当前位置：网站首页>PostgreSQL uses pgpool II to realize read-write separation + load balancing

PostgreSQL uses pgpool II to realize read-write separation + load balancing

2022-07-06 23:53:00 【Hua Weiyun】

One 、Pgpool-II brief introduction

Official website ：https://www.pgpool.net/mediawiki/index.php/Main_Page

Pgpool-II It's a PostgreSQL The server and PostgreSQL Middleware working between database clients . It is based on BSD Licensed . It provides the following functions .

Connection pool

Pgpool-II To preserve with PostgreSQL Server connection , And have the same properties （ Use account , database , Protocol version ） Reuse them when new connections come in . It reduces connection overhead , And improve the overall throughput of the system .

Copy

Pgpool-II Can manage multiple PostgreSQL The server . The copy function can be used in 2 Create real-time backups on physical disks or more , So that in the event of a disk failure the service can continue to run without stopping the server .

Load balancing

If the database is copied , On any server SELECT All queries will return the same result .Pgpool-II Use the copy function , By allocating... Between multiple servers SELECT Query to reduce each PostgreSQL Server load , So as to improve the overall throughput of the system . at best , Performance and PostgreSQL The number of servers increased proportionally . In the case of many users executing many queries at the same time , Load balancing works best .

Limit out of connection

PostgreSQL The maximum number of concurrent connections is limited , Connection rejected after so many connections . however , Setting the maximum number of connections increases resource consumption and affects system performance .pgpool-II There are also restrictions on the maximum number of connections , But additional connections will be queued , Instead of returning an error immediately .

House dogs

Watchdog Can coordinate multiple Pgpool-II, Create a powerful cluster system , Avoid a single point of failure or brain crack . The watchdog can be on other pgpool-II Node performs a life check , To detect Pgpoll-II Failure of . If the activity Pgpool-II failure , Then you can use the spare Pgpool-II Promote to active , And take over the virtual IP.

In memory query cache

Querying the cache in memory allows you to save a pair of SELECT Statement and its result . If there is the same SELECT,Pgpool-II Will return the value... From the cache . Because it doesn't involve SQL Parse or access PostgreSQL, So it's very fast in memory cache . On the other hand , In some cases , It may be slower than the normal path , Because it increases the cost of storing cached data .

Pgpool-II speak PostgreSQL The back-end and front-end protocols , And pass messages between the back end and the front end . therefore , Database applications （ front end ） Think Pgpool-II It's practical PostgreSQL The server , The server （ Back end ） take Pgpool-II Think of it as one of its clients . because Pgpool-II Transparent to both the server and the client , So existing database applications can work with Pgpool-II Use it together .Pgpool-II about PostgreSQL The back-end and front-end protocols , And pass connections between them . therefore , Database applications （ front end ） Think Pgpool-II It's practical PostgreSQL The server , The server （ Back end ） take Pgpool-II Think of it as one of its clients . because Pgpool-II Transparent to both the server and the client , So existing database applications can work with Pgpool-II Use it together , Almost no need to change its source code .

Two 、 High availability architecture diagram

PG build 1 Lord 2 The process of copying from stream can refer to “【DB treasure 60】PG12 Highly available 1 Lord 2 From the stream replication environment to build and switch test ”, This paper is based on this article , Use Pgpool-II To achieve read-write separation and load balancing function .

The overall structure is as follows ：

3、 ... and 、 Use Docker Rapid deployment Pgpool-II middleware

--  Download mirroring docker pull registry.cn-hangzhou.aliyuncs.com/lhrbest/lhrpgpool:4.2.2--  Create a private network docker network create --subnet=172.72.6.0/24 pg-network--  establish Pgpool-II The container of docker rm -f lhrpgpooldocker run -d --name lhrpgpool -h lhrpgpool \  --net=pg-network --ip 172.72.6.6 \  -p 9999:9999 -p 9898:9898 \  -v /sys/fs/cgroup:/sys/fs/cgroup \  --privileged=true lhrbest/lhrpgpool:4.2.2 \  /usr/sbin/initdocker network connect bridge lhrpgpooldocker restart lhrpgpooldocker exec -it lhrpgpool bash

Four 、 To configure Pgpool-II

All of the following operations are in pgsql User execution .

