:11 ?        00:00:00 kingbasecluster: worker process  
> kingbase 32062 30708  0 15:12 ?        00:00:00 kingbase: SUPERMANAGER\_V8ADMIN TEST 192.168.4.127(19332) idle

**1.2 查看主备复制信息**

===保证集群中主备复制状态正常。===

**1.3 在线增加新的备库节点**  
===为保证测试环境和接近于生产环境，在创建备库时，在主库端进行大的事务处理，当备库创建完成后查看是否合主库数据一致。===

> 在主库进行事务处理（模拟生产库）：
> 
> prod=\# insert into t10 values (generate\_series(1,10000000),'tom');  
>    
> prod=\# select count(\*) from t10;  
>   count  
> \----------  
>  11000000  
> (1 row)

**1.3.1 在新备库上通过sys\_basebackup在线创建备库**

**1） 创建备库数据存储目录**  
*===注意：目录位置尽量和其他已存在节点的目录位置一致，这样对于后期配置改动较少。===*

> \[kingbase@srv159 v8r3\]$ mkdir -p /home/kingbase/cluster/ESHA/db/data  
> \[kingbase@srv159 v8r3\]$ chmod 0700 /home/kingbase/cluster/ESHA/db/data

**2）通过sys\_basebackup创建备库**

> \[kingbase@srv159 v8r3\]$sys\_basebackup -h 192.168.4.127 -D /home/kingbase/cluster/ESHA/db/data -F p -x -v -P -U SYSTEM -W 123456 -p 54321  
> 事务日志起始于时间点: 0/B20060D0, 基于时间表3  
> 2079583/2079583 kB (100%), 1/1 表空间  
> transaction log end point: 0/D5C48538  
> sys\_basebackup: base backup completed

**1.3.2 配置备库recovery.conf文件**

> \[kingbase@srv159 data\]$ pwd  
> /home/kingbase/cluster/ESHA/db/data  
> \[kingbase@srv159 data\]$ cat recovery.conf  
> standby\_mode='on'  
> primary\_conninfo='port=54321 host=192.168.4.127 user=SYSTEM password=MTIzNDU2 application\_name=srv159'  
> recovery\_target\_timeline='latest'  
> primary\_slot\_name ='slot\_srv159'

**1.3.3 配置备库kingbase.conf文件（部分内容)**

> \# Add settings for extensions here  
> max\_wal\_senders=4  
> wal\_keep\_segments=256  
> hot\_standby\_feedback=on  
> max\_prepared\_transactions=100  
> port=54321  
> control\_file\_copy='/home/kingbase/cluster/ESHA/template.bk'  
> fsync=on  
> wal\_log\_hints=on  
> archive\_mode=on  
> hot\_standby=on  
> wal\_level=replica  
> \#\#\#synchronous\_standby\_names='1(srv129)'  
> archive\_dest='/home/kingbase/cluster/ESHA/archivedir/'  
> max\_replication\_slots=4  
> log\_directory='/home/kingbase/cluster/ESHA/db/data/sys\_log/'

**1.3.4 配置所有节点sys\_hba.conf文件**

> \# "local" is for Unix domain socket connections only  
> local   all             all                                     md5  
> \# IPv4 local connections:  
> host    all             all             127.0.0.1/32            md5  
> host    all             all             0.0.0.0/0               md5  
> \# IPv6 local connections:  
> host    all             all             ::1/128                 md5  
> host    all             all             ::0/0                   md5  
> \# Allow replication connections from localhost, by a user with the  
> \# replication private  
> host all SYSTEM 192.168.4.127/16 md5  
> host all SYSTEM 192.168.4.129/16 md5  
> host all SYSTEM 192.168.4.159/16 md5  
>    
> host replication SYSTEM 192.168.4.127/16 md5  
> host replication SYSTEM 192.168.4.129/16 md5  
> host replication SYSTEM 192.168.4.159/16 md5

*===注意：修改已有节点的sys\_hba.conf文件后，需要reload数据库进程才能使配置生效。===*

> \[kingbase@srv129 data\]$ sys\_ctl reload -D /home/kingbase/cluster/ESHA/db/data  
> 服务器进程发出信号

**1.3.5 从主库拷贝以下目录和文件到新备库**

通过scp远程拷贝以下目录和文件到新增节点相同的位置：

===说明：“/home/kingbase/cluster/ESHA”是集群部署的目录，其中“ESHA”是集群名称。对于archivedir也可以创建空目录。===

