大家好， 今天大表哥和大家分享的是 PG 复制的明星产品工具 REPMGR。
REPMGR 是一款 PG 复制的管理工具。 如果小伙伴们有ORACLE data guard 的经验， REPMGR 和 DG 的 broker 有功能上是十分类似的。
支持复制节点的创建和管理，手动switch over, fail over 的命令。
REPMGR 最早是 业界著名的 第二象限公司开发的， 后来第二象限公司被EDB （另一家业界著名的企业版公司）收购后，现在属于EDB的一个当家的产品。 EDB 会提供付费的 7* 24 的 online 技术支持。
官网网址是： https://repmgr.org/ 官网上提供了 软件的下载的链接和 文档的入口。
这里下载软件前，需要注意一下 repmgr 和 pg 数据库软件版本兼容如下图：
https://repmgr.org/docs/5.2/install-requirements.html#INSTALL-COMPATIBILITY-MATRIX
我们这里的PG 版是 PG 13, 我们选择 对应支持的 repmgr 5.2 版本。
1.下载软件 version 5.2
INFRA [postgres@wqdcsrv3352 ~]# wget https://repmgr.org/download/repmgr-5.2.0.tar.gz --no-check-certificate
--2022-08-08 17:32:25-- https://repmgr.org/download/repmgr-5.2.0.tar.gz
Resolving repmgr.org (repmgr.org)... 88.99.121.168, 2a01:4f8:c17:9eaf::1
Connecting to repmgr.org (repmgr.org)|88.99.121.168|:443... connected.
WARNING: cannot verify repmgr.org's certificate, issued by ‘/C=US/O=Let's Encrypt/CN=R3’:
Issued certificate has expired.
HTTP request sent, awaiting response... 200 OK
Length: 446137 (436K) [application/x-gzip]
Saving to: ‘repmgr-5.2.0.tar.gz’
100%[===================================================================================================================================================>] 446,137 472KB/s in 0.9s
2022-08-08 17:32:26 (472 KB/s) - ‘repmgr-5.2.0.tar.gz’ saved [446137/446137]
2.源码编译安装
INFRA [postgres@wqdcsrv3352 postgreSQL]# tar -xvf repmgr-5.2.0.tar.gz
...
INFRA [postgres@wqdcsrv3352 repmgr-5.2.0]# ./configure
checking for a sed that does not truncate output... /bin/sed
checking for pg_config... /opt/postgreSQL/pg12/bin/pg_config
configure: building against PostgreSQL 12.3
checking for gnused... no
checking for gsed... no
checking for sed... yes
configure: creating ./config.status
config.status: creating Makefile
config.status: creating Makefile.global
config.status: creating config.h
INFRA [postgres@wqdcsrv3352 repmgr-5.2.0]# make install
...
/usr/bin/mkdir -p '/opt/postgreSQL/pg12/lib/postgresql'
/usr/bin/mkdir -p '/opt/postgreSQL/pg12/share/postgresql/extension'
/usr/bin/mkdir -p '/opt/postgreSQL/pg12/share/postgresql/extension'
/usr/bin/mkdir -p '/opt/postgreSQL/pg12/bin'
/usr/bin/install -c -m 755 repmgr.so '/opt/postgreSQL/pg12/lib/postgresql/repmgr.so'
/usr/bin/install -c -m 644 .//repmgr.control '/opt/postgreSQL/pg12/share/postgresql/extension/'
/usr/bin/install -c -m 644 .//repmgr--unpackaged--4.0.sql .//repmgr--unpackaged--5.1.sql .//repmgr--unpackaged--5.2.sql .//repmgr--4.0.sql .//repmgr--4.0--4.1.sql .//repmgr--4.1.sql .//repmgr--4.1--4.2.sql .//repmgr--4.2.sql .//repmgr--4.2--4.3.sql .//repmgr--4.3.sql .//repmgr--4.3--4.4.sql .//repmgr--4.4.sql .//repmgr--4.4--5.0.sql .//repmgr--5.0.sql .//repmgr--5.0--5.1.sql .//repmgr--5.1.sql .//repmgr--5.1--5.2.sql .//repmgr--5.2.sql '/opt/postgreSQL/pg12/share/postgresql/extension/'
/usr/bin/install -c -m 755 repmgr repmgrd '/opt/postgreSQL/pg12/bin/'
3.测试安装软件的版本
INFRA [postgres@wqdcsrv3352 repmgr-5.2.0]# repmgr --version
repmgr 5.2.0
4.配置节点之间的SSH免密配置并进行测试：
我们实验的节点是3节点： 10.67.39.49， 10.67.39.149，10.67.38.50
INFRA [postgres@wqdcsrv3352 repmgr-5.2.0]# ssh 10.67.39.49 hostname -i
10.67.39.49
INFRA [postgres@wqdcsrv3352 repmgr-5.2.0]# ssh 10.67.39.149 hostname -i
10.67.39.149
INFRA [postgres@wqdcsrv3352 repmgr-5.2.0]# ssh 10.67.38.50 hostname -i
10.67.38.50
5.PG 实例复制集的规划
6.在节点10.67.38.50 初始化主库：
构建数据路径：
INFRA [postgres@wqdcsrv3352 repmgr-5.2.0]# mkdir -p /data/postgreSQL/1998/{data,backups,scripts,archive_wals}
INFRA [postgres@wqdcsrv3352 repmgr-5.2.0]# chown -R postgres:postgres /data/postgreSQL/1998
初始化数据库：
INFRA [postgres@wqdcsrv3352 repmgr-5.2.0]# /opt/postgreSQL/pg12/bin/initdb -D /data/postgreSQL/1998/data/ -W
The files belonging to this database system will be owned by user "postgres".
This user must also own the server process.
The database cluster will be initialized with locale "en_US.UTF-8".
The default database encoding has accordingly been set to "UTF8".
The default text search configuration will be set to "english".
Data page checksums are disabled.
