Oracle 10gR2 Dataguard搭建(非duplicate方式)
我的實驗環境：
源生產庫(主庫)：
IP地址：192.168.1.30
Oracle 10.2.0.5 單實例
新DG庫（備庫）：
IP地址：192.168.1.31
Oracle 10.2.0.5 單實例
- 1.源生產庫開啟歸檔
- 2.rman備份源生產庫
- 3.修改源生產庫參數
- 4.配置tnsnames.ora
- 5.同步密碼文件
- 6.配置pfile文件
- 7.創建備庫控制文件
- 8.還原備庫
- 9.開啟日志應用
- 10.switchover測試
- 11.創建還原點，激活備庫測試
- 12.failover測試
1.源生產庫開啟歸檔
部署Dataguard環境，要求主庫必須開啟歸檔模式；如果沒有開啟，需要先申請停機開啟歸檔。shutdown immediate
startup mount
alter database archivelog;
alter database open;
archive log list;
注意：歸檔日志存放位置，並制定歸檔日志刪除策略；
歸檔日志刪除舉例(刪除7天前歸檔)：
vi delarch.sql
crosscheck archivelog all;
delete noprompt archivelog all completed before "sysdate-7";
vi delarch.sh
#!/bin/bash
#Oracle ENV (you may need to change it.)
export ORACLE_BASE=/u01/app/oracle
export ORACLE_HOME=/u01/app/oracle/product/10.2.0/db_1
export ORACLE_SID=ora10
export PATH=$ORACLE_HOME/bin:$PATH
rman target / @/home/oracle/rman/delarch.sql log=/home/oracle/rman/delarch.log
如果之前主庫開啟了歸檔，且有合理的備份策略；則實際上很可能不需要此步驟單獨清除歸檔。
2.rman備份源生產庫
將主庫rman全備，視具體情況選擇是否壓縮備份集，我這里空間有限選擇壓縮備份集。備份腳本：
[oracle@oradb30 rman]$ cat backup.sql
run {
allocate channel d1 type disk;
allocate channel d2 type disk;
backup as compressed backupset database format '/orabak/rman/data_%d_%T_%s.bak' plus archivelog format '/orabak/rman/log_%d_%T_%s.bak';
release channel d1;
release channel d2;
}
[oracle@oradb30 rman]$ cat backup.sh
#!/bin/bash
export ORACLE_BASE=/u01/app/oracle
export ORACLE_HOME=/u01/app/oracle/product/10.2.0/db_1
export ORACLE_SID=ora10
export PATH=$ORACLE_HOME/bin:$PATH
rman target / @backup.sql log backup.log
執行備份：
[oracle@oradb30 rman]$ nohup sh backup.sh &
[oracle@oradb30 rman]$ tail -200f backup.log
Recovery Manager: Release 10.2.0.5.0 - Production on Wed Jan 4 21:47:23 2017
Copyright (c) 1982, 2007, Oracle. All rights reserved.