4.1、 Copy parameter file

[root@lhrpgpool etc]# su - pgsqlLast login: Mon Apr 26 14:28:35 CST 2021 on pts/0[pgsql@lhrpgpool /]# cp /postgresql/pgpool/etc/pgpool.conf.sample /postgresql/pgpool/etc/pgpool.conf[pgsql@lhrpgpool /]# cp /postgresql/pgpool/etc/pool_hba.conf.sample /postgresql/pgpool/etc/pool_hba.conf[pgsql@lhrpgpool /]# cp /postgresql/pgpool/etc/pcp.conf.sample /postgresql/pgpool/etc/pcp.conf

4.2、 To configure pgpool.conf

vi Enter column operation ：Ctrl + v Enter block selection mode , Then move the cursor to select the line you want to comment on , Then press capital I Enter the insert mode at the beginning of the line , Enter a comment symbol such as // or #, After input , Press two ESC,Vim Will automatically add comments to all the line capitals you selected , Save exit completion comment .

vi /postgresql/pgpool/etc/pgpool.conf# - pgpool Connection Settings -listen_addresses = '*'# - Backend Connection Settings -backend_hostname0 = '172.72.6.2'backend_port0 = 5432backend_weight0 = 1backend_data_directory0 = '/var/lib/postgresql/data'backend_flag0 = 'ALLOW_TO_FAILOVER'backend_application_name0 = 'lhrpg64302'backend_hostname1 = '172.72.6.3'backend_port1 = 5432backend_weight1 = 1backend_data_directory1 = '/var/lib/postgresql/data'backend_flag1 = 'ALLOW_TO_FAILOVER'backend_application_name1 = 'lhrpg64303'backend_hostname2 = '172.72.6.4'backend_port2 = 5432backend_weight2 = 1backend_data_directory2 = '/var/lib/postgresql/data'backend_flag2 = 'ALLOW_TO_FAILOVER'backend_application_name2 = 'lhrpg64304'# - Authentication -enable_pool_hba = on# - Where to log -log_destination = 'syslog'log_connections = on#------------------------------------------------------------------------------# FILE LOCATIONS#------------------------------------------------------------------------------pid_file_name = '/postgresql/pgpool/pgpool.pid'logdir = '/tmp'#------------------------------------------------------------------------------# LOAD BALANCING MODE#------------------------------------------------------------------------------load_balance_mode = on

4.3、 Turn on the system log （root operation ）

pgpool By default, the log of is written to stderr, It can be set to syslog in . But because it's written by default LOCAL0 Level logs will be discarded , So it's in /etc/rsyslog.conf Modify the syslog Specified output path ：

echo "local0.* /postgresql/pgpool/pgpool.log" >> /etc/rsyslog.confsystemctl restart rsyslog.service

4.4、 To configure pool_hba.conf

su -  pgsqlecho "host all all 0.0.0.0/0 md5" >> /postgresql/pgpool/etc/pool_hba.conf

4.5、 Generate pool_passwd file

pg_md5 --md5auth --username=nobody "lhr"pg_md5 --md5auth --username=lhr "lhr"pg_md5 --md5auth --username=pgpool "lhr"pg_md5 --md5auth --username=postgres "lhr"

Will generate pool_passwd file ：

[pgsql@lhrpgpool ~]$ pg_md5 --md5auth --username=nobody "lhr"[pgsql@lhrpgpool ~]$ pg_md5 --md5auth --username=lhr "lhr"[pgsql@lhrpgpool ~]$ pg_md5 --md5auth --username=pgpool "lhr"[pgsql@lhrpgpool ~]$ pg_md5 --md5auth --username=postgres "lhr"[pgsql@lhrpgpool ~]$ cd /postgresql/etc[pgsql@lhrpgpool etc]$ cat pool_passwdnobody:md5188b0dfd531e1734c1e5bb4d57053d64lhr:md53e5c401ee2b9f28db1bb075b1b99e0adpgpool:md547b6c1f1700de696bcb6b98dc6a21ab6postgres:md5da3edeb741de62d06ab73785ed222494

4.6、 To configure pcp.conf file

[pgsql@lhrcentos76 ~]$ pg_md5  -u=pgpool "lhr"3996643de967b80174e48fb45d7227b1echo  "pgpool:3996643de967b80174e48fb45d7227b1" >> /postgresql/pgpool/etc/pcp.conf

4.7、 Create users in the main library

psql -U postgres -h 192.168.66.35 -p64302create role nobody login encrypted password 'lhr';create role lhr login encrypted password 'lhr';create role pgpool login encrypted password 'lhr';grant postgres to nobody,lhr,pgpool;

stay PostgreSQL There's no concept of distinguishing between users and roles ,CREATE USER by CREATE ROLE Another name for , The two commands are almost identical , The only difference is ”CREATE USER” Command to create the user's LOGIN Property defaults to on , and CREATE ROLE Command to create the user's NOLOGIN Property defaults to on.