Enter new superuser password:
Enter it again:
fixing permissions on existing directory /data/postgreSQL/1998/data ... ok
creating subdirectories ... ok
selecting dynamic shared memory implementation ... posix
selecting default max_connections ... 100
selecting default shared_buffers ... 128MB
selecting default time zone ... Asia/Shanghai
creating configuration files ... ok
running bootstrap script ... ok
performing post-bootstrap initialization ... ok
syncing data to disk ... ok
initdb: warning: enabling "trust" authentication for local connections
You can change this by editing pg_hba.conf or using the option -A, or
--auth-local and --auth-host, the next time you run initdb.
Success. You can now start the database server using:
/opt/postgreSQL/pg12/bin/pg_ctl -D /data/postgreSQL/1998/data/ -l logfile start
修改参数文件 postgresql.conf （这里我们只修改一些必须的参数，非生产环境最优化的设置）
listen_addresses = '*' port = 1998 max_connections = 2000 wal_level = replica archive_mode = on archive_command = 'test ! -f /data/postgreSQL/1998/data/%f && cp %p /data/postgreSQL/1998/archive_wals/%f' max_wal_senders = 10 wal_keep_segments = 512 hot_standby = on shared_preload_libraries = 'repmgr'
修改访问权限文件 pg_hba.conf , 假设我们将要创建的账号是 repmgr , 存放相关复制信息元数据的数据库也是叫 repmgr
pg_hba.conf 追加内容如下：
local replication repmgr trust host replication repmgr 127.0.0.1/32 trust host replication repmgr 10.67.38.50/32 trust host replication repmgr 10.67.39.149/32 trust host replication repmgr 10.67.39.49/32 trust local repmgr repmgr trust host repmgr repmgr 127.0.0.1/32 trust host repmgr repmgr 10.67.38.50/32 trust host repmgr repmgr 10.67.39.149/32 trust host repmgr repmgr 10.67.39.49/32 trust
启动主库：
INFRA [postgres@wqdcsrv3352 data]# /opt/postgreSQL/pg12/bin/pg_ctl -D /data/postgreSQL/1998/data/ -l logfile start
waiting for server to start.... done
server started
7.在主库上，创建数据库repmgr 和相关的账户 repmgr
postgres@[local:/tmp]:1998=#83348 create user repmgr superuser password 'repmgr';
CREATE ROLE
postgres@[local:/tmp]:1998=#83348 ALTER USER repmgr SET search_path TO repmgr, "$user", public;
ALTER ROLE
postgres@[local:/tmp]:1998=#83348 create database repmgr owner repmgr;
CREATE DATABASE
postgres@[local:/tmp]:1998=#83348 \l
List of databases
Name | Owner | Encoding | Collate | Ctype | Access privileges
-----------+----------+----------+-------------+-------------+-----------------------
postgres | postgres | UTF8 | en_US.UTF-8 | en_US.UTF-8 |
repmgr | repmgr | UTF8 | en_US.UTF-8 | en_US.UTF-8 |
template0 | postgres | UTF8 | en_US.UTF-8 | en_US.UTF-8 | =c/postgres +
| | | | | postgres=CTc/postgres
template1 | postgres | UTF8 | en_US.UTF-8 | en_US.UTF-8 | =c/postgres +
| | | | | postgres=CTc/postgres
(4 rows)
8.创建repmgr 的配置文件： repmgr.conf
node_id=1 node_name=pg50 conninfo='host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5' data_directory='/data/postgreSQL/1998/data' pg_bindir='/opt/postgreSQL/pg12/bin'
9.注册主节点到 repmgr 的管理工具中
INFRA [postgres@wqdcsrv3352 postgreSQL]# repmgr -f /opt/postgreSQL/repmgr.conf primary register
INFO: connecting to primary database...
NOTICE: attempting to install extension "repmgr"
NOTICE: "repmgr" extension successfully installed
NOTICE: primary node record (ID: 1) registered
查看注册节点的状态：
INFRA [postgres@wqdcsrv3352 postgreSQL]# repmgr -f /opt/postgreSQL/repmgr.conf cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+------+---------+-----------+----------+----------+----------+----------+----------------------------------------------------------------------------------------
1 | pg50 | primary | * running | | default | 100 | 1 | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
我们也可以 metadata 的元数据库中 查看节点的信息：
repmgr@[local:/tmp]:1998=#94078 \x
Expanded display is on.
