Oracle DG测试failover和后续恢复报告
Oracle DG测试failover和后续恢复报告
一、概述
二、验证过程：
三、结论
一、概述
本文是针对在DG灾备环境进行failover操作以及后续恢复的报告。
**我这里的测试环境是：**
数据库版本：Oracle 11.2.0.4
Site A：主库 db_unique_name=jyzhao
Site B：备库（实时应用）db_unique_name=mynas
Site C：备库（延迟1小时应用）db_unique_name=jyzhao_s
以下章节涉及到的简称注释：
A库 => Site A：主库
B库 => Site B：备库（实时应用）
C库 => Site C：备库（延迟1小时应用）
验证：
当A库crash后，在B库进行failover将B切换为新的主库，确认failover之后，A库和C库应该如何处理才可以成为新的备库继续使用？是否需要重建？重建的话，是否需要重新备份来恢复，以前的备份是否可以用来创建备库？
二、验证过程：
2.1 A库异常关闭
A库：
SQL> shutdown abort
2.2 B库进行failover切换为新主库
failover 标准步骤如下：
#取消DG应用
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE CANCEL;
#重启下数据库（建议）
shutdown immediate;
startup
#操作不可逆，确定实际情况需要failover
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE FINISH force;
SELECT OPEN_MODE, DATABASE_ROLE, SWITCHOVER_STATUS, FORCE_LOGGING, DATAGUARD_BROKER, GUARD_STATUS FROM V$DATABASE;
#尝试常规切换为主库
ALTER DATABASE COMMIT TO SWITCHOVER TO PRIMARY WITH SESSION SHUTDOWN;
如果这一步的常规切换失败，提示需要介质恢复，那么：
1）恢复备库 recover standby database until cancel;
2）激活备库 alter database activate standby database;
#最后重新启动数据库
shutdown immediate;
startup
查看此时B库的信息：
SQL> select name, database_role, open_mode from gv$database;
NAME DATABASE_ROLE OPEN_MODE
--------- ---------------- --------------------
JYZHAO PRIMARY READ WRITE
SQL> select * from v$log;
GROUP# THREAD# SEQUENCE# BYTES BLOCKSIZE MEMBERS ARC STATUS FIRST_CHANGE# FIRST_TIME NEXT_CHANGE# NEXT_TIME
---------- ---------- ---------- ---------- ---------- ---------- --- ---------------- ------------- ------------ ------------ ------------
1 1 3 52428800 512 2 NO CURRENT 17882720 03-SEP-17 2.8147E+14
2 1 2 52428800 512 2 YES ACTIVE 17882488 03-SEP-17 17882720 03-SEP-17
3 2 0 52428800 512 2 YES UNUSED 0 0
4 2 0 52428800 512 2 YES UNUSED 0 0
可以看到，目前B库已成为新的主库，redo日志的sequence重新开始。
2.3 要求C库成为新主库的备库
现在要求C库成为新主库的备库。是否需要重建C库呢？答案是不需要。下面具体来看下验证过程。
**C库的alert日志：**
Sun Sep 03 14:09:58 2017
Archived Log entry 9 added for thread 1 sequence 1227 ID 0x97764e10 dest 1:
ARC2: Archive log thread 1 sequence 1227 available in 60 minute(s)
Sun Sep 03 14:09:58 2017
RFS[3]: Selected log 11 for thread 1 sequence 1228 dbid -1785877518 branch 919999037
Sun Sep 03 14:19:24 2017
RFS[3]: Possible network disconnect with primary database
Sun Sep 03 14:19:24 2017
RFS[2]: Possible network disconnect with primary database
Sun Sep 03 14:19:24 2017
RFS[4]: Assigned to RFS process 10190
RFS[4]: Possible network disconnect with primary database
可以看到，在A库crash之后，C库收到网络无法连接到A库的告警，说明C库目前没有新的操作。
接下来想要C库成为B库（新主库）的备库，就需要尝试在B库上配置DG参数，使得B库的归档可以传输到C库。
show parameter log_archive_config
show parameter log_archive_dest_3
alter system set log_archive_config = 'DG_CONFIG=(jyzhao,mynas,jyzhao_s)';
alter system set log_archive_dest_3 = 'SERVICE=jyzhao_s LGWR ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) db_unique_name=jyzhao_s';
同时在B库的tnsnames.ora文件中增加到C库的连接：
#Standby Single Instance