connected to target database: ORA10 (DBID=914008358)
RMAN> run {
2> allocate channel d1 type disk;
3> allocate channel d2 type disk;
4> backup as compressed backupset database format '/orabak/rman/data_%d_%T_%s.bak' plus archivelog format '/orabak/rman/log_%d_%T_%s.bak';
5> release channel d1;
6> release channel d2;
7> }
8>
using target database control file instead of recovery catalog
allocated channel: d1
channel d1: sid=143 devtype=DISK
allocated channel: d2
channel d2: sid=142 devtype=DISK
Starting backup at 04-JAN-17
current log archived
channel d1: starting compressed archive log backupset
channel d1: specifying archive log(s) in backup set
input archive log thread=1 sequence=46 recid=45 stamp=932415090
input archive log thread=1 sequence=47 recid=46 stamp=932420845
channel d1: starting piece 1 at 04-JAN-17
channel d2: starting compressed archive log backupset
channel d2: specifying archive log(s) in backup set
input archive log thread=1 sequence=43 recid=42 stamp=932414314
input archive log thread=1 sequence=44 recid=43 stamp=932414362
input archive log thread=1 sequence=45 recid=44 stamp=932415036
channel d2: starting piece 1 at 04-JAN-17
channel d1: finished piece 1 at 04-JAN-17
piece handle=/orabak/rman/log_ORA10_20170104_102.bak tag=TAG20170104T214725 comment=NONE
channel d1: backup set complete, elapsed time: 00:00:02
channel d2: finished piece 1 at 04-JAN-17
piece handle=/orabak/rman/log_ORA10_20170104_103.bak tag=TAG20170104T214725 comment=NONE
channel d2: backup set complete, elapsed time: 00:00:02
Finished backup at 04-JAN-17
Starting backup at 04-JAN-17
channel d1: starting compressed full datafile backupset
channel d1: specifying datafile(s) in backupset
input datafile fno=00001 name=/oradata/ora10/ORA10/datafile/o1_mf_system_d5tw48bw_.dbf
input datafile fno=00002 name=/oradata/ora10/ORA10/datafile/o1_mf_undotbs1_d5tw48d8_.dbf
input datafile fno=00004 name=/oradata/ora10/ORA10/datafile/o1_mf_users_d5tw48dg_.dbf
channel d1: starting piece 1 at 04-JAN-17
channel d2: starting compressed full datafile backupset
channel d2: specifying datafile(s) in backupset
input datafile fno=00003 name=/oradata/ora10/ORA10/datafile/o1_mf_sysaux_d5tw48c3_.dbf
input datafile fno=00007 name=/oradata/ora10/ORA10/datafile/o1_mf_forhapoc_d5x5bm2b_.dbf
input datafile fno=00005 name=/oradata/ora10/ORA10/datafile/o1_mf_dbs_d_ji_d5x45hbj_.dbf
input datafile fno=00006 name=/oradata/ora10/ORA10/datafile/o1_mf_dbs_i_ji_d5x45jd9_.dbf
channel d2: starting piece 1 at 04-JAN-17
channel d2: finished piece 1 at 04-JAN-17
piece handle=/orabak/rman/data_ORA10_20170104_105.bak tag=TAG20170104T214728 comment=NONE
channel d2: backup set complete, elapsed time: 00:00:35
channel d1: finished piece 1 at 04-JAN-17
piece handle=/orabak/rman/data_ORA10_20170104_104.bak tag=TAG20170104T214728 comment=NONE
channel d1: backup set complete, elapsed time: 00:00:42
Finished backup at 04-JAN-17
Starting backup at 04-JAN-17
current log archived
channel d1: starting compressed archive log backupset
channel d1: specifying archive log(s) in backup set
input archive log thread=1 sequence=48 recid=47 stamp=932420890
channel d1: starting piece 1 at 04-JAN-17
channel d1: finished piece 1 at 04-JAN-17
piece handle=/orabak/rman/log_ORA10_20170104_106.bak tag=TAG20170104T214810 comment=NONE
channel d1: backup set complete, elapsed time: 00:00:02
Finished backup at 04-JAN-17
Starting Control File and SPFILE Autobackup at 04-JAN-17
piece handle=/orabak/rman/20170104/controlfilec-914008358-20170104-06 comment=NONE
Finished Control File and SPFILE Autobackup at 04-JAN-17
released channel: d1
released channel: d2
Recovery Manager complete.
