前言:
postgresql数据库只用自身的一些配置是无法做到最优的优化的,需要通过一些外置插件(中间件)来提高服务器的整体性能,通俗的说就是数据库仅仅依靠自身是无法达到性能最优的,很多时候需要更改数据库的整体架构,使用一些目前来说成熟的技术,比如,读写分离技术,负载均衡技术,高速缓存技术等等集群方面的技术。
下图是一些较为成熟的集群方案:
从上表可以看到,pgpool是比较全面的一个中间件,什么连接池,负载均衡都有,还有没有写出来的缓存功能,其实使用此中间件的不能拒绝的诱惑就是负载均衡和缓存了,其它的功能倒是没有什么。
本文将就pgpool的负载均衡和高速缓存功能做一个详细的介绍。
一,
pgpool的简单介绍
Pgpool-II是一个在PostgreSQL服务器和PostgreSQL数据库客户端之间工作的中间件。它是根据BSD许可证授权的。它提供以下功能。
连接池
Pgpool-II保存与PostgreSQL服务器的连接,并在具有相同属性(即用户名,数据库,协议版本)的新连接进入时重用它们。它减少了连接开销,并提高了系统的整体吞吐量。
复制
Pgpool-II可以管理多个PostgreSQL服务器。使用复制功能可以在2个或更多物理磁盘上创建实时备份,以便在磁盘发生故障时服务可以继续运行而不会停止服务器。
负载均衡
如果复制了数据库,则在任何服务器上执行SELECT查询都将返回相同的结果。Pgpool-II利用复制功能,通过在多个服务器之间分配SELECT查询来减少每个PostgreSQL服务器的负载,从而提高系统的整体吞吐量。充其量,性能与PostgreSQL服务器的数量成比例地提高。在许多用户同时执行许多查询的情况下,负载平衡最有效。
限制超出连接
PostgreSQL的最大并发连接数有限制,连接在这么多连接后被拒绝。但是,设置最大连接数会增加资源消耗并影响系统性能。pgpool-II对最大连接数也有限制,但额外连接将排队,而不是立即返回错误。
看门狗
Watchdog可以协调多个Pgpool-II,创建一个强大的集群系统,避免单点故障或脑裂。看门狗可以对其他pgpool-II节点执行生命检查,以检测Pgpoll-II的故障。如果活动Pgpool-II发生故障,则可以将备用Pgpool-II提升为活动状态,并接管虚拟IP。
查询缓存
在内存中查询缓存允许保存一对SELECT语句及其结果。如果有相同的SELECT,Pgpool-II将从缓存中返回值。由于不涉及SQL解析或访问PostgreSQL,因此在内存缓存中使用速度非常快。另一方面,在某些情况下,它可能比正常路径慢,因为它增加了存储缓存数据的一些开销。
Pgpool-II讲PostgreSQL的后端和前端协议,并在后端和前端之间传递消息。因此,数据库应用程序(前端)认为Pgpool-II是实际的PostgreSQL服务器,服务器(后端)将Pgpool-II视为其客户端之一。因为Pgpool-II对服务器和客户端都是透明的,所以现有的数据库应用程序可以与Pgpool-II一起使用。Pgpool-II讲述PostgreSQL的后端和前端协议,并在它们之间传递连接。因此,数据库应用程序(前端)认为Pgpool-II是实际的PostgreSQL服务器,服务器(后端)将Pgpool-II视为其客户端之一。因为Pgpool-II对服务器和客户端都是透明的,所以现有的数据库应用程序可以与Pgpool-II一起使用,不需要对现有的业务系统进行更改。
那么,架构方面应该是通过看门狗,利用一个虚拟IP(也就是VIP)代理一个自身就是主从流复制的集群,VIP可以看做是前端,数据库可以看做后端,而主从流复制的数据库集群是具有这么一个特点:主服务器可读可写,从服务只读不可写。VIP通过pgpool的负载均衡功能就可以即可读也可以写了,负载均衡的策略是主从根据特定算法分配读任务,写任务仍然是交给主服务器完成。
这样,集群的使用率就自然的增高了,一些常用的查询语句通过pgpool的缓存功能,缓存起来,自然的整个集群的查询效率就提高了。
OK,下面开始讲述如何搭建pgpool。
二,
pgpool的官方网址:pgpool Wiki
下载和安装教程都有,里面也提供了一些比较新的rpm安装包,rpm仓库地址是:Index of /yum/rpms/4.4/redhat/rhel-7-x86_64
直接将该地址配置成yum仓库就可以了
postgresql的版本和大体情况如下:
11服务器是主服务器,12是从服务器
主从复制的搭建见我的博客:postgresql|数据库|【postgresql-12的基于pg_basebackup的主从复制部署】_postgresql12 主备_晚风_END的博客-CSDN博客
