슬라이드 1

Size: px

Start display at page:

Download "슬라이드 1"

정윤 필
7 years ago
Views:

1 사례를통해본 RMAN (RMAN Case Study) Seungtaek Lee( 放浪 A)

RMAN 백업 Server-Managed Backup(RMAN) Components Target : 백업대상

Restore하여여러가지용도로사용하는 Channel : Target 에 Server Session을할당하여실제백업 /

백업시필요한 Library Catalog Auxiliary Target Server Server Server

2 RMAN 백업 Server-Managed Backup(RMAN) Components Target : 백업대상 Catalog : 백업정보가저장되는 RMAN Repository Auxiliary : 백업또는 Target 를 Restore하여여러가지용도로사용하는 Channel : Target 에 Server Session을할당하여실제백업 / 복구를수행하게하는역할대부분의 RMAN Command는 Channel을통하여수행함 MML : RMAN Tape 백업시필요한 Library Catalog Auxiliary Target Server Server Server Session Session Session Connect channel channel channel Connect Connect RMAN MML 2

3 RMAN 백업 Server-Managed Backup(RMAN) 의내부작업 Restored Data File Header Block 1 Block 2 Block 3 Block 4 Block 5 Backup Set Online Data File Header Block 1 Block 2 Block 3 Block 4 Block Header Block 1 Block 2 Block 3 Block 4 Block Header Block 1 Block 2 Block 3 Block 4 Block 5 Block 1 Block 2 Block 3 Block Block 1 Block 2 Block 3 Block 5 Header Block 4 Block Block 1 Block 2 Block 3 Block Block 1 Block Checkpoint Absolute Fuzzy 2 Restore Header Block 1 Block 2 Block 3 Block 4 Block Recovery RMAN 은 BEGIN/END Command 없이백업을수행하기때문에 Block Image 같은추가 Redo 가발생하지않음 RMAN 은 Full Backup 뿐만아니라 Block 레벨의 Incremental Backup, Compressed Backup 등다양한방식의백업을지원 RMAN 은백업시에 Data File 의 Header Block 부터백업을수행하기때문 Header Block 1 Block 2 Block 3 Block 4 Block 5 Block 1 Header 에 User-Managed Backup 방식과달리 Header Block 의 Checkpoint 을고정시키지않음. 복구시 Redo Log 적용을 Header Block 의 Header Block 1 Block 2 Block 3 Block 4 Block 5 1 Online Backup Large Pool Checkpoint 시점부터복구를시작 RMAN 백업은 End Backup 을수행하지않기때문에백업시에 Redo Log 최소적용시점정보인 Absolute Fuzzy 값을백업 Data File Header 에기록 Absolute Fuzzy 은백업 Data File 의 Block SC 값중가장큰 3

로설정하고 Catalog 로접속한다음 register database로 Catalog 에 정보를등록 Backup 수행시에는 Standby 를 Target 로설정하고 Catalog 에접속한후 Backup을수행

4 RMAN 백업 Online Standby Backup Online Standby Backup을수행하기위해서는 Backup 정보를저장하는 Catalog 가필요 RMAN은 Catalog 가구성안되어있을시에는Catalog 대신에 Target 의 Controlfile을사용 Controlfile을사용하면 Primary 와 Standby 의 Control file이다르기때문에 Backup 정보를서로공유할수없음 Target 를 Primary 로설정하고 Catalog 로접속한다음 register database로 Catalog 에 정보를등록 Backup 수행시에는 Standby 를 Target 로설정하고 Catalog 에접속한후 Backup을수행 Standby 와 Primary 의 ID가동일하기때문에 Primary 에서 Standby 의백업을사용할수있음 Primary Standby Instance Tape Drive Instance Primary Catalog Standby 4

