rac如何部署dataguard到单机文件系统

环境:
节点1(ASM):
pbulic ip: 192.168.50.199
node1-vip: 192.168.50.251
sid:racdb1
节点2(ASM) 
public ip: 192.168.50.200
node2-vip: 192.168.50.252
sid:racdb2
备份节点(文件系统):
ip: 192.168.50.202
sid:racbak

---------------------------------------------rac主机部署-----------------------
1.设置主库为force logging模式(在其中一个节点执行即可)
SQL> alter database force logging;

2.配置两个节点上的tnsnames.ora
添加如下内容:
myrac1 =
  (DESCRIPTION =
    (ADDRESS_LIST =
      (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.50.251)(PORT = 1521))
    )
    (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = racdb)
      )
  )

myrac2 =
  (DESCRIPTION =
    (ADDRESS_LIST =
      (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.50.252)(PORT = 1521))
    )
    (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = racdb)
      )
  )

racbak =
  (DESCRIPTION =
    (ADDRESS_LIST =
      (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.50.202)(PORT = 1521))
    )
    (CONNECT_DATA =
      (SERVICE_NAME = racbak)
    )
  )

myrac =
  (DESCRIPTION =
    (ADDRESS_LIST =
      (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.50.251)(PORT = 1521))
      (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.50.252)(PORT = 1521))     
    )
    (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = racdb)
      (FAILOVER_MODE =
        (TYPE = SELECT)
        (METHOD = BASIC)
        (RETRIES = 180)
        (DELAY = 5)
      )
    )
  )

3.修改数据库参数
alter system set db_unique_name=racdb scope=spfile sid='*';
alter system set log_archive_config='dg_config=(racdb,racbak)' scope=both sid='*';
我的环境归档路径1和归档路径2都已经使用了,所以这里配置归档路径3
alter system set log_archive_dest_3= 'service=racbak async  valid_for=(online_logfiles,primary_role) db_unique_name=racbak' scope=both sid='*';
alter system set log_archive_dest_state_3=enable scope=both sid='*';
alter system set standby_file_management=auto scope=both sid='*';
alter system set fal_server=racbak scope=both sid='*';
alter system set fal_client=myrac1 scope=both sid='racdb1';
alter system set fal_client=myrac2 scope=both sid='racdb2';
alter system set db_file_name_convert='/u01/app/oracle/oradata/racbak/','+DATA01/racdb/datafile/' scope=spfile sid='*';
alter system set log_file_name_convert='/u01/app/oracle/oradata/racbak/','+DATA01/racdb/onlinelog/' scope=spfile sid='*';

4.备份主库
在其中一个节点上运行,注意该节点需要能访问每个rac节点的归档日志
run{
allocate channel dup type disk;
backup format '/bak/rman_bak/df_t%t_s%s_p%p' database;
sql 'alter system archive log current';
backup format '/bak/rman_bak/archive_log_t%t_s%s_p%p' archivelog all;
backup spfile format '/bak/rman_bak/spfile_%u_%T.bak';
release channel dup;
}

6.生成standby控制文件
rman>copy current controlfile for standby to '/bak/rman_bak/standby.ctl';

7.将以上生成的备份集,密码文件传到备库

scp archive_log_t837368189_s53_p1     oracle@192.168.50.202:/bak
scp archive_log_t837368197_s54_p1     oracle@192.168.50.202:/bak
scp archive_log_t837368199_s55_p1     oracle@192.168.50.202:/bak
scp archive_log_t837368200_s56_p1     oracle@192.168.50.202:/bak
scp archive_log_t837368216_s57_p1     oracle@192.168.50.202:/bak
scp archive_log_t837368217_s58_p1     oracle@192.168.50.202:/bak
scp df_t837367891_s51_p1              oracle@192.168.50.202:/bak
scp df_t837368124_s52_p1              oracle@192.168.50.202:/bak
scp spfile_1rouiedf_20140120.bak      oracle@192.168.50.202:/bak
scp standby.ctl                       oracle@192.168.50.202:/bak
scp orapwracdb1                       oracle@192.168.50.202:/bak
scp tnsnames.ora                      oracle@192.168.50.202:/bak

--------------------------------------------备库上的操作------------------------------
1.从备份集恢复spfile
RMAN>set dbid 834465720  --在主库上获取数据的dbid:Select dbid From v$database
RMAN>startup nomount
RMAN>restore spfile to pfile '/u01/export/home/oracle/pfile.ora' from '/bak/spfile_1rouiedf_20140120.bak';

2.修改pfile,最后的内容如下
racbak.__java_pool_size=4194304
racbak.__large_pool_size=4194304
racbak.__oracle_base='/u01/app/oracle'#ORACLE_BASE set from environment
racbak.__pga_aggregate_target=209715200
racbak.__sga_target=213909504
racbak.__shared_io_pool_size=0
racbak.__shared_pool_size=138412032
racbak.__streams_pool_size=0

