構建Dataguard邏輯備庫指南

一直想做一個logical standby,每次都是以失敗告終，在屢敗屢戰之後，終於迎來了勝利！據說可以使用網格控制器點幾下，簡單的把物理備庫轉換成邏輯備庫，下次有工夫的時候可以測試下！相信一切問題都是時間問題，一切煩惱都是自尋煩惱！下面來描述下如何將physical standby轉換成logical standby,在開始之前，請先配好dataguard 的physical standby

physical standby配置請參考：
http://ylw6006.blog.51cto.com/470441/661105
dataguard各參數含義請參考：
http://space.itpub.net/519536/viewspace-578352
http://hi.baidu.com/edeed/blog/item/f54c6e06b92bc77703088174.html

環境介紹：
主庫：
IP地址: 192.168.227.20/24
OS版本：Rhel5.4 32位
SID名稱： orcl
db_name: orcl
db_unique_name: primary

物理備庫：(轉換前)
IP地址: 192.168.227.30/24
OS版本：Rhel5.4 32位
SID名稱： orcl
db_name: orcl
db_unique_name: standby

邏輯備庫：(轉換後)
IP地址: 192.168.227.20/24
OS版本：Rhel5.4 32位
SID名稱： orcl
db_name: logdb
db_unique_name: logdb

一：在物理備庫上確認歸檔日誌apply正常

SQL> SELECT ARCH.THREAD# "Thread", ARCH.SEQUENCE# "Last Sequence Received",
          APPL.SEQUENCE# "Last Sequence Applied"
         FROM
         (SELECT THREAD# ,SEQUENCE# FROM V$ARCHIVED_LOG
         WHERE (THREAD#,FIRST_TIME )
          IN
         (SELECT THREAD#,MAX(FIRST_TIME) FROM V$ARCHIVED_LOG GROUP BY THREAD#)) ARCH,
          (SELECT THREAD# ,SEQUENCE# FROM V$LOG_HISTORY WHERE (THREAD#,FIRST_TIME )
          IN
         (SELECT THREAD#,MAX(FIRST_TIME) FROM V$LOG_HISTORY GROUP BY THREAD#)) APPL
     WHERE ARCH.THREAD# = APPL.THREAD# ORDER BY 1

Thread Last Sequence Received Last Sequence Applied
---------- ---------------------- ---------------------
1 10 10

二:在備庫上停止接收和應用歸檔日誌

SQL> alter database recover managed standby database cancel;
Database altered.

三:配置主庫

SQL> show parameter log_archive_config;

NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
log_archive_config string DG_CONFIG=(primary,logdb)

SQL> show parameter dest_1;

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
db_create_online_log_dest_1          string
log_archive_dest_1                   string      LOCATION=/u01/arch/orcl valid_
                                                 for=(all_logfiles,all_roles) d
                                                 b_unique_name=primary

SQL> show parameter dest_2; //service=standby，這裏standby指的是tns串

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
db_create_online_log_dest_2          string
log_archive_dest_2                   string      service=standby lgwr async val
                                                 id_for=(online_logfiles,primar
                                                 y_role) db_unique_name=logdb

SQL> !mkdir -p /u01/arch/standby/orcl
SQL> alter system set log_archive_dest_3='LOCATION=/u01/arch/standby/orcl VALID_FOR=(STANDBY_LOGFILES,STANDBY_ROLE)';
System altered.

在主庫生成LogMiner字典,切換日誌
SQL> execute dbms_logstdby.build;
PL/SQL procedure successfully completed.

四：配置備庫
將物理備庫轉換成邏輯備庫，在這之前請確保物理備庫使用spfile啓動

SQL> alter database recover to logical standby logdb;
Database altered.

SQL> shutdown immediate
SQL> startup mount
SQL> show parameter log_archive_config; //dg_config後面的參數指的是db_unique_name

NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
log_archive_config string DG_CONFIG=(primary,logdb)

SQL> alter system set db_unique_name='logdb' scope=spfile;
System altered

SQL> alter system set log_archive_dest_1='LOCATION=/u01/arch/orcl valid_for=(online_logfiles,all_roles) db_unique_name=logdb' scope=spfile;
System altered

SQL> !mkdir -p /u01/arch/standby/orcl

SQL> alter system set log_archive_dest_3='LOCATION=/u01/arch/standby/orcl VALID_FOR=(STANDBY_LOGFILES,STANDBY_ROLE)';
System altered.