5 RMAN 백업 Catalog 생성및 Database 등록 $ rman catalog rman/rman01@rman RMAN> create catalog; rman target / catalog rman/rman01@rman RMAN> register database; Catalog 에 Primary 만등록하면됨 DataGuard 환경에서 RMAN Configuration CONFIGURE _UNIQUE_NAME PRI CONNECT IDENTIFIER PRI_SYN'; CONFIGURE _UNIQUE_NAME STB CONNECT IDENTIFIER STB_SYN'; DataGuard 환경에서는 RMAN CONFIGURE 설정값은 CATALOG 에서 Local Control File 로 Reverse Resync 가발생 각 고유의 CONFIGURE 설정을위해 _UNIQUE_NAME 생성후 FOR _UNIQUE_NAME 구문으로각각설정할수있음 CONFIGURE CONTROLFILE AUTOBACKUP ON; 백업수행시자동으로 Control File 과 spfile 백업수행 (OFF 시에는 Datafile 1 백업시에만포함됨 ) CONFIGURE BACKUP OPTIMIZATION ON; 동일 Checkpoint 에같은백업존재시백업 Skip 함 CONFIGURE ARCHIVELOG DELETION POLICY TO APPLIED ALL STANY; Standby 에 Apply 된 Archive Log 만삭제 CONFIGURE SNAPSHOT CONTROLFILE NAME TO '+FS_DG/PRI/CONTROLFILE/snapcf_PRI.f ; RAC 일경우 Snapshot Controfile 이 Shared Device 에존재해야함 ( 이상 ) 로컬로설정할경우 ORA 에러가발생 (Bug ) Standby 와 Primary 의 Snapshot Controlfile 은경로와파일명을같게설정해야함 (Bug ) 5

6 RMAN 백업 RAC Multi Instance RMAN Backup Target 가 RAC일경우 Channel을 RAC의여러 Instance에할당하여자원을분산해서사용함으로백업수행속도를빠르게할수있음 Channel할당시 TAF나 Service를사용하면가용성이나자원사용을효율적으로할수있음 Catalog connect target /; connect rcvcat Target Instance 1 Connect Server Server channel Server Session channel Session channel Session Connect RMAN run{ allocate channel t1 type 'SBT_TAPE' connect 'sys/q1w2e3r4@mos15k_lb'; allocate channel t2 type 'SBT_TAPE' connect 'sys/q1w2e3r4@mos15k_lb'; allocate channel t3 type 'SBT_TAPE' connect 'sys/q1w2e3r4@mos15k_lb'; allocate channel t4 type 'SBT_TAPE' connect 'sys/q1w2e3r4@mos15k_lb'; send 'NSR_ENV=(NSR_SERVER=kmesbk, NSR_CLIENT=dmos15d1, NSR_DATA_VOLUME_POOL=MOS)'; backup as backupset incremental level 0 filesperset 4 tag 'STBY_TAPE_BKST_FULL' format '%d_%u' database; Target Instance 2 Server Server Server Session Session Session channel channel channel release channel t1; release channel t2; release channel t3; release channel t4; } 6

7 RMAN 백업 Block Change Tracking File 과거 Backup 이후변경된 block 정보를비트맵으로저장하여 Incremental Backup 수행시변경된 Block만 Read하여빠르게 Backup(Fast Incremental Backup) 할수있게지원하는 File 각 Datafile마다최대 8개버전 (_bct_bitmaps_per_file) 의 Backup에대한변경된 block 정보를 bitmap extent에 switch하면서저장 (bitmap extent 1개에는 1개의버전정보만저장됨 ) 변경된 block은 32k(_bct_chunk_size) 크기의 Chunk 단위로관리 RAC 환경하에서는 BCT File을 ASM/Raw Device/CFS 같은공유Device에생성하여야함 11g부터 Standby 에도적용가능 SGA List of changed blocks CTWR Redo generation Change tracking file Redo log 7