*.audit_file_dest='/u01/product/oracle/admin/racbak/adump'
*.audit_trail='db'
*.compatible='11.2.0.0.0'
*.control_files='/u01/app/oracle/oradata/racbak/control01.ctl'
*.db_block_size=8192
*.db_domain=''
*.db_name='racdb'
*.db_unique_name='racbak'
*.diagnostic_dest='/u01/product/oracle'
*.dispatchers='(PROTOCOL=TCP) (SERVICE=racbakXDB)'
*.log_archive_format='%t_%s_%r.dbf'
*.open_cursors=300
*.processes=150
*.remote_login_passwordfile='exclusive'
*.undo_tablespace='UNDOTBS1'

*.log_archive_config='dg_config=(racdb,racbak)'
*.log_archive_dest_1=
 'location=/u01/archive_log/
  valid_for=(all_logfiles,all_roles)
  db_unique_name=racbak'
*.log_archive_dest_2=
 'service=myrac1 async
  valid_for=(online_logfiles,primary_role)
  db_unique_name=racdb'
*.log_archive_dest_state_1=enable
*.log_archive_dest_state_2=enable
*.standby_file_management=auto
*.fal_server=myrac1,myrac2
*.fal_client=racbak
*.thread=1
*.db_file_name_convert='+DATA01/racdb/datafile/','/u01/app/oracle/oradata/racbak/'
*.log_file_name_convert='+DATA01/racdb/onlinelog/','/u01/app/oracle/oradata/racbak/','+FRA/racdb/onlinelog/','/u01/app/oracle/oradata/racbak/'

创建参数文件中定义的目录(仔细检查每个目录,没有创建的话需要创建)
mkdir -p /u01/product/oracle/admin/racbak/adump
mkdir -p /u01/app/oracle/oradata/racbak

3.将主库传过来的tnsnames文件,密码文件和standby控制文件传到指定目录
cp tnsnames.ora /u01/app/oracle/product/11.2.0/db_1/network/admin/
cp standby.ctl /u01/app/oracle/oradata/racbak/
cd cd /u01/app/oracle/oradata/racbak/
mv standby.ctl control01.ctl
cp orapwracdb1 /u01/app/oracle/product/11.2.0/db_1/dbs/
cd /u01/app/oracle/product/11.2.0/db_1/dbs/
mv orapwracdb1 orapwracbak

4.启动数据库到nomount状态
SQL> connect / as sysdba
Connected to an idle instance.
SQL> startup nomount pfile=/u01/export/home/oracle/pfile.ora
ORACLE instance started.

Total System Global Area  497995776 bytes
Fixed Size                  1337464 bytes
Variable Size             260048776 bytes
Database Buffers          230686720 bytes
Redo Buffers                5922816 bytes

5.启动到mount状态
SQL> alter database mount;

Database altered.

6.指定备份集路径,bak目录下是从主库传过来的备份集
rman>catalog start with '/bak/';

7.利用rman恢复出备库
set dbid 834465720
RMAN>
run{
set newname for datafile 1 to  '/u01/app/oracle/oradata/racbak/system01.dbf';
set newname for datafile 2 to  '/u01/app/oracle/oradata/racbak/sysaux01.dbf';
set newname for datafile 3 to  '/u01/app/oracle/oradata/racbak/undotbs01.dbf';
set newname for datafile 4 to  '/u01/app/oracle/oradata/racbak/users01.dbf';
set newname for datafile 5 to  '/u01/app/oracle/oradata/racbak/undotbs02.dbf';
set newname for datafile 6 to  '/u01/app/oracle/oradata/racbak/tps_hxl01.dbf';
set newname for tempfile 1 to '/u01/app/oracle/oradata/temp01.dbf';
restore database;
switch datafile all;
}

8.应用日志
SQL> alter database recover managed standby database using current logfile disconnect from session;
alter database recover managed standby database using current logfile disconnect from session
*
ERROR at line 1:
ORA-38500: USING CURRENT LOGFILE option not available without standby redo logs

9.添加standby日志组
Standby redo log的组数一般为(N +1)* thread# (N分别为每个thread#的联机日志组数).
在本例,每一个thread的联机日志都是3组,所以需要添加6组standby redo log:

alter database add standby logfile thread 1 group 7 ('/u01/app/oracle/oradata/racbak/stdbyredo_1_07.log') size 50m;
alter database add standby logfile thread 1 group 8 ('/u01/app/oracle/oradata/racbak/stdbyredo_1_08.log') size 50m;
alter database add standby logfile thread 1 group 9 ('/u01/app/oracle/oradata/racbak/stdbyredo_1_09.log') size 50m;
alter database add standby logfile thread 1 group 10 ('/u01/app/oracle/oradata/racbak/stdbyredo_1_10.log') size 50m;

alter database add standby logfile thread 2 group 11 ('/u01/app/oracle/oradata/racbak/stdbyredo_2_11.log') size 50m;
alter database add standby logfile thread 2 group 12 ('/u01/app/oracle/oradata/racbak/stdbyredo_2_12.log') size 50m;
alter database add standby logfile thread 2 group 13 ('/u01/app/oracle/oradata/racbak/stdbyredo_2_13.log') size 50m;
alter database add standby logfile thread 2 group 14 ('/u01/app/oracle/oradata/racbak/stdbyredo_1_14.log') size 50m;

9.重新应用日志
SQL> alter database recover managed standby database using current logfile disconnect from session;

Database altered.