SQL> shutdown immediate

修改tnsnames.ora文件，主庫上也需要修改
[oracle@orcl u01]$ tnsping standby
TNS Ping Utility for Linux: Version 10.2.0.1.0 - Production on 28-SEP-2011 11:01:31
Copyright (c) 1997, 2005, Oracle. All rights reserved.
Used parameter files:

Used TNSNAMES adapter to resolve the alias
) ERVICE_NAME = logdb.herostart.com) 192.168.227.30)(PORT = 1521)) OK (20 msec)

將備庫打開，置爲應用sql模式
SQL> conn /as sysdba
Connected to an idle instance.
SQL> startup mount
SQL> ALTER DATABASE OPEN RESETLOGS;
Database altered.

SQL> alter database start logical standby apply immediate;
alter database start logical standby apply immediate
*
ERROR at line 1:
ORA-16239: IMMEDIATE option not available without standby redo logs

SQL> alter database add standby logfile group 4 '/u01/app/oracle/oradata/orcl/standby04.dbf' size 50M;
Database altered.

SQL> alter database add standby logfile group 5 '/u01/app/oracle/oradata/orcl/standby05.dbf' size 50M;
Database altered.

SQL> alter database add standby logfile group 6 '/u01/app/oracle/oradata/orcl/standby06.dbf' size 50M;
Database altered.

SQL> alter database start logical standby apply immediate;
Database altered

SQL> select name, database_role from v$database;

NAME DATABASE_ROLE
--------- ----------------
LOGDB LOGICAL STANDBY

五：測試

主庫更新：
SQL> select sysdate from dual;

SYSDATE
-------------------
2011-09-28:14:07:04

SQL> update hr.employees set salary=200 where employee_id=100;
1 row updated.

備庫查詢：
SQL> select salary from hr.employees where employee_id=100;

SALARY
----------
100

SQL> select sysdate from dual;

SYSDATE
-------------------
2011-09-28:14:08:38

主庫提交，切換日誌
SQL> commit;
Commit complete.

SQL> alter system switch logfile;
System altered.

備庫再次查詢
SQL> select salary from hr.employees where employee_id=100;

SALARY
----------
200

六：驗證

SQL> SELECT SEQUENCE#, FIRST_TIME, NEXT_TIME, DICT_BEGIN, DICT_END FROM DBA_LOGSTDBY_LOG ORDER BY SEQUENCE#;

SEQUENCE# FIRST_TIME          NEXT_TIME           DIC DIC
---------- ------------------- ------------------- --- ---
        24 2011-09-28:13:19:50 2011-09-28:13:46:09 NO NO
        25 2011-09-28:13:46:09 2011-09-28:14:09:11 NO NO

SQL> select file_name,applied from DBA_LOGSTDBY_LOG ;

FILE_NAME                                     APPLIED
--------------------------------------------- --------
/u01/arch/standby/orcl/1_24_760125331.arc     YES
/u01/arch/standby/orcl/1_25_760125331.arc     YES


SQL> SELECT EVENT_TIME, STATUS, EVENT FROM DBA_LOGSTDBY_EVENTS ORDER BY EVENT_TIME, COMMIT_SCN;

EVENT_TIME           STATUS               EVENT
-------------------- -------------------- --------------------------------
2011-09-28:11:09:32 ORA-16111: log minin
                     g and apply setting
                     up

2011-09-28:11:28:04 ORA-16128: User init
iated stop apply suc
cessfully completed

2011-09-28:11:28:04 ORA-16112: log minin
g and apply stopping

EVENT_TIME STATUS EVENT
-------------------- -------------------- --------------------------------

2011-09-28:11:28:14 ORA-16111: log minin
g and apply setting
up

2011-09-28:11:53:23 ORA-16226: DDL skipp ALTER DATABASE OPEN
ed due to lack of su
pport

2011-09-28:12:49:31 ORA-16128: User init
iated stop apply suc

EVENT_TIME STATUS EVENT
-------------------- -------------------- --------------------------------
cessfully completed

2011-09-28:12:50:11 ORA-16111: log minin
g and apply setting
up

2011-09-28:13:23:59 ORA-16226: DDL skipp ALTER DATABASE OPEN
ed due to lack of su
pport

8 rows selected.