repmgr@[local:/tmp]:1998=#94078 SELECT * FROM repmgr.nodes;
-[ RECORD 1 ]----+---------------------------------------------------------------------------------------
node_id | 1
upstream_node_id |
active | t
node_name | pg50
type | primary
location | default
priority | 100
conninfo | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
repluser | repmgr
slot_name |
config_file | /opt/postgreSQL/repmgr.conf
10.克隆standby 节点 ， 从节点克隆之前，我们同样需要：
a）安装repmgr 软件 : 同样的步骤，参考上面的部分
b）创建数据库实例的基本路径：
INFRA [postgres@wqdcsrv3353 ~]# mkdir -p /data/postgreSQL/1998/{data,backups,scripts,archive_wals}
INFRA [postgres@wqdcsrv3353 ~]# chown -R postgres:postgres /data/postgreSQL/1998
c)编辑repmgr的配置文件repmgr.conf，每个节点上添加各自的节点相关的信息
node_id=1 node_name=pg50 conninfo='host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5' data_directory='/data/postgreSQL/1998/data' pg_bindir='/opt/postgreSQL/pg12/bin' node_id=2 node_name=pg49 conninfo='host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5' data_directory='/data/postgreSQL/1998/data' pg_bindir='/opt/postgreSQL/pg12/bin' node_id=3 node_name=pg149 conninfo='host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5' data_directory='/data/postgreSQL/1998/data' pg_bindir='/opt/postgreSQL/pg12/bin'
d)执行clone 从节点的命令， 可以加上参数-- dry-run 测试一下命令
注意这条命令需要在从节点上运行： 我们可以看到 --dry-run 会进行一下extension 检查和一些必要的参数的检查：
全部符合条件会输出： INFO: all prerequisites for “standby clone” are met
INFRA [postgres@wqdcsrv3354 postgreSQL]# repmgr -h 10.67.38.50 -U repmgr -p 1998 -d repmgr -f /opt/postgreSQL/repmgr.conf standby clone --dry-run
NOTICE: destination directory "/data/postgreSQL/1998/data" provided
INFO: connecting to source node
DETAIL: connection string is: host=10.67.38.50 user=repmgr port=1998 dbname=repmgr
DETAIL: current installation size is 31 MB
INFO: "repmgr" extension is installed in database "repmgr"
INFO: replication slot usage not requested; no replication slot will be set up for this standby
INFO: parameter "max_wal_senders" set to 10
NOTICE: checking for available walsenders on the source node (2 required)
INFO: sufficient walsenders available on the source node
DETAIL: 2 required, 10 available
NOTICE: checking replication connections can be made to the source server (2 required)
INFO: required number of replication connections could be made to the source server
DETAIL: 2 replication connections required
WARNING: data checksums are not enabled and "wal_log_hints" is "off"
DETAIL: pg_rewind requires "wal_log_hints" to be enabled
NOTICE: standby will attach to upstream node 1
HINT: consider using the -c/--fast-checkpoint option
INFO: all prerequisites for "standby clone" are met
我们去掉 – dry-run ： 正式执行 clone 命令 repmgr -h 10.67.38.50 -U repmgr -p 1998 -d repmgr -f /opt/postgreSQL/repmgr.conf standby clone。
我们可以看到 clone 的过程：
a)测试主库的连接性
b)check 是否有 复制槽
c)check walsenders, replication connections
d)pg_rewind 提示我们需要打开参数 wal_log_hints
e)检查实例的文件夹权限： "/data/postgreSQL/1998/data
f)生成数据库的备份命令： /opt/postgreSQL/pg12/bin/pg_basebackup -l “repmgr base backup” -D /data/postgreSQL/1998/data -h 10.67.38.50 -p 1998 -U repmgr -X stream
g)提示数据库的命令： pg_ctl -D /data/postgreSQL/1998/data start
h)提示数据库启动之后，standby 注册到repmgr 的命令 ：repmgr standby register
INFRA [postgres@wqdcsrv3354 postgreSQL]# repmgr -h 10.67.38.50 -U repmgr -p 1998 -d repmgr -f /opt/postgreSQL/repmgr.conf standby clone
NOTICE: destination directory "/data/postgreSQL/1998/data" provided
INFO: connecting to source node
DETAIL: connection string is: host=10.67.38.50 user=repmgr port=1998 dbname=repmgr
DETAIL: current installation size is 31 MB
INFO: replication slot usage not requested; no replication slot will be set up for this standby
NOTICE: checking for available walsenders on the source node (2 required)
NOTICE: checking replication connections can be made to the source server (2 required)
WARNING: data checksums are not enabled and "wal_log_hints" is "off"
DETAIL: pg_rewind requires "wal_log_hints" to be enabled
INFO: checking and correcting permissions on existing directory "/data/postgreSQL/1998/data"
NOTICE: starting backup (using pg_basebackup)...
HINT: this may take some time; consider using the -c/--fast-checkpoint option
INFO: executing:
/opt/postgreSQL/pg12/bin/pg_basebackup -l "repmgr base backup" -D /data/postgreSQL/1998/data -h 10.67.38.50 -p 1998 -U repmgr -X stream
NOTICE: standby clone (using pg_basebackup) complete
NOTICE: you can now start your PostgreSQL server
HINT: for example: pg_ctl -D /data/postgreSQL/1998/data start
HINT: after starting the server, you need to register this standby with "repmgr standby register"
我们按照提示，启动standby 数据库实例 ：
INFRA [postgres@wqdcsrv3354 postgreSQL]# /opt/postgreSQL/pg12/bin/pg_ctl -D /data/postgreSQL/1998/data -l logfile start
waiting for server to start.... done
server started
注册standby database 到 repmgr 的元数据库：
INFRA [postgres@wqdcsrv3354 postgreSQL]# repmgr -f /opt/postgreSQL/repmgr.conf standby register
INFO: connecting to local node "pg149" (ID: 3)
INFO: connecting to primary database
WARNING: --upstream-node-id not supplied, assuming upstream node is primary (node ID 1)
INFO: standby registration complete
NOTICE: standby node "pg149" (ID: 3) successfully registered
11）查看集群的状态以及数据库中元数据