JYZHAO_S =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.1.111)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = jyzhao_s)
)
)
在B库设置完成后，观察B库的告警：
Sun Sep 03 14:37:41 2017
ALTER SYSTEM SET log_archive_dest_3='SERVICE=jyzhao_s LGWR ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) db_unique_name=jyzhao_s' SCOPE=BOTH;
ARC3: Standby redo logfile selected for thread 1 sequence 3 for destination LOG_ARCHIVE_DEST_3
Thread 1 cannot allocate new log, sequence 5
Checkpoint not complete
Current log# 2 seq# 4 mem# 0: +DATA/mynas/onlinelog/group_2.278.953484865
Current log# 2 seq# 4 mem# 1: +FRA/mynas/onlinelog/group_2.714.953484869
Thread 1 advanced to log sequence 5 (LGWR switch)
Current log# 1 seq# 5 mem# 0: +DATA/mynas/onlinelog/group_1.277.953484853
Current log# 1 seq# 5 mem# 1: +FRA/mynas/onlinelog/group_1.720.953484859
******************************************************************
LGWR: Setting 'active' archival for destination LOG_ARCHIVE_DEST_3
******************************************************************
LNS: Standby redo logfile selected for thread 1 sequence 5 for destination LOG_ARCHIVE_DEST_3
Archived Log entry 382 added for thread 1 sequence 4 ID 0x978ff56d dest 1:
Sun Sep 03 14:37:55 2017
ARC3: Standby redo logfile selected for thread 1 sequence 4 for destination LOG_ARCHIVE_DEST_3
Sun Sep 03 14:39:49 2017
Thread 1 advanced to log sequence 6 (LGWR switch)
Current log# 2 seq# 6 mem# 0: +DATA/mynas/onlinelog/group_2.278.953484865
Current log# 2 seq# 6 mem# 1: +FRA/mynas/onlinelog/group_2.714.953484869
Sun Sep 03 14:39:49 2017
LNS: Standby redo logfile selected for thread 1 sequence 6 for destination LOG_ARCHIVE_DEST_3
Sun Sep 03 14:39:49 2017
Archived Log entry 388 added for thread 1 sequence 5 ID 0x978ff56d dest 1:
然后返回C库操作，将C库开启实时日志应用：
SQL> alter database recover managed standby database cancel;
Database altered.
SQL> alter database recover managed standby database using current logfile disconnect from session;
Database altered.
此时再观察C库的告警日志：
Sun Sep 03 14:38:10 2017
Clearing online log 11 of thread 1 sequence number 0
Sun Sep 03 14:38:19 2017
MRP0: Incarnation has changed! Retry recovery...
RFS[7]: Selected log 11 for thread 1 sequence 4 dbid -1785877518 branch 953735009
Errors in file /u01/app/oracle/diag/rdbms/jyzhao_s/jyzhao/trace/jyzhao_pr00_10161.trc:
ORA-19906: recovery target incarnation changed during recovery
Recovery interrupted!
Sun Sep 03 14:38:19 2017
Archived Log entry 11 added for thread 1 sequence 4 ID 0x978ff56d dest 1:
Sun Sep 03 14:38:20 2017
started logmerger process
Sun Sep 03 14:38:20 2017
Managed Standby Recovery not using Real Time Apply
Parallel Media Recovery started with 2 slaves
Media Recovery start incarnation depth : 1, target inc# : 3, irscn : 17882484
Waiting for all non-current ORLs to be archived...
All non-current ORLs have been archived.
