月归档：三月 2011

lag() lead()函数使用

发表于 2011 年 3 月 18 日 由 惜分飞

1、语法
LAG (value_expression [,offset] [,default]) OVER ([query_partition_clause] order_by_clause)
LEAD (value_expression [,offset] [,default]) OVER ([query_partition_clause] order_by_clause)

2、说明
1）lag 和lead 可以 获取结果集中，按一定排序所排列的当前行的上下相邻若干offset 的某个行的某个列(不用结果集的自关联）
2）lag 和lead 有三个参数，第一个参数是列名，第二个参数是偏移的offset，第三个参数是超出记录窗口时的默认值
3）lag ，lead 分别是向前，向后

3、准备试验环境

CREATE TABLE t(t_NAME VARCHAR2(10),t_order NUMBER);
INSERT INTO t VALUES('aaaa',1);
INSERT INTO t VALUES('bbbb',2);
INSERT INTO t VALUES('cccc',3);
INSERT INTO t VALUES('dddd',6);
INSERT INTO t VALUES('eeee',4);
INSERT INTO t VALUES('ffff',5);
SELECT * FROM t;
T_NAME        T_ORDER
---------- ----------
aaaa                1
bbbb                2
cccc                3
dddd                6
eeee                4
ffff                5
6 rows selected

4、测试

SELECT T_ORDER,
LAG(T_NAME, 1,'lag_fei') OVER(ORDER BY T_ORDER) "LAG",
T_NAME,
LEAD(T_NAME, 1,'lead_fei') OVER(ORDER BY T_ORDER)"LEAD"
FROM T
ORDER BY T_ORDER;

T_ORDER LAG        T_NAME     LEAD
---------- ---------- ---------- ----------
1 lag_fei    aaaa       bbbb
2 aaaa       bbbb       cccc
3 bbbb       cccc       eeee
4 cccc       eeee       ffff
5 eeee       ffff       dddd
6 ffff       dddd       lead_fei
6 rows selected
发表在 Oracle 开发 | 评论关闭

Data Gurad switchover切换（Physical Standby Databases）

发表于 2011 年 3 月 17 日 由 惜分飞

一、主库
查看当前数据库状态
select database_role,switchover_status from v$database;

当switchover_status为SESSIONS ACTIVE时
alter database commit to switchover to physical standby with session shutdown;
否则执行
alter database commit to switchover to physical standby;

切换成功关闭数据库
shutdown immediate;

startup mount;
note：11g 可以直接置于 read only 使用命令startup

开启实时应用
alter database recover managed standby database using current logfile disconnect from session;

恢复备库未利用日志
alter database recover managed standby database disconnect from session;

二、备库
查看当前数据库状态
select database_role,switchover_status from v$database;

如果是switchover_status为recovery needed或switchover latent 恢复数据库
alter database recover managed standby database disconnect from session;

切换数据库
alter database commit to switchover to primary with session shutdown;

开启数据库
startup

–查询除自己进程外的其他进程，并kill掉

SELECT 'kill -9 '||p.spid,'alter system kill session '''||v.sid||','||v.SERIAL#||''''
FROM V$SESSION v ,v$process p  
WHERE v.TYPE = 'USER' 
AND v.SID <> (SELECT DISTINCT SID FROM V$MYSTAT) 
AND v.paddr=p.addr;

三、补充11g官方文档处理顺序和操作语句
1、主库切换
SELECT SWITCHOVER_STATUS FROM V$DATABASE;
ALTER DATABASE COMMIT TO SWITCHOVER TO PHYSICAL STANDBY WITH SESSION SHUTDOWN;
SHUTDOWN ABORT;
STARTUP MOUNT;

2、备库切换
SELECT SWITCHOVER_STATUS FROM V$DATABASE;
ALTER DATABASE COMMIT TO SWITCHOVER TO PRIMARY WITH SESSION SHUTDOWN;
ALTER DATABASE OPEN;

3、开启应用（新备库–原主库）
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE USING CURRENT LOGFILE DISCONNECT FROM SESSION;

发表在 Data Guard | 评论关闭

11g data guard（PHYSICAL STANDBY）配置

发表于 2011 年 3 月 16 日 由 惜分飞

一、 Preparing the Primary Database for Standby Database Creation
1、安装redhat 5.5系统和11g r2数据库，相关参数如下

Database DB_UNIQUE_NAME TNS Ip
Primary first first 192.168.1.2
Physical standby sencode sencode 192.168.1.3

2、Enable Forced Logging
2.1）查看数据库是否开启force log
select force_logging from v$database;
2.2）如果得到结果是YES，跳过2.3）
2.3）ALTER DATABASE FORCE LOGGING;

3、Enable Archiving
3.1）查看数据库是否启动归档模式
archive log list;
3.2）如果得到类此下面结果跳过步骤3.3）
Database log mode Archive Mode
Automatic archival Enabled
3.3）开启归档模式
SHUTDOWN IMMEDIATE;
STARTUP MOUNT;
ALTER DATABASE ARCHIVELOG;
ALTER DATABASE OPEN;

4、创建Primary Database和Standby Database的tns，确保两者能够相互tnsping通
Note：tns配置成dedicated server模式

二、Step-by-Step Instructions for Creating a Physical Standby Database
1、Create a Backup Copy of the Primary Database Datafiles
方法一、冷备份
关闭Primary Database数据库（immediate）拷贝datafile、redo、undo、temp到Standby Database位置
Note：如果确保自己有能力在没有redo、undo、temp文件的情况下恢复数据库，可以只拷贝datafile（建议全拷）
方法二、rman备份
Rman 备份，然后在Standby端恢复

