在資料檔案損壞,資料庫無法正常啟動的時候,Oracle可透過資料庫備份+歸檔日誌+線上日誌實現資料庫的完整恢復,與之類似,PostgreSQL也可以透過資料庫備份+歸檔WAL日誌+線上WAL日誌進行完整恢復.
#### 一、場景
1.執行備份
2.構造測試資料
3.刪除資料檔案(保留WAL日誌檔案)
4.使用備份+線上WAL日誌檔案進行完整恢復

#### 二、模擬完整恢復
**引數配置**
```
archive_mode = on 
archive_command = '/home/xdb/archive.sh'
wal_level = replica
max_wal_size = 4GB
min_wal_size = 1024MB
```
詳細請參考[Backup&Recovery#1(基本操作)](http://blog.itpub.net/6906/viewspace-2638060/)

**執行備份**
檢視當前的LSN
```
testdb=# select pg_current_wal_lsn();
 pg_current_wal_lsn 
--------------------
 0/39A63C78
(1 row)

```

使用工具pg_basebackup對資料庫進行備份
```
testdb=# \q
[xdb@localhost testdb]$ pg_basebackup -D /data/backup/0312-1/ -l 0312-1 -v -F tar -z
pg_basebackup: initiating base backup, waiting for checkpoint to complete
pg_basebackup: checkpoint completed
pg_basebackup: write-ahead log start point: 0/3A000108 on timeline 15
pg_basebackup: starting background WAL receiver
pg_basebackup: created temporary replication slot "pg_basebackup_2978"
pg_basebackup: write-ahead log end point: 0/3A0001D8
pg_basebackup: waiting for background process to finish streaming ...
pg_basebackup: base backup completed
[xdb@localhost testdb]$ 
[xdb@localhost testdb]$ ll /data/backup/0312-1/
total 44384
-rw-------. 1 xdb xdb 45427619 Mar 12 17:30 base.tar.gz
-rw-------. 1 xdb xdb    18927 Mar 12 17:30 pg_wal.tar.gz
[xdb@localhost testdb]$ 
####
[xdb@localhost ~]$ psql -d testdb
psql (11.2)
Type "help" for help.

testdb=# checkpoint;
CHECKPOINT
testdb=#