8 RMAN 백업 Incrementally Updated Backup 최초 Backup 시에는 Image Copy Full Backup 을수행하고이후 Backup 수행시에는 Incremental Backup 을 수행하여이때만들어지는 Incremental Backup 을 Image Copy Full Backup 본에 Update 하는 Backup 대용량 환경하에서 Incremental Backup 수행만으로 Full Backup 본을생성할수있으므로 Disk Full Backup 을좀더빠르게수행할수있는 Backup 방식 Tag 를설정하여여러시점의백업본을구성할수도있음 /recover1 Time A1 Time A2 Time A3 Full Backup (Image Copy) Full Backup (Image Copy) Full Backup (Image Copy) Database Update Incremental Backup (Backup Set) Update Incremental Backup (Backup Set) /recover2 Time B1 Time B2 Time B3 Full Backup (Image Copy) Full Backup (Image Copy) Full Backup (Image Copy) Update Incremental Backup (Backup Set) Update Incremental Backup (Backup Set) 8

9 RMAN 백업 Incrementally Updated Backup Script (/recover1) connect target / connect rcvcat rman/rman01@rman run { allocate channel f1d1 device type disk format '/recover1/rman/datafile/%u'; allocate channel f1d2 device type disk format '/recover1/rman/datafile/%u'; allocate channel f1d3 device type disk format '/recover1/rman/datafile/%u'; allocate channel f1d4 device type disk format '/recover1/rman/datafile/%u'; crosscheck backupset tag='stby_disk_copy_full1'; crosscheck copy like '/recover1/rman/%' ; delete noprompt expired backupset tag='stby_disk_copy_full1'; delete noprompt expired copy like '/recover1/rman/%' ; delete noprompt copy like '/recover1/rman/%' completed before 'sysdate-7'; delete noprompt backupset tag 'STBY_DISK_COPY_FULL1' completed before 'sysdate-2'; backup incremental level 1 for recover of copy with tag 'STBY_DISK_COPY_FULL1' database; recover copy of database with tag 'STBY_DISK_COPY_FULL1'; backup as copy current controlfile tag 'STBY_DISK_COPY_FULL1' format '/recover1/rman/controlfile/stbyctl1.cpy' reuse; backup as copy current controlfile tag 'STBY_DISK_COPY_FULL1' format '/recover2/arch/controlfile/stbyctl1.cpy' reuse; backup as backupset current controlfile tag 'STBY_DISK_COPY_FULL1' format '/recover1/rman/controlfile/stbyctl1.bkp.%u'; backup as backupset current controlfile tag 'STBY_DISK_COPY_FULL1' format '/recover2/arch/controlfile/stbyctl1.bkp.%u'; backup as backupset current controlfile tag 'STBY_DISK_COPY_FULL1' format '/recover1/rman/controlfile/stbyctl1.bkp' reuse; backup as backupset current controlfile tag 'STBY_DISK_COPY_FULL1' format '/recover2/arch/controlfile/stbyctl1.bkp' reuse; release channel f1d1; release channel f1d2; release channel f1d3; release channel f1d4; } 9

