共计 6287 个字符，预计需要花费 16 分钟才能阅读完成。
本篇内容主要讲解“怎么使用 pg_rewind”，感兴趣的朋友不妨来看看。本文介绍的方法操作简单快捷，实用性强。下面就让丸趣 TV 小编来带大家学习“怎么使用 pg_rewind”吧!
pg_rewind
是 postgresql 主丛数据库之同步数据目录的工具。需要目标服务器在 postgresql.conf 中允许 wal_log_hints，或者在 initdb 初始化集群时允许 checksums，full_page_writes 也必须为 on
pg_rewind 只复制表数据文件中更改的块; 所有其他文件都被完整复制，包括配置文件。pg_rewind 相对于使用 pg_basebackup 备份或 rsync 等工具的优势在于，pg_rewind 不需要读取数据库中未更改的块。这使得在数据库很大且之间只有一小部分块不同的情况下，速度会快得多。
pg_rewind [option…] {-D | –target-pgdata} directory {–source-pgdata=directory | –source-server=connstr
参数：
-D directory –target-pgdata=directory
此选项指定与源同步的目标数据目录。在运行 pg_rewind 之前，必须干净关闭目标服务器
–source-pgdata=directory
指定要与之同步的源服务器的数据目录的文件系统路径。此选项要求干净关闭源服务器
–source-server=connstr
指定要连接到源 PostgreSQL 服务器的 libpq 连接字符串。连接必须是具有超级用户访问权限的正常 (非复制) 连接。此选项要求源服务器正在运行，而不是处于恢复模式。
-n –dry-run
除了实际修改目标目录之外，执行所有操作。
-P –progress
使进展报告。
实验使用两台主机，都安装 postgresql-10.7，已配置流复制
主：192.168.56.5 m1
丛：192.168.56.25 m7
m1(主)：创建测试表和数据
postgres=# create table test (id int,e_name varchar(100),e_mail varchar(100),d_id int);
CREATE TABLE
postgres=# \d
List of relations
Schema | Name | Type | Owner
--------+------+-------+----------
public | test | table | postgres
(1 row)
postgres=# insert into test values(1, zbs , 123@126.com ,10);
INSERT 0 1
postgres=# select * from test;
id | e_name | e_mail | d_id
----+--------+-------------+------
1 | zbs | 123@126.com | 10
(1 row)m7 (丛)：查询数据复制成功
[postgres@z_leader ~]$ psql postgres
psql (10.7)
Type help for help.
postgres=# \d
List of relations
Schema | Name | Type | Owner
--------+------+-------+----------
public | test | table | postgres
(1 row)
postgres=# select * from test;
id | e_name | e_mail | d_id
----+--------+-------------+------
1 | zbs | 123@126.com | 10
(1 row)提升丛库为新主库
[postgres@z_leader data]$ pg_ctl promote -D /usr/local/pg/data
waiting for server to promote.... done
server promoted
[postgres@z_leader data]$ psql postgres
psql (10.7)
Type help for help.
postgres=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
(1 row)m1(原主库)插入一条记录，模拟原主库上的数据没有复制到原丛库上
postgres=# insert into test values(2, zbs1 , 124@126.com ,10);
INSERT 0 1
postgres=# select * from test;
id | e_name | e_mail | d_id
----+--------+-------------+------
1 | zbs | 123@126.com | 10
2 | zbs1 | 124@126.com | 10
(2 rows)m7：在原丛库上 (已提升为主库) 插入一条记录并查看结果
postgres=# insert into test values(3, zbs2 , 124@126.com ,10);
INSERT 0 1
postgres=# select * from test;
id | e_name | e_mail | d_id
----+--------+-------------+------
1 | zbs | 123@126.com | 10
3 | zbs2 | 124@126.com | 10
(2 rows)m1 将原主库变为新主库的丛库
[postgres@localhost ~]$ kill -INT `head -1 /usr/local/pg/data/postmaster.pid`– 配置流复制文件和参数
[postgres@localhost data]$ mv recovery.done recovery.conf
[postgres@localhost data]$ cat recovery.conf
standby_mode = on
restore_command = cp /usr/local/pg/arch/%f
primary_conninfo = host=192.168.56.25 port=5432 user=rep
recovery_target_timeline = latest
[postgres@localhost data]$– 启动数据库
[postgres@localhost ~]$ /usr/local/pg/bin/pg_ctl -D /usr/local/pg/data -l logfile start
waiting for server to start.... done
server started
[postgres@localhost data]$ psql postgres
psql (10.7)
Type help for help.