三,
pgpool的部署
该部署工作是比较繁琐的,难度是比较高的,主要是权限问题需要小心处理,其次是参数众多,很多地方需要根据实际的情况调整,最后是pgpool的功能比较多,如何配置好是需要比较多的耐心。
首先,大概介绍一下pgpool的组件,该中间件的管理组件有三个,一个是数据库侧使用的pool工具,该工具是以插件的形式安装在postgresql数据库内的,第二个是操作系统侧的pcp工具,这个工具需要在pgpool的主配置文件内配置,第三个是pgpoolAdm,此工具是PHP编写的web端管理工具,可以在web端方便的查看pgpool并且管理配置pgpool,目前的版本应该是需要高版本的PHP支持,暂时没有使用。
1,
管理工具的安装
本案例中,只安装数据库侧的管理工具pool和pcp,pool工具在源码包内。
pgpool-II-4.4.4.tar.gz这个文件上传到服务器解压后,和普通的postgresql插件没什么区别,一样的make && make install 就可以了,前提是环境变量里有定义PGHOME和PGDATA这两个变量。
2,
yum安装
配置好本地仓库和上面提到的官网仓库就可以运行以下命令安装了,这里安装了一个memcached服务,该服务后面作为缓存服务使用的。
yum install pgpool-II-pg12-debuginfo-4.4.2 pgpool-II-pg12-4.4.2 pgpool-II-pg12-devel-4.4.2 pgpool-II-pg12-extensions-4.4.2 -y
yum install memcached -y &&systemctl enable memcached && systemctl enable pgpool && systemctl start pgpool memcached
安装完毕后将会在/etc目录下看到pgpool-II,此目录里是pgpool的配置文件和一些高可用故障转移脚本,本案例中这些脚本不打算使用,只配置pgpool服务,另外需要注意,两个服务器都需要安装,memcached只在一个服务器安装就可以了
3,
配置文件
可以看到,这些文件都是postgres的属组,这些一定要注意哦
[root@node1 pgpool-II]# ls -al
total 144
drwxr-xr-x. 3 root root 202 Sep 18 06:18 .
drwxr-xr-x. 83 root root 8192 Sep 17 19:16 ..
-rw------- 1 postgres postgres 900 Sep 17 11:15 pcp.conf
-rw-------. 1 postgres postgres 858 Jan 22 2023 pcp.conf.sample
-rw------- 1 postgres postgres 52960 Sep 18 02:01 pgpool.conf
-rw-------. 1 postgres postgres 52964 Jan 22 2023 pgpool.conf.sample
-rw------- 1 postgres postgres 2 Sep 17 10:21 pgpool_node_id
-rw------- 1 postgres postgres 3537 Sep 17 11:54 pool_hba.conf
-rw-------. 1 postgres postgres 3476 Jan 22 2023 pool_hba.conf.sample
-rw-------. 1 postgres postgres 45 Sep 17 11:05 pool_passwd
drwxr-xr-x. 2 root root 4096 Sep 17 10:02 sample_scripts
pcp.conf 的配置
该文件是存放pgpool的管理密码,此密码可以和postgresql数据库的密码不一样,也就是说随便定,定义方式非常简单,用户名:密码的形式添加在该文件末尾即可,只是需要注意一点,密码是md5加密的,不能明文(两种方式都可以,嫌麻烦的话就第三行那个命令,用户是postgres,密码是123456)
[root@node1 pgpool-II]# pg_md5 123456
e10adc3949ba59abbe56e057f20f883e
[root@node1 pgpool-II]# echo "postgres:e10adc3949ba59abbe56e057f20f883e">>./pcp.conf
