您好,登錄后才能下訂單哦!
Oracle 12C 新特性 -RMAN 恢復表
oracle 如何恢復單張表數據?
場景一:處理簡單,恢復速度快
1 如果表被 delete 、 update 、 insert ,并已經 commit ,在時間和 undo 允許的情況下,可以通過閃回將數據迅速找回;
alter table t1 enable row movement;
--- 基于 Scn 閃回
flashback table t1 to scn 1726953;
--- 基于時間閃回
flashback table t1 to timestamp to_timestamp('2020-02-08 12:00:00','yyyy-mm-dd hh34:mi:ss');
或者通過CTAS 方式將數據保存到另一張表里
create table t1_recv as select * from t1 as of timestamp (systimestamp - interval '2' minute);
2 如果表被 drop ,并且沒有使用 purge 清空回收站,可以通過回收站找回被刪除的表 (user_recyclebin);
--- 將回收站的數據通過 CTAS 方式保存到另一張表里
show recyclebin
select object_name,ORIGINAL_NAME from user_recyclebin;
select original_name,operation,droptime from recyclebin;
create table t1 as select * from "BIN$HKCe4+VkDVXgU8UDqMAWvQ==$0";
--- 或者直接閃回表
flashback table t1 to before drop;
場景二:通過備份進行恢復
處理復雜,恢復速度慢的場景
1 如果表被 delete 、 update 、 insert ,并已經 commit ,時間已過去很久, undo 信息已被覆蓋:
或者 表被truncate 或 drop 并清空了回收站。
如果有expdp 或 exp 備份,可以用備份恢復,會丟失數據。
如果有rman 備份, 12C 之前,可以將備份基于時間點不完全恢復到另一個數據庫里,然后通過 expdp 導出丟失數據的表,在導入到生產庫,缺點是操作繁瑣,時間較長。
從12C 開始,通過 rman 恢復單張表,操作命令變的更簡單,但恢復時間還是會比較長。
場景三:處理難度大,不一定能完全恢復的場景
表被誤操作(delete 、 update 、 truncate 、 drop) ,并且沒有有效的備份,回收站已被清空, undo 數據已被覆蓋:
可以考慮bbed,ODU 等工具進行數據恢復。
本次實驗模擬pdb 中一張表誤被 delete ,并通過 rman 進行單表恢復。
實驗過程如下:
一 查看數據庫基本信息
1 數據庫版本為19C( 相當于 12.2.0.3 版本 )
SQL> select banner_full from v$version;
BANNER_FULL
---------------------------------------------------------------------
Oracle Database 19c Enterprise Edition Release 19.0.0.0.0 - Production
Version 19.3.0.0.0
二 檢查并 啟動歸檔模式
SQL> archive log list;
Database log mode No Archive Mode
Automatic archival Disabled
Archive destination /u01/app/oracle/product/19.0.0/dbhome_1/dbs/arch
Oldest online log sequence 12
Current log sequence 14
[root@cjcos ~]# mkdir /arch
[root@cjcos ~]# chown oracle.oinstall /arch
SQL> alter system set log_archive_dest_1='location=/arch' scope=both;
SQL> alter system set log_archive_format = "cjcpdb_arch_%t_%s_%r.arc" scope=spfile;
SQL> shutdown immediate
SQL> startup mount
SQL> alter database archivelog;
SQL> alter database open;
SQL> archive log list;
Database log mode Archive Mode
Automatic archival Enabled
Archive destination /arch
Oldest online log sequence 12
Next log sequence to archive 14
Current log sequence 14
SQL> alter system switch logfile;
[oracle@cjcos ~]$ cd /arch/
[oracle@cjcos arch]$ ll -rth
total 74M
-rw-r----- 1 oracle oinstall 74M Feb 8 11:04 cjcpdb_arch_1_14_1030641846.arc
三 創建測試數據
SQL> conn sys/oracle@cjcpdb as sysdba
Connected.