2、Create a Control File for the Standby Database
ALTER DATABASE CREATE STANDBY CONTROLFILE AS ‘/tmp/control01.ctl’;
Scp /tmp/control01.ctl oracle@192.168.1.3:/u01/oradata/first/

3、Create a Parameter File for the Standby Database
3.1）Create a parameter file
CREATE PFILE=’/tmp/pfile’ FROM SPFILE;

3.2）Modify the Parameter File
3.2.1）Primary Database
first.__db_cache_size=71303168
first.__java_pool_size=4194304
first.__large_pool_size=4194304
first.__oracle_base=’/u01′#ORACLE_BASE set from environment
first.__pga_aggregate_target=121634816
first.__sga_target=192937984
first.__shared_io_pool_size=0
first.__shared_pool_size=100663296
first.__streams_pool_size=4194304
*.audit_file_dest=’/u01/admin/first/adump’
*.audit_trail=’db’
*.compatible=’11.2.0.0.0′
*.control_files=’/u01/oradata/first/control01.ctl’,
‘/u01/flash_recovery_area/first/control02.ctl’
*.db_block_size=8192
*.db_domain=’dg’
*.db_name=’first’
*.db_unique_name=’first’
*.db_recovery_file_dest=’/u01/flash_recovery_area’
*.db_recovery_file_dest_size=4039114752
*.diagnostic_dest=’/u01′
*.dispatchers=’(PROTOCOL=TCP) (SERVICE=firstXDB)’
*.log_archive_config=’dg_config=(first,second)’
*.log_archive_dest_1=’LOCATION=/u01/archive valid_for=(all_logfiles,all_roles)
*.log_archive_dest_2=’service=second lgwr async valid_for=(online_logfiles,primary_role) db_unique_name=second’
*.log_archive_format=’%t_%s_%r.arc’
*.log_archive_dest_state_1=enable
*.log_archive_dest_state_2=enable
fal_server=second
*.memory_target=314572800
*.nls_language=’SIMPLIFIED CHINESE’
*.nls_territory=’CHINA’
*.open_cursors=300
*.processes=150
*.remote_login_passwordfile=’EXCLUSIVE’
*.undo_tablespace=’UNDOTBS1′
standby_file_management=auto

3.2.2）Standby Database
second.__db_cache_size=71303168
second.__java_pool_size=4194304
second.__large_pool_size=4194304
second.__oracle_base=’/u01′#ORACLE_BASE set from environment
second.__pga_aggregate_target=121634816
second.__sga_target=192937984
second.__shared_io_pool_size=0
second.__shared_pool_size=100663296
second.__streams_pool_size=4194304
*.audit_file_dest=’/u01/admin/second/adump’
*.audit_trail=’db’
*.compatible=’11.2.0.0.0′
*.control_files=’/u01/oradata/first/control01.ctl’
*.db_block_size=8192
*.db_domain=’dg’
*.db_name=’first’
*.db_unique_name=’second’
*.db_recovery_file_dest=’/u01/flash_recovery_area’
*.db_recovery_file_dest_size=4039114752
*.diagnostic_dest=’/u01′
*.dispatchers=’(PROTOCOL=TCP) (SERVICE=firstXDB)’
*.log_archive_config=’dg_config=(first,second)’
*.log_archive_dest_1=’LOCATION=/u01/archive valid_for=(all_logfiles,all_roles) db_unique_name=second’
*.log_archive_dest_2=’service=first lgwr async valid_for=(online_logfiles,primary_role)
*.log_archive_format=’%t_%s_%r.arc’
*.log_archive_dest_state_1=enable
*.log_archive_dest_state_2=enable
fal_server=first
*.memory_target=314572800
*.nls_language=’SIMPLIFIED CHINESE’
*.nls_territory=’CHINA’
*.open_cursors=300
*.processes=150
*.remote_login_passwordfile=’EXCLUSIVE’
*.undo_tablespace=’UNDOTBS1′
standby_file_management=auto

scp /tmp/pfile oracle@192.168.1.3:/tmp/
拷贝主库密码文件到备库，确保sys密码一致
scp /u01/oracle/dbs/orapwfirst oracle@192.168.1.3:/u01/oracle/dbs/orapwsecond

4、Start the Physical Standby Database
4.1）Startup pfile=’/tmp/pfile’ mount（可以直接open，oracle会自动置于read only模式）
4.2）Add STANDBY LOGFILE
ALTER DATABASE ADD STANDBY LOGFILE (‘/u01/oradata/first/slog1.rdo’) SIZE 50M;
ALTER DATABASE ADD STANDBY LOGFILE (‘/u01/oradata/first/slog1.rdo’) SIZE 50M;
ALTER DATABASE ADD STANDBY LOGFILE (‘/u01/oradata/first/slog1.rdo’) SIZE 50M;
4.3）Start Redo Apply
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE USING CURRENT LOGFILE DISCONNECT FROM SESSION;

5、Start Primary Database
Startup pfile=’/tmp/pfile’;

6、Verify the Physical Standby Database Is Performing Properly
6.1）standby database,
SELECT SEQUENCE#, FIRST_TIME, NEXT_TIME FROM V$ARCHIVED_LOG ORDER BY SEQUENCE#;
6.2）Primary Database
ALTER SYSTEM SWITCH LOGFILE;
SELECT SEQUENCE#, FIRST_TIME, NEXT_TIME FROM V$ARCHIVED_LOG ORDER BY SEQUENCE#;
6.3）standby database
SELECT SEQUENCE#, FIRST_TIME, NEXT_TIME FROM V$ARCHIVED_LOG ORDER BY SEQUENCE#;
观察SEQUENCE#在Primary Database和standby database是否一致

7、Create Spfile
create spfile from pfile=’/tmp/pfile’;(Primary Database和standby database)

发表在 Data Guard | 评论关闭