[1]+ Done nohup sh backup.sh
[oracle@oradb30 rman]$
備份完成后,將備份集拷貝到備機。
3.修改源生產庫參數
查看主庫文件(數據文件、臨時文件、重做日志文件)存放目錄：select name from v$datafile union all
select name from v$tempfile union all
select member from v$logfile;
修改源生產庫參數：
--設置convert參數（我這里源端和目標端都設置了db_create_file_dest，沒有設置這兩個參數）
alter system set log_file_name_convert='','' scope=spfile;
alter system set db_file_name_convert='','' scope=spfile;
--設置數據庫為force logging
alter database force logging;
--設置db_unique_name, log_archive_config(這里主庫的db_unique_name，出於對現有的生產環境最小影響考慮沒有改)
alter system set db_unique_name='ora10' scope=spfile;
alter system set log_archive_config='DG_CONFIG=(ora10,ora10dg)';
--歸檔日志目錄
alter system set log_archive_dest_1='LOCATION=/orabak/arch VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=ora10';
alter system set log_archive_dest_2='SERVICE=ora10dg ASYNC LGWR VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=ora10dg';
--歸檔日志文件命名規則
alter system set log_archive_format='arch_%r_%t_%s.arc' scope=spfile;
--DG的fal_server和fal_client
alter system set fal_server='ora10dg';
alter system set fal_client='ora10';
--設置standby_file_management為自動
alter system set standby_file_management=AUTO;
--設置備庫日志文件組，數量一般為目標庫日志文件組+1
alter database add standby logfile group 11 size 52428800;
alter database add standby logfile group 12 size 52428800;
alter database add standby logfile group 13 size 52428800;
alter database add standby logfile group 14 size 52428800;
4.配置tnsnames.ora
Dataguard環境，log_archive_config的配置，就是用到tnsnames.ora配置文件中的別名。--配置tnsnames.ora
在主庫所有節點上的tnsnames.ora添加相應的連接串，並傳給備庫。
---主庫tnsnames.ora添加
vi $ORACLE_HOME/network/admin/tnsnames.ora
ORA10 =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.1.30)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = ora10)
)
)
ORA10DG =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.1.31)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = ora10dg)
)
)
將tnsnames.ora上傳到備庫，如果之前tnsnames有其他內容，可以選擇添加內容，只要最終保證主備庫可以互相訪問即可。
5.同步密碼文件
--密碼文件從主庫拷貝到備庫
[oracle@oradb30 dbs]$ scp orapwora10 192.168.1.31:/u01/app/oracle/product/10.2.0/db_1/dbs/
oracle@192.168.1.31's password:
orapwora10 100% 1536 1.5KB/s 00:00
[oracle@oradb30 dbs]$
6.配置pfile文件
Dataguard環境下，需要將主庫的參數文件傳到備庫，進行修改，尤其注意路徑在備庫都存在且有相應權限。--創建pfile文件
從主庫的spfile中導出pfile文件，上傳到備庫，並做適當的修改調整。
--在主庫上創建pfile文件，並上傳到備庫
create pfile='/tmp/init.ora' from spfile;
SQL> create pfile='/tmp/init.ora' from spfile;
File created.
SQL> exit
Disconnected from Oracle Database 10g Enterprise Edition Release 10.2.0.5.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
[oracle@oradb30 dbs]$ scp /tmp/init.ora 192.168.1.31:/tmp/
reverse mapping checking getaddrinfo for bogon failed - POSSIBLE BREAK-IN ATTEMPT!