10 RMAN 백업 Incrementally Updated Backup Script (/recover2) connect target / connect rcvcat rman/rman01@rman run { allocate channel f2d1 device type disk format '/recover2/rman/datafile/%u'; allocate channel f2d2 device type disk format '/recover2/rman/datafile/%u'; allocate channel f2d3 device type disk format '/recover2/rman/datafile/%u'; allocate channel f2d4 device type disk format '/recover2/rman/datafile/%u'; crosscheck backupset tag='stby_disk_copy_full2'; crosscheck copy like '/recover2/rman/%' ; delete noprompt expired backupset tag='stby_disk_copy_full2'; delete noprompt expired copy like '/recover2/rman/%' ; delete noprompt copy like '/recover2/rman/%' completed before 'sysdate-7'; delete noprompt backupset tag 'STBY_DISK_COPY_FULL2' completed before 'sysdate-2'; backup incremental level 1 for recover of copy with tag 'STBY_DISK_COPY_FULL2' database; recover copy of database with tag 'STBY_DISK_COPY_FULL2'; backup as copy current controlfile tag 'STBY_DISK_COPY_FULL2' format '/recover2/rman/controlfile/stbyctl2.cpy' reuse; backup as copy current controlfile tag 'STBY_DISK_COPY_FULL2' format '/recover2/arch/controlfile/stbyctl2.cpy' reuse; backup as backupset current controlfile tag 'STBY_DISK_COPY_FULL2' format '/recover2/rman/controlfile/stbyctl2.bkp.%u'; backup as backupset current controlfile tag 'STBY_DISK_COPY_FULL2' format '/recover2/arch/controlfile/stbyctl2.bkp.%u'; backup as backupset current controlfile tag 'STBY_DISK_COPY_FULL2' format '/recover2/rman/controlfile/stbyctl2.bkp' reuse; backup as backupset current controlfile tag 'STBY_DISK_COPY_FULL2' format '/recover2/arch/controlfile/stbyctl2.bkp' reuse; release channel f2d1; release channel f2d2; release channel f2d3; release channel f2d4; } 10

11 Standby Image Copy 본을이용한 복구 Standby 의백업본을 Primary 서버에 NFS 로 Mount 함 Primary Control File 에는 Standby 에서수행한디스크백업정보가없기때문에 Catalog 를연결필요 Primary 에서 Standby 백업정보를접근하기위해 set backup files for device type disk to accessible 설정 Standby 에서수행한 Tape Backup 백업정보는 Catalog 만연결하면특별한설정없이 Primary 에서 Default 로볼수있음 Catalog 에있는 Standby 백업정보를이용하여 Full/Tablespace/Datafile/Block 복구를할수있음 Data Guard Primary Standby Instance Block Datafile Tablespace Full 3 RMAN Recovery Online Compress Parallel Image 1 RMAN Backup Instance Primary RMAN Image Standby 2 FS Mount/Cataloging 11

12 STEP 1. Primary 서버에 Standby Image Copy 본 Mount NFS 로 Mount 시 OS 별 Mount 옵션확인할것 (Note ID 참조 ) NFS 문제시 ORA-17500: ODM err:permission denied 에러발생함 $ df kmmepz01-priv1:/export/data % /recover2/rman/datafile/data1 kmmepz02-priv1:/export/data % /recover2/rman/datafile/data2 kmmepz01-priv1:/export/data % /recover2/rman/datafile/data3 kmmepz02-priv1:/export/data % /recover2/rman/datafile/data4 STEP 2. set backup files for device type disk to accessible 설정 10gR2 까지는 Standby 에서디스크백업수행시반드시 Nocatalog 로백업 / 복구를수행하여야했고 Primary 에서복구하기위해서는 Catalog 작업이별도로필요 11gR1 부터는 Standby 에서디스크백업수행시 Catalog 로백업을수행하고 Primary 에서복구시에는 set backup files for device type disk to accessible 구문으로별도의 Catalog 작업없이복구가능 # rman target / catalog rman/rman01@rman RMAN> set backup files for device type disk to accessible RMAN> list copy; 12

13 STEP 3. Restore and Recovery Database/Tablespace/Datafile/Block 단위로 Restore 및 Recovery 가능 $ rman target / catalog rman/rman01@rman run { allocate channel disk1 device type disk ; allocate channel disk2 device type disk ; restore database from tag 'STBY_DISK_COPY_FULL2' ; release channel disk1; release channel disk2; } Block Recover 예 $ rman target / catalog rman/rman01@rman RMAN> blockrecover datafile 6 block 9,10 datafile 8 block 11,15 from tag 'STBY_DISK_COPY_FULL2' ; 13