> \[kingbase@\#localhost ESHA\]$ pwd  
> /home/kingbase/cluster/ESHA  
> \[kingbase@\#localhost ESHA\]$ ls -lh  
> 总用量 8.0K  
> drwxrwxr-x 2 kingbase kingbase    6 10月 30 14:23 archivedir  
> drwxrwxr-x 7 kingbase kingbase   84 10月 30 17:00 db  
> drwxrwxr-x 3 kingbase kingbase  181 10月 30 17:23 log  
> drwxrwxr-x 3 kingbase kingbase   29 10月 30 14:55 run  
> \-rw------- 1 kingbase kingbase 8.0K 10月 31 13:19 template.bk  
>    
> \[kingbase@srv127 db\]$ pwd  
> /home/kingbase/cluster/ESHA/db  
> \[kingbase@srv127 db\]$ ls -lh  
> 总用量 20K  
> drwxrwxr-x  2 kingbase kingbase 4.0K 10月 29 17:29 bin  
> drwxrwxr-x  2 kingbase kingbase  295 10月 29 17:26 etc  
> \-rw-------  1 kingbase kingbase  151 8月  18 20:13 kingbase.log  
> drwxrwxr-x  2 kingbase kingbase 4.0K 7月  27 10:41 lib  
> drwxrwxr-x  9 kingbase kingbase 4.0K 7月  27 10:41 share
> 
>   
> *===注意：目录和文件拷贝不要拷贝kingbasecluster目录到新增节点，在V8R3的集群中，cluster只支持主备两个节点，新增节点只能是data node。===*

**拷贝主库/etc/cron.d/KINGBASECRON文件到备库相同目录下：**

![watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16 1][]

**1.3.6 在主库端添加replication\_slot**

**1） 查看已经创建的replication\_slot:**

**2） 创建新增备库的replication\_slot**

> TEST=\# select \* from sys\_create\_physical\_replication\_slot('slot\_srv159');  
>   slot\_name  | xlog\_position  
> \-------------+---------------  
>  slot\_srv159 |  
> (1 row)

**3）查看主备复制的replication\_slots**

**1.3.7 启动备库实例加入主备复制架构**

> \[kingbase@srv159 data\]$ sys\_ctl start -D /home/kingbase/cluster/ESHA/db/data  
> 正在启动服务器进程  
> \[kingbase@srv159 data\]$ 日志:  日志输出重定向到日志收集进程  
> 提示:  后续的日志输出将出现在目录 "/home/kingbase/cluster/ESHA/db/data/sys\_log"中.  
>    
> 查看备库进程：  
> \[kingbase@srv159 data\]$ ps -ef |grep kingbase  
> kingbase  5071     1  0 20:01 pts/3    00:00:00 /opt/Kingbase/ES/V8R3/Server/bin/kingbase -D /home/kingbase/cluster/ESHA/db/data  
> kingbase  5072  5071  0 20:01 ?        00:00:00 kingbase: logger process  
> kingbase  5073  5071 90 20:01 ?        00:00:19 kingbase: startup process   waiting for 0000000300000000000000D6  
> kingbase  5077  5071  0 20:01 ?        00:00:00 kingbase: checkpointer process  
> kingbase  5078  5071  0 20:01 ?        00:00:00 kingbase: writer process  
> kingbase  5199  5071  0 20:01 ?        00:00:00 kingbase: stats collector process  
> kingbase  5200  5071  0 20:01 ?        00:00:00 kingbase: wal receiver process  
> kingbase  5201  5071  0 20:01 ?        00:00:00 kingbase: wal sender process SYSTEM 192.168.4.159(62391) idle

** 1.3.8 查看主备复制信息**

**1）查看主备复制状态信息**

> TEST=\# select \* from sys\_stat\_replication;  
>   pid  | usesysid | usename | application\_name |  client\_addr  | client\_hostname | client\_port |         backend\_start         | backend\_xmin |   state   | sent\_lo  
> cation | write\_location | flush\_location | replay\_location | sync\_priority | sync\_state  
> \-------+----------+---------+------------------+---------------+-------------  
>  16851 |       10 | SYSTEM  | srv129           | 192.168.4.129 |                 |       52967 | 2020-10-23 15:11:17.836557+08 |              | streaming | 0/ED000  
> 000    | 0/ED000000     | 0/ED000000     | 0/ECFFF418      |             0 | async  
>  12202 |       10 | SYSTEM  | srv159           | 192.168.4.159 |                 |       47993 | 2020-10-23 20:10:32.521960+08 |         2059 | streaming | 0/ED000  
> 000    | 0/ED000000     | 0/ED000000     | 0/ECFFF418      |             0 | async  
> (2 rows)