INFRA [postgres@wqdcsrv3352 data]# repmgr -f /opt/postgreSQL/repmgr.conf cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+-------+---------+-----------+----------+----------+----------+----------+-----------------------------------------------------------------------------------------
1 | pg50 | primary | * running | | default | 100 | 1 | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
2 | pg49 | standby | running | pg50 | default | 100 | 1 | host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
3 | pg149 | standby | running | pg50 | default | 100 | 1 | host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
repmgr@[local:/tmp]:1998=#100555 SELECT * FROM repmgr.nodes;
node_id | upstream_node_id | active | node_name | type | location | priority | conninfo | repluser | slot_
name | config_file
---------+------------------+--------+-----------+---------+----------+----------+-----------------------------------------------------------------------------------------+----------+------
-----+-----------------------------
1 | | t | pg50 | primary | default | 100 | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5 | repmgr |
| /opt/postgreSQL/repmgr.conf
3 | 1 | t | pg149 | standby | default | 100 | host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5 | repmgr |
| /opt/postgreSQL/repmgr.conf
2 | 1 | t | pg49 | standby | default | 100 | host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5 | repmgr |
| /opt/postgreSQL/repmgr.conf
(3 rows)
12)测试switch over 命令
我们假设主库需要进行存储级别的维护，需要关机。 我们需要把主节点切换到 10.67.39.49 这个节点上。
我们这个时候只需要在 10.67.39.49 执行 switch over 命令即可：
switch over 的大致流程许下：
1）关闭老的 primary pg50 :/opt/postgreSQL/pg12/bin/pg_ctl -D ‘/data/postgreSQL/1998/data’ -W -m fast stop
2)等待老的主库彻底关闭后，在 server “pg49” ，进行 pg_promote()
3）重启启动老的 primary pg50， 降级成 standby 数据库， 指向复制源 server “pg49”
4)sibling nodes 兄弟节点 同样进行了 重定向 复制源 server “pg49”
5）整个switch over 过程结束
INFRA [postgres@wqdcsrv3353 data]# repmgr -f /opt/postgreSQL/repmgr.conf standby switchover --siblings-follow
NOTICE: executing switchover on node "pg49" (ID: 2)
NOTICE: local node "pg49" (ID: 2) will be promoted to primary; current primary "pg50" (ID: 1) will be demoted to standby
NOTICE: stopping current primary node "pg50" (ID: 1)
NOTICE: issuing CHECKPOINT on node "pg50" (ID: 1)
DETAIL: executing server command "/opt/postgreSQL/pg12/bin/pg_ctl -D '/data/postgreSQL/1998/data' -W -m fast stop"
INFO: checking for primary shutdown; 1 of 60 attempts ("shutdown_check_timeout")
INFO: checking for primary shutdown; 2 of 60 attempts ("shutdown_check_timeout")
NOTICE: current primary has been cleanly shut down at location 0/6000028
NOTICE: promoting standby to primary
DETAIL: promoting server "pg49" (ID: 2) using pg_promote()
NOTICE: waiting up to 60 seconds (parameter "promote_check_timeout") for promotion to complete
NOTICE: STANDBY PROMOTE successful
DETAIL: server "pg49" (ID: 2) was successfully promoted to primary
INFO: local node 1 can attach to rejoin target node 2
DETAIL: local node's recovery point: 0/6000028; rejoin target node's fork point: 0/60000A0
NOTICE: setting node 1's upstream to node 2
WARNING: unable to ping "host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5"
DETAIL: PQping() returned "PQPING_NO_RESPONSE"
NOTICE: starting server using "/opt/postgreSQL/pg12/bin/pg_ctl -w -D '/data/postgreSQL/1998/data' start"
NOTICE: NODE REJOIN successful
DETAIL: node 1 is now attached to node 2
NOTICE: node "pg49" (ID: 2) promoted to primary, node "pg50" (ID: 1) demoted to standby
NOTICE: executing STANDBY FOLLOW on 1 of 1 siblings
INFO: STANDBY FOLLOW successfully executed on all reachable sibling nodes
NOTICE: switchover was successful
DETAIL: node "pg49" is now primary and node "pg50" is attached as standby
NOTICE: STANDBY SWITCHOVER has completed successfully
我们再次观察整个复制集的状态： pg49 promote 成为了 primary DB
INFRA [postgres@wqdcsrv3352 data]# repmgr -f /opt/postgreSQL/repmgr.conf cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+-------+---------+-----------+----------+----------+----------+----------+-----------------------------------------------------------------------------------------
1 | pg50 | standby | running | pg49 | default | 100 | 1 | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
2 | pg49 | primary | * running | | default | 80 | 2 | host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
3 | pg149 | standby | running | pg49 | default | 20 | 1 | host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
13）测试 failover , 手动 promote 节点 。
当前复制集的状态：
INFRA [postgres@wqdcsrv3352 ~]# repmgr -f /opt/postgreSQL/repmgr.conf cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+-------+---------+-----------+----------+----------+----------+----------+-----------------------------------------------------------------------------------------
1 | pg50 | standby | running | pg49 | default | 100 | 2 | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
2 | pg49 | primary | * running | | default | 100 | 2 | host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
3 | pg149 | standby | running | pg49 | default | 100 | 2 | host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
手动关闭 kill -9 掉 主节点的 进程， 模拟主库 crash
INFRA [postgres@wqdcsrv3353 ~]# kill -9 62513
再次查看复制集的状态： 这个时候 原来的主节点已经变成了 unreachable 的状态了。
INFRA [postgres@wqdcsrv3352 ~]# repmgr -f /opt/postgreSQL/repmgr.conf cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+-------+---------+---------------+----------+----------+----------+----------+-----------------------------------------------------------------------------------------