[root@node1 pgpool-II]# echo "postgres:`pg_md5 123456`">>./pcp.conf
[root@node1 pgpool-II]# cat pcp.conf
# PCP Client Authentication Configuration File
# ============================================
#
# This file contains user ID and his password for pgpool
# communication manager authentication.
#
# Note that users defined here do not need to be PostgreSQL
# users. These users are authorized ONLY for pgpool
# communication manager.
#
# File Format
# ===========
#
# List one UserID and password on a single line. They must
# be concatenated together using ':' (colon) between them.
# No spaces or tabs are allowed anywhere in the line.
#
# Example:
# postgres:e8a48653851e28c69d0506508fb27fc5
#
# Be aware that there will be no spaces or tabs at the
# beginning of the line! although the above example looks
# like so.
#
# Lines beginning with '#' (pound) are comments and will
# be ignored. Again, no spaces or tabs allowed before '#'.# USERID:MD5PASSWD
postgres:e10adc3949ba59abbe56e057f20f883e
pgpool.conf文件的配置:
该文件是pgpool的主配置文件,其中注释的行已经全部去掉了,只保留了放开的内容
说明:该配置文件内定义的文件路径需要手动建立,/var/run/postgresql 属组是postgres
sr_check_user = 'nobody' 这个nobody用户需要在主数据库创建,创建命令为create role nobody login replication encrypted password 'replica';
为什么是主数据库呢?因为是流复制,主数据库创建了 ,从数据库自然就有了嘛,上面提到的插件也是如此的哦。
[root@node1 pgpool-II]# sed -e '/^$/d' pgpool.conf |grep -v "\#"
backend_clustering_mode = 'streaming_replication'
listen_addresses = '*'
port = 15433
unix_socket_directories = '/var/run/postgresql'
pcp_listen_addresses = '*'
pcp_port = 19999
pcp_socket_dir = '/var/run/postgresql'
backend_hostname0 = '192.168.123.11'
backend_port0 = 5432
backend_weight0 = 1
backend_data_directory0 = '/usr/local/pgsql/data'
backend_flag0 = 'ALLOW_TO_FAILOVER'
backend_application_name0 = 'node1'
backend_hostname1 = '192.168.123.12'
backend_port1 = 5432
backend_weight1 = 1
backend_data_directory1 = '/usr/local/pgsql/data'
backend_flag1 = 'ALLOW_TO_FAILOVER'
backend_application_name1 = 'node2'
enable_pool_hba = on
pool_passwd = 'pool_passwd'
process_management_mode = dynamic
num_init_children = 32
min_spare_children = 5
max_spare_children = 10
max_pool = 4
child_life_time = 5min
log_destination = 'stderr'
log_connections = on
log_disconnections = on
log_hostname = on
log_statement = on
log_per_node_statement = on
log_client_messages = on
logging_collector = on
log_directory = '/var/log/pgpool_log'
log_filename = 'pgpool-%a.log'
log_file_mode = 0600
log_truncate_on_rotation = on
log_rotation_age = 1d
log_rotation_size = 0
pid_file_name = '/var/run/postgresql/pgpool.pid'
logdir = '/tmp'
connection_cache = on
reset_query_list = 'ABORT; DISCARD ALL'
load_balance_mode = on
database_redirect_preference_list = 'postgres:1'
sr_check_period = 10
sr_check_user = 'nobody'
sr_check_password = 'replica'
sr_check_database = 'postgres'
delay_threshold = 1
delay_threshold_by_time = 1
prefer_lower_delay_standby = on
use_watchdog = on
hostname0 = '192.168.123.11'
wd_port0 = 9000
pgpool_port0 = 15433
hostname1 = '192.168.123.12'
wd_port1 = 9000
pgpool_port1 = 15433
wd_ipc_socket_dir = '/var/run/postgresql'
delegate_ip = '192.168.123.222'
if_cmd_path = '/sbin'
if_up_cmd = 'ip addr add $_IP_$/24 dev ens33 label ens33:0'
if_down_cmd = 'ip addr del $_IP_$/24 dev ens33'
arping_path = '/usr/sbin'
arping_cmd = 'arping -U $_IP_$ -w 1 -I ens33'
wd_monitoring_interfaces_list = ''
wd_lifecheck_method = 'heartbeat'
wd_interval = 10
heartbeat_hostname0 = '192.168.123.11'
heartbeat_port0 = 19694
heartbeat_device0 = 'ens33'
heartbeat_hostname1 = '192.168.123.12'
heartbeat_port1 = 19694
heartbeat_device1 = 'ens33'
wd_life_point = 3
wd_lifecheck_query = 'SELECT 1'
memory_cache_enabled = off
memqcache_method = 'memcached'
memqcache_memcached_host = '192.168.123.11'
memqcache_memcached_port = 11211
memqcache_total_size = 64MB
memqcache_max_num_cache = 1000000
memqcache_expire = 0
memqcache_cache_block_size = 1MB
pool_passwd文件的配置:
重要:
su - postgres
pg_md5 -m -p -u postgres pool_passwd
#此时会提示输入密码,此密码是postgresql服务器的postgres用户的密码,一会会用此命令登录postgresql数据库的哦
[root@node1 pgpool-II]# su - postgres
Last login: Mon Sep 18 06:34:54 CST 2023 on pts/1
[postgres@node1 ~]$ pg_md5 -m -p -u postgres pool_passwd
password:
[postgres@node1 ~]$ logout
[root@node1 pgpool-II]# cat pool_passwd
postgres:md5a3556571e93b0d20722ba62be61e8c2d
pool_hab.conf文件的配置
该文件的作用是定义pgpool哪些用户可以访问哪些后端的postgresql数据库,功能和postgresql数据库的pg_hba.conf文件类似
如果不想太麻烦(也就是不太安全),那么,如下配置即可:
# IPv4 local connections:
host all all 127.0.0.1/32 trust
host all all ::1/128 trust
host all all 0.0.0.0/0 md5
pgpool_node_id文件的配置
此文件是标识文件,表明该pgpool 对应于哪个后端,因此,11服务器上此文件内容就一个0,12服务器上次文件内容就一个1即可,当然了,如果有其它的节点,就依次增加数字即可,最多好像是127个节点。
创建用户
Pcp.conf 文件内的用户和密码可以自己随意指定,pool_pass文件内的用户必须是数据库的真实用户和密码,目前只使用了postgres用户
配置pcp.conf文件[root@node1 pgpool-II]# pg_md5 123456e10adc3949ba59abbe56e057f20f883e[root@node1 pgpool-II]# echo "postgres:e10adc3949ba59abbe56e057f20f883e">>./pcp.conf[root@node1 pgpool-II]# echo "postgres:`pg_md5 123456`">>./pcp.conf[root@node1 pgpool-II]# cat pcp.conf# PCP Client Authentication Configuration File# ============================================## This file contains user ID and his password for pgpool# communication manager authentication.## Note that users defined here do not need to be PostgreSQL# users. These users are authorized ONLY for pgpool# communication manager.## File Format# ===========## List one UserID and password on a single line. They must# be concatenated together using ':' (colon) between them.# No spaces or tabs are allowed anywhere in the line.## Example:# postgres:e8a48653851e28c69d0506508fb27fc5## Be aware that there will be no spaces or tabs at the# beginning of the line! although the above example looks# like so.## Lines beginning with '#' (pound) are comments and will# be ignored. Again, no spaces or tabs allowed before '#'.# USERID:MD5PASSWDpostgres:e10adc3949ba59abbe56e057f20f883e重要—配置pool_passwd:su - postgrespg_md5 -m -p -u postgres pool_passwd#此时会提示输入密码,此密码是postgresql服务器的postgres用户的密码,一会会用此命令登录postgresql数据库的哦[root@node1 pgpool-II]# su - postgresLast login: Mon Sep 18 06:34:54 CST 2023 on pts/1[postgres@node1 ~]$ pg_md5 -m -p -u postgres pool_passwdpassword:[postgres@node1 ~]$ logout[root@node1 pgpool-II]# cat pool_passwdpostgres:md5a3556571e93b0d20722ba62be61e8c2dpool_hab.conf文件的配置该文件的作用是定义pgpool哪些用户可以访问哪些后端的postgresql数据库,功能和postgresql数据库的pg_hba.conf文件类似如果不想太麻烦(也就是不太安全),那么,建议是直接把数据库的pg_hba.conf文件复制过来即可配置文件内的命令需要有粘滞特殊权限,ip和arping命令,因此,命令如下:
chmod u+s /sbin/ip
chmod u+s /usr/sbin/arping
启动服务和停止服务
根据以上的配置文件,我们需要把配置文件里用到的文件夹手动创建出来,并赋予postgres属组:mkdir /var/run/postgresqlchown -Rf postgres. /var/run/postgresqlpgpool的启动和停止该中间件的启停是比较特殊的,既可以使用systemctl管理器管理也可以直接二进制启停,为了规范操作,就使用systemctl管理进行吧:systemctl enable pgpool && systemctl start pgpool服务正常启动的样子:[root@node1 pgpool-II]# systemctl status pgpool● pgpool.service - Pgpool-IILoaded: loaded (/usr/lib/systemd/system/pgpool.service; enabled; vendor preset: disabled)Active: active (running) since Mon 2023-09-18 01:04:32 CST; 19h agoProcess: 58354 ExecStop=/usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf $STOP_OPTS stop (code=exited, status=0/SUCCESS)Process: 45217 ExecReload=/usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf reload (code=exited, status=0/SUCCESS)Main PID: 58360 (pgpool)Tasks: 20Memory: 14.1MCGroup: /system.slice/pgpool.service├─58360 /usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf -n├─58361 pgpool: PgpoolLogger├─58362 pgpool: watchdog├─58371 pgpool: lifecheck├─58372 pgpool: postgres postgres 192.168.123.1(50284) idle├─58373 pgpool: heartbeat receiver├─58374 pgpool: wait for connection request├─58375 pgpool: heartbeat sender├─58376 pgpool: wait for connection request├─58377 pgpool: wait for connection request├─58378 pgpool: postgres postgres 192.168.123.1(50279) idle├─58379 pgpool: wait for connection request├─58380 pgpool: wait for connection request├─58381 pgpool: wait for connection request├─58382 pgpool: wait for connection request├─58383 pgpool: wait for connection request├─58385 pgpool: PCP: wait for connection request├─58386 pgpool: worker process├─58387 pgpool: health check process(0)└─58388 pgpool: health check process(1)Sep 18 01:20:32 node1 systemd[1]: Reloading Pgpool-II.Sep 18 01:20:32 node1 systemd[1]: Reloaded Pgpool-II.Sep 18 01:25:42 node1 systemd[1]: Reloading Pgpool-II.Sep 18 01:25:42 node1 systemd[1]: Reloaded Pgpool-II.Sep 18 01:37:32 node1 systemd[1]: Reloading Pgpool-II.Sep 18 01:37:32 node1 systemd[1]: Reloaded Pgpool-II.Sep 18 01:47:36 node1 systemd[1]: Reloading Pgpool-II.Sep 18 01:47:36 node1 systemd[1]: Reloaded Pgpool-II.Sep 18 02:01:32 node1 systemd[1]: Reloading Pgpool-II.Sep 18 02:01:32 node1 systemd[1]: Reloaded Pgpool-II.直接二进制形式启停的命令:[root@node2 ~]# pgpool ###启动[root@node2 ~]# pgpool -m fast stop2023-09-18 21:01:30.055: main pid 44987: LOG: stop request sent to pgpool (pid: 39228). waiting for termination...2023-09-18 21:01:30.055: main pid 44987: LOCATION: main.c:546.done.(这里需要注意一点,systemctl和二进制不能混用,否则另一个是不会生效的,下面是混用后的演示)[root@node2 ~]# pgpool[root@node2 ~]# systemctl status pgpool● pgpool.service - Pgpool-IILoaded: loaded (/usr/lib/systemd/system/pgpool.service; disabled; vendor preset: disabled)Active: failed (Result: exit-code) since Mon 2023-09-18 21:01:30 CST; 22s agoProcess: 44988 ExecStop=/usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf $STOP_OPTS stop (code=exited, status=3)Process: 44599 ExecReload=/usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf reload (code=exited, status=0/SUCCESS)Process: 39228 ExecStart=/usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf $OPTS (code=exited, status=0/SUCCESS)Main PID: 39228 (code=exited, status=0/SUCCESS)Sep 18 01:25:38 node2 systemd[1]: Reloaded Pgpool-II.Sep 18 01:37:24 node2 systemd[1]: Reloading Pgpool-II.Sep 18 01:37:24 node2 systemd[1]: Reloaded Pgpool-II.Sep 18 01:47:31 node2 systemd[1]: Reloading Pgpool-II.Sep 18 01:47:31 node2 systemd[1]: Reloaded Pgpool-II.Sep 18 02:01:29 node2 systemd[1]: Reloading Pgpool-II.Sep 18 02:01:29 node2 systemd[1]: Reloaded Pgpool-II.Sep 18 21:01:30 node2 systemd[1]: pgpool.service: control process exited, code=exited status=3Sep 18 21:01:30 node2 systemd[1]: Unit pgpool.service entered failed state.Sep 18 21:01:30 node2 systemd[1]: pgpool.service failed.[root@node2 ~]# systemctl start pgpool[root@node2 ~]# systemctl status pgpool● pgpool.service - Pgpool-IILoaded: loaded (/usr/lib/systemd/system/pgpool.service; disabled; vendor preset: disabled)Active: failed (Result: exit-code) since Mon 2023-09-18 21:02:01 CST; 1s agoProcess: 45637 ExecStop=/usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf $STOP_OPTS stop (code=exited, status=3)Process: 44599 ExecReload=/usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf reload (code=exited, status=0/SUCCESS)Process: 45635 ExecStart=/usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf $OPTS (code=exited, status=3)Main PID: 45635 (code=exited, status=3)Sep 18 21:02:01 node2 systemd[1]: Started Pgpool-II.Sep 18 21:02:01 node2 pgpool[45635]: 2023-09-18 21:02:01.310: main pid 45635: FATAL: could not open pid file "/var/run/postgresql/pgpool.pid"Sep 18 21:02:01 node2 systemd[1]: pgpool.service: main process exited, code=exited, status=3/NOTIMPLEMENTEDSep 18 21:02:01 node2 pgpool[45637]: 2023-09-18 21:02:01.316: main pid 45637: FATAL: could not read pid fileSep 18 21:02:01 node2 pgpool[45637]: 2023-09-18 21:02:01.316: main pid 45637: LOCATION: main.c:532Sep 18 21:02:01 node2 systemd[1]: pgpool.service: control process exited, code=exited status=3Sep 18 21:02:01 node2 systemd[1]: Unit pgpool.service entered failed state.Sep 18 21:02:01 node2 systemd[1]: pgpool.service failed.