SQL> select name from v$dbfile;
NAME
---------------------------------------------------------------------
/u01/app/oracle/oradata/CJCDB/cjcpdb/system01.dbf
/u01/app/oracle/oradata/CJCDB/cjcpdb/sysaux01.dbf
/u01/app/oracle/oradata/CJCDB/cjcpdb/undotbs01.dbf
/u01/app/oracle/oradata/CJCDB/cjcpdb/users01.dbf
SQL> create tablespace cjctbs datafile '/u01/app/oracle/oradata/CJCDB/cjcpdb/cjctbs01.dbf' size 10M autoextend on;
SQL> create user cjc identified by cjc default tablespace cjctbs;
SQL> grant connect,resource,dba to cjc;
SQL> conn cjc/cjc@cjcpdb
SQL> create table t1 as select * from dba_objects;
SQL> insert into t1 select* from t1;
72397 rows created.
SQL> /
144794 rows created.
SQL> commit;
SQL> update t1 set object_id=rownum;
289588 rows updated.
SQL> commit;
Commit complete.
四 :全庫備份
[root@cjcos ~]# mkdir /backup/rman -p
[root@cjcos ~]# chown oracle.oinstall /backup -R
[oracle@cjcos ~]$ rman target /
Recovery Manager: Release 19.0.0.0.0 - Production on Sat Feb 8 11:21:00 2020
Version 19.3.0.0.0
Copyright (c) 1982, 2019, Oracle and/or its affiliates. All rights reserved.
connected to target database: CJCDB (DBID=3744777523)
RMAN>
run
{
allocate channel ch2 type disk;
allocate channel ch3 type disk;
backup database tag 'full' format '/backup/rman/%d_full_%T_%U.bak';
sql 'alter system archive log current';
backup archivelog all tag 'arch' format '/backup/rman/%d_arch_%T_%U.bak';
backup current controlfile tag 'ctl' format '/backup/rman/%d_ctl_%T_%U.bak';
release channel ch2;
release channel ch3;
}
查看備份信息
RMAN> list backup of database;
[oracle@cjcos rman]$ pwd
/backup/rman
[oracle@cjcos rman]$ ll -rth
total 2.4G
-rw-r----- 1 oracle oinstall 415M Feb 8 11:48 CJCDB_full_20200208_13uo0vma_1_1.bak
-rw-r----- 1 oracle oinstall 789M Feb 8 11:48 CJCDB_full_20200208_12uo0vma_1_1.bak
-rw-r----- 1 oracle oinstall 272M Feb 8 11:49 CJCDB_full_20200208_15uo0vp0_1_1.bak
-rw-r----- 1 oracle oinstall 302M Feb 8 11:49 CJCDB_full_20200208_14uo0vp0_1_1.bak
-rw-r----- 1 oracle oinstall 305M Feb 8 11:50 CJCDB_full_20200208_16uo0vqd_1_1.bak
-rw-r----- 1 oracle oinstall 252M Feb 8 11:50 CJCDB_full_20200208_17uo0vqe_1_1.bak
-rw-r----- 1 oracle oinstall 6.3M Feb 8 11:50 CJCDB_arch_20200208_1auo0vrt_1_1.bak
-rw-r----- 1 oracle oinstall 74M Feb 8 11:50 CJCDB_arch_20200208_19uo0vrt_1_1.bak
-rw-r----- 1 oracle oinstall 125K Feb 8 11:50 CJCDB_arch_20200208_1buo0vs1_1_1.bak
-rw-r----- 1 oracle oinstall 18M Feb 8 11:50 CJCDB_ctl_20200208_1cuo0vs2_1_1.bak
五 :查看當前SCN 等信息
SQL> conn cjc/cjc@cjcpdb
Connected.