1 | pg50 | standby | running | ? pg49 | default | 100 | 2 | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
2 | pg49 | primary | ? unreachable | ? | default | 100 | | host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
3 | pg149 | standby | running | ? pg49 | default | 100 | 2 | host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
WARNING: following issues were detected
- unable to connect to node "pg50" (ID: 1)'s upstream node "pg49" (ID: 2)
- unable to determine if node "pg50" (ID: 1) is attached to its upstream node "pg49" (ID: 2)
- unable to connect to node "pg49" (ID: 2)
- node "pg49" (ID: 2) is registered as an active primary but is unreachable
- unable to connect to node "pg149" (ID: 3)'s upstream node "pg49" (ID: 2)
- unable to determine if node "pg149" (ID: 3) is attached to its upstream node "pg49" (ID: 2)
HINT: execute with --verbose option to see connection error messages
我们现在用 repmgr的 promote 命令来把 pg50 这个节点提升为 主节点：
INFRA [postgres@wqdcsrv3352 ~]# repmgr -f /opt/postgreSQL/repmgr.conf standby promote --siblings-follow
NOTICE: promoting standby to primary
DETAIL: promoting server "pg50" (ID: 1) using pg_promote()
NOTICE: waiting up to 60 seconds (parameter "promote_check_timeout") for promotion to complete
NOTICE: STANDBY PROMOTE successful
DETAIL: server "pg50" (ID: 1) was successfully promoted to primary
NOTICE: executing STANDBY FOLLOW on 1 of 1 siblings
INFO: STANDBY FOLLOW successfully executed on all reachable sibling nodes
我们再次查看复制集的状态： pg50 已经被提升为 主库 ， 原来的主库 pg49 已经标记为 failed 的转态
INFRA [postgres@wqdcsrv3352 ~]# repmgr -f /opt/postgreSQL/repmgr.conf cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+-------+---------+-----------+----------+----------+----------+----------+-----------------------------------------------------------------------------------------
1 | pg50 | primary | * running | | default | 100 | 3 | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
2 | pg49 | primary | - failed | ? | default | 100 | | host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
3 | pg149 | standby | running | pg50 | default | 100 | 2 | host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
老的主库 pg49 我们可以通过命令： repmgr node rejoin
对于主库的 rejoin 的命令， 需要调用 pg_rewind(), 所以参数 wal_log_hints 必须要打开
否则会报错误：
ERROR: --force-rewind specified but pg_rewind cannot be used
DETAIL: “wal_log_hints” is set to “off” and data checksums are disabled
INFRA [postgres@wqdcsrv3353 data]# repmgr -f /opt/postgreSQL/repmgr.conf node rejoin -d 'host=10.67.38.50 dbname=repmgr user=repmgr password=repmgr port=1998' --force-rewind --dry-run
INFO: replication connection to the rejoin target node was successful
INFO: local and rejoin target system identifiers match
DETAIL: system identifier is 7129704771708963420
NOTICE: pg_rewind execution required for this node to attach to rejoin target node 1
DETAIL: rejoin target server's timeline 3 forked off current database system timeline 2 before current recovery point 0/70000A0
ERROR: --force-rewind specified but pg_rewind cannot be used
DETAIL: "wal_log_hints" is set to "off" and data checksums are disabled
我们主库开启 参数 wal_log_hints = on 后，再次执行rejoin 命令：
INFRA [postgres@wqdcsrv3353 data]# repmgr -f /opt/postgreSQL/repmgr.conf node rejoin -d 'host=10.67.38.50 dbname=repmgr user=repmgr password=repmgr port=1998' --force-rewind
NOTICE: pg_rewind execution required for this node to attach to rejoin target node 1
DETAIL: rejoin target server's timeline 3 forked off current database system timeline 2 before current recovery point 0/70000A0
NOTICE: executing pg_rewind
DETAIL: pg_rewind command is "/opt/postgreSQL/pg12/bin/pg_rewind -D '/data/postgreSQL/1998/data' --source-server='host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5'"
NOTICE: 0 files copied to /data/postgreSQL/1998/data
NOTICE: setting node 2's upstream to node 1
WARNING: unable to ping "host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5"
DETAIL: PQping() returned "PQPING_NO_RESPONSE"
NOTICE: starting server using "/opt/postgreSQL/pg12/bin/pg_ctl -w -D '/data/postgreSQL/1998/data' start"
NOTICE: NODE REJOIN successful
DETAIL: node 2 is now attached to node 1
我们再次查看集群的状态： pg49 节点恢复了正常
INFRA [postgres@wqdcsrv3352 data]# repmgr -f /opt/postgreSQL/repmgr.conf cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+-------+---------+-----------+----------+----------+----------+----------+-----------------------------------------------------------------------------------------
1 | pg50 | primary | * running | | default | 100 | 3 | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
2 | pg49 | standby | running | pg50 | default | 100 | 2 | host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
3 | pg149 | standby | running | pg50 | default | 100 | 3 | host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
14)repmgrd 自动 failover 测试
repmgr 同样支持自动的 failover 操作， 需要启动一个 repmgrd 的后台进程，需要在配置文件/opt/postgreSQL/repmgr.conf 中添加如下的参数：
location='default'
failover='automatic'
promote_command='/opt/postgreSQL/pg12/bin/repmgr standby promote -f /opt/postgreSQL/repmgr.conf --log-to-file'
follow_command='/opt/postgreSQL/pg12/bin/repmgr standby follow -f /opt/postgreSQL/repmgr.conf --log-to-file --upstream-node-id=%n'