```

**構造測試資料**
建立資料表,插入資料
```
testdb=# create table tbl01(id int,c1 char(200),c2 char(200));
CREATE TABLE
testdb=# insert into tbl01 select f,f||'c1',f||'c2' from generate_series(1,100000) f;
INSERT 0 100000
testdb=# select pg_current_wal_lsn();
 pg_current_wal_lsn 
--------------------
 0/3DD39618
(1 row)
testdb=# create table tbl02(id int,c1 char(200),c2 char(200));
CREATE TABLE
testdb=# insert into tbl02 select f,f||'c1',f||'c2' from generate_series(1,100000) f;
INSERT 0 100000
testdb=# select pg_current_wal_lsn();
 pg_current_wal_lsn 
--------------------
 0/40A62F20
(1 row)

testdb=# 
```
當前線上日誌檔案為0000000F0000000000000040
```
[xdb@localhost testdb]$ ll $PGDATA/pg_wal
total 196632
-rw-------. 1 xdb xdb       42 Mar 12 17:10 00000008.history
-rw-------. 1 xdb xdb       85 Mar 12 17:10 0000000C.history
-rw-------. 1 xdb xdb 16777216 Mar 12 17:10 0000000E0000000000000039.partial
-rw-------. 1 xdb xdb      129 Mar 12 17:10 0000000E.history
-rw-------. 1 xdb xdb 16777216 Mar 12 17:30 0000000F000000000000003A
-rw-------. 1 xdb xdb      323 Mar 12 17:30 0000000F000000000000003A.00000108.backup
-rw-------. 1 xdb xdb 16777216 Mar 12 17:32 0000000F000000000000003B
-rw-------. 1 xdb xdb 16777216 Mar 12 17:32 0000000F000000000000003C
-rw-------. 1 xdb xdb 16777216 Mar 12 17:32 0000000F000000000000003D
-rw-------. 1 xdb xdb 16777216 Mar 12 17:32 0000000F000000000000003E
-rw-------. 1 xdb xdb 16777216 Mar 12 17:32 0000000F000000000000003F
-rw-------. 1 xdb xdb 16777216 Mar 12 17:32 0000000F0000000000000040
-rw-------. 1 xdb xdb 16777216 Mar 12 17:10 0000000F0000000000000041
-rw-------. 1 xdb xdb 16777216 Mar 12 17:10 0000000F0000000000000042
-rw-------. 1 xdb xdb 16777216 Mar 12 17:10 0000000F0000000000000043
-rw-------. 1 xdb xdb 16777216 Mar 12 17:30 0000000F0000000000000044
-rw-------. 1 xdb xdb      173 Mar 12 17:11 0000000F.history
drwx------. 2 xdb xdb     4096 Mar 12 17:32 archive_status
[xdb@localhost testdb]$ 
```
歸檔日誌檔案資訊
```
[xdb@localhost testdb]$ ll /data/archivelog/20190312/
total 245772
-rw-------. 1 xdb xdb 16777216 Mar 12 17:06 0000000E0000000000000032
-rw-------. 1 xdb xdb 16777216 Mar 12 17:06 0000000E0000000000000033
-rw-------. 1 xdb xdb      323 Mar 12 17:06 0000000E0000000000000033.00000028.backup
-rw-------. 1 xdb xdb 16777216 Mar 12 17:07 0000000E0000000000000034
-rw-------. 1 xdb xdb 16777216 Mar 12 17:07 0000000E0000000000000035
-rw-------. 1 xdb xdb 16777216 Mar 12 17:07 0000000E0000000000000036
-rw-------. 1 xdb xdb 16777216 Mar 12 17:07 0000000E0000000000000037
-rw-------. 1 xdb xdb 16777216 Mar 12 17:07 0000000E0000000000000038
-rw-------. 1 xdb xdb 16777216 Mar 12 17:11 0000000E0000000000000039.partial
-rw-------. 1 xdb xdb 16777216 Mar 12 17:30 0000000F0000000000000039
-rw-------. 1 xdb xdb 16777216 Mar 12 17:30 0000000F000000000000003A
-rw-------. 1 xdb xdb      323 Mar 12 17:30 0000000F000000000000003A.00000108.backup
-rw-------. 1 xdb xdb 16777216 Mar 12 17:32 0000000F000000000000003B
-rw-------. 1 xdb xdb 16777216 Mar 12 17:32 0000000F000000000000003C
-rw-------. 1 xdb xdb 16777216 Mar 12 17:32 0000000F000000000000003D
-rw-------. 1 xdb xdb 16777216 Mar 12 17:32 0000000F000000000000003E
-rw-------. 1 xdb xdb 16777216 Mar 12 17:32 0000000F000000000000003F
-rw-------. 1 xdb xdb      173 Mar 12 17:11 0000000F.history
[xdb@localhost testdb]$ 
```

**模擬資料檔案損壞**
刪除資料檔案目錄,kill postgres程式
```
[xdb@localhost testdb]$ rm -rf ./base
[xdb@localhost testdb]$ psql -d testdb
psql: FATAL:  database "testdb" does not exist
DETAIL:  The database subdirectory "base/16384" is missing.
[xdb@localhost testdb]$ ps -ef|grep postgres
xdb       2914     1  0 17:11 pts/2    00:00:00 /appdb/xdb/pg11.2/bin/postgres
xdb       2915  2914  0 17:11 ?        00:00:00 postgres: logger   
xdb       2918  2914  0 17:11 ?        00:00:00 postgres: checkpointer   
xdb       2919  2914  0 17:11 ?        00:00:00 postgres: background writer   
xdb       2921  2914  0 17:11 ?        00:00:00 postgres: stats collector   
xdb       2925  2914  0 17:11 ?        00:00:00 postgres: walwriter   
xdb       2926  2914  0 17:11 ?        00:00:00 postgres: autovacuum launcher  
xdb       2927  2914  0 17:11 ?        00:00:00 postgres: archiver   last was 0000000F000000000000003F
xdb       2928  2914  0 17:11 ?        00:00:00 postgres: logical replication launcher  
xdb       2977  2914  0 17:30 ?        00:00:00 postgres: xdb testdb [local] idle
xdb       3014  2519  0 17:33 pts/2    00:00:00 grep --color=auto postgres
[xdb@localhost testdb]$ kill -9 2914
[xdb@localhost testdb]$ ps -ef|grep postgres
xdb       3016  2519  0 17:34 pts/2    00:00:00 grep --color=auto postgres
```

