再做 Oracle 11gR2 RAC+DG4
第3部分已完成RAC主库到dg备库的日志同步. 这部分将配置dg备库 应用同步过来的日志.
接着完成RAC和dg角色互换.
查看备库是否应用日志
SQL> select sequence# ,name ,applied from v$archived_log;
SEQUENCE# NAME APPLIED
———- ———————————————————————- ———
24 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_24.264.915256125 NO
22 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_22.262.915256125 NO
23 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_23.263.915256125 NO
26 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_26.265.915256129 NO
25 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_25.266.915256129 NO
27 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_27.267.915256129 NO
28 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_28.268.915256129 NO
29 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_29.269.915256129 NO
32 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_32.270.915256129 NO
30 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_30.272.915256129 NO
31 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_31.271.915256129 NO
applied属性为NO,dg备库还没有应用同步过来的重做日志.
SQL> alter database recover managed standby database using current logfile disconnect
from session;
Database altered.
再次执行select sequence# ,name ,applied from v$archived_log; 发现已经同步
SEQUENCE# NAME APPLIED
———- ———————————————————————- ———
24 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_24.264.915256125 YES
22 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_22.262.915256125 YES
23 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_23.263.915256125 YES
26 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_26.265.915256129 YES
25 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_25.266.915256129 YES
27 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_27.267.915256129 YES
28 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_28.268.915256129 YES
29 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_29.269.915256129 YES
32 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_32.270.915256129 YES
30 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_30.272.915256129 YES
31 +FLASH/phydb/archivelog/2016_06_23/thread_1_seq_31.271.915256129 YES
将dg从mount状态切换到open状态, 但在切换之前必须关闭dg应用这些日志.
SQL> alter database open;
Database altered.
查看dg状态
SQL > select dbid,name,switchover_status,db_unique_name,database_role,
open_mode,current_scn from v$database;
重新应用日志
SQL> alter database recover managed standby database using current logfile disconnect
from session;
Database altered.
测试RAC主库和dg备库的数据
node1和gd都执行select file_name,tablespace_name from dba_data_files 发现数据文件是一致的.
SQL> select file_name,tablespace_name from dba_data_files;
FILE_NAME TABLESPACE_NAME
———————————————————— ——————————
+DATA/devdb/datafile/users.259.914028331 USERS
+DATA/devdb/datafile/undotbs1.258.914028331 UNDOTBS1
+DATA/devdb/datafile/sysaux.257.914028331 SYSAUX
+DATA/devdb/datafile/system.256.914028329 SYSTEM
+DATA/devdb/datafile/example.264.914028635 EXAMPLE
+DATA/devdb/datafile/undotbs2.265.914028947 UNDOTBS2
6 rows selected.
RAC查看原表空间, 创建表空间, 查看表空间路径,验证表空间
SQL >create tablespace dg_tbs datafile size 5m;
SQL >show parameter db_create_file;
SQL> select file_name,tablespace_name from dba_data_files;
SQL> select file_name,tablespace_name from dba_data_files;
FILE_NAME TABLESPACE_NAME
———————————————————— ——————————
+DATA/devdb/datafile/users.259.914028331 USERS
+DATA/devdb/datafile/undotbs1.258.914028331 UNDOTBS1
+DATA/devdb/datafile/sysaux.257.914028331 SYSAUX
+DATA/devdb/datafile/system.256.914028329 SYSTEM
+DATA/devdb/datafile/example.264.914028635 EXAMPLE
+DATA/devdb/datafile/undotbs2.265.914028947 UNDOTBS2
+DATA/devdb/datafile/dg_tbs.269.265.917493135 DG_TBS
7 rows selected.
dg备库执行 select file_name,tablespace_name from dba_data_files;
说明dg备库将表空件同步过来了.
RAC主库删除该表空间
SQL >drop tablespace dg_tbs including contents and datafiles;
Tablespace dropped
dg备库检查一下现有表空间
上图所示,当主库删除表空间后,备库对应的表空间也消失了.
RAC主库和dg备库的角色互换
RAC上执行
SQL > alter system set fal_client=’devdb’ sid=’devdb1’;
System altered.
SQL > alter system set fal_client=’devdb’ sid=’devdb2’;
System altered.