oracle@192.168.1.31's password:
init.ora 100% 1430 1.4KB/s 00:00
[oracle@oradb30 dbs]$
--備庫最終pfile參數文件
[oracle@oradb31 admin]$ vi /tmp/init.ora
*.audit_file_dest='/u01/app/oracle/admin/ora10/adump'
*.background_dump_dest='/u01/app/oracle/admin/ora10/bdump'
*.compatible='10.2.0.5.0'
*.core_dump_dest='/u01/app/oracle/admin/ora10/cdump'
*.db_block_size=8192
*.db_create_file_dest='/oradata'
*.db_domain=''
*.db_file_multiblock_read_count=16
*.db_name='ora10'
*.db_recovery_file_dest='/orabak/flash_recovery_area'
*.db_recovery_file_dest_size=2147483648
*.db_unique_name='ora10dg'
*.dispatchers='(PROTOCOL=TCP) (SERVICE=ora10XDB)'
*.fal_client='ora10dg'
*.fal_server='ora10'
*.job_queue_processes=10
*.log_archive_config='DG_CONFIG=(ora10dg,ora10)'
*.log_archive_dest_1='LOCATION=/orabak/arch VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=ora10dg'
*.log_archive_dest_2='SERVICE=ora10 ASYNC LGWR VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=ora10'
*.log_archive_format='arch_%r_%t_%s.arc'
*.open_cursors=300
*.pga_aggregate_target=96468992
*.processes=150
*.remote_login_passwordfile='EXCLUSIVE'
*.sga_target=290455552
*.standby_file_management='AUTO'
*.undo_management='AUTO'
*.undo_tablespace='UNDOTBS1'
*.user_dump_dest='/u01/app/oracle/admin/ora10/udump'
啟動備庫到nomount狀態：
SQL> startup nomount pfile='/tmp/init.ora';
ORA-01261: Parameter db_recovery_file_dest destination string cannot be translated
ORA-01262: Stat failed on a file destination directory
Linux-x86_64 Error: 2: No such file or directory
SQL> exit
Disconnected
[oracle@oradb31 admin]$ mkdir -p /orabak/flash_recovery_area
[oracle@oradb31 admin]$ sqlplus / as sysdba
SQL*Plus: Release 10.2.0.5.0 - Production on Wed Jan 4 23:24:25 2017
Copyright (c) 1982, 2010, Oracle. All Rights Reserved.
Connected to an idle instance.
SQL> startup nomount pfile='/tmp/init.ora';
ORACLE instance started.
Total System Global Area 293601280 bytes
Fixed Size 2095768 bytes
Variable Size 146802024 bytes
Database Buffers 138412032 bytes
Redo Buffers 6291456 bytes
SQL> exit
7.創建備庫控制文件
Dataguard環境下，需要在主庫創建備庫的控制文件並傳輸到備庫。--創建控制文件
主庫執行，創建備庫控制文件
alter database create standby controlfile as '/tmp/control01.ctlbak';
[oracle@oradb30 dbs]$ scp /tmp/control01.ctlbak 192.168.1.31:/tmp/
reverse mapping checking getaddrinfo for bogon failed - POSSIBLE BREAK-IN ATTEMPT!
oracle@192.168.1.31's password:
control01.ctlbak 100% 6928KB 6.8MB/s 00:00
8.還原備庫
8.1 在備庫上啟動數據庫到nomount狀態
#創建SPFILE
SQL>create SPFILE from pfile='/tmp/init.ora';
#啟動到nomount狀態
startup nomount
8.2 啟動數據庫到mount狀態
restore standby controlfile from '/tmp/control01.ctlbak'; alter database mount;可能需要手工注冊備份集；
crosscheck backupset;
catalog start with '/orabak/rman/';
8.3 還原備份
vi /home/oracle/scripts/restore.shrman target / <<EOF! > ora10_restore.log
run {
allocate channel d1 type disk;
allocate channel d2 type disk;
restore database;
release channel d1;
release channel d2;
}
exit;
EOF!
nohup sh restore.sh &
vi /home/oracle/scripts/recover.sh
rman target / <<EOF! > ora10_recover.log
run {
allocate channel d1 type disk;
allocate channel d2 type disk;
recover database;
release channel d1;
release channel d2;
}
exit;
EOF!