[root@node2 ~]# pgpool -m fast stop2023-09-18 21:02:13.716: main pid 45805: LOG: stop request sent to pgpool (pid: 45161). waiting for termination...2023-09-18 21:02:13.716: main pid 45805: LOCATION: main.c:546.done.[root@node2 ~]# systemctl start pgpool[root@node2 ~]# systemctl status pgpool● pgpool.service - Pgpool-IILoaded: loaded (/usr/lib/systemd/system/pgpool.service; disabled; vendor preset: disabled)Active: active (running) since Mon 2023-09-18 21:02:21 CST; 2s agoProcess: 45637 ExecStop=/usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf $STOP_OPTS stop (code=exited, status=3)Process: 44599 ExecReload=/usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf reload (code=exited, status=0/SUCCESS)Main PID: 46021 (pgpool)Tasks: 20Memory: 6.3MCGroup: /system.slice/pgpool.service├─46021 /usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf -n├─46022 pgpool: PgpoolLogger├─46023 pgpool: watchdog├─46025 pgpool: lifecheck├─46026 pgpool: heartbeat receiver├─46027 pgpool: heartbeat sender├─46028 pgpool: wait for connection request├─46029 pgpool: wait for connection request├─46030 pgpool: wait for connection request├─46031 pgpool: wait for connection request├─46032 pgpool: wait for connection request
测试读写分离,负载均衡
pgpool的管理工具简单的使用pcp是一套管理工具,也就是说有很多pcp名称开始的一族命令[root@node1 pgpool-II]# pcp_pcp_attach_node pcp_health_check_stats pcp_node_info pcp_proc_count pcp_promote_node pcp_reload_config pcp_watchdog_info pcp_detach_node pcp_node_count pcp_pool_status pcp_proc_info pcp_recovery_node pcp_stop_pgpool 查看有几个pgpool节点:注意,pcp定义的端口需要写哦,如果你更改过了的话,本文使用的是19999,这个命令可以随意定义,定义在pcp.conf 文件内,只是记得要md5加密哦。本例是 用户是postgres,密码是123456[root@node2 pgpool-II]# echo "postgres:`pg_md5 123456`" >>pcp.conf[root@node2 pgpool-II]# pcp_node_count -U postgres -p 19999Password:2查询pgpool集群的信息:[root@node2 pgpool-II]# pcp_node_info -Upostgres -p19999Password:192.168.123.11 5432 1 0.500000 waiting unknown primary unknown 0 none none 2023-09-18 21:06:40192.168.123.12 5432 1 0.500000 waiting unknown standby unknown 0.000000 none none 2023-09-18 21:06:40重新加载所有配置:[root@node2 pgpool-II]# pcp_reload_config -Upostgres -p19999Password:pcp_reload_config -- Command Successful查看看门狗的状态:[root@node2 pgpool-II]# pcp_watchdog_info -U postgres -p19999Password:2 2 NO 192.168.123.11:15433 Linux node1 192.168.123.11192.168.123.12:15433 Linux node2 192.168.123.12 15433 9000 7 STANDBY 0 MEMBER192.168.123.11:15433 Linux node1 192.168.123.11 15433 9000 4 LEADER 0 MEMBER数据库内使用插件管理pgpool:查看pool的状态:[root@node2 pgpool-II]# su - postgres -c "psql -Upostgres -p 15433 -h 192.168.123.222"Password for user postgres:psql (12.5)Type "help" for help.postgres=# show pool_nodes;node_id | hostname | port | status | pg_status | lb_weight | role | pg_role | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change ---------+----------------+------+--------+-----------+-----------+---------+---------+------------+-------------------+-------------------+-------------------+------------------------+---------------------0 | 192.168.123.11 | 5432 | up | unknown | 0.500000 | primary | unknown | 0 | false | 0 | | | 2023-09-18 21:53:081 | 192.168.123.12 | 5432 | up | unknown | 0.500000 | standby | unknown | 0 | true | 0.000000 second | | | 2023-09-18 21:53:08(2 rows)查看缓存命中率:目前暂时是关闭的postgres=# show pool_cache;num_cache_hits | num_selects | cache_hit_ratio | num_hash_entries | used_hash_entries | num_cache_entries | used_cache_entries_size | free_cache_entries_size | fragment_cache_entries_size----------------+-------------+-----------------+------------------+-------------------+-------------------+-------------------------+-------------------------+-----------------------------0 | 0 | 0.00 | 0 | 0 | 0 | 0 | 0 | 0(1 row)测试以及一些需要注意的地方OK,开始测试啦,这里需要着重说明,VIP是非常非常重要的,测试的时候是使用VIP登录的哦:VIP登录数据库:[root@node1 pgpool-II]# su - postgres -c "psql -Upostgres -p 15433 -h 192.168.123.222"Password for user postgres:psql (12.5)Type "help" for help.