5、 ... and 、 start-up Pgpool-II

cat >> /lib/systemd/system/pgpool.service <<"EOF"[Unit]Description=Pgpool-IIAfter=syslog.target network.target[Service]User=pgsqlGroup=pgsqlEnvironmentFile=-/etc/sysconfig/pgpoolExecStart=/postgresql/pgpool/bin/pgpool -f /postgresql/pgpool/etc/pgpool.conf -nExecStop=/postgresql/pgpool/bin/pgpool -f /postgresql/pgpool/etc/pgpool.conf -m fast stopExecReload=/postgresql/pgpool/bin/pgpool -f /postgresql/pgpool/etc/pgpool.conf reloadLimitNOFILE=65536KillMode=processKillSignal=SIGINTRestart=on-abnormalRestartSec=30sTimeoutSec=0[Install]WantedBy=multi-user.targetEOFchmod +x /lib/systemd/system/pgpool.servicesystemctl enable pgpoolsystemctl start pgpoolsystemctl status pgpool

result ：

[root@lhrpgpool etc]# systemctl status pgpool● pgpool.service - Pgpool-II   Loaded: loaded (/usr/lib/systemd/system/pgpool.service; enabled; vendor preset: disabled)   Active: active (running) since Mon 2021-04-26 14:35:27 CST; 1s ago Main PID: 1954 (pgpool)   CGroup: /docker/c3f0c9600069f719f49e81e3602f642925feecb5e134c8aedb96753728730013/system.slice/pgpool.service           ├─1954 /postgresql/pgpool/bin/pgpool -f /postgresql/pgpool/etc/pgpool.conf -n           ├─1956 pgpool: wait for connection request           ├─1957 pgpool: wait for connection request           ├─1958 pgpool: wait for connection request           ├─1959 pgpool: wait for connection request           ├─1960 pgpool: wait for connection request           ├─1961 pgpool: wait for connection request           ├─1962 pgpool: wait for connection request           ├─1963 pgpool: wait for connection request           ├─1964 pgpool: wait for connection request           ├─1965 pgpool: wait for connection request           ├─1966 pgpool: wait for connection request           ├─1967 pgpool: wait for connection request           ├─1968 pgpool: wait for connection request           ├─1969 pgpool: wait for connection request           ├─1970 pgpool: wait for connection request           ├─1971 pgpool: wait for connection request           ├─1972 pgpool: wait for connection request           ├─1973 pgpool: wait for connection request           ├─1974 pgpool: wait for connection request           ├─1975 pgpool: wait for connection request           ├─1976 pgpool: wait for connection request           ├─1977 pgpool: wait for connection request           ├─1978 pgpool: wait for connection request           ├─1979 pgpool: wait for connection request           ├─1980 pgpool: wait for connection request           ├─1981 pgpool: wait for connection request           ├─1982 pgpool: wait for connection request           ├─1983 pgpool: wait for connection request           ├─1984 pgpool: wait for connection request           ├─1985 pgpool: wait for connection request           ├─1986 pgpool: wait for connection request           ├─1987 pgpool: wait for connection request           ├─1988 pgpool: PCP: wait for connection request           ├─1989 pgpool: worker process           ├─1990 pgpool: health check process(0)           ├─1991 pgpool: health check process(1)           └─1992 pgpool: health check process(2)Apr 26 14:35:28 lhrpgpool pgpool[1954]: [14-1] 2021-04-26 14:35:28: pid 1954: LOG:  find_primary_node: standby node is 2Apr 26 14:35:28 lhrpgpool pgpool[1988]: [15-1] 2021-04-26 14:35:28: pid 1988: LOG:  PCP process: 1988 startedApr 26 14:35:28 lhrpgpool pgpool[1990]: [15-1] 2021-04-26 14:35:28: pid 1990: LOG:  process startedApr 26 14:35:28 lhrpgpool pgpool[1991]: [15-1] 2021-04-26 14:35:28: pid 1991: LOG:  process startedApr 26 14:35:28 lhrpgpool pgpool[1989]: [15-1] 2021-04-26 14:35:28: pid 1989: LOG:  process startedApr 26 14:35:28 lhrpgpool pgpool[1992]: [15-1] 2021-04-26 14:35:28: pid 1992: LOG:  process startedApr 26 14:35:28 lhrpgpool pgpool[1954]: [15-1] 2021-04-26 14:35:28: pid 1954: LOG:  pgpool-II successfully started. version 4.2.2 (chichiriboshi)Apr 26 14:35:28 lhrpgpool pgpool[1954]: [16-1] 2021-04-26 14:35:28: pid 1954: LOG:  node status[0]: 1Apr 26 14:35:28 lhrpgpool pgpool[1954]: [17-1] 2021-04-26 14:35:28: pid 1954: LOG:  node status[1]: 2Apr 26 14:35:28 lhrpgpool pgpool[1954]: [18-1] 2021-04-26 14:35:28: pid 1954: LOG:  node status[2]: 2

