NBU上Oracle数据库恢复演练手册-526互联

3.11|11.1.1 安装新的客户端

在备份服务器上添加hosts信息

IP地址客户端
{NBU Master IP地址} {NBU Master主机名}

解压NBU客户端安装包并运行安装程序,安装客户端

3.21|21.1.2 新建异机恢复文件

在备份服务器上建立支持异机恢复要求的文件。
备份服务器建立No.Restrictions空文件
Windows platform: InstallPath\netbackup\db\altnames\No.Restrictions
UNIX platform: /usr/openv/netbackup/db/altnames/No.Restrictions
3.31|31.1.3 新建数据库实例

对于需要恢复的机器，需要新建立一个数据库实例进行数据恢复。

本次恢复针到的实例为test

3.41|41.1.4 建立spfile文件

原spfile导出文件为：
[root@backup-test testtest]# cat inittest1.ora.old
test2.__db_cache_size=7918845952
test1.__db_cache_size=6643777536
test2.__java_pool_size=134217728
test1.__java_pool_size=134217728
test2.__large_pool_size=167772160
test1.__large_pool_size=167772160
test1.__oracle_base='/oracle/'#ORACLE_BASE set from environment
test2.__oracle_base='/oracle/'#ORACLE_BASE set from environment
test2.__pga_aggregate_target=6442450944
test1.__pga_aggregate_target=5637144576
test2.__sga_target=9663676416
test1.__sga_target=10468982784
test2.__shared_io_pool_size=0
test1.__shared_io_pool_size=0
test2.__shared_pool_size=1342177280
test1.__shared_pool_size=3422552064
test2.__streams_pool_size=0
test1.streams_pool_size=0
.audit_file_dest='/oracle/admin/test/adump'
.audit_trail='db'
.cluster_database=true
.compatible='11.2.0.4.0'
.control_files='+DATA_DG/test/controlfile/current.256.834077963','+DATA_DG/test/controlfile/current.275.834330051','+FLASH_DG/test/controlfile/current.278.834330403'
.db_block_size=8192
.db_create_file_dest='+DATA_DG'
.db_domain=''
.db_name='test'
.diagnostic_dest='/oracle'
.dispatchers='(PROTOCOL=TCP) (SERVICE=testXDB)'
test1.instance_number=1
test2.instance_number=2
.lock_sga=FALSE
.log_archive_dest_1='LOCATION=+FLASH_DG'
.log_archive_format='%t%s%r.dbf'
.memory_max_target=16106127360
.memory_target=16106127360
.open_cursors=500
.processes=1000
.remote_listener='test-scan:1521'
.remote_login_passwordfile='exclusive'
.session_cached_cursors=100
.sessions=1105
test2.thread=2
test1.thread=1
.undo_retention=7200
test1.undo_tablespace='UNDOTBS1'
test2.undo_tablespace='UNDOTBS2'

由于需要恢复到单机，删除RAC相关信息。另外由于硬件资源的不同，删除需要恢复的资源配置。
新的配置文件为
[root@backup-test testtest]# cat inittest1.ora
.__oracle_base='/oradata/testtest/'#ORACLE_BASE set from environment
.__pga_aggregate_target=5637144576
.sga_target=10468982784
.audit_file_dest='/oradata/testtest/adump'
.audit_trail='db'
.compatible='11.2.0.4.0'
.control_files='/oradata/testtest/controlfile/current.256.834077963'
.db_block_size=8192
.db_create_file_dest='/oradata/testtest/'
.db_domain=''
.db_name='test'
.diagnostic_dest='/oradata/testtest/'
.dispatchers='(PROTOCOL=TCP) (SERVICE=testXDB)'
.lock_sga=FALSE
.log_archive_dest_1='LOCATION=/oradata/testtest/archive/'
.log_archive_format='%t%s%r.dbf'
.open_cursors=500
.processes=1000
.remote_login_passwordfile='exclusive'
.session_cached_cursors=100
.sessions=1105
.undo_retention=7200
.undo_tablespace='UNDOTBS1'