Media Recovery Delayed for 60 minute(s) (thread 1 sequence 1226)
Sun Sep 03 14:38:31 2017
RFS[8]: Assigned to RFS process 10704
RFS[8]: Opened log for thread 1 sequence 1228 dbid -1785877518 branch 919999037
Archived Log entry 12 added for thread 1 sequence 1228 rlc 919999037 ID 0x97764e10 dest 3:
RFS[8]: Opened log for thread 1 sequence 1 dbid -1785877518 branch 953735009
Sun Sep 03 14:38:36 2017
RFS[7]: Opened log for thread 1 sequence 2 dbid -1785877518 branch 953735009
Archived Log entry 13 added for thread 1 sequence 1 rlc 953735009 ID 0x978ff56d dest 3:
Archived Log entry 14 added for thread 1 sequence 2 rlc 953735009 ID 0x978ff56d dest 3:
Sun Sep 03 14:40:13 2017
Archived Log entry 15 added for thread 1 sequence 5 ID 0x978ff56d dest 1:
Sun Sep 03 14:40:13 2017
RFS[6]: Selected log 11 for thread 1 sequence 6 dbid -1785877518 branch 953735009
Sun Sep 03 14:40:37 2017
alter database recover managed standby database cancel
Sun Sep 03 14:40:38 2017
MRP0: Background Media Recovery cancelled with status 16037
Errors in file /u01/app/oracle/diag/rdbms/jyzhao_s/jyzhao/trace/jyzhao_pr00_10688.trc:
ORA-16037: user requested cancel of managed recovery operation
Recovery interrupted!
Sun Sep 03 14:40:38 2017
MRP0: Background Media Recovery process shutdown (jyzhao)
Managed Standby Recovery Canceled (jyzhao)
Completed: alter database recover managed standby database cancel
Sun Sep 03 14:40:53 2017
alter database recover managed standby database using current logfile disconnect from session
Attempt to start background Managed Standby Recovery process (jyzhao)
Sun Sep 03 14:40:53 2017
MRP0 started with pid=20, OS id=10747
MRP0: Background Managed Standby Recovery process started (jyzhao)
started logmerger process
Sun Sep 03 14:40:58 2017
Managed Standby Recovery starting Real Time Apply
Parallel Media Recovery started with 2 slaves
Media Recovery start incarnation depth : 1, target inc# : 3, irscn : 17882484
Waiting for all non-current ORLs to be archived...
All non-current ORLs have been archived.
Managed Standby Recovery started with USING CURRENT LOGFILE
Ignoring previously specified DELAY 60 minutes for thread 1 sequence 1226
Media Recovery Log /u01/oradata/JYZHAO_S/archivelog/2017_09_03/o1_mf_1_1226_dtq74r91_.arc
Managed Standby Recovery started with USING CURRENT LOGFILE
Ignoring previously specified DELAY 60 minutes for thread 1 sequence 1227
Media Recovery Log /u01/oradata/JYZHAO_S/archivelog/2017_09_03/o1_mf_1_1227_dtq75p6j_.arc
Media Recovery Log /u01/oradata/JYZHAO_S/archivelog/2017_09_03/o1_mf_1_1228_dtq8v7c7_.arc
Identified End-Of-Redo (failover) for thread 1 sequence 1228 at SCN 0x0.110dd74
Completed: alter database recover managed standby database using current logfile disconnect from session
Resetting standby activation ID 2541112848 (0x97764e10)
Media Recovery End-Of-Redo indicator encountered
Media Recovery Continuing
Media Recovery Log /u01/oradata/JYZHAO_S/archivelog/2017_09_03/o1_mf_1_1_dtq8vdpr_.arc
Media Recovery Log /u01/oradata/JYZHAO_S/archivelog/2017_09_03/o1_mf_1_2_dtq8vdpy_.arc
Media Recovery Log /u01/oradata/JYZHAO_S/archivelog/2017_09_03/o1_mf_1_3_dtq8tgvl_.arc
Sun Sep 03 14:41:11 2017