nohup sh recover.sh &
9.開啟日志應用
備庫開啟日志應用SQL> alter database recover managed standby database using current logfile disconnect from session;
10. switchover測試
switchover,主備角色互換。 注意，主機的crontab 定時任務確認也正確遷移。--源生產庫：
SELECT SWITCHOVER_STATUS FROM V$DATABASE;
alter database commit to switchover to physical standby with session shutdown;
--備庫（switchover為新生產庫）：
SELECT SWITCHOVER_STATUS FROM V$DATABASE;
ALTER DATABASE COMMIT TO SWITCHOVER TO PRIMARY WITH SESSION SHUTDOWN;
ALTER DATABASE OPEN;
--返回源生產庫執行（switchover為新備庫）：
shutdown immediate
startup mount
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE USING CURRENT LOGFILE DISCONNECT;
如果發現備庫沒有實時恢復，重置鏈路的狀態再次切換日志試試：
SQL> alter system set log_archive_dest_state_2=defer;
SQL> alter system set log_archive_dest_state_2=enable;
SQL> alter system switch logfile;
11. 創建還原點，激活備庫測試
11.1 創建restore point
**1) 在備庫上設置flashback區域及大小**SQL> alter system set db_recovery_file_dest='/orabak/flash_recovery_area';
SQL> alter system set db_recovery_file_dest_size=1000g;
SQL> select current_scn||'' from v$database;
2) Standby庫取消redo日志應用並創建一個數據庫閃回點
SQL> ALTER DATABASE RECOVER MANAGED STANDBY DATABASE CANCEL;
--這里注意下，如果你的備庫環境是RAC，需要關閉其他節點，然后才可以創建restore point：
SQL> drop restore point before_application_test;
SQL> CREATE RESTORE POINT before_application_test GUARANTEE FLASHBACK DATABASE;
Restore point created.
11.2 激活備庫
**1) 停止主庫的日志傳到備庫**SQL> ALTER SYSTEM SET LOG_ARCHIVE_DEST_STATE_2=defer;
2) 激活備庫
SQL> alter system set job_queue_processes = 0 scope=both sid='*';
SQL> ALTER DATABASE ACTIVATE STANDBY DATABASE;
SQL> shutdown immediate
SQL> startup
3) 檢查dblinks
select * from dba_db_links;
如果有dblink需要刪除或其他方式禁用；
4）通知新的應用IP，開始測試應用工作
注意：為了預防測試過程中有job或者dblink修改了其他數據庫信息，一般要修改job_queue_processes參數，刪除dblink。
--參數
alter system set job_queue_processes = 0 scope=both sid='*';
--dblinks
select * from dba_db_links;
11.3 閃回恢復備庫
**1) 恢復備庫**SQL> shutdown immediate
SQL> startup mount;
SQL> FLASHBACK DATABASE TO RESTORE POINT before_application_test;
SQL> ALTER DATABASE CONVERT TO PHYSICAL STANDBY;
SQL> shutdown immediate
SQL> startup mount
SQL> alter database recover managed standby database using current logfile disconnect;
恢復參數job_queue_processes設置值：
alter system set job_queue_processes = 10 scope=both sid='*';
恢復參數db_recovery_file_dest_size設置值：
alter system set db_recovery_file_dest_size=2g;
2) 恢復日志傳輸
恢復主庫的日志傳到備庫
ALTER SYSTEM SET LOG_ARCHIVE_DEST_STATE_2=enable;
12. failover測試
failover，源生產庫不再可用，備庫強制激活為主庫；#取消DG應用
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE CANCEL;
#重啟下數據庫（建議）
shutdown immediate;
startup
#操作不可逆，確定實際情況需要failover
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE FINISH force;
SELECT OPEN_MODE, DATABASE_ROLE, SWITCHOVER_STATUS, FORCE_LOGGING, DATAGUARD_BROKER, GUARD_STATUS FROM V$DATABASE;
#嘗試常規切換為主庫
ALTER DATABASE COMMIT TO SWITCHOVER TO PRIMARY WITH SESSION SHUTDOWN;
如果這一步的常規切換失敗，提示需要介質恢復，那么：
1）恢復備庫 recover standby database until cancel;
2）激活備庫 alter database activate standby database;
#最后重新啟動數據庫
shutdown immediate;
startup
failover之后，即便原主庫恢復也要重新搭建DG了。