SQL > alter system set fal_server=’phydb’ sid=’*’;
System altered.
node1和node2执行
show parameter spfile;
RAC主库创建standby logfile;
在创建之前查看一下本地日志组
SQL> select * from v$logfile;
GROUP# STATUS TYPE MEMBER IS_
———- ——- ——- ———————————————————— —
2 ONLINE +DATA/phydb/onlinelog/group_2.270.915262415 NO
2 ONLINE +FLASH/phydb/onlinelog/group_2.289.915262423 YES
1 ONLINE +DATA/phydb/onlinelog/group_1.269.915262411 NO
1 ONLINE +FLASH/phydb/onlinelog/group_1.288.915262415 YES
3 ONLINE +DATA/phydb/onlinelog/group_3.271.915262423 NO
3 ONLINE +FLASH/phydb/onlinelog/group_3.290.915262429 YES
4 ONLINE +DATA/phydb/onlinelog/group_4.272.915262429 NO
4 ONLINE +FLASH/phydb/onlinelog/group_4.291.915262433 YES
可见本地日志组是1,2,3,4 ; 因此创建standby logfile时不能和上述日志重复
SQL > alter database add standby logfile thread 1 group 5 size 50m, group 6 size 50m,
group 7 size 50m;
Database altered.
SQL > alter database add standby logfile thread 2 group 8 size 50m, group 9 size 50m,
group 10 size 50m;
Database altered.
RAC主库修改参数db_file_name_convert和 log_file_name_convert
参数作用是处理主库和备库磁盘结构不一致的问题,比如: 主库磁盘组data备库磁盘组datadg,
将来在data磁盘组创建表空间,
而备库得到这个表空间,表空间含有data磁盘组, 而备库没有data磁盘组, 因此需要这几个参数将旧磁盘组
data转换到新的磁盘组datadg.否则因为备库上没有data磁盘组而不能创建这个表空间.
SQL >alter system set db_file_name_convert=’+DATA/phydb/’ , ‘+DATA/devdb/’ sid=’*’ scope=spfile;
Database altered.
SQL >alter system set log_file_name_convert=’+DATA/phydb/’ , ‘+DATA/devdb/’ sid=’*’ scope=spfile;
Database altered.
下面进入主备切换环节
停掉RAC的其中一个节点,这里停止node2
SQL > shutdown immediate;
Datebase closed.
Datebase dismounted.
ORACLE instance shut down.
将node1切换成备库
SQL > alter database commit to switchover to physical standby with session shutdown;
Database altered.
查看node1状态
SQL > select dbid,name,current_scn,protection_mode,database_role,force_logging,open_mode,
switchover_status from gv$database;
目前的情况 RAC和dg都是standby状态了
dg切换成主库
SQL > alter database commit to switchover to primary;
Database altered.
查看一下dg状态
SQL > select dbid,name,current_scn,protection_mode,database_role,force_logging,open_mode,
switchover_status from gv$database;
停止node1和node2
SQL > shutdown immediate;
Datebase closed.
Datebase dismounted.
ORACLE instance shut down.
退出sqlplus ,切换到grid用户,然后停止实例devdb
node1和node2切换到mount状态
node1 -> srvctl start database -d devdb -o mount;
node1和node2切换到oracle用户,sqlplus
SQL > select dbid,name,current_scn,protection_mode,database_role,force_logging,open_mode,
switchover_status from gv$database;
发现node1和node2的switchover_status都是session active的; DATABASE_ROLE都变成了
PHYSICAL STANDBY
RAC备库应用日志
SQL > alter database recover managed standby database using current logfile disconnect
from session;
Database altered.
将dg主库从mount状态切换到open状态
SQL > alter database open;
Database altered.
此时dg主库 OPEN_MODE是READ WRITE, SWITCHOVER_STATUS是RESOLVABLE GAP
RAC主库的SWITCHOVER_STATUS状态改为READ ONLY
SQL > alter database recover managed standby database cancel;
Database altered.
SQL > alter database open;
Database altered.
SQL > select dbid,name,current_scn,protection_mode,database_role,force_logging,open_mode,
switchover_status from gv$database;
RAC备库打开日志应用
SQL > alter database recover managed standby database using current logfile disconnect
from session;
Database altered.
最终状态确认
dg主库状态和rac备库状态
至此rac库和dg库角色切换操作成功.
oracle 11gR2 RAC+DG搭建测试完成.