**執行恢復**
備份線上日誌
```
[xdb@localhost ~]$ mkdir /data/backup/wal
[xdb@localhost testdb]$ cp -R ./pg_wal/* /data/backup/wal/
[xdb@localhost testdb]$ 
[xdb@localhost testdb]$  ll /data/backup/wal/
total 196632
-rw-------. 1 xdb xdb       42 Mar 12 17:34 00000008.history
-rw-------. 1 xdb xdb       85 Mar 12 17:34 0000000C.history
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000E0000000000000039.partial
-rw-------. 1 xdb xdb      129 Mar 12 17:34 0000000E.history
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000F000000000000003A
-rw-------. 1 xdb xdb      323 Mar 12 17:34 0000000F000000000000003A.00000108.backup
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000F000000000000003B
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000F000000000000003C
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000F000000000000003D
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000F000000000000003E
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000F000000000000003F
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000F0000000000000040
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000F0000000000000041
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000F0000000000000042
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000F0000000000000043
-rw-------. 1 xdb xdb 16777216 Mar 12 17:34 0000000F0000000000000044
-rw-------. 1 xdb xdb      173 Mar 12 17:34 0000000F.history
drwx------. 2 xdb xdb     4096 Mar 12 17:34 archive_status
[xdb@localhost testdb]$ 
```
從資料庫備份中恢復
```
[xdb@localhost testdb]$ rm -rf *
[xdb@localhost testdb]$ cp /data/backup/0312-1/base.tar.gz ./
[xdb@localhost testdb]$ tar zxf base.tar.gz

```
恢復線上日誌
```
cp -R /data/backup/wal/0000000F0000000000000040 ./pg_wal
[xdb@localhost testdb]$ ll ./pg_wal
total 16384
-rw-------. 1 xdb xdb 16777216 Mar 12 17:35 0000000F0000000000000040
drwx------. 2 xdb xdb        6 Mar 12 17:30 archive_status
```
建立recovery.conf檔案
```
[xdb@localhost testdb]$ vim recovery.conf
[xdb@localhost testdb]$ cat recovery.conf 
#Recovery
restore_command='cp /data/archivelog/20190312/%f %p'
#restore_target=XX