**2）查看数据库数据信息**

> 主库：  
> prod=\# select count(\*) from t10;  
>   count  
> \----------  
>  21000000  
> (1 row)  
>    
> 备库：  
> prod=\# select count(\*) from t10;  
>   count  
> \----------  
>  21000000  
> (1 row)

**备注：**

> 1） 在主备状态信息中已经查看到新增节点（srv159）的复制状态信息。  
> 2） 查看主库和备库同步数据一致，说明在创建基础备份期间，备库和主库保持了数据的一致性。  
> 3） 由以上可知，备库创建完成。

**二、配置新节点加入集群服务一键启停**

**2.1 编辑所有节点HAmodule.conf**

![watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16 2][]

> 注意：在KB\_ALL\_IP参数中加入新的备库节点的ip。

**2.2 在新增节点修改HAmodule.conf**

![watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16 3][]

> 注意：将KB\_LOCALHOST\_IP参数配置为本机ip。

**2.3 配置已经存在节点（Cluster的主备库）kingbase\_cluster.conf文件**

![watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16 4][]

> 注意：图中黄色部分为新增节点的配置，仔细修改，否则在show pool\_nodes时无法发现新增节点。

**2.4 测试kingbase\_monitor.sh一键启停集群：**

> \[kingbase@srv129 bin\]$ ./kingbase\_monitor.sh restart  
> \-----------------------------------------------------------------------  
> 2020-10-29 17:32:33 KingbaseES automation beging...  
> 2020-10-29 17:32:33 stop kingbasecluster \[192.168.4.127\] ...  
> DEL VIP NOW AT 2020-10-29 17:32:32 ON ens192  
> No VIP on my dev, nothing to do.  
> 2020-10-29 17:32:33 Done...  
> 2020-10-29 17:32:33 stop kingbasecluster \[192.168.4.129\] ...  
> DEL VIP NOW AT 2020-10-29 17:32:34 ON ens192  
> No VIP on my dev, nothing to do.  
> 2020-10-29 17:32:34 Done...  
> 2020-10-29 17:32:34 stop kingbase \[192.168.4.127\] ...  
> set /home/kingbase/cluster/ESHA/db/data down now...  
> 2020-10-29 17:32:37 Done...  
> 2020-10-29 17:32:38 Del kingbase VIP \[192.168.4.99/24\] ...  
> DEL VIP NOW AT 2020-10-29 17:32:37 ON ens192  
> No VIP on my dev, nothing to do.  
> 2020-10-29 17:32:38 Done...  
> 2020-10-29 17:32:38 stop kingbase \[192.168.4.129\] ...  
> set /home/kingbase/cluster/ESHA/db/data down now...  
> 2020-10-29 17:32:40 Done...  
> 2020-10-29 17:32:41 Del kingbase VIP \[192.168.4.99/24\] ...  
> DEL VIP NOW AT 2020-10-29 17:32:42 ON ens192  
> execute: \[/sbin/ip addr del 192.168.4.99/24 dev ens192\]  
> Oprate del ip cmd end.  
> 2020-10-29 17:32:42 Done...  
> 2020-10-29 17:32:42 stop kingbase \[192.168.4.159\] ...  
> 2020-10-29 17:32:42 Done...  
> 2020-10-29 17:32:43 Del kingbase VIP \[192.168.4.99/24\] ...  
> DEL VIP NOW AT 2020-10-29 17:32:43 ON ens192  
> No VIP on my dev, nothing to do.  
> 2020-10-29 17:32:43 Done...  
> ......................  
> all stop..  
>    
> start crontab kingbase position : \[3\]  
> Redirecting to /bin/systemctl restart crond.service  
> start crontab kingbase position : \[3\]  
> Redirecting to /bin/systemctl restart crond.service  
> ADD VIP NOW AT 2020-10-29 17:32:51 ON ens192  
> execute: \[/sbin/ip addr add 192.168.4.99/24 dev ens192 label ens192:2\]  
> execute: /usr/sbin/arping -U 192.168.4.99 -I ens192 -w 1  
> ARPING 192.168.4.99 from 192.168.4.99 ens192  
> Sent 1 probes (1 broadcast(s))  
> Received 0 response(s)  
> Redirecting to /bin/systemctl restart crond.service  
> wait kingbase recovery 5 sec...  
> start crontab kingbasecluster line number: \[4\]  
> Redirecting to /bin/systemctl restart crond.service  
> start crontab kingbasecluster line number: \[4\]  
> Redirecting to /bin/systemctl restart crond.service  
> ......................  
> all started..  
> ...  
> now we check again  
> ==================================================================  
> |             ip |                       program|              \[status\]  
> \[  192.168.4.127\]|             \[kingbasecluster\]|              \[active\]  
> \[  192.168.4.129\]|             \[kingbasecluster\]|              \[active\]  
> \[  192.168.4.127\]|                    \[kingbase\]|              \[active\]  
> \[  192.168.4.129\]|                    \[kingbase\]|              \[active\]  
> \[  192.168.4.159\]|                    \[kingbase\]|              \[active\]  
> ==================================================================  
> 注意：从以上可以获得，集群架构为2个节点的cluster，3个节点的data node。