3.51|51.1.5 建立数据文件夹

根据spfile建立相关文件
[root@backup-test testtest]# ll -a
total 52
drwxrwxrwx 7 oracle oinstall 4096 Sep 25 16:26 .
drwxrwxr-x 29 oracle oinstall 12288 Sep 28 10:03 ..
drwxr-xr-x 2 oracle oinstall 4096 Sep 28 10:37 adump
drwxr-xr-x 2 oracle oinstall 4096 Sep 25 16:02 archive
drwxr-xr-x 2 oracle oinstall 4096 Sep 28 10:10 controlfile
drwxrwxr-x 3 oracle oinstall 4096 Sep 25 16:20 diag
-rwxrwxrwx 1 oracle oinstall   760 Sep 25 16:26 inittest1.ora
-rwxr-xr-x 1 oracle oinstall 1578 Sep 25 15:45 inittest1.ora.old
drwxrwxrwx 2 oracle oinstall 4096 Sep 25 12:04 testtest
-rwxrwxrwx 1 oracle oinstall 6520 Sep 25 14:44 utlpwdmg.sql

3.61|61.1.6 启动数据库到nomount状态
sqlplus /nolog
idle> conn / as sysdba
idle> startup nomount pfile='/oradata/testtest/inittest1.ora'

idle> startup nomount pfile='/oradata/testtest/inittest1.ora'
ORACLE instance started.

Total System Global Area 396726272 bytes
Fixed Size                  2253504 bytes
Variable Size             318770496 bytes
Database Buffers           67108864 bytes
Redo Buffers                8593408 bytes

3.71|71.1.7 列出备份信息
在备份服务器上列出相关备份信息
bplist -C test-db01 -s 09/27/2015 -t 4 -R -b -l /
。。。。。。。

3.81|81.1.8 还原controlfile，并mount数据库
RMAN还原
RMAN>
RMAN>connect target /
RMAN> run{
2> allocate channel ch0 type 'sbt_tape';
3> send 'NB_ORA_CLIENT=test-db01';
4> send 'nb_ora_serv=nbu-b';
5> restore controlfile from '/cntrl_3175_1_891580028';
6> release channel ch0;
7> }

using target database control file instead of recovery catalog
allocated channel: ch0
channel ch0: SID=578 device type=SBT_TAPE
channel ch0: Veritas NetBackup for Oracle - Release 7.6 (2014102721)

sent command to channel: ch0

sent command to channel: ch0

Starting restore at 2015-09-28 10:10:36

channel ch0: restoring control file
channel ch0: restore complete, elapsed time: 00:00:15
output file name=/oradata/testtest/controlfile/current.256.834077963
Finished restore at 2015-09-28 10:10:52

released channel: ch0

RMAN> alter database mount;

database mounted

3.91|91.1.9 查找数据文件信息
sqlplus /nolog
idle> conn / as sysdba

idle> select name from v$datafile;

NAME

+DATA_DG/test/datafile/system.259.834077963
+DATA_DG/test/datafile/sysaux.260.834077967
+DATA_DG/test/datafile/undotbs1.261.834077971
+DATA_DG/test/datafile/undotbs2.263.834077979
+DATA_DG/test/datafile/users.264.834077981
+DATA_DG/test/datafile/zhos_dataspace.268.889894605
+DATA_DG/test/datafile/zhos_dataspace.272.834956161
+DATA_DG/test/datafile/zhos_dataspace.271.834956199
+DATA_DG/test/datafile/spcred_dataspace.269.834956241
+DATA_DG/test/datafile/zhos_indexspace.273.834317857
+DATA_DG/test/datafile/audit_dataspace.280.846839715

11 rows selected.

idle> select member from v$logfile;

MEMBER

+DATA_DG/test/onlinelog/group_1.257.834077963
+DATA_DG/test/onlinelog/group_2.258.834077963
+DATA_DG/test/onlinelog/group_3.265.834080437
+DATA_DG/test/onlinelog/group_4.266.834080437
+DATA_DG/test/onlinelog/group_5.276.834332677
+DATA_DG/test/onlinelog/group_6.277.834332723
+DATA_DG/test/onlinelog/group_7.278.834332751
+DATA_DG/test/onlinelog/group_8.279.834332783

8 rows selected.