SQL> show con_name user
CON_NAME
------------------------------
CJCPDB
USER is "CJC"
SQL> select count(*) from t1;
COUNT(*)
----------
289588
SQL> select to_char(sysdate,'yyyy-mm-dd hh34:mi:ss') from dual;
TO_CHAR(SYSDATE,'YY
-------------------
2020-02-08 11:53:51
SQL> select current_scn from v$database;
CURRENT_SCN
-----------
2301721
六 :模擬誤刪除操作
SQL> delete t1 where object_id>10;
289578 rows deleted.
SQL> commit;
Commit complete.
SQL> select count(*) from t1;
COUNT(*)
----------
10
--- 針對 delete 級別的誤刪除,在時間和 undo 允許的情況下,可以通過閃回將數據迅速找回
SQL> create table t2 as select * from t1 as of timestamp (systimestamp - interval '3' minute);
Table created.
SQL> select count(*) from t2;
COUNT(*)
----------
289588
本案例使用rman 進行 t1 表恢復,恢復到刪除操作之前的時刻。
七 :使用rman 備份恢復 t1 表
1 創建輔助數據庫恢復目錄 /auxiliary/recover ,創建 expdp 導出目錄 /auxiliary/dumpfiles
[root@cjcos ~]# mkdir /auxiliary/{recover,dumpfiles} -p
[root@cjcos ~]# chown oracle.oinstall /auxiliary -R
[root@cjcos ~]# ll -rth /auxiliary/
total 0
drwxr-xr-x 2 oracle oinstall 6 Feb 8 12:14 recover
drwxr-xr-x 2 oracle oinstall 6 Feb 8 12:14 dumpfiles
2 rman執行恢復操作
[oracle@cjcos ~]$ rman target /
Recovery Manager: Release 19.0.0.0.0 - Production on Sat Feb 8 12:15:53 2020
Version 19.3.0.0.0
Copyright (c) 1982, 2019, Oracle and/or its affiliates. All rights reserved.
connected to target database: CJCDB (DBID=3744777523)
RMAN>
run{
RECOVER TABLE CJC.T1 OF PLUGGABLE DATABASE CJCPDB
UNTIL SCN 2301721
AUXILIARY DESTINATION '/auxiliary/recover'
DATAPUMP DESTINATION '/auxiliary/dumpfiles'
DUMP FILE 't1_scn_2301721.dmp'
REMAP TABLE 'CJC'.'T1':'T1_RECVR';
}
1 恢復 CJCPDB 數據庫下 CJC 用戶下的 T1 表。
2 恢復到 SCN 2301721 時刻。
3 恢復過程中創建的輔助實例恢復位置 /auxiliary/recover 。
4 恢復過程中通過 expdp 導出的備份位置 /auxiliary/dumpfiles ,備份名為 t1_scn_2301721.dmp 。
5 并將恢復數據插入到 T1_RECVR 表中,驗證數據沒問題后再手動將 T1_RECVR 數據插回到 T1 中。
下面是截取恢復過程中部分信息,均是自動完成,無需人為干預。
......
---1 創建輔助實例 DiFo 并恢復數據
Creating automatic instance, with SID='DiFo'
initialization parameters used for automatic instance:
db_name=CJCDB
db_unique_name=DiFo_pitr_CJCPDB_CJCDB
compatible=19.0.0
db_block_size=8192
db_files=200
diagnostic_dest=/u01/app/oracle
_system_trig_enabled=FALSE
sga_target=1104M
processes=200
db_create_file_dest=/auxiliary/recover
log_archive_dest_1='location=/auxiliary/recover'
enable_pluggable_database=true
_clone_one_pdb_recovery=true
#No auxiliary parameter file used
starting up automatic instance CJCDB
......