14 Standby Image Copy 본을이용한 Test 구성 Image Copy 본은 Datafile 과구조가같기때문에 Restore 작업없이 Copy 본으로경로를수정하여 를 Open 할수있음 Open 을위해 Controlfile, Datafile, Tempfile, Redo Log File, Archive Log File, BCT File, Flashback Log 에대한경로변경작업필요 Image Copy 본은 Primary 서버나 Standby 서버가아닌별도의복구 / 테스트서버에구성할것을권고 Primary 서버나 Standby 서버에서구성할경우 Service 등록과데이터베이스파일에대한중복여부 확인작업이필요하기때문에신중하게작업을하여야함 Data Guard Primary Standby 복구 /TEST Instance Instance Instance Primary Standby RMAN Image1 RMAN Image2 복구용 Test 용 14

15 Standby Image Copy 본을이용한 Test 구성 STEP 1. Test 정보확인 Archive log mode 필수 #. oraenv ORACLE_SID = [TEST]? TEST The Oracle base remains unchanged with value /u01/app/oracle # /oracle/ora11g>env grep ORA ORACLE_BASE=/u01/app/oracle ORACLE_SID=TEST ORACLE_HOME=/u01/app/oracle/product/11.2.0/dbhome_1 # sqlplus / as sysdba SQL> startup nomount STEP 2. Controlfile Restore Test 생성을위해서는별도의 Oracle Home 을설치하고 를 Nomount 함 Nomount 에서 Controlfile Restore 수행 Controlfile Restore 는 Copy 본 (stbyctl2.cpy) 이아닌 Standby 에서수행한 Backupset Controlfile(stbyctl2.bkp) 을사용할것 # export ORACLE_SID=TEST # rman target / RMAN> restore controlfile from '/recover2/rman1/controlfile/stbyctl2.bkp'; RMAN> sql 'alter database mount'; sql statement: alter database mount released channel: ORA_DISK_1 15

16 # export ORACLE_SID=TEST # sqlplus / as sysdba SQL> select CONTROLFILE_TYPE from v$database; CONTROL BACKUP STANY가아닌 BACKUP이어야함 SQL> select name from v$controlfile; STEP 3. Datafile Copy Backup 확인 # rman target / RMAN> list datafilecopy tag='stby_disk_copy_full2'; List of Datafile Copies ======================= Key File S Completion Time Ckp Ckp Time A 02-DEC :54: DEC :23:27 Name: /recover2/rman/datafile/data_d-kmmes_i _ts-system_fno-1_9jnrrbhm Tag: STBY_DISK_COPY_FULL A 02-DEC :55: DEC :23:27 Name: /recover2/rman/datafile/data_d-kmmes_i _ts-sysaux_fno-2_9inrrbhk Tag: STBY_DISK_COPY_FULL

17 STEP 4. Datafile Rename switch database to copy 명령어로 Datafile Rename 수행 # export ORACLE_SID=TEST # rman target / RMAN> switch database to copy; datafile 1 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-system_fno-1_9jnrrbhm" datafile 2 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-sysaux_fno-2_9inrrbhk" datafile 3 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-undotbs1_fno-3_98nrrb24" datafile 4 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-undotbs2_fno-4_99nrrb24" datafile 5 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-users_fno-5_9knrrbij" datafile 6 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-undotbs1_fno-6_9anrrb24" datafile 7 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-undotbs1_fno-7_9bnrrb24" datafile 8 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-undotbs1_fno-8_9cnrrbad" datafile 9 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-undotbs1_fno-9_9dnrrbag" datafile 10 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-undotbs2_fno-10_9enrrbai" datafile 11 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-undotbs2_fno-11_9fnrrbal" datafile 12 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-undotbs2_fno-12_9gnrrbhc" datafile 13 switched to datafile copy "/recover2/rman/datafile/data_d-kmmes_i _ts-undotbs2_fno-13_9hnrrbhi" 17