6、 ... and 、 Query node status

[pgsql@lhrpgpool ~]$ psql -U nobody -h localhost  -p 9999 -d postgrespsql (13.2, server 12.2 (Debian 12.2-2.pgdg100+1))Type "help" for help.postgres=> show  pool_nodes; node_id |  hostname  | port | status | lb_weight |  role   | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change  ---------+------------+------+--------+-----------+---------+------------+-------------------+-------------------+-------------------+------------------------+--------------------- 0       | 172.72.6.2 | 5432 | up     | 0.333333  | primary | 0          | false             | 0                 |                   |                        | 2021-04-26 14:35:27 1       | 172.72.6.3 | 5432 | up     | 0.333333  | standby | 0          | false             | 0                 |                   |                        | 2021-04-26 14:35:27 2       | 172.72.6.4 | 5432 | up     | 0.333333  | standby | 0          | true              | 0                 |                   |                        | 2021-04-26 14:35:27(3 rows)

see pgpool Backend state ：

pcp_node_info -U pgpool -h localhost -p 9898 -n 0 -vpcp_node_info -U pgpool -h localhost -p 9898 -n 1 -vpcp_node_info -U pgpool -h localhost -p 9898 -n 2 -v

result ：

[pgsql@lhrpgpool ~]$ pcp_node_info -U pgpool -h localhost -p 9898 -n 0 -vPassword: Hostname               : 172.72.6.2Port                   : 5432Status                 : 2Weight                 : 0.333333Status Name            : upRole                   : primaryReplication Delay      : 0Replication State      : Replication Sync State : Last Status Change     : 2021-04-26 14:35:27[pgsql@lhrpgpool ~]$ pcp_node_info -U pgpool -h localhost -p 9898 -n 1 -vPassword: Hostname               : 172.72.6.3Port                   : 5432Status                 : 2Weight                 : 0.333333Status Name            : upRole                   : standbyReplication Delay      : 0Replication State      : Replication Sync State : Last Status Change     : 2021-04-26 14:35:27[pgsql@lhrpgpool ~]$ pcp_node_info -U pgpool -h localhost -p 9898 -n 2 -vPassword: Hostname               : 172.72.6.4Port                   : 5432Status                 : 2Weight                 : 0.333333Status Name            : upRole                   : standbyReplication Delay      : 0Replication State      : Replication Sync State : Last Status Change     : 2021-04-26 14:35:27

7、 ... and 、 Test read-write separation + Load balancing

Before the test , Consider modifying the file pgpool.conf The following parameters in ：

log_statement=alllog_per_node_statement =onclient_min_messages =loglog_min_messages = info

take effect ：

pgpool reload

After the test is completed , Change back to the original value ：

log_statement=offlog_per_node_statement = off# client_min_messages =notice# log_min_messages = warning