```
執行恢復,啟動資料庫
```
[xdb@localhost testdb]$ pg_ctl start
waiting for server to start....2019-03-12 17:36:21.310 CST [3030] LOG:  listening on IPv4 address "0.0.0.0", port 5432
2019-03-12 17:36:21.310 CST [3030] LOG:  listening on IPv6 address "::", port 5432
2019-03-12 17:36:21.328 CST [3030] LOG:  listening on Unix socket "/tmp/.s.PGSQL.5432"
2019-03-12 17:36:21.341 CST [3030] LOG:  redirecting log output to logging collector process
2019-03-12 17:36:21.341 CST [3030] HINT:  Future log output will appear in directory "pg_log".
 done
server started
```
日誌輸出
```
2019-03-12 17:36:21.341 CST,,,3030,,5c877d95.bd6,1,,2019-03-12 17:36:21 CST,,0,LOG,00000,"ending log output to stderr",,"Future log output will go to log destination ""csvlog"".",,,,,,,""
2019-03-12 17:36:21.350 CST,,,3032,,5c877d95.bd8,1,,2019-03-12 17:36:21 CST,,0,LOG,00000,"database system was interrupted; last known up at 2019-03-12 17:30:44 CST",,,,,,,,,""
2019-03-12 17:36:21.358 CST,,,3032,,5c877d95.bd8,2,,2019-03-12 17:36:21 CST,,0,LOG,00000,"starting archive recovery",,,,,,,,,""
2019-03-12 17:36:21.372 CST,,,3032,,5c877d95.bd8,3,,2019-03-12 17:36:21 CST,,0,LOG,00000,"restored log file ""0000000F.history"" from archive",,,,,,,,,""
2019-03-12 17:36:21.486 CST,,,3032,,5c877d95.bd8,4,,2019-03-12 17:36:21 CST,,0,LOG,00000,"restored log file ""0000000F000000000000003A"" from archive",,,,,,,,,""
2019-03-12 17:36:21.693 CST,,,3032,,5c877d95.bd8,5,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"redo starts at 0/3A000108",,,,,,,,,""
2019-03-12 17:36:21.696 CST,,,3032,,5c877d95.bd8,6,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"consistent recovery state reached at 0/3A0001D8",,,,,,,,,""
2019-03-12 17:36:21.696 CST,,,3030,,5c877d95.bd6,2,,2019-03-12 17:36:21 CST,,0,LOG,00000,"database system is ready to accept read only connections",,,,,,,,,""
2019-03-12 17:36:21.826 CST,,,3032,,5c877d95.bd8,7,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"restored log file ""0000000F000000000000003B"" from archive",,,,,,,,,""
2019-03-12 17:36:22.245 CST,,,3032,,5c877d95.bd8,8,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"restored log file ""0000000F000000000000003C"" from archive",,,,,,,,,""
2019-03-12 17:36:22.614 CST,,,3032,,5c877d95.bd8,9,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"restored log file ""0000000F000000000000003D"" from archive",,,,,,,,,""
2019-03-12 17:36:23.039 CST,,,3032,,5c877d95.bd8,10,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"restored log file ""0000000F000000000000003E"" from archive",,,,,,,,,""
2019-03-12 17:36:23.342 CST,,,3032,,5c877d95.bd8,11,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"restored log file ""0000000F000000000000003F"" from archive",,,,,,,,,""
2019-03-12 17:36:23.874 CST,,,3032,,5c877d95.bd8,12,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"invalid record length at 0/40A63B08: wanted 24, got 0",,,,,,,,,""
2019-03-12 17:36:23.874 CST,,,3032,,5c877d95.bd8,13,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"redo done at 0/40A63AD0",,,,,,,,,""
2019-03-12 17:36:23.874 CST,,,3032,,5c877d95.bd8,14,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"last completed transaction was at log time 2019-03-12 17:33:32.760492+08",,,,,,,,,""
2019-03-12 17:36:23.879 CST,,,3032,,5c877d95.bd8,15,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"selected new timeline ID: 16",,,,,,,,,""
2019-03-12 17:36:24.773 CST,,,3032,,5c877d95.bd8,16,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"archive recovery complete",,,,,,,,,""
2019-03-12 17:36:24.777 CST,,,3032,,5c877d95.bd8,17,,2019-03-12 17:36:21 CST,1/0,0,LOG,00000,"restored log file ""0000000F.history"" from archive",,,,,,,,,""
2019-03-12 17:36:25.589 CST,,,3030,,5c877d95.bd6,3,,2019-03-12 17:36:21 CST,,0,LOG,00000,"database system is ready to accept connections",,,,,,,,,""
```
驗證資料
```
[xdb@localhost testdb]$ psql -d testdb
psql (11.2)
Type "help" for help.

testdb=# 
testdb=# select count(*) from tbl01;
 count  
--------
 100000
(1 row)

testdb=# select count(*) from t02;
 count  
--------
 100000
(1 row)

testdb=# 
```
時間線歷史檔案,在歸檔目錄和pg_wal目錄下均存在以當前時間線命名的history,該檔案說明了該Cluster的歷史.
```
[xdb@localhost ~]$ cat /data/archivelog/20190312/00000010.history
7	0/27000000	no recovery target specified

8	0/2A0012E8	no recovery target specified

12	0/32000000	no recovery target specified

14	0/39A63BD0	no recovery target specified

15	0/40A63B08	no recovery target specified
[xdb@localhost ~]$ cat $PGDATA/pg_wal/00000010.history 
7	0/27000000	no recovery target specified

8	0/2A0012E8	no recovery target specified

12	0/32000000	no recovery target specified

14	0/39A63BD0	no recovery target specified

15	0/40A63B08	no recovery target specified
[xdb@localhost ~]$ 
```
DONE!

#### 三、參考資料
[Base Backup & Point-in-Time Recovery](http://www.interdb.jp/pg/pgsql10.html)
[Write Ahead Log](https://www.postgresql.org/docs/current/runtime-config-wal.html)
[pgbasebackup](https://www.postgresql.org/docs/current/app-pgbasebackup.html)
[PostgreSQL DBA(28) - Backup&Recovery#1(基本操作)](http://blog.itpub.net/6906/viewspace-2638060/)

PostgreSQL DBA(30) - Backup&Recovery#3(資料檔案損壞恢復)

相關文章