log_file='/opt/postgreSQL/repmgrd.log'
monitoring_history=true
monitor_interval_secs=5
reconnect_attempts=6
reconnect_interval=5
启动 remgrd 进程：
INFRA [postgres@wqdcsrv3352 ~]# repmgrd -f /opt/postgreSQL/repmgr.conf --pid-file /tmp/repmgrd.pid
[2022-08-10 13:46:31] [NOTICE] redirecting logging output to "/opt/postgreSQL/repmgrd.log"
查看后天进程：
INFRA [postgres@wqdcsrv3352 ~]# ps -ef | grep repmgrd
postgres 80814 1 0 13:46 ? 00:00:00 repmgrd -f /opt/postgreSQL/repmgr.conf --pid-file /tmp/repmgrd.pid
postgres 86653 74937 0 13:55 pts/1 00:00:00 grep --color=auto repmgrd
测试自动 fail over 的过程， 根据节点的优先级 priority来选择 提升的主节点。
我们手动的设置节点的优先级如下： pg149 priority = 20 , pg49 priority = 80
在配置文件中，添加参数 priority=20 。
node_id=3 node_name=pg149 conninfo='host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5' data_directory='/data/postgreSQL/1998/data' pg_bindir='/opt/postgreSQL/pg12/bin' priority=20
对从节点进行重新注册操作（unregister | register)
INFRA [postgres@wqdcsrv3354 data]# repmgr -f /opt/postgreSQL/repmgr.conf standby unregister
INFO: connecting to local standby
INFO: connecting to primary database
NOTICE: unregistering node 3
INFO: standby unregistration complete
INFRA [postgres@wqdcsrv3354 data]# repmgr -f /opt/postgreSQL/repmgr.conf standby register --upstream-node-id=1
INFO: connecting to local node "pg149" (ID: 3)
INFO: connecting to primary database
INFO: standby registration complete
我们再次查看 节点的 priority: 可以看到 10.67.39.49 的优先级是 80 ， 10.67.39.149 的 优先级是20，
INFRA [postgres@wqdcsrv3352 data]# repmgr -f /opt/postgreSQL/repmgr.conf cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+-------+---------+-----------+----------+----------+----------+----------+-----------------------------------------------------------------------------------------
1 | pg50 | primary | * running | | default | 100 | 1 | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
2 | pg49 | standby | running | pg50 | default | 80 | 1 | host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
3 | pg149 | standby | running | pg50 | default | 20 | 1 | host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
我们接下来，kill 掉主库的进程， 测试节点自动 fail over:
INFRA [postgres@wqdcsrv3352 ~]# kill -9 108471
这个时候，我们去访问 pg49 的节点查看复制集的状态， pg49 已经 提升为主库。
INFRA [postgres@wqdcsrv3353 postgreSQL]# repmgr -f /opt/postgreSQL/repmgr.conf cluster show
ID | Name | Role | Status | Upstream | Location | Priority | Timeline | Connection string
----+-------+---------+-----------+----------+----------+----------+----------+-----------------------------------------------------------------------------------------
1 | pg50 | primary | - failed | ? | default | 100 | | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
2 | pg49 | primary | * running | | default | 80 | 4 | host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
3 | pg149 | standby | running | pg49 | default | 20 | 3 | host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5
WARNING: following issues were detected
- unable to connect to node "pg50" (ID: 1)
HINT: execute with --verbose option to see connection error messages
我们可以通过rejoin 命令， 把fail 掉的节点 再次加入到 复制集中：
INFRA [postgres@wqdcsrv3352 ~]# repmgr -f /opt/postgreSQL/repmgr.conf node rejoin -d 'host=10.67.39.49 dbname=repmgr user=repmgr password=repmgr port=1998' --force-rewind --dry-run
ERROR: database is not shut down cleanly
DETAIL: pg_rewind will not be able to run
HINT: database should be restarted then shut down cleanly after crash recovery completes
由于我们是 kill -9 杀掉数据库的进程，我们需要先启动/关闭一下数据库，进行必要的恢复。
INFRA [postgres@wqdcsrv3352 ~]# /opt/postgreSQL/pg12/bin/pg_ctl -D /data/postgreSQL/1998/data start -l logfile
pg_ctl: another server might be running; trying to start server anyway
waiting for server to start.... done
server started
INFRA [postgres@wqdcsrv3352 ~]# /opt/postgreSQL/pg12/bin/pg_ctl -D /data/postgreSQL/1998/data stop -m fast
waiting for server to shut down.... done
server stopped
再次执行 rejoin 命令
INFRA [postgres@wqdcsrv3352 ~]# repmgr -f /opt/postgreSQL/repmgr.conf node rejoin -d 'host=10.67.39.49 dbname=repmgr user=repmgr password=repmgr port=1998' --force-rewind
NOTICE: pg_rewind execution required for this node to attach to rejoin target node 2
DETAIL: rejoin target server's timeline 4 forked off current database system timeline 3 before current recovery point 0/8000028
NOTICE: executing pg_rewind
DETAIL: pg_rewind command is "/opt/postgreSQL/pg12/bin/pg_rewind -D '/data/postgreSQL/1998/data' --source-server='host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5'"
NOTICE: 0 files copied to /data/postgreSQL/1998/data
NOTICE: setting node 1's upstream to node 2
WARNING: unable to ping "host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5"
DETAIL: PQping() returned "PQPING_NO_RESPONSE"
NOTICE: starting server using "/opt/postgreSQL/pg12/bin/pg_ctl -w -D '/data/postgreSQL/1998/data' start"
NOTICE: NODE REJOIN successful
DETAIL: node 1 is now attached to node 2
此时复制集的状态全部恢复正常：
ID | Name | Role | Status | Upstream | repmgrd | PID | Paused? | Upstream last seen
----+-------+---------+-----------+----------+---------+-------+---------+--------------------
1 | pg50 | standby | running | pg49 | running | 80814 | no | 2 second(s) ago
2 | pg49 | primary | * running | | running | 46755 | no | n/a
3 | pg149 | standby | running | pg49 | running | 27960 | no | 2 second(s) ago
15） 暂停 repmgrd 自动 failover 的操作：