[oracle@orcl orcl]$ tail -f /u01/app/oracle/admin/orcl/bdump/alert_orcl.log
Primary database is in MAXIMUM PERFORMANCE mode
Wed Sep 28 14:09:13 2011
LOGMINER: Log Auto Delete - deleting: /u01/arch/orcl/1_22_760125331.arc
Deleted file /u01/arch/orcl/1_22_760125331.arc
Wed Sep 28 14:09:13 2011
LOGMINER: Log Auto Delete - deleting: /u01/arch/orcl/1_23_760125331.arc
Deleted file /u01/arch/orcl/1_23_760125331.arc
Wed Sep 28 14:09:13 2011
RFS[3]: Successfully opened standby log 4: '/u01/app/oracle/oradata/orcl/standby04.dbf'
Wed Sep 28 14:09:15 2011
LOGMINER: Begin mining logfile: /u01/app/oracle/oradata/orcl/standby04.dbf
Wed Sep 28 14:09:15 2011
RFS LogMiner: Registered logfile [/u01/arch/standby/orcl/1_25_760125331.arc] to LogMiner session id [1]

[oracle@orcl ~]$ ll /u01/arch/standby/orcl/*
-rw-r----- 1 oracle oinstall 772608 Sep 28 13:46 /u01/arch/standby/orcl/1_24_760125331.arc
-rw-r----- 1 oracle oinstall 1597952 Sep 28 14:09 /u01/arch/standby/orcl/1_25_760125331.arc

SQL> SELECT NAME, VALUE FROM V$LOGSTDBY_STATS;

NAME                                                         VALUE
------------------------------------------------------------ ----------
number of preparers                                          1
number of appliers                                           5
maximum SGA for LCR cache                                    30
parallel servers in use                                      9
maximum events recorded                                      100
preserve commit order                                        TRUE
transaction consistency                                      FULL
record skip errors                                           Y
record skip DDL                                              Y
record applied DDL                                           N
record unsupported operations                                N

NAME                                                         VALUE
------------------------------------------------------------ ----------
coordinator state                                            IDLE
transactions ready                                           364
transactions applied                                         364
coordinator uptime                                           10071
realtime logmining                                           Y
apply delay                                                  0
Log Miner session ID                                         1
txns delivered to client                                     3466
DML txns delivered                                           1498
DDL txns delivered                                           11
CTAS txns delivered                                          0

NAME                                                         VALUE
------------------------------------------------------------ ----------
Recursive txns delivered                                     1957
Rolled back txns seen                                        6
LCRs delivered to client                                     11908
bytes of redo processed                                      11022032
bytes paged out                                              0
seconds spent in pageout                                     0
bytes checkpointed                                           0
seconds spent in checkpoint                                  0
bytes rolled back                                            0
seconds spent in rollback                                    0
seconds system is idle                                       2509

33 rows selected.

SQL> SELECT APPLIED_SCN, LATEST_SCN, MINING_SCN, RESTART_SCN FROM V$LOGSTDBY_PROGRESS;

APPLIED_SCN LATEST_SCN MINING_SCN RESTART_SCN
----------- ---------- ---------- -----------
550947 550963 550948 550948

七：排錯

先前一直出現如下錯誤
SQL> SELECT TYPE, STATUS_CODE, STATUS FROM V$LOGSTDBY_PROCESS;

TYPE STATUS_CODE STATUS
------------------------------ -----------------------------------
COORDINATOR 1291 ORA-01291: missing logfile

SQL> SELECT * FROM V$LOGSTDBY_STATE;

PRIMARY_DBID SESSION_ID REALTIME_APPLY STATE
------------ -------------------------------------------------------
1287906064 1 Y WAITING FOR DICTIONARY LOGS

後來發現是因爲日誌文件未從主庫傳送到備庫，手動傳送並註冊後正常，傳送不成功的原因是主庫的log_archive_dest_2參數設置不正確導致
SQL> ALTER DATABASE REGISTER LOGICAL LOGFILE '/u01/arch/orcl/1_11_760125331.arc';
Database altered.

SQL> ALTER DATABASE REGISTER LOGICAL LOGFILE '/u01/arch/orcl/1_12_760125331.arc';
Database altered.

SQL> ALTER DATABASE REGISTER LOGICAL LOGFILE '/u01/arch/orcl/1_13_760125331.arc';
Database altered.

SQL> ALTER DATABASE REGISTER LOGICAL LOGFILE '/u01/arch/orcl/1_14_760125331.arc';
Database altered.