3.101|101.1.10 数据文件恢复

编辑还原脚本如下：
export ORACLE_SID=test
rman <<EOF
connect target /
run {
allocate channel ch0 type 'SBT_TAPE';
allocate channel ch1 type 'SBT_TAPE';
send 'nb_ora_serv=nbu-b';
send 'nb_ora_client=test-db01';
set newname for datafile '+DATA_DG/test/datafile/system.259.834077963' to '/oradata/testtest/system.259.834077963';
set newname for datafile '+DATA_DG/test/datafile/sysaux.260.834077967' to '/oradata/testtest/sysaux.260.834077967';
set newname for datafile '+DATA_DG/test/datafile/undotbs1.261.834077971' to '/oradata/testtest/undotbs1.261.834077971';
set newname for datafile '+DATA_DG/test/datafile/undotbs2.263.834077979' to '/oradata/testtest/undotbs2.263.834077979';
set newname for datafile '+DATA_DG/test/datafile/users.264.834077981' to '/oradata/testtest/users.264.834077981';
set newname for datafile '+DATA_DG/test/datafile/zhos_dataspace.268.889894605' to '/oradata/testtest/zhos_dataspace.268.889894605';
set newname for datafile '+DATA_DG/test/datafile/zhos_dataspace.272.834956161' to '/oradata/testtest/zhos_dataspace.272.834956161';
set newname for datafile '+DATA_DG/test/datafile/zhos_dataspace.271.834956199' to '/oradata/testtest/zhos_dataspace.271.834956199';
set newname for datafile '+DATA_DG/test/datafile/spcred_dataspace.269.834956241' to '/oradata/testtest/spcred_dataspace.269.834956241';
set newname for datafile '+DATA_DG/test/datafile/zhos_indexspace.273.834317857' to '/oradata/testtest/zhos_indexspace.273.834317857';
set newname for datafile '+DATA_DG/test/datafile/audit_dataspace.280.846839715' to '/oradata/testtest/audit_dataspace.280.846839715';
restore database;
switch datafile all;
release channel ch0;
release channel ch1;
}
exit
EOF

运行数据还原
recover database;

恢复完成

3.111|111.1.11 归档日志恢复

idle> select THREAD#,SEQUENCE#,FIRST_TIME,COMPLETION_TIME from v$archived_log where COMPLETION_TIME>to_date('20150927 20:00:00','yyyymmdd hh24:mi:ss') and COMPLETION_TIME<to_date('20150928 6:00:00','yyyymmdd hh24:mi:ss') order by SEQUENCE#;

   THREAD# SEQUENCE# FIRST_TIME          COMPLETION_TIME

1 27587 2015-09-27 19:59:31 2015-09-27 21:01:28
......

39 rows selected.
RMAN还原归档日志
节点1还原
RMAN> connect target /

connected to target database: test (DBID=825185095, not open)

RMAN> run{
set archivelog destination to '/oradata/testtest/archive';
2> 3> allocate channel ch0 type 'SBT_TAPE';
send 'NB_ORA_CLIENT=test-db01';
send 'nb_ora_serv=nbu-b';
restore archivelog sequence between 27587 and 27597 thread 1;
4> 5> 6> 7> release channel ch0;
8> }

executing command: SET ARCHIVELOG DESTINATION

......

released channel: ch0
节点2还原
RMAN> run{
set archivelog destination to '/oradata/testtest/archive';
2> 3> allocate channel ch0 type 'SBT_TAPE';
4> send 'NB_ORA_CLIENT=test-db01';
5> send 'nb_ora_serv=nbu-b';
6> restore archivelog sequence between 45287 and 45314 thread 2;
7> release channel ch0;
8> }

executing command: SET ARCHIVELOG DESTINATION

allocated channel: ch0
channel ch0: SID=673 device type=SBT_TAPE
channel ch0: Veritas NetBackup for Oracle - Release 7.6 (2014102721)

sent command to channel: ch0

sent command to channel: ch0

Starting restore at 2015-09-28 14:14:04

channel ch0: starting archived log restore to user-specified destination
......
released channel: ch0

3.121|121.1.12 联机日志路径修改
Sqlplus 修改
idle>
alter database rename file '+DATA_DG/test/onlinelog/group_1.257.834077963' to'/oradata/testtest/group_1.257.834077963';
......

3.131|131.1.13 Recover并打开数据库
Sqlplus启动数据库
idle> recover database until cancel using backup controlfile;
ORA-00279: change 10315737932005 generated at 09/27/2015 22:03:45 needed for thread 1
ORA-00289: suggestion : /oradata/testtest/archive/1_27589_834077959.dbf
ORA-00280: change 10315737932005 for thread 1 is in sequence #27589

Specify log: {=suggested | filename | AUTO | CANCEL}
AUTO
ORA-00308: cannot open archived log '/oradata/testtest/archive/1_27589_834077959.dbf'
ORA-27037: unable to obtain file status
Linux-x86_64 Error: 2: No such file or directory
Additional information: 3

..............

ORA-00308: cannot open archived log '/oradata/testtest/archive/2_45315_834077959.dbf'
ORA-27037: unable to obtain file status
Linux-x86_64 Error: 2: No such file or directory
Additional information: 3

idle>
idle> alter database open resetlogs;

Database altered.

3.141|141.1.14 验证数据库

idle> select open_mode from v$database;

OPEN_MODE

READ WRITE