Testing process ：

--  open 3 A window psql -U postgres -h 192.168.66.35 -p 9999 -d sbtestcreate table test(id int);insert into test values(1);select * from test;

window 1：

C:\Users\lhrxxt>psql -U postgres -h 192.168.66.35 -p 9999 -d sbtestPassword for user postgres:psql (13.2, server 12.2 (Debian 12.2-2.pgdg100+1))Type "help" for help.sbtest=# show pool_nodes;LOG:  statement: show pool_nodes;LOG:  DB node id: 0 backend pid: 10106 statement: SELECT version() node_id |  hostname  | port | status | lb_weight |  role   | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change---------+------------+------+--------+-----------+---------+------------+-------------------+-------------------+-------------------+------------------------+--------------------- 0       | 172.72.6.2 | 5432 | up     | 0.333333  | primary | 0          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 1       | 172.72.6.3 | 5432 | up     | 0.333333  | standby | 0          | true              | 0                 |                   |                        | 2021-04-26 09:54:44 2       | 172.72.6.4 | 5432 | up     | 0.333333  | standby | 0          | false             | 0                 |                   |                        | 2021-04-26 09:54:44(3 rows)sbtest=# create table test(id int);LOG:  statement: create table test(id int);LOG:  DB node id: 0 backend pid: 10106 statement: create table test(id int);CREATE TABLEsbtest=# insert into test values (1);LOG:  statement: insert into test values (1);LOG:  DB node id: 0 backend pid: 10106 statement: insert into test values (1);INSERT 0 1sbtest=# select * from test;LOG:  statement: select * from test;LOG:  DB node id: 0 backend pid: 10106 statement: SELECT count(*) FROM pg_class AS c, pg_namespace AS n WHERE c.oid = pg_catalog.to_regclass('"test"') AND c.relnamespace = n.oid AND n.nspname = 'pg_catalog'LOG:  DB node id: 0 backend pid: 10106 statement: SELECT count(*) FROM pg_catalog.pg_class AS c, pg_namespace AS n WHERE c.relname = 'test' AND c.relnamespace = n.oid AND n.nspname ~ '^pg_temp_'LOG:  DB node id: 0 backend pid: 10106 statement: SELECT count(*) FROM pg_catalog.pg_class AS c WHERE c.oid = pg_catalog.to_regclass('"test"') AND c.relpersistence = 'u'LOG:  DB node id: 1 backend pid: 6069 statement: select * from test; id----  1(1 row)sbtest=# show pool_nodes;LOG:  statement: show pool_nodes; node_id |  hostname  | port | status | lb_weight |  role   | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change---------+------------+------+--------+-----------+---------+------------+-------------------+-------------------+-------------------+------------------------+--------------------- 0       | 172.72.6.2 | 5432 | up     | 0.333333  | primary | 0          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 1       | 172.72.6.3 | 5432 | up     | 0.333333  | standby | 1          | true              | 0                 |                   |                        | 2021-04-26 09:54:44 2       | 172.72.6.4 | 5432 | up     | 0.333333  | standby | 0          | false             | 0                 |                   |                        | 2021-04-26 09:54:44(3 rows)

window 2：

C:\Users\lhrxxt>psql -U postgres -h 192.168.66.35 -p 9999 -d sbtestPassword for user postgres:psql (13.2, server 12.2 (Debian 12.2-2.pgdg100+1))Type "help" for help.sbtest=# show pool_nodes;LOG:  statement: show pool_nodes;LOG:  DB node id: 0 backend pid: 10116 statement: SELECT version() node_id |  hostname  | port | status | lb_weight |  role   | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change---------+------------+------+--------+-----------+---------+------------+-------------------+-------------------+-------------------+------------------------+--------------------- 0       | 172.72.6.2 | 5432 | up     | 0.333333  | primary | 0          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 1       | 172.72.6.3 | 5432 | up     | 0.333333  | standby | 1          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 2       | 172.72.6.4 | 5432 | up     | 0.333333  | standby | 0          | true              | 0                 |                   |                        | 2021-04-26 09:54:44(3 rows)sbtest=# select * from test;LOG:  statement: select * from test;LOG:  DB node id: 0 backend pid: 10116 statement: SELECT count(*) FROM pg_class AS c, pg_namespace AS n WHERE c.oid = pg_catalog.to_regclass('"test"') AND c.relnamespace = n.oid AND n.nspname = 'pg_catalog'LOG:  DB node id: 0 backend pid: 10116 statement: SELECT count(*) FROM pg_catalog.pg_class AS c, pg_namespace AS n WHERE c.relname = 'test' AND c.relnamespace = n.oid AND n.nspname ~ '^pg_temp_'LOG:  DB node id: 0 backend pid: 10116 statement: SELECT count(*) FROM pg_catalog.pg_class AS c WHERE c.oid = pg_catalog.to_regclass('"test"') AND c.relpersistence = 'u'LOG:  DB node id: 2 backend pid: 6085 statement: select * from test; id----  1(1 row)sbtest=# show pool_nodes;LOG:  statement: show pool_nodes; node_id |  hostname  | port | status | lb_weight |  role   | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change---------+------------+------+--------+-----------+---------+------------+-------------------+-------------------+-------------------+------------------------+--------------------- 0       | 172.72.6.2 | 5432 | up     | 0.333333  | primary | 0          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 1       | 172.72.6.3 | 5432 | up     | 0.333333  | standby | 1          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 2       | 172.72.6.4 | 5432 | up     | 0.333333  | standby | 1          | true              | 0                 |                   |                        | 2021-04-26 09:54:44(3 rows)sbtest=# insert into test values(2);LOG:  statement: insert into test values(2);LOG:  DB node id: 0 backend pid: 10116 statement: insert into test values(2);INSERT 0 1sbtest=# show pool_nodes;LOG:  statement: show pool_nodes; node_id |  hostname  | port | status | lb_weight |  role   | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change---------+------------+------+--------+-----------+---------+------------+-------------------+-------------------+-------------------+------------------------+--------------------- 0       | 172.72.6.2 | 5432 | up     | 0.333333  | primary | 0          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 1       | 172.72.6.3 | 5432 | up     | 0.333333  | standby | 1          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 2       | 172.72.6.4 | 5432 | up     | 0.333333  | standby | 1          | true              | 0                 |                   |                        | 2021-04-26 09:54:44(3 rows)