18 STEP 5. Tempfile Rename switch tempfile 명령어로 Tempfile Rename 수행 # rman target / run { SET NEWNAME FOR TEMPFILE 1 TO '/recover2/rman1/datafile/temp01.dbf'; SET NEWNAME FOR TEMPFILE 2 TO '/recover2/rman1/datafile/temp02.dbf'; SET NEWNAME FOR TEMPFILE 3 TO '/recover2/rman1/datafile/temp03.dbf'; SET NEWNAME FOR TEMPFILE 4 TO '/recover2/rman1/datafile/temp04.dbf'; SET NEWNAME FOR TEMPFILE 5 TO '/recover2/rman1/datafile/temp05.dbf'; switch tempfile all; } STEP 6. Redo Log File Rename alter database rename file 명령어로 Redo Log File Rename 수행 # export ORACLE_SID=TEST SQL> alter database rename file '/oracle/kmmes/redolog/1_redo01.log' to '/recover2/rman/datafile/1_redo01.log'; SQL> alter database rename file '/oracle/kmmes/mirrlog/1_redo01_m.log' to '/recover2/rman/datafile/1_redo01_m.lo g ; SQL> alter database rename file '/oracle/kmmes/redolog/stby_1_redo31.log' to '/recover2/rman/datafile/stby_1_red o31.log'; SQL> alter database rename file '/oracle/kmmes/redolog/stby_1_redo32.log' to '/recover2/rman/datafile/stby_1_red o32.log'; 18

19 STEP 7. Active Redo Log Drop drop logfile 에러발생 (Active Standby Log) 시해당 Log Clear 후 drop SQL> alter database clear logfile group 32; alter database clear logfile group 32 * ERROR at line 1: ORA-00344: unable to re-create online log '/oracle/kmmes/redolog/stby_1_redo32.log' ORA-27040: file create error, unable to create file IBM AIX RISC System/6000 Error: 2: No such file or directory Additional information: 1 SQL> alter database drop logfile group 32; STEP 8. BCT File Disable # export ORACLE_SID=TEST # rman target / SQL> ALTER DATABASE DISABLE BLOCK CHANGE TRACKING; SQL> select * from V$BLOCK_CHANGE_TRACKING; 19

20 STEP 9. Database File Rename 최종확인 # export ORACLE_SID=TEST # sqlplus / as sysdba SQL> select name from v$controlfile; SQL> select name from v$datafile; SQL> select file#,name from v$tempfile; SQL> select * from V$BLOCK_CHANGE_TRACKING; SQL> select member from v$logfile; STEP 10. Datafile 복구및 Open noarchivelog 로 Open RAC 백업을 Single 로 resetlog 로 Open 하기위해서 init 파일에 _no_recovery_through_resetlogs=true 설정할것 (Bug ) # export ORACLE_SID=TEST # sqlplus / as sysdba SQL> recover database until cancel using backup controlfile; ORA-00279: change generated at 12/01/ :23:27 needed for thread 1 ORA-00289: suggestion : /recover3/archive/1_74_ dbf ORA-00280: change for thread 1 is in sequence #74 Specify log: {<RET>=suggested filename AUTO CANCEL} CANCEL Media recovery cancelled. SQL> select * from v$recover_file; SQL> alter database noarchivelog; SQL> alter database open resetlogs; 20

21 ZFS Appliance 를이용한 Test 구성 ZFS Appliance 의 Snapshot 기능을이용하여 RMAN Image Copy 본을여러시점의 Snapshot 백업을수행함 필요한 Snapshot 을 Read Write 가가능한 Clone 으로복제하여테스트서버에 Mount 함 테스트 의 Controlfile 에변경된경로정보를다시 Catalog Datafile/Redo Log File/Temp File 등을 rename 함 BCT/Flashback Log 등을 Disable 함 Reset Log 로 Open 함 Data Guard Primary Standby 복구 /TEST Instance Instance Instance Primary Standby RMAN Image Snap 1 Snap 2 Snap 3 Snap 4 Snap 5 Clone 2 Clone 5-1 Clone 5-2 복구용 Test 용 21