postgres=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
(1 row)
postgres=# select * from test;
id | e_name | e_mail | d_id
----+--------+-------------+------
1 | zbs | 123@126.com | 10
2 | zbs1 | 124@126.com | 10
(2 rows)– 在 m7 上插入的记录未能复制过来
— 日志信息
2019-03-02 09:15:17.415 CST [2492] LOG: consistent recovery state reached at 0/D000098
2019-03-02 09:15:17.415 CST [2492] LOG: invalid record length at 0/D000098: wanted 24, got 0
2019-03-02 09:15:17.415 CST [2490] LOG: database system is ready to accept read only connections
2019-03-02 09:15:17.429 CST [2500] LOG: fetching timeline history file for timeline 6 from primary server
2019-03-02 09:15:17.460 CST [2500] FATAL: could not start WAL streaming: ERROR: requested starting point 0/D000000 on timeline 5 is not in this
server s history
DETAIL: This server s history forked from timeline 5 at 0/C003168.
cp: missing destination file operand after `/usr/local/pg/arch/00000006.history
Try `cp --help for more information.
cp: missing destination file operand after `/usr/local/pg/arch/00000007.history
Try `cp --help for more information.
cp: missing destination file operand after `/usr/local/pg/arch/00000006.history
Try `cp --help for more information.
2019-03-02 09:15:17.469 CST [2492] LOG: new timeline 6 forked off current database system timeline 5 before current recovery point 0/D000098
cp: missing destination file operand after `/usr/local/pg/arch/00000005000000000000000D[postgres@localhost ~]$ kill -INT `head -1 /usr/local/pg/data/postmaster.pid`— 使得 pg_rewind 同步数据库时间线
[[postgres@localhost ~]$ pg_rewind --target-pgdata /usr/local/pg/data --source-server= host=192.168.56.25 port=5432 user=postgres dbname=postgres -P
connected to server
servers diverged at WAL location 0/C003168 on timeline 5
rewinding from last common checkpoint at 0/C003010 on timeline 5
reading source file list
reading target file list
reading WAL in target
need to copy 100 MB (total source directory size is 118 MB)
102599/102599 kB (100%) copied
creating backup label and updating control file
syncing target data directory
Done!–pg_rewind 后此文件需要重新配置
[postgres@localhost data]$ cat recovery.conf
standby_mode = on
restore_command = cp /usr/local/pg/arch/%f
primary_conninfo = host=192.168.56.25 port=5432 user=rep
recovery_target_timeline = latest[postgres@localhost ~]$ /usr/local/pg/bin/pg_ctl -D /usr/local/pg/data -l logfile start
waiting for server to start.... done
server started
[postgres@localhost ~]$ psql postgres
psql (10.7)
Type help for help.
postgres=# select * from test;
id | e_name | e_mail | d_id
----+--------+-------------+------
1 | zbs | 123@126.com | 10
3 | zbs2 | 124@126.com | 10
(2 rows)
postgres=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
(1 row)– 原主库没有复制到丛库的记录消失，在新主库上插入的记录已同步
m7(新主库)
[postgres@z_leader ~]$ psql postgres
psql (10.7)
Type help for help.
postgres=# insert into test values(4, zbs2 , 124@126.com ,10);
INSERT 0 1
postgres=# select * from test;
id | e_name | e_mail | d_id
----+--------+-------------+------
1 | zbs | 123@126.com | 10
3 | zbs2 | 124@126.com | 10
4 | zbs2 | 124@126.com | 10
(3 rows)m1(新丛库)
postgres=# select * from test;
id | e_name | e_mail | d_id
----+--------+-------------+------
1 | zbs | 123@126.com | 10
3 | zbs2 | 124@126.com | 10
4 | zbs2 | 124@126.com | 10
(3 rows)到此，相信大家对“怎么使用 pg_rewind”有了更深的了解，不妨来实际操作一番吧！这里是丸趣 TV 网站，更多相关内容可以进入相关频道进行查询，关注我们，继续学习！