window 3：

C:\Users\lhrxxt>psql -U postgres -h 192.168.66.35 -p 9999 -d sbtestPassword for user postgres:psql (13.2, server 12.2 (Debian 12.2-2.pgdg100+1))Type "help" for help.sbtest=# show pool_nodes;LOG:  statement: show pool_nodes;LOG:  DB node id: 0 backend pid: 10127 statement: SELECT version() node_id |  hostname  | port | status | lb_weight |  role   | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change---------+------------+------+--------+-----------+---------+------------+-------------------+-------------------+-------------------+------------------------+--------------------- 0       | 172.72.6.2 | 5432 | up     | 0.333333  | primary | 0          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 1       | 172.72.6.3 | 5432 | up     | 0.333333  | standby | 1          | true              | 0                 |                   |                        | 2021-04-26 09:54:44 2       | 172.72.6.4 | 5432 | up     | 0.333333  | standby | 1          | false             | 0                 |                   |                        | 2021-04-26 09:54:44(3 rows)sbtest=# select * from test;LOG:  statement: select * from test;LOG:  DB node id: 0 backend pid: 10127 statement: SELECT count(*) FROM pg_class AS c, pg_namespace AS n WHERE c.oid = pg_catalog.to_regclass('"test"') AND c.relnamespace = n.oid AND n.nspname = 'pg_catalog'LOG:  DB node id: 0 backend pid: 10127 statement: SELECT count(*) FROM pg_catalog.pg_class AS c, pg_namespace AS n WHERE c.relname = 'test' AND c.relnamespace = n.oid AND n.nspname ~ '^pg_temp_'LOG:  DB node id: 0 backend pid: 10127 statement: SELECT count(*) FROM pg_catalog.pg_class AS c WHERE c.oid = pg_catalog.to_regclass('"test"') AND c.relpersistence = 'u'LOG:  DB node id: 1 backend pid: 6084 statement: select * from test; id----  1  2(2 rows)sbtest=# show pool_nodes;LOG:  statement: show pool_nodes; node_id |  hostname  | port | status | lb_weight |  role   | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change---------+------------+------+--------+-----------+---------+------------+-------------------+-------------------+-------------------+------------------------+--------------------- 0       | 172.72.6.2 | 5432 | up     | 0.333333  | primary | 0          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 1       | 172.72.6.3 | 5432 | up     | 0.333333  | standby | 2          | true              | 0                 |                   |                        | 2021-04-26 09:54:44 2       | 172.72.6.4 | 5432 | up     | 0.333333  | standby | 1          | false             | 0                 |                   |                        | 2021-04-26 09:54:44(3 rows)sbtest=# insert into test values(3);LOG:  statement: insert into test values(3);LOG:  DB node id: 0 backend pid: 10127 statement: insert into test values(3);INSERT 0 1sbtest=# show pool_nodes;LOG:  statement: show pool_nodes; node_id |  hostname  | port | status | lb_weight |  role   | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change---------+------------+------+--------+-----------+---------+------------+-------------------+-------------------+-------------------+------------------------+--------------------- 0       | 172.72.6.2 | 5432 | up     | 0.333333  | primary | 0          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 1       | 172.72.6.3 | 5432 | up     | 0.333333  | standby | 2          | true              | 0                 |                   |                        | 2021-04-26 09:54:44 2       | 172.72.6.4 | 5432 | up     | 0.333333  | standby | 1          | false             | 0                 |                   |                        | 2021-04-26 09:54:44(3 rows)--  Do a few more tests postgres=# show pool_nodes;LOG:  statement: show pool_nodes; node_id |  hostname  | port | status | lb_weight |  role   | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change---------+------------+------+--------+-----------+---------+------------+-------------------+-------------------+-------------------+------------------------+--------------------- 0       | 172.72.6.2 | 5432 | up     | 0.333333  | primary | 6          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 1       | 172.72.6.3 | 5432 | up     | 0.333333  | standby | 7          | false             | 0                 |                   |                        | 2021-04-26 09:54:44 2       | 172.72.6.4 | 5432 | up     | 0.333333  | standby | 8          | true              | 0                 |                   |                        | 2021-04-26 09:54:44(3 rows)