Media Recovery Log /u01/oradata/JYZHAO_S/archivelog/2017_09_03/o1_mf_1_4_dtq8tvy5_.arc
Media Recovery Log /u01/oradata/JYZHAO_S/archivelog/2017_09_03/o1_mf_1_5_dtq8yfr5_.arc
Media Recovery Waiting for thread 1 sequence 6 (in transit)
Recovery of Online Redo Log: Thread 1 Group 11 Seq 6 Reading mem 0
Mem# 0: /u01/oradata/standbylog/standby_group_11.log
实际看到，C库已经可以正常应用日志。说明C库不需要重建即可通过简单配置成为新主库B库的新备库。
2.4 要求A库成为新主库的备库
此时A库启动的话，是一个独立运行的数据库，如果想将A库也设置为主库的话，那么，通过新主库的最新备份肯定是可行的，但是如果数据量很大，之前A库自己本身有历史的备份，能否不再耗时备份新主库，直接通过历史的备份恢复呢？其实这个从上面的C库不再需要重建直接成为新主库的备库，也可以推断出，是可以的。只需要确认这个备份是在failover之前完成的。下面我们来具体实验验证下可行性。
在B库创建新的备库控制文件，并传输到A库相同路径下：
backup current controlfile for standby format '/tmp/std_control02.ctl';
在A库启动到nomount，恢复新的备库控制文件
restore standby controlfile from '/tmp/std_control02.ctl';
在A库查看数据文件头的检查点，确认是在failover之前：
SYS@jyzhao1 >select name from v$datafile;
NAME
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
+DATA1/jyzhao/datafile/system.258.951608183
+DATA1/jyzhao/datafile/sysaux.257.951608183
+DATA1/jyzhao/datafile/undotbs1.259.951608185
+DATA1/jyzhao/datafile/users.265.951608205
+DATA1/jyzhao/datafile/undotbs2.261.951608185
+DATA1/jyzhao/datafile/dbs_d_jingyu.262.951608185
+DATA1/jyzhao/datafile/dbs_i_jingyu.263.951608185
+DATA1/jyzhao/datafile/test.264.951608185
+DATA1/jyzhao/datafile/test2.260.951608185
+DATA1/jyzhao/datafile/dbadata.276.952933931
10 rows selected.
SYS@jyzhao1 >select checkpoint_change# from v$database;
CHECKPOINT_CHANGE#
------------------
17892035
SYS@jyzhao1 >select checkpoint_change# from v$datafile;
CHECKPOINT_CHANGE#
------------------
17892035
17892035
17892035
17892035
17892035
17892035
17892035
17892035
17892035
17892035
10 rows selected.
SYS@jyzhao1 >select checkpoint_change# from v$datafile_header;
CHECKPOINT_CHANGE#
------------------
0
0
0
0
0
0
0
0
0
0
10 rows selected.
上面这个数据文件头的检查点是0，说明数据文件没有正确获取到，实际上是由于OMF的名字有变化，直接将数据文件路径catalog到备份集中，再switch即可。
catalog start with '+DATA1/jyzhao/datafile/';
switch database to copy;
再次查询：
SYS@jyzhao1 >select current_scn||'' from v$database;
CURRENT_SCN||''
----------------------------------------
17892467
SYS@jyzhao1 >select checkpoint_change# from v$database;
CHECKPOINT_CHANGE#
------------------
17892035
SYS@jyzhao1 >select checkpoint_change# from v$datafile;
CHECKPOINT_CHANGE#
------------------
17892035
17892035
17892035
17892035
17892035
17892035
17892035
17892035
17892035
17892035
10 rows selected.
SYS@jyzhao1 >select checkpoint_change# from v$datafile_header;
CHECKPOINT_CHANGE#
------------------
17882484
17882484
17882484
17882484
17882484
17882484
17882484
17882484
17882484
17882484
10 rows selected.
此时在mount状态下开启日志应用：
alter database recover managed standby database disconnect from session;
从告警日志观察，确认应用到最新时，取消日志应用：
alter database recover managed standby database cancel;
打开数据库，开启实时应用：
alter database recover managed standby database USING CURRENT LOGFILE disconnect from session;
最终查询可以正常实时应用。
三、结论
一般来说，在A库crash之后，B库failover成为新的主库，那么原来设置为延迟1小时应用的C 库是可以直接配置成为新主库的备库。A库修复后，也可以通过failover之前的现有备份集来恢复到failover之前的状态，而不需要在新主库重新去备份。
Oracle DG测试failover和后续恢复报告的更多相关文章
- oracle dg failover灾难切换
oracle dg failover灾难切换SQL> alter database recover managed standby database finish force;SQL> a ...