随便建立一个测试用的表,并向表内些一个测试数据,这个就不班门弄斧了,非常的简单,然后查询pool的状态:postgres=# show pool_nodes;node_id | hostname | port | status | pg_status | lb_weight | role | pg_role | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change ---------+----------------+------+--------+-----------+-----------+---------+---------+------------+-------------------+-------------------+-------------------+------------------------+---------------------0 | 192.168.123.11 | 5432 | up | unknown | 0.500000 | primary | unknown | 11 | false | 0 | | | 2023-09-18 21:53:081 | 192.168.123.12 | 5432 | up | unknown | 0.500000 | standby | unknown | 12 | true | 0.000000 second | | | 2023-09-18 21:53:08(2 rows)可以看到,负载均衡是生效的,select_cnt 是查询次数,主节点查询了11次,从节点查询了12次利用navicat再次查询,或者使用pgbench压测工具:编辑postgres=# show pool_nodes;node_id | hostname | port | status | pg_status | lb_weight | role | pg_role | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change ---------+----------------+------+--------+-----------+-----------+---------+---------+------------+-------------------+-------------------+-------------------+------------------------+---------------------0 | 192.168.123.11 | 5432 | up | unknown | 0.500000 | primary | unknown | 32 | false | 0 | | | 2023-09-18 21:53:081 | 192.168.123.12 | 5432 | up | unknown | 0.500000 | standby | unknown | 33 | true | 0.000000 second | | | 2023-09-18 21:53:08(2 rows)查看pgpool的日志(截取相关部分日志,如下):2023-09-18 22:41:50.701: Navicat pid 39838: LOCATION: pool_proto_modules.c:27122023-09-18 22:41:50.701: Navicat pid 39838: LOG: statement: select * from test11312023-09-18 22:41:50.701: Navicat pid 39838: LOCATION: pool_proto_modules.c:2112023-09-18 22:41:50.701: Navicat pid 39838: LOG: DB node id: 1 backend pid: 8540 statement: select * from test11312023-09-18 22:41:50.701: Navicat pid 39838: LOCATION: pool_proto_modules.c:35692023-09-18 22:41:50.702: Navicat pid 39845: LOG: Query message from frontend.2023-09-18 22:41:50.702: Navicat pid 39845: DETAIL: query: "SELECT c.conkey FROM pg_constraint c WHERE c.contype = 'p' and c.conrelid = 16398"2023-09-18 22:41:50.702: Navicat pid 39845: LOCATION: pool_proto_modules.c:27122023-09-18 22:41:50.702: Navicat pid 39845: LOG: statement: SELECT c.conkey FROM pg_constraint c WHERE c.contype = 'p' and c.conrelid = 163982023-09-18 22:41:50.702: Navicat pid 39845: LOCATION: pool_proto_modules.c:2112023-09-18 22:41:50.702: Navicat pid 39845: LOG: DB node id: 0 backend pid: 61668 statement: SELECT c.conkey FROM pg_constraint c WHERE c.contype = 'p' and c.conrelid = 163982023-09-18 22:41:50.702: Navicat pid 39845: LOCATION: pool_proto_modules.c:35692023-09-18 22:41:52.695: Navicat pid 39838: LOG: Query message from frontend.2023-09-18 22:41:52.695: Navicat pid 39838: DETAIL: query: "show pool_nodes"2023-09-18 22:41:52.695: Navicat pid 39838: LOCATION: pool_proto_modules.c:27122023-09-18 22:41:52.695: Navicat pid 39838: LOG: statement: show pool_nodes2023-09-18 22:41:52.695: Navicat pid 39838: LOCATION: pool_proto_modules.c:2112023-09-18 22:42:28.368: psql pid 39844: LOG: Query message from frontend.2023-09-18 22:42:28.368: psql pid 39844: DETAIL: query: "show pool_nodes;"2023-09-18 22:42:28.368: psql pid 39844: LOCATION: pool_proto_modules.c:27122023-09-18 22:42:28.368: psql pid 39844: LOG: statement: show pool_nodes;2023-09-18 22:42:28.368: psql pid 39844: LOCATION: pool_proto_modules.c:211可以看到,负载均衡功能完美运行缓存服务的测试:首先,查看memcache服务是否正常,确认正常后,修改主配置文件,打开缓存功能:[root@node1 pgpool-II]# systemctl status memcached● memcached.service - MemcachedLoaded: loaded (/usr/lib/systemd/system/memcached.service; enabled; vendor preset: disabled)Active: active (running) since Sun 2023-09-17 19:16:26 CST; 1 day 3h agoMain PID: 1443 (memcached)Tasks: 6Memory: 4.4MCGroup: /system.slice/memcached.service└─1443 /usr/bin/memcached -u memcached -p 11211 -m 64 -c 1024Sep 17 19:16:26 node1 systemd[1]: Started Memcached.