**2.5 查看新增节点信息**

![watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16 5][]

**三、集群切换测试**

**3.1 测试目标**

>   
> 当原备库数据库服务停止或宕机，在主库服务或宕机时，新的备库切换为主库。

**3.2 测试步骤**

> 1、先停止原备库的kingbaseES服务。  
> 2、再停止主库的kingbaseES服务。  
> 3、查看新的备库是否被切换为主库。

**3.3 实施主备切换**

>   
> **1、停止原备库的kingbaseES服务**
> 
> \[kingbase@srv129 bin\]$ sys\_ctl stop -D /home/kingbase/cluster/ESHA/db/data  
> 等待服务器进程关闭 .... 完成  
> 服务器进程已经关闭
> 
>   
> **2、查看复制状态信息**
> 
> **1）复制状态信息**
> 
> TEST=\# select \* from sys\_stat\_replication;  
>  pid  | usesysid | usename | application\_name |  client\_addr  | client\_hostname | client\_port |         backend\_start         | backend\_xmin |   stat  
> e   | sent\_location | write\_location | flush\_location | replay\_location | sync\_priority | sync\_state  
> \------+----------+---------+------------------+---------------+--------------  
>  1556 |       10 | SYSTEM  | srv129           | 192.168.4.129 |                 |       20766 | 2020-10-31 15:17:03.946716+08 |              | stream  
> ing | 1/5603A788    | 1/5603A788     | 1/5603A788     | 1/5603A750      |             1 | sync  
>  3387 |       10 | SYSTEM  | srv159           | 192.168.4.159 |                 |       47333 | 2020-10-31 15:21:05.275911+08 |              | stream  
> ing | 1/5603A788    | 1/5603A788     | 1/5603A788     | 1/5603A750      |             2 | potential  
> (2 rows)  
> 注意：在一主两从的架构中，新增节点的sync\_stat状态为”potential”
> 
> **2）原备库kingbaseES被停止**
> 
> TEST=\# select \* from sys\_stat\_replication;  
>  pid  | usesysid | usename | application\_name |  client\_addr  | client\_hostname | client\_port |         backend\_start         | backend\_xmin |   stat  
> e   | sent\_location | write\_location | flush\_location | replay\_location | sync\_priority | sync\_state  
> \------+----------+---------+------------------+---------------+---------------  
>  3387 |       10 | SYSTEM  | srv159           | 192.168.4.159 |                 |       47333 | 2020-10-31 15:21:05.275911+08 |              | stream  
> ing | 1/5603A980    | 1/5603A980     | 1/5603A980     | 1/5603A8D8      |             0 | async  
> (1 row)  
> 注意：在一主一从的架构中，新增节点的sync\_stat状态为“async”。
> 
> **3）查看pool node信息**
> 
> TEST=\# show pool\_nodes;  
>  node\_id |   hostname    | port  | status | lb\_weight |  role   | select\_cnt | load\_balance\_node | replication\_delay  
> \---------+---------------+-------+--------+-----------+---------+------------+-------------------+-------------------  
>  0       | 192.168.4.127 | 54321 | up     | 0.333333  | primary | 5          | false             | 0  
>  1       | 192.168.4.129 | 54321 | down   | 0.333333  | standby | 0          | false             | 0  
>  2       | 192.168.4.159 | 54321 | up     | 0.333333  | standby | 0          | true              | 0  
> (3 rows)
> 
>   
> **3、停止主库kingbaseES数据库服务**
> 
> \[kingbase@srv127 etc\]$ sys\_ctl stop -D /home/kingbase/cluster/ESHA/db/data  
> 等待服务器进程关闭 .... 完成  
> 服务器进程已经关闭
> 
>   
> ** 4、切换完成查看备库信息**  
> **1）查看新备库的kingbaseES进程**
> 
> \[kingbase@\#localhost log\]$ ps -ef |grep kingbase  
> kingbase 11680     1  0 15:21 pts/1    00:00:01 /home/kingbase/cluster/ESHA/db/bin/kingbase -D /home/kingbase/cluster/ESHA/db/data  
> kingbase 11681 11680  0 15:21 ?        00:00:00 kingbase: logger process  
> kingbase 11686 11680  0 15:21 ?        00:00:00 kingbase: checkpointer process  
> kingbase 11687 11680  0 15:21 ?        00:00:00 kingbase: writer process  
> kingbase 11689 11680  0 15:21 ?        00:00:00 kingbase: stats collector process  
> kingbase 14821 31212  0 10月30 pts/0  00:00:00 ../bin/kingbasecluster -n -d  
> kingbase 14840 14821  0 10月30 pts/0  00:00:35 kingbasecluster: watchdog  
> root     14850 14675  0 09:48 pts/1    00:00:00 su - kingbase  
> kingbase 14851 14850  0 09:48 pts/1    00:00:00 -bash  
> kingbase 15518 11680  0 15:33 ?        00:00:00 kingbase: wal writer process  
> kingbase 15519 11680  0 15:33 ?        00:00:00 kingbase: autovacuum launcher process  
> kingbase 15520 11680  0 15:33 ?        00:00:00 kingbase: archiver process   last was 00000005.history  
> kingbase 15521 11680  0 15:33 ?        00:00:00 kingbase: bgworker: syslogical supervisor  
> kingbase 16075 11680  0 15:35 ?        00:00:00 kingbase: SUPERMANAGER\_V8ADMIN TEST 192.168.4.127(34260) idle
> 
>   
> **2）连接实例查看数据库状态**
> 
> \[kingbase@\#localhost data\]$ ksql -U system -W 123456 TEST  
> ksql (V008R003C002B0160)  
> Type "help" for help.  
>    
> TEST=\# select sys\_is\_in\_recovery();  
>  sys\_is\_in\_recovery  
> \--------------------  
>  f  
> (1 row)  
> 注意：备库已经切换为主库状态。
> 
> **3）查看recovery日志**
> 
> \[kingbase@\#localhost log\]$ tail -f recovery.log  
> \---------------------------------------------------------------------  
> 2020-10-31 15:36:02 recover beging...  
> 2020-10-31 15:36:02 check if the network is ok  
> ping trust ip 192.168.4.1 scuccess ping times :\[3\], scuccess times:\[3\]  
> determine if i am master or standby  
> I,m node is primary, determine whether there is a standby db can be set to synchronous standby  
> *===由以上可以获得，备库已经切换为主库。===*