- RHEL6.4 + Oracle 11g DG测试环境快速搭建参考
环境现状: 两台虚拟主机A和B: 1. A机器已安装ASM存储的Oracle 11g 实例 参考:http://www.cnblogs.com/jyzhao/p/4332410.html 2 ...
- 【DATAGUARD】物理dg的failover切换(六)
[DATAGUARD]物理dg的failover切换(六) 一.1 BLOG文档结构图 一.2 前言部分 一.2.1 导读 各位技术爱好者,看完本文后,你可以掌握如下的技能,也可以学到一些其它你 ...
- oracle DG 主备切换语句整理
今日花了一下午时间进行了Oracle DataGuard的切换练习,参考了网上好多文章,最后将一些语句进行摘录,以备以后查询使用.之后有时间会带来Oracle DG的搭建和切换全过程文章. DataG ...
- 某控股公司OA系统ORACLE DG搭建
*此处安装ORACLE DATAGUARD是利用ORACLE RMAN DUPLICATE方式安装.*可以搭建好ORACLE DG再来impdp生产数据,也可以先导入主库数据再来做DG*注意看下面的配 ...
- oracle DG搭建
Oracle DG 搭建1. 环境 OS IP hostname db_name DB_UNIQUE_NAME主库 RHEL 5.4 192.168.12.20 edgzrip1.oracle.com ...
- oracle dg状态检查及相关命令
oracle dg 状态检查 先检查备库的归档日志同步情况 SELECT NAME,applied FROM v$archived_log; alter database recover manage ...
- ORACLE DG 库参数db_file_name_convert和log_file_name_convert的作用
https://www.cnblogs.com/xqzt/p/5089826.html ORACLE DG 库参数db_file_name_convert和log_file_name_convert的 ...
- oracle DG查看延时时间
oracle DG查看延时时间 SQL> select value from v$dataguard_stats where name='apply lag'; 例如: SQL> sele ...
随机推荐
- oracle分组-神奇的cube和rollup
先看代码: 表结构如下: emp表 EMPNO NOT NULL NUMBER(4) ENAME ...
- Ubuntu彻底删除/卸载mysql,php,apache
一.卸载删除 mysql 1 sudo apt-get autoremove --purge mysql-server-5.02 sudo apt-get remove mysql-server3 s ...
- JAVA 笔记 ClassLoader.getResourceAsStream() 与 Class.getResourceAsStream()的区别
Class.getResourceAsStream() 会指定要加载的资源路径与当前类所在包的路径一致. 例如你写了一个MyTest类在包com.test.mycode 下,那么MyTest.c ...
- R语言成功加载rJava方法
加载rJava的同时,要下载JAVA的JRE并且配置环境变量JAVA_HOME,因为rJava的调用需要java运行环境. 1.下载JRE 64位版本的JRE官网下载:http://www.java. ...
- [Usaco2007 Open]Fliptile 翻格子游戏
[Usaco2007 Open]Fliptile 翻格子游戏 题目 Farmer John knows that an intellectually satisfied cow is a happy ...
- 5.npm scripts 使用指南
简单介绍 scripts里面的 "start": "node app" npm run start 相当于 node app { "name" ...
- CODE大全告诉你java是否开始没落了
CODE大全告诉你java是否开始没落了! 22 岁,对于一个技术人来说可谓正当壮年.但对于一门编程语言来说,情况可能又有不同.各类编程语言横空出世,纷战不休,然而 TIOBE 的语言排行榜上,Jav ...
- AndroidTv开发中常用的adb命令
盒子应用开发时,调试比手机上的开发比较麻烦一点,而且需要经常跟 adb 打交道,不管是 wifi 连接调试,还是应用删除安装等.这里记录一些常用的操作,方便查阅. adb wifi连接调试 方法一:需 ...
- Cocoapods 应用第二部分-私有库相关
我们在这里,使用的是 第一部分使用pod lib create YohunlUtilsPod 创建的framework工程来说明.其创建过程在此就不重复了,当然你也可以下载我已经创建好的demo ht ...
- python实战第一天-pymysql模块并练习
操作系统 Ubuntu 15.10 IDE & editor JetBrains PyCharm 5.0.2 ipython3 Python版本 python-3.4.3 安装pymysql模 ...