---2 基于 scn 恢復
contents of Memory Script:
{
# set requested point in time
set until scn 2301721;
# online the datafiles restored or switched
sql clone 'CJCPDB' "alter database datafile 13 online";
# recover and open resetlogs
recover clone database tablespace "CJCPDB":"CJCTBS", "SYSTEM", "CJCPDB":"SYSTEM", "UNDOTBS1", "CJCPDB":"UNDOTBS1", "SYSAUX", "CJCPDB":"SYSAUX" delete archivelog;
alter clone database open resetlogs;
}
executing Memory Script
......
--- 3 expdp 導出 T1 表數據,到 /auxiliary/dumpfiles/t1_scn_2301721.dmp
contents of Memory Script:
{
# create directory for datapump import
sql 'CJCPDB' "create or replace directory
TSPITR_DIROBJ_DPDIR as ''
/auxiliary/dumpfiles''";
# create directory for datapump export
sql clone 'CJCPDB' "create or replace directory
TSPITR_DIROBJ_DPDIR as ''
/auxiliary/dumpfiles''";
}
executing Memory Script
.....
Performing export of tables...
EXPDP> Starting "SYS"."TSPITR_EXP_DiFo_aBzr":
EXPDP> Processing object type TABLE_EXPORT/TABLE/TABLE_DATA
EXPDP> Processing object type TABLE_EXPORT/TABLE/STATISTICS/TABLE_STATISTICS
EXPDP> Processing object type TABLE_EXPORT/TABLE/STATISTICS/MARKER
EXPDP> Processing object type TABLE_EXPORT/TABLE/TABLE
EXPDP> . . exported "CJC"."T1" 38.18 MB 289588 rows
EXPDP> Master table "SYS"."TSPITR_EXP_DiFo_aBzr" successfully loaded/unloaded
EXPDP> ******************************************************************************
EXPDP> Dump file set for SYS.TSPITR_EXP_DiFo_aBzr is:
EXPDP> /auxiliary/dumpfiles/t1_scn_2301721.dmp
EXPDP> Job "SYS"."TSPITR_EXP_DiFo_aBzr" successfully completed at Sat Feb 8 12:32:57 2020 elapsed 0 00:01:45
Export completed
......
--- 4 impdp 將 T1 數據導入到正式庫 T1_RECVR 表中
Performing import of tables...
IMPDP> Master table "SYS"."TSPITR_IMP_DiFo_BABf" successfully loaded/unloaded
IMPDP> Starting "SYS"."TSPITR_IMP_DiFo_BABf":
IMPDP> Processing object type TABLE_EXPORT/TABLE/TABLE
IMPDP> Processing object type TABLE_EXPORT/TABLE/TABLE_DATA
IMPDP> . . imported "CJC"."T1_RECVR" 38.18 MB 289588 rows
IMPDP> Processing object type TABLE_EXPORT/TABLE/STATISTICS/TABLE_STATISTICS
IMPDP> Processing object type TABLE_EXPORT/TABLE/STATISTICS/MARKER
IMPDP> Job "SYS"."TSPITR_IMP_DiFo_BABf" successfully completed at Sat Feb 8 12:34:17 2020 elapsed 0 00:00:49
Import completed
八 :查看數據已經恢復到t1_recvr
---對比數據沒問題后可以 選擇手動 插回到t1 中
SQL> conn cjc/cjc@cjcpdb
SQL> select count(*) from t1;
COUNT(*)
----------
10
SQL> select count(*) from t1_recvr;
COUNT(*)
----------
289588
詳細信息可參考官方文檔:
https://docs.oracle.com/en/database/oracle/oracle-database/12.2/bradv/rman-recovering-tables-partitions.html#GUID-04D1B794-25D0-4C1E-9A98-F08AD361B41E
......
歡迎關注我的微信公眾號"IT小Chen",共同學習,共同成長!!!
免責聲明:本站發布的內容(圖片、視頻和文字)以原創、轉載和分享為主,文章觀點不代表本網站立場,如果涉及侵權請聯系站長郵箱:is@yisu.com進行舉報,并提供相關證據,一經查實,將立刻刪除涉嫌侵權內容。