You can see , Read operations are evenly allocated to 0、1 and 2 this 3 A node , Write operations are assigned only to 0 Node .

in addition , Relevant information will also be recorded in the alarm log （/postgresql/pgpool/pgpool.log）.

Apr 26 10:01:06 lhrpgpool pgpool[32486]: [20-1] 2021-04-26 10:01:06: pid 32486: LOG:  statement: show pool_nodes;Apr 26 10:01:23 lhrpgpool pgpool[32486]: [21-1] 2021-04-26 10:01:23: pid 32486: LOG:  statement: insert into test values(3);Apr 26 10:01:23 lhrpgpool pgpool[32486]: [22-1] 2021-04-26 10:01:23: pid 32486: LOG:  DB node id: 0 backend pid: 10127 statement: insert into test values(3);Apr 26 10:01:26 lhrpgpool pgpool[32486]: [23-1] 2021-04-26 10:01:26: pid 32486: LOG:  statement: show pool_nodes;Apr 26 10:02:02 lhrpgpool pgpool[32486]: [24-1] 2021-04-26 10:02:02: pid 32486: LOG:  statement: select * from test;Apr 26 10:02:02 lhrpgpool pgpool[32486]: [25-1] 2021-04-26 10:02:02: pid 32486: LOG:  DB node id: 1 backend pid: 6084 statement: select * from test;Apr 26 10:02:04 lhrpgpool pgpool[32485]: [25-1] 2021-04-26 10:02:04: pid 32485: LOG:  statement: select * from test;Apr 26 10:02:04 lhrpgpool pgpool[32485]: [26-1] 2021-04-26 10:02:04: pid 32485: LOG:  DB node id: 1 backend pid: 6069 statement: select * from test;Apr 26 10:02:06 lhrpgpool pgpool[32485]: [27-1] 2021-04-26 10:02:06: pid 32485: LOG:  statement: select * from test;Apr 26 10:02:06 lhrpgpool pgpool[32485]: [28-1] 2021-04-26 10:02:06: pid 32485: LOG:  DB node id: 1 backend pid: 6069 statement: select * from test;Apr 26 10:02:09 lhrpgpool pgpool[32476]: [24-1] 2021-04-26 10:02:09: pid 32476: LOG:  statement: select * from test;Apr 26 10:02:09 lhrpgpool pgpool[32476]: [25-1] 2021-04-26 10:02:09: pid 32476: LOG:  DB node id: 2 backend pid: 6085 statement: select * from test;Apr 26 10:02:11 lhrpgpool pgpool[32486]: [26-1] 2021-04-26 10:02:11: pid 32486: LOG:  statement: select * from test;Apr 26 10:02:11 lhrpgpool pgpool[32486]: [27-1] 2021-04-26 10:02:11: pid 32486: LOG:  DB node id: 1 backend pid: 6084 statement: select * from test;Apr 26 10:02:13 lhrpgpool pgpool[32486]: [28-1] 2021-04-26 10:02:13: pid 32486: LOG:  statement: select * from test;Apr 26 10:02:13 lhrpgpool pgpool[32486]: [29-1] 2021-04-26 10:02:13: pid 32486: LOG:  DB node id: 1 backend pid: 6084 statement: select * from test;