**四、总结**

> 1、kingbaseES V8R3读写分离集群可以在线手工方式增加新的节点，进行集群架构的横向扩展。  
> 2、在修改kingbasecluster.conf后需要重新启动kingbasecluster服务，可以先在cluster备节点通过命令(kingbasecluster –m fast stop关闭服务），然后再关闭cluster主节点服务；先启动cluster主节点服务(kingbasecluster –n >kingbasecluster.log 2>&1 &)方式启动集群服务，主节点启动完成后再启动cluster备节点。  
> 3、增加节点前做好测试，应该在业务低峰时候来完成。  
> 4、本测试文档只是在CentOS 7环境下，其他版本使用前请在线下测试。

【本文正在参与炫“库”行动-人大金仓有奖征文】

[icon-default.png_t_L892]: /images/20220828/08f0fd7d2730492193688cf666bc6ca1.png
[CSDN_icon-default.png_t_L892_https_marketing.csdn.net_p_98bd30353e7cb998b6070a89e8b91edb]: https://marketing.csdn.net/p/98bd30353e7cb998b6070a89e8b91edb
[watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16]: /images/20220828/a61cd1eb7bfa4fbd95b640f906e36ee7.png
[watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16 1]: /images/20220828/b4253e13b2854d44968acd275e7a1bd6.png
[watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16 2]: /images/20220828/9a86a5f0368b4f989ce56bf09aa86ac9.png
[watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16 3]: /images/20220828/70662a2af4fc4b6b86d8a57db0e96179.png
[watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16 4]: /images/20220828/09230ac82ea3425ab31abd6d8395a597.png
[watermark_type_ZHJvaWRzYW5zZmFsbGJhY2s_shadow_50_text_Q1NETiBA5a6i5bGF5aSp5rav_size_20_color_FFFFFF_t_70_g_se_x_16 5]: /images/20220828/360a6141dd0c449a9abfe8aabcc1b61f.png