现实的环境中，有这么一种情况：就是按照计划内可以down机进行维护，这个时候，我们是不希望发生 auto failover 的。
我们通过命令 repmgr -f /opt/postgreSQL/repmgr.conf service pause 把这个 auto failover 暂停掉：
INFRA [postgres@wqdcsrv3352 ~]# repmgr -f /opt/postgreSQL/repmgr.conf service status
ID | Name | Role | Status | Upstream | repmgrd | PID | Paused? | Upstream last seen
----+-------+---------+-----------+----------+---------+-------+---------+--------------------
1 | pg50 | standby | running | pg49 | running | 80814 | no | 0 second(s) ago
2 | pg49 | primary | * running | | running | 46755 | no | n/a
3 | pg149 | standby | running | pg49 | running | 27960 | no | 4 second(s) ago
INFRA [postgres@wqdcsrv3352 ~]# repmgr -f /opt/postgreSQL/repmgr.conf service pause
NOTICE: node 1 (pg50) paused
NOTICE: node 2 (pg49) paused
NOTICE: node 3 (pg149) paused
INFRA [postgres@wqdcsrv3352 ~]# repmgr -f /opt/postgreSQL/repmgr.conf service status
ID | Name | Role | Status | Upstream | repmgrd | PID | Paused? | Upstream last seen
----+-------+---------+-----------+----------+---------+-------+---------+--------------------
1 | pg50 | standby | running | pg49 | running | 80814 | yes | 1 second(s) ago
2 | pg49 | primary | * running | | running | 46755 | yes | n/a
3 | pg149 | standby | running | pg49 | running | 27960 | yes | 0 second(s) ago
测试手动关闭一下主库：
INFRA [postgres@wqdcsrv3353 postgreSQL]# /opt/postgreSQL/pg12/bin/pg_ctl -D /data/postgreSQL/1998/data stop -m fast
waiting for server to shut down.... done
server stopped
查看复制集状态： 2个standby 是正常的running , 没有新的提升的 primary 节点
INFRA [postgres@wqdcsrv3352 ~]# repmgr -f /opt/postgreSQL/repmgr.conf service status
ID | Name | Role | Status | Upstream | repmgrd | PID | Paused? | Upstream last seen
----+-------+---------+---------------+----------+---------+-------+---------+--------------------
1 | pg50 | standby | running | ? pg49 | running | 80814 | yes | 32 second(s) ago
2 | pg49 | primary | ? unreachable | ? | n/a | n/a | n/a | n/a
3 | pg149 | standby | running | ? pg49 | running | 27960 | yes | 32 second(s) ago
WARNING: following issues were detected
- unable to connect to node "pg50" (ID: 1)'s upstream node "pg49" (ID: 2)
- unable to determine if node "pg50" (ID: 1) is attached to its upstream node "pg49" (ID: 2)
- unable to connect to node "pg49" (ID: 2)
- node "pg49" (ID: 2) is registered as an active primary but is unreachable
- unable to connect to node "pg149" (ID: 3)'s upstream node "pg49" (ID: 2)
- unable to determine if node "pg149" (ID: 3) is attached to its upstream node "pg49" (ID: 2)
HINT: execute with --verbose option to see connection error messages
15）元数据库的一些表信息：
4张表和2张试图： 你可以利用这些表和试图中的信息，进行一些 trouble shutting 或者 定制一些监控的任务
repmgr@[local:/tmp]:1998=#24783 \d
List of relations
Schema | Name | Type | Owner
--------+--------------------+-------+--------
repmgr | events | table | repmgr
repmgr | monitoring_history | table | repmgr
repmgr | nodes | table | repmgr
repmgr | replication_status | view | repmgr
repmgr | show_nodes | view | repmgr
repmgr | voting_term | table | repmgr
(6 rows)
select * from events; --一些重要的事件，包括一些手动的switch over， register， clone .. 等命令
repmgr@[local:/tmp]:1998=#24783 select * from events;
node_id | event | successful | event_timestamp | details
---------+---------------------------+------------+-------------------------------+--------------------------------------------------------------------------------------
1 | cluster_created | t | 2022-08-09 11:32:37.132352+08 |
1 | primary_register | t | 2022-08-09 11:32:37.135251+08 |
3 | standby_clone | t | 2022-08-09 14:21:05.635694+08 | cloned from host "10.67.38.50", port 1998; backup method: pg_basebackup; --force: N
3 | standby_register | t | 2022-08-09 14:37:23.449443+08 | standby registration succeeded; upstream node ID is 1
3 | standby_clone | t | 2022-08-09 14:42:37.715999+08 | cloned from host "10.67.38.50", port 1998; backup method: pg_basebackup; --force: N
3 | standby_register | t | 2022-08-09 14:43:54.61979+08 | standby registration succeeded; upstream node ID is 1 (-F/--force option was used)
3 | standby_unregister | t | 2022-08-09 14:47:36.042825+08 |
3 | standby_register | t | 2022-08-09 14:53:44.67895+08 | standby registration succeeded; upstream node ID is 1
2 | standby_register | t | 2022-08-09 14:55:23.007685+08 | standby registration succeeded; upstream node ID is 1
2 | standby_unregister | t | 2022-08-09 15:17:55.424511+08 |
2 | standby_register | t | 2022-08-09 15:18:00.450973+08 | standby registration succeeded; upstream node ID is 1
3 | standby_unregister | t | 2022-08-09 15:18:11.450469+08 |
3 | standby_register | t | 2022-08-09 15:18:33.773641+08 | standby registration succeeded; upstream node ID is 1
2 | standby_promote | t | 2022-08-09 15:30:53.706084+08 | server "pg49" (ID: 2) was successfully promoted to primary
2 | standby_switchover | t | 2022-08-09 15:30:53.926593+08 | node "pg49" (ID: 2) promoted to primary, node "pg50" (ID: 1) demoted to standby
3 | standby_follow | t | 2022-08-09 15:30:59.287742+08 | standby attached to upstream node "pg49" (ID: 2)
1 | standby_promote | t | 2022-08-10 10:30:23.283646+08 | server "pg50" (ID: 1) was successfully promoted to primary
3 | standby_follow | t | 2022-08-10 10:30:28.632501+08 | standby attached to upstream node "pg50" (ID: 1)
2 | node_rejoin | t | 2022-08-10 11:11:29.473095+08 | node 2 is now attached to node 1