22 STEP 1. ZFS Clone 백업본 Mount /clone/recover2/rman/datafile/ STEP 2. Controlfile Restore Test 생성을위해서는별도의 Oracle Home 을설치하고 를 Nomount 함 Nomount 에서 Controlfile Restore 수행 Controlfile Restore 는 Copy 본 (stbyctl2.cpy) 이아닌 Standby 에서수행한 Backupset Controlfile(stbyctl2.bkp) 을사용할것 # sqlplus / as sysdba SQL> startup nomount # rman target / RMAN> restore controlfile from '/clone/recover2/rman1/controlfile/stbyctl2.bkp'; RMAN> sql 'alter database mount'; # sqlplus / as sysdba SQL> select CONTROLFILE_TYPE from v$database; CONTROL BACKUP STANY가아닌 BACKUP이어야함 SQL> select name from v$controlfile; 22

23 STEP 3. Image Copy 본 Uncatalog/Catalog Backup 당시경로와달라졌기때문에 (/recover2 /clone/recover2) 백업정보삭제 (Uncatalog) 후등록 (Catalog ) # rman target / RMAN> list datafilecopy tag='stby_disk_copy_full2'; RMAN> change datafilecopy tag='stby_disk_copy_full2' uncatalog; RMAN> list datafilecopy tag='stby_disk_copy_full2'; RMAN> catalog start with '/clone/recover2/' noprompt; FILE# 로한개의 ImageCopy만존재하는지확인 RMAN> list datafilecopy tag='stby_disk_copy_full2'; RMAN> list incarnation; List of Database Incarnations Key Inc Key Name ID STATUS Reset Reset Time KMQC PARENT 1 10-DEC :56: KMQC PARENT DEC :17: KMQC CURRENT DEC :30:57 # 운영 의 RESETLOGS_CHANGE# 과 Current Incarnation 의 Reset 과 Mismatch 시해당 Incarnation# 로 Reset RMAN> reset database to incarnation 번호 ; 23

24 STEP 4. Datafile Rename switch database to copy 명령어로 Datafile Rename 수행 # rman target / RMAN> switch database to copy; datafile 1 switched to datafile copy "/clone/recover2/rman/datafile/data1/data_d-kmqc_i _ts-system_fno-1 _03nspu7f" datafile 2 switched to datafile copy "/clone/recover2/rman/datafile/data4/data_d-kmqc_i _ts-sysaux_fno-2 _12nsq0kc"... STEP 5. Tempfile Rename switch tempfile 명령어로 Tempfile Rename 수행 # rman target / run { SET NEWNAME FOR TEMPFILE 1 TO '/clone/recover2/rman1/datafile/temp01.dbf'; switch tempfile all; } STEP 6. Redo Log File Rename alter database rename file 명령어로 Redo Log File Rename 수행 alter database rename file '+DATA_EXA2/kmqcdg/onlinelog/group_ ' to '/clone/recover2/rman/datafile /1redo1.log'; alter database rename file '+DATA_EXA2/kmqcdg/onlinelog/group_ ' to '/clone/recover2/rman/datafile /1redo2.log ; 24

25 STEP 7. Standby Log File Drop alter database drop logfile group 31; alter database drop logfile group 32;. drop logfile 에러발생 (Active Standby Log) 시해당 Log Clear 후 drop TEST: SYS> alter database drop logfile group 32 * ERROR at line 1: TEST: SYS> alter database clear logfile group 32; '+data_exa * ORA-29701: unable to connect to Cluster Synchronization Service SQL> alter database drop logfile group 32; STEP 8. Block Change Tracking File Disable SQL> ALTER DATABASE DISABLE BLOCK CHANGE TRACKING; SQL> select * from V$BLOCK_CHANGE_TRACKING; 25