SQL> select session_id,state from v$logstdby_state;

SESSION_ID STATE
---------- -------------------
1 LOADING DICTIONARY

值得一提的是邏輯備庫會自動刪除已經應用的日誌
[oracle@orcl u01]$ tail -f /u01/app/oracle/admin/orcl/bdump/alert_orcl.log
Deleted file /u01/arch/orcl/1_13_760125331.arc
Wed Sep 28 11:53:54 2011
LOGMINER: End mining logfile: /u01/arch/orcl/1_15_760125331.arc
Wed Sep 28 11:53:54 2011
LOGMINER: Begin mining logfile: /u01/arch/orcl/1_16_760125331.arc
Wed Sep 28 11:53:55 2011
LOGMINER: End mining logfile: /u01/arch/orcl/1_16_760125331.arc
Wed Sep 28 11:53:55 2011
LOGMINER: Log Auto Delete - deleting: /u01/arch/orcl/1_14_760125331.arc
Deleted file /u01/arch/orcl/1_14_760125331.arc

SQL> select file_name,applied from DBA_LOGSTDBY_LOG ;

FILE_NAME APPLIED
-------------------------------- --------
/u01/arch/orcl/1_15_760125331.arc CURRENT
/u01/arch/orcl/1_16_760125331.arc CURRENT

八：備註

以下數據類型不被邏輯備庫支持：
NCLOB
LONG
LONG RAW
BFILE
ROWID
UROWID

檢查不被邏輯備庫支持的表：
SQL> SELECT DISTINCT OWNER,TABLE_NAME FROM DBA_LOGSTDBY_UNSUPPORTED ORDER BY OWNER,TABLE_NAME;

以下操作不被邏輯備庫支持：
ALTER DATABASE
ALTER SESSION
ALTER SNAPSHOT
ALTER SNAPSHOT LOG
ALTER SYSTEM SWITCH LOG
CREATE CONTROL FILE
CREATE DATABASE
CREATE DATABASE LINK
CREATE PFILE FROM SPFILE
CREATE SCHEMA AUTHORIZATION
CREATE SNAPSHOT
CREATE SNAPSHOT LOG
CREATE SPFILE FROM PFILE
CREATE TABLE AS SELECT FROM A CLUSTER TABLE
DROP DATABASE LINK
DROP SNAPSHOT
DROP SNAPSHOT LOG
EXPLAIN
LOCK TABLE
RENAME
SET CONSTRAINTS
SET ROLE
SET TRANSACTION

高級隊列的管理和物化視圖的刷新不被支持（DBMS_AQADM，DBMS_MVIEW_REFRESH）

Logical Standby要求每張表應該有Primary key或者Unique index，如果必須有沒有唯一性標示的表，那麼可以激活Primary庫的supplemental logging屬性，但是這樣將會在redo log中記錄該表中每一條記錄的所有字段作爲標示，會大大增加redo log的記錄量。

檢查沒有唯一性標示的表：
SQL> SELECT OWNER, TABLE_NAME,BAD_COLUMN FROM DBA_LOGSTDBY_NOT_UNIQUE WHERE TABLE_NAME NOT IN (SELECT TABLE_NAME FROM DBA_LOGSTDBY_UNSUPPORTED);

如果可以確認表中有唯一標示的字段，但是又不想真正創建主鍵或者索引來增加更新的開銷，那麼可以創建disabled RELY類型的約束，比如：
SQL> ALTER TABLE mytab ADD PRIMARY KEY (id, name) RELY DISABLE;

檢查supplemental logging是否已經被激活。
SQL> SELECT SUPPLEMENTAL_LOG_DATA_PK, SUPPLEMENTAL_LOG_DATA_UI FROM V$DATABASE;

用以下方法激活Primary庫的supplemental logging屬性：
SQL> ALTER DATABASE ADD SUPPLEMENTAL LOG DATA (PRIMARY KEY, UNIQUE INDEX) COLUMNS;

激活supplemental logging屬性以後，作一次日誌切換，保證當前日誌中不包含非追加日誌數據（nonsupplemental log）。
SQL> ALTER SYSTEM ARCHIVE LOG CURRENT;

創建一個新的表空間，用於LogMiner，否則Logical Standby需要的對象將默認創建在SYSTEM表空間中。
SQL> CREATE TABLESPACE logmnrts DATAFILE '/disk1/oracle/dbs/logmnrts.dbf' SIZE25MAUTOEXTEND ON MAXSIZE UNLIMITED;
SQL> EXECUTE DBMS_LOGMNR_D.SET_TABLESPACE('logmnrts');

無論是物理備庫還是邏輯備庫，都應該先啓動備庫，再啓動主庫；先關閉主庫，後關閉備庫！

參考文檔：
http://avdeo.com/2010/05/04/converting-physical-standby-to-logical-oracle-dataguard-10g/

構建Dataguard邏輯備庫指南

使用Python腳本進行es數據清理

使用Helm簡化K8S應用管理

使用init-containers滿足程序猿剛需

使用rancher滿足程序猿剛需

K8S問題集錦

Mac下配置sublime實現LaTeX

https://yachay.unat.edu.pe/blog/index.php?comment_area=format_blog&comment_component=blog&comment_co

linux以太網驅動總結