到这里整个dataguard环境部署完成 .


--------------------------------------------问题列表------------------------------
1.主库因某个日志文件没有传输到备份,导致日志无法应用
备库日志信息如下:
Archived Log entry 50 added for thread 2 sequence 36 ID 0x31c44f6e dest 1:
RFS[15]: Selected log 12 for thread 2 sequence 31 dbid 834465720 branch 837290977
CORRUPTION DETECTED: In redo blocks starting at block 1920count 977 for thread 2 sequence 31
Errors in file /u01/app/oracle/diag/rdbms/racbak/racbak/trace/racbak_rfs_6215.trc:
ORA-00272: error writing archive log /u01/app/oracle/oradata/racbak/stdbyredo_2_12.log
ORA-00354: corrupt redo log block header

在备库的归档目录没有发现thread 2 sequence 36对应的日志文件,不知道什么原因没有传输到备库.

手工传到备库:
scp 2_31_837290977.dbf oracle@192.168.50.202:/u01/archive_log/

手工注册:
alter database register physical logfile '/u01/archive_log/2_31_837290977.dbf';

查看备库日志 :
ALTER DATABASE REGISTER [PHYSICAL] LOGFILE
Completed: alter database register physical logfile '/u01/archive_log/2_31_837290977.dbf'
Tue Jan 21 23:19:35 2014
Media Recovery Log /u01/archive_log/2_31_837290977.dbf
Media Recovery Log /u01/archive_log/1_26_837290977.dbf
Media Recovery Log /u01/archive_log/2_32_837290977.dbf
Media Recovery Log /u01/archive_log/1_27_837290977.dbf
Media Recovery Log /u01/archive_log/2_33_837290977.dbf
Media Recovery Log /u01/archive_log/2_34_837290977.dbf
Media Recovery Log /u01/archive_log/1_28_837290977.dbf
Media Recovery Log /u01/archive_log/2_35_837290977.dbf
Media Recovery Log /u01/archive_log/1_29_837290977.dbf
Media Recovery Log /u01/archive_log/1_30_837290977.dbf
Media Recovery Log /u01/archive_log/1_31_837290977.dbf
Media Recovery Log /u01/archive_log/2_36_837290977.dbf
Media Recovery Log /u01/archive_log/2_37_837290977.dbf
Media Recovery Log /u01/archive_log/2_38_837290977.dbf
Media Recovery Waiting for thread 1 sequence 32 (in transit)
Recovery of Online Redo Log: Thread 1 Group 7 Seq 32 Reading mem 0
  Mem# 0: /u01/app/oracle/oradata/racbak/stdbyredo_1_07.log
Media Recovery Waiting for thread 2 sequence 39

可以看到备库正常使用主库传过来的日志了 。
2.备库临时表空间对应的数据文件依然是asm的存储方式，导致备库用到临时表空间的会话报错误
备库日志文件错误:
Wed Jan 22 07:26:18 2014
Errors in file /u01/app/oracle/diag/rdbms/racbak/racbak/trace/racbak_dbw0_6048.trc:
ORA-01157: cannot identify/lock data file 201 - see DBWR trace file
ORA-01110: data file 201: '+DATA01/racdb/tempfile/temp.261.836849023'
ORA-17503: ksfdopn:2 Failed to open file +DATA01/racdb/tempfile/temp.261.836849023
ORA-15001: diskgroup "DATA01" does not exist or is not mounted
ORA-15077: could not locate ASM instance serving a required diskgroup
ORA-29701: unable to connect to Cluster Synchronization Service
Errors in file /u01/app/oracle/diag/rdbms/racbak/racbak/trace/racbak_dbw0_6048.trc:
ORA-01186: file 201 failed verification tests
ORA-01157: cannot identify/lock data file 201 - see DBWR trace file
ORA-01110: data file 201: '+DATA01/racdb/tempfile/temp.261.836849023'
File 201 not verified due to error ORA-01157
Successfully added datafile 7 to media recovery
Datafile #7: '/u01/app/oracle/oradata/racbak/tps_hxl02.278.837513931'
Wed Jan 22 07:32:56 2014
ERROR: slave communication error with ASM; terminating process 8121
Errors in file /u01/app/oracle/diag/rdbms/racbak/racbak/trace/racbak_ora_8121.trc:

解决办法:

为该临时数据文件对应的默认表空间添加另外一个数据文件
SQL>alter tablespace temp add tempfile '/u01/app/oracle/oradata/racbak/temp01.dbf' size 100m;

删除ASM的数据文件
SQL> alter database tempfile '+DATA01/racdb/tempfile/temp.261.836849023' drop;

3.rman过来的时候备库会自动生成rac每个节点的在线日志组, 因为是单实例的备库,只需要一个实例的在线日志组就可以了,如下将禁用实例2,同时可以删除实例2对应的在线日志组.

alter database disable thread 2;
alter database drop logfile group 2 thread 2;
alter database drop logfile group 4 thread 2;
alter database drop logfile group 6 thread 2;

-- The End --