编辑重新加载pgpool服务,两个服务器都要修改,重新加载:可以看到部分查询落在了主节点,但没有缓存,部分查询落在了从节点,走的是缓存2023-09-18 22:50:54.481: Navicat pid 47139: LOG: fetch from memory cache2023-09-18 22:50:54.342: Navicat pid 47139: DETAIL: query result fetched from cache. statement: select * from test11312023-09-18 22:50:54.185: Navicat pid 47138: LOG: DB node id: 0 backend pid: 25501 statement: SELECT c.conkey FROM pg_constraint c WHERE c.contype = 'p' and c.conrelid = 163982023-09-18 22:50:54.185: Navicat pid 47138: LOCATION: pool_proto_modules.c:35692023-09-18 22:50:54.342: Navicat pid 47139: LOG: Query message from frontend.2023-09-18 22:50:54.342: Navicat pid 47139: DETAIL: query: "select * from test1131"2023-09-18 22:50:54.342: Navicat pid 47139: LOCATION: pool_proto_modules.c:27122023-09-18 22:50:54.342: Navicat pid 47139: LOG: statement: select * from test11312023-09-18 22:50:54.342: Navicat pid 47139: LOCATION: pool_proto_modules.c:2112023-09-18 22:50:54.342: Navicat pid 47139: LOG: fetch from memory cache2023-09-18 22:50:54.342: Navicat pid 47139: DETAIL: query result fetched from cache. statement: select * from test11312023-09-18 22:50:54.342: Navicat pid 47139: LOCATION: pool_memqcache.c:8212023-09-18 22:50:54.343: Navicat pid 47138: LOG: Query message from frontend.2023-09-18 22:50:54.343: Navicat pid 47138: DETAIL: query: "SELECT c.conkey FROM pg_constraint c WHERE c.contype = 'p' and c.conrelid = 16398"2023-09-18 22:50:54.343: Navicat pid 47138: LOCATION: pool_proto_modules.c:27122023-09-18 22:50:54.343: Navicat pid 47138: LOG: statement: SELECT c.conkey FROM pg_constraint c WHERE c.contype = 'p' and c.conrelid = 163982023-09-18 22:50:54.343: Navicat pid 47138: LOCATION: pool_proto_modules.c:2112023-09-18 22:50:54.343: Navicat pid 47138: LOG: DB node id: 0 backend pid: 25501 statement: SELECT c.conkey FROM pg_constraint c WHERE c.contype = 'p' and c.conrelid = 163982023-09-18 22:50:54.343: Navicat pid 47138: LOCATION: pool_proto_modules.c:35692023-09-18 22:50:54.480: Navicat pid 47139: LOG: Query message from frontend.2023-09-18 22:50:54.481: Navicat pid 47139: DETAIL: query: "select * from test1131"2023-09-18 22:50:54.481: Navicat pid 47139: LOCATION: pool_proto_modules.c:27122023-09-18 22:50:54.481: Navicat pid 47139: LOG: statement: select * from test11312023-09-18 22:50:54.481: Navicat pid 47139: LOCATION: pool_proto_modules.c:2112023-09-18 22:50:54.481: Navicat pid 47139: LOG: fetch from memory cache2023-09-18 22:50:54.481: Navicat pid 47139: DETAIL: query result fetched from cache. statement: select * from test11312023-09-18 22:50:54.481: Navicat pid 47139: LOCATION: pool_memqcache.c:821查看缓存命中率:可以看到查询了53次,总命中率是0.5(cache_hit_ratio就是缓存命中率,可以知道,该数据库集群的性能会得到极大的提升)postgres=# show pool_cache;num_cache_hits | num_selects | cache_hit_ratio | num_hash_entries | used_hash_entries | num_cache_entries | used_cache_entries_size | free_cache_entries_size | fragment_cache_entries_size----------------+-------------+-----------------+------------------+-------------------+-------------------+-------------------------+-------------------------+-----------------------------53 | 52 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0(1 row)postgres=# show pool_nodes;node_id | hostname | port | status | pg_status | lb_weight | role | pg_role | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change ---------+----------------+------+--------+-----------+-----------+---------+---------+------------+-------------------+-------------------+-------------------+------------------------+---------------------0 | 192.168.123.11 | 5432 | up | unknown | 0.500000 | primary | unknown | 52 | false | 0 | | | 2023-09-18 22:48:371 | 192.168.123.12 | 5432 | up | unknown | 0.500000 | standby | unknown | 0 | true | 0.000000 second | | | 2023-09-18 22:48:37
|
主备切换(这个功能用不上)
安装及使用总结
首先,需要明白一点,pgpool并不是十分容易配置的服务,可能在配置和调试的过程中会有非常多的错误,但是不需要害怕这些错误,善于利用百度等等搜索引擎,其次,主配置文件里的很多地方配置好后可能是需要重启服务的,而服务的重启顺序是先从节点重启在主节点重启,这么做的原因是可能会配置故障恢复脚本等情形下的高可用,防止主节点乱跑,从而给自己造成不必要的麻烦。
最后pool_hba.conf 里最好使用数据库内的pg_hab.conf的所有内容,一个是安全,一个是防止两者不一致造成的麻烦(毕竟报错后的问题分析还是比较麻烦的嘛)