pg_basebackup

从port 5435恢复到 port 5436的一次实验

———————————————————————————#

目录

1、root用户建立好新的备份数据实例

2、切回postgres用户,修改参数文件,重启

3、写pg_basebackup命令

4、检查

5、创建表,插入2000行数据,commit提交,记录xact_id

6、插入1000行数据,commit提交,记录xact_id

7、去新实例数据路径:pgdata_basebackup

8、创建recovery.signal

9、启动

报错,--3、pg_basebackup -p 5435 -U postgres -D /pgccc/pgdata_basebackup/ -Pv 

---

1、root用户建立好新的备份数据实例

root:
mkdir pgdata_basebackup
chown postgres:root pgdata_basebackup/
chmod +R 0750 pgdata_basebackup/
———————————————————————————#

2、切回postgres用户,修改参数文件,重启

vim postgres.conf
    wal_level='replica'
    archive_mode='on'
    archive_command='cp %p /xx/xx/%f'
pg_ctl restart
———————————————————————————#

3、写pg_basebackup命令

pg_basebackup -p 5435 -U postgres -D /pgccc/pgdata_basebackup/ -Pv 

执行结果为:
        pg_basebackup -p 5435 -U postgres -D /pgccc/pgdata_basebackup/ -Pv 
        pg_basebackup: initiating base backup, waiting for checkpoint to complete
        pg_basebackup: checkpoint completed
        pg_basebackup: write-ahead log start point: 0/24000028 on timeline 1
        pg_basebackup: starting background WAL receiver
        pg_basebackup: created temporary replication slot "pg_basebackup_3162"
        63149/63149 kB (100%), 1/1 tablespace                                         
        pg_basebackup: write-ahead log end point: 0/24000100
        pg_basebackup: waiting for background process to finish streaming ...
        pg_basebackup: syncing data to disk ...
        pg_basebackup: renaming backup_manifest.tmp to backup_manifest
        pg_basebackup: base backup completed
        
———————————————————————————#

 

4、检查

~~
pg_verifybackup  /pgccc/pgdata_basebackup/  
    backup successfully verified
~~
cat /pgccc/pgdata_basebackup/backup_label 
    START WAL LOCATION: 0/24000028 (file 000000010000000000000024)
    CHECKPOINT LOCATION: 0/24000060
    BACKUP METHOD: streamed
    BACKUP FROM: primary
    START TIME: 2023-06-02 04:29:52 CST
    LABEL: pg_basebackup base backup
    START TIMELINE: 1
~~
more /pgccc/pgdata_basebackup/backup_manifest
———————————————————————————#

5、创建表,插入2000行数据,commit提交,记录xact_id

postgres=# create table zcjbakcs(id int,name text);
CREATE TABLE

postgres=# insert into zcjbakcs select n,n||'_zcj' from generate_series(1,2000) n;
INSERT 0 2000
postgres=# commit;
WARNING:  there is no transaction in progress
COMMIT
postgres=# select pg_switch_wal();
 pg_switch_wal 
---------------
 0/250234F0
(1 row)

postgres=# select pg_current_xact_id();
 pg_current_xact_id 
--------------------
               1045
(1 row)

 

———————————————————————————#

6、插入1000行数据,commit提交,记录xact_id

postgres=# insert into zcjbakcs select n,n||'2_zcj' from generate_series(1,1000) n; 
INSERT 0 1000
postgres=# commit;
WARNING:  there is no transaction in progress
COMMIT
postgres=# select pg_switch_wal();                                                 
 pg_switch_wal 
---------------
 0/26018CF0
(1 row)

postgres=# select pg_current_xact_id();                                            
 pg_current_xact_id 
--------------------
               1048
(1 row)

 

———————————————————————————#

7、去新实例数据路径:pgdata_basebackup

修改postgresql.conf
    port改为5436
    restore_command='cp /pgccc/pgdata_5432/pg_archive/%f %p '
    recovery_target_xid='1000'
    recovery_target_action = 'promote' 

 
———————————————————————————#

8、创建recovery.signal

这新实例数据路径下创建recovery.signal

———————————————————————————#

9、启动

./pg_ctl -D /pgccc/pgdata_basebackup/ start

查行数

———————————————————————————#

 

报错,--3、
pg_basebackup -p 5435 -U postgres -D /pgccc/pgdata_basebackup/ -Pv 

[postgres@pgcs pgdata_5432]$ pg_basebackup -p 5435 -U postgres -D /pgccc/pgdata_basebackup/ -Pv                   
pg_basebackup: initiating base backup, waiting for checkpoint to complete
pg_basebackup: checkpoint completed
pg_basebackup: write-ahead log start point: 0/7000028 on timeline 1
pg_basebackup: error: directory "/pgccc/pgdata_5432/zcjtablespace" exists but is not empty
pg_basebackup: removing contents of data directory "/pgccc/pgdata_basebackup/"

--这说明你的pg有个表空间位置不对