1 | repmgrd_start | t | 2022-08-10 13:46:31.323124+08 | monitoring cluster primary "pg50" (ID: 1)
2 | repmgrd_start | t | 2022-08-10 13:52:07.285655+08 | monitoring connection to upstream node "pg50" (ID: 1)
3 | repmgrd_start | t | 2022-08-10 13:52:11.080653+08 | monitoring connection to upstream node "pg50" (ID: 1)
2 | standby_promote | t | 2022-08-10 14:05:24.727227+08 | server "pg49" (ID: 2) was successfully promoted to primary
2 | repmgrd_failover_promote | t | 2022-08-10 14:05:24.733076+08 | node "pg49" (ID: 2) promoted to primary; old primary "pg50" (ID: 1) marked as failed
2 | repmgrd_reload | t | 2022-08-10 14:05:24.734995+08 | monitoring cluster primary "pg49" (ID: 2)
3 | standby_follow | t | 2022-08-10 14:05:25.844297+08 | standby attached to upstream node "pg49" (ID: 2)
3 | repmgrd_failover_follow | t | 2022-08-10 14:05:25.853657+08 | node "pg149" (ID: 3) now following new upstream node "pg49" (ID: 2)
2 | child_node_new_connect | t | 2022-08-10 14:05:29.745089+08 | new standby "pg149" (ID: 3) has connected
1 | node_rejoin | t | 2022-08-10 15:23:45.44888+08 | node 1 is now attached to node 2
1 | repmgrd_standby_reconnect | t | 2022-08-10 15:23:48.37959+08 | node restored as standby after 25 seconds, monitoring connection to upstream node 2
2 | child_node_new_connect | t | 2022-08-10 15:23:48.472264+08 | new standby "pg50" (ID: 1) has connected
(31 rows）
repmgr@[local:/tmp]:1998=#24783 select * from monitoring_history; -- 各个节点监控的信息
primary_node_id | standby_node_id | last_monitor_time | last_apply_time | last_wal_primary_location | last_wal_standby_location | replication_lag | apply_lag
-----------------+-----------------+-------------------------------+-------------------------------+---------------------------+---------------------------+-----------------+-----------
1 | 2 | 2022-08-10 13:52:07.552218+08 | 2022-08-10 13:52:07.28584+08 | 0/7023A50 | 0/7023A50 | 0 | 0
1 | 3 | 2022-08-10 13:52:18.815005+08 | 2022-08-10 13:52:11.080848+08 | 0/70272E0 | 0/70272E0 | 0 | 0
1 | 2 | 2022-08-10 13:52:12.560008+08 | 2022-08-10 13:52:11.084817+08 | 0/70273C0 | 0/70273C0 | 0 | 0
1 | 3 | 2022-08-10 13:52:23.825027+08 | 2022-08-10 13:52:12.299763+08 | 0/70274A0 | 0/70274A0 | 0 | 0
1 | 2 | 2022-08-10 13:52:17.566368+08 | 2022-08-10 13:52:16.094428+08 | 0/7027580 | 0/7027580 | 0 | 0
1 | 3 | 2022-08-10 13:52:28.832021+08 | 2022-08-10 13:52:17.306121+08 | 0/7027660 | 0/7027660 | 0 | 0
1 | 2 | 2022-08-10 13:52:22.573409+08 | 2022-08-10 13:52:21.101802+08 | 0/7027740 | 0/7027740 | 0 | 0
repmgr@[local:/tmp]:1998=#24783 select * from nodes; -- 各个节点的信息
node_id | upstream_node_id | active | node_name | type | location | priority | conninfo | repluser | slot_
name | config_file
---------+------------------+--------+-----------+---------+----------+----------+-----------------------------------------------------------------------------------------+----------+------
-----+-----------------------------
2 | | t | pg49 | primary | default | 80 | host=10.67.39.49 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5 | repmgr |
| /opt/postgreSQL/repmgr.conf
3 | 2 | t | pg149 | standby | default | 20 | host=10.67.39.149 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5 | repmgr |
| /opt/postgreSQL/repmgr.conf
1 | 2 | t | pg50 | standby | default | 100 | host=10.67.38.50 port=1998 user=repmgr password=repmgr dbname=repmgr connect_timeout=5 | repmgr |
| /opt/postgreSQL/repmgr.conf
(3 rows)
repmgr@[local:/tmp]:1998=#24783 select * from replication_status; -- 各个节点的复制信息，可用于节点的延时监控
-[ RECORD 1 ]-------------+------------------------------
primary_node_id | 2
standby_node_id | 1
standby_name | pg50
node_type | standby
active | t
last_monitor_time | 2022-08-10 15:46:19.87505+08
last_wal_primary_location | 0/714CC28
last_wal_standby_location | 0/714CC28
replication_lag | 0 bytes
replication_time_lag | 00:00:00
apply_lag | 0 bytes
communication_time_lag | 00:00:01.822418
-[ RECORD 2 ]-------------+------------------------------
primary_node_id | 2
standby_node_id | 3
standby_name | pg149
node_type | standby
active | t
last_monitor_time | 2022-08-10 15:46:28.107282+08
last_wal_primary_location | 0/714CD08
last_wal_standby_location | 0/714CD08
replication_lag | 0 bytes
replication_time_lag | 00:00:00
apply_lag | 0 bytes
communication_time_lag | 00:00:01.822418
-[ RECORD 3 ]-------------+------------------------------
primary_node_id | 1
standby_node_id | 2
standby_name | pg49
node_type | primary
active | t
last_monitor_time | 2022-08-10 14:05:13.685229+08
last_wal_primary_location |
last_wal_standby_location | 0/7052030
replication_lag |
replication_time_lag |
apply_lag |
communication_time_lag | 00:00:01.822418
repmgr@[local:/tmp]:1998=#24783 select * from show_nodes; -- 各个节点的信息与表nodes信息基本一致
repmgr@[local:/tmp]:1998=#24783 select * from voting_term; -- 投票选举的次数
term
------
2
(1 row)
结束语的思考：
此时可能还会存在2个问题：
1）Standby 节点复制如果出现延时的情况下，auto fail over 就会发生数据丢失的情况。
这个时候我们可以设置优先级高的节点为 sync 模式， sync 模式保证了数据的强一致性， 但是牺牲了性能为代价。需要权衡你的业务数据的重要性，来看是否需要配置成 sync 模式 。
2） 在发生了 auto failover 之后， 应用是否能自动感知，是否支持 TAF (Transportation application failover) 应用自动故障转移？ 如何支持VIP的漂移？
下一篇，我们将讨论一下 PG 的应用自动故障转移方案。
Have a fun 🙂 ！
「喜欢这篇文章，您的关注和赞赏是给作者最好的鼓励」
关注作者
【版权声明】本文为墨天轮用户原创内容，转载时必须标注文章的来源（墨天轮），文章链接，文章作者等基本信息，否则作者和墨天轮有权追究责任。如果您发现墨天轮中有涉嫌抄袭或者侵权的内容，欢迎发送邮件至：contact@modb.pro进行举报，并提供相关证据，一经查实，墨天轮将立刻删除相关内容。