26 STEP 9. Database File Rename 최종확인 SQL> select member from v$logfile; SQL> select file#,name from v$tempfile; SQL> select * from V$BLOCK_CHANGE_TRACKING; SQL> select name from v$controlfile; SQL> select name from v$datafile; STEP 10. Recover 및 Open noarchivelog 로 Open RAC 백업을 Single 로 resetlog 로 Open 하기위해서 init 파일에 _no_recovery_through_resetlogs=true 설정할것 (Bug ) SQL> recover database until time ' :12:30:00' using backup controlfile; SQL> alter database noarchivelog; SQL> alter database open resetlogs; 26

27 Duplicate Database 를이용한 Test 구성 Primary 또는 Standby 에서수행한 VTL 백업을 Restore 하여별도의 Test 를구성할수있음 Duplicate 는원본 와다른 ID 로생성됨 Data Guard Primary Standby TEST Instance Instance Instance Primary Standby Duplicate 27

28 Duplicate Database 를이용한 Test 구성 STEP 1. Startup Nomount Duplicate 생성을위해서는별도의 Oracle Home 을설치하고 를 Nomount 함 kmmepr02[#/oracle/ora11g] sqlplus / as sysdba TEST2: SYS> startup nomount STEP 2. Control File Restore -- Catalog 을이용한 Rman 접속 kmmepr02[#/oracle/ora11g] rman target / catalog rman/rman01@rman run { allocate channel ch00 type 'sbt_tape'; restore controlfile to '/recover3/test2/datafile/test2.ctl' from 'c c'; release channel ch00; } --> '/c c' 해당정보는 RMAN의 'list backupset;' 명령을통해해당시점확인후 "Handle" 번호입력. Handle: bk_7372_1_ ==> Datafie Handle: c a ==> Control File & spfile Handle: al_7377_1_ ==> Archivelog 28

29 STEP 3. Duplicate 생성 Duplicate 생성을위해서는 Taget/Catalog/Auxiliary 에접속하고 Auxiliary 에 Duplicate 가생성됨 kmmepr02[#/oracle/ora11g] rman target catalog auxiliary / run { allocate auxiliary channel ch00 type 'SBT_TAPE' parms="env=(nb_ora_policy=kmmepe04_ora_full_kmqcdg2,nb_ora_clien T=kmmepe04_exa,NB_ORA_SERV=kmmepr01)"; allocate auxiliary channel ch01 type 'SBT_TAPE' parms="env=(nb_ora_policy=kmmepe04_ora_full_kmqcdg2,nb_ora_clien T=kmmepe04_exa,NB_ORA_SERV=kmmepr01)"; allocate auxiliary channel ch02 type 'SBT_TAPE' parms="env=(nb_ora_policy=kmmepe04_ora_full_kmqcdg2,nb_ora_clien T=kmmepe04_exa,NB_ORA_SERV=kmmepr01)"; allocate auxiliary channel ch03 type 'SBT_TAPE' parms="env=(nb_ora_policy=kmmepe04_ora_full_kmqcdg2,nb_ora_clien T=kmmepe04_exa,NB_ORA_SERV=kmmepr01)"; set until time "TO_DATE('2012/12/15 18:00:00','YYYY/MM/DD HH24:MI:SS')"; set newname for database to '/recover3/test2/datafile/%b'; duplicate target database to "TEST2" nofilenamecheck logfile '/recover3/test2/onlinelog/redo01.dbf' size 512M, '/recover3/test2/onlinelog/redo02.dbf' size 512M, '/recover3/test2/onlinelog/redo03.dbf' size 512M ; release channel ch00; release channel ch01; release channel ch02; release channel ch03; } 29

30 RMAN Issues STEP 4. Open -- Duplicate 생성후, TEST2 재기동 kmmepr02[#/oracle/ora11g] sqlplus / as sysdba TEST2: SYS> shutdown immediate; TEST2: SYS> startup; 30

슬라이드 1

슬라이드 1 RMAN Backup and Recovery Seungtaek Lee( 放浪 DBA) User-managed Backup vs Server-managed Backup 단계 Inconsistency 현상 User-managed Backup Server-managed Backup (RMAN) 시작. File Inconsistency : Backup 수행시 File