ARCH和LGWR进程同步DG日志的区别
ARCH和LGWR进程同步DG日志的区别
我在做Standby RAC实验时，起初使用的是ARCH传输，后来将其改为LGWR传输（实际是LGWR分出的小工进程LNS）：
--之前的设置
alter system set log_archive_dest_2='SERVICE=mynas ARCH VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=mynas';
--修改设置，可以在线修改：
alter system set log_archive_dest_2='SERVICE=mynas VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=mynas';
最直观的就是LGWR进程传输可以延迟很低，甚至基本是实时的，即使是ASYNC，另外还有一些细微的差别。
1.ARCH进程传输
主库：
SYS@jyzhao1 >select process, client_process, sequence#, status from v$managed_standby;
PROCESS CLIENT_P SEQUENCE# STATUS
--------- -------- ---------- ------------
ARCH ARCH 146 CLOSING
ARCH ARCH 137 CLOSING
ARCH ARCH 147 CLOSING
ARCH ARCH 147 CLOSING
在主库查询可以看到，只有几个归档进程。
备库：
SQL> select process, client_process, sequence#, status from v$managed_standby;
PROCESS CLIENT_P SEQUENCE# STATUS
--------- -------- ---------- ------------
ARCH ARCH 146 CLOSING
ARCH ARCH 113 CLOSING
ARCH ARCH 0 CONNECTED
ARCH ARCH 111 CLOSING
RFS ARCH 0 IDLE
RFS UNKNOWN 0 IDLE
RFS UNKNOWN 0 IDLE
RFS ARCH 0 IDLE
RFS UNKNOWN 0 IDLE
RFS UNKNOWN 0 IDLE
MRP0 N/A 147 WAIT_FOR_LOG
11 rows selected.
SQL> select * from v$standby_log;
GROUP# DBID THREAD# SEQUENCE# BYTES BLOCKSIZE USED ARC STATUS FIRST_CHANGE# FIRST_TIME NEXT_CHANGE# NEXT_TIME LAST_CHANGE# LAST_TIME
---------- ---------------------------------------- ---------- ---------- ---------- ---------- ---------- --- ---------- ------------- ------------ ------------ ------------ ------------ ------------
11 UNASSIGNED 1 0 52428800 512 0 NO UNASSIGNED
12 UNASSIGNED 1 0 52428800 512 0 NO UNASSIGNED
13 UNASSIGNED 1 0 52428800 512 0 NO UNASSIGNED
21 UNASSIGNED 2 0 52428800 512 0 NO UNASSIGNED
22 UNASSIGNED 2 0 52428800 512 0 NO UNASSIGNED
23 UNASSIGNED 2 0 52428800 512 0 NO UNASSIGNED
6 rows selected.
在备库查询可以看到，RFS中也只有ARCH进程，SRLs也都没有使用，即使指定使用using current logfile。
2.LGWR进程传输
主库：
SYS@jyzhao1 >select process, client_process, sequence#, status from v$managed_standby;
PROCESS CLIENT_P SEQUENCE# STATUS
--------- -------- ---------- ------------
ARCH ARCH 149 CLOSING
ARCH ARCH 137 CLOSING
ARCH ARCH 148 CLOSING
ARCH ARCH 149 CLOSING
LNS LNS 150 WRITING
我们发现，使用LGWR进程传输，在主库查询可以看到差异，比之前ARCH传输多了一个LNS的进程，这就是LGWR分出来的小工进程。
备库：
SQL> select process, client_process, sequence#, status from v$managed_standby;
PROCESS CLIENT_P SEQUENCE# STATUS
--------- -------- ---------- ------------
ARCH ARCH 148 CLOSING
ARCH ARCH 149 CLOSING
ARCH ARCH 0 CONNECTED
ARCH ARCH 115 CLOSING
RFS ARCH 0 IDLE
RFS UNKNOWN 0 IDLE
RFS UNKNOWN 0 IDLE
RFS ARCH 0 IDLE
RFS UNKNOWN 0 IDLE
RFS LGWR 116 IDLE
MRP0 N/A 116 APPLYING_LOG
PROCESS CLIENT_P SEQUENCE# STATUS
--------- -------- ---------- ------------
RFS LGWR 150 IDLE
RFS UNKNOWN 0 IDLE
13 rows selected.
SQL> select * from v$standby_log;
GROUP# DBID THREAD# SEQUENCE# BYTES BLOCKSIZE USED ARC STATUS FIRST_CHANGE# FIRST_TIME NEXT_CHANGE# NEXT_TIME LAST_CHANGE# LAST_TIME
---------- ---------------------------------------- ---------- ---------- ---------- ---------- ---------- --- ---------- ------------- ------------ ------------ ------------ ------------ ------------
11 UNASSIGNED 1 0 52428800 512 0 NO UNASSIGNED
12 2509089778 1 150 52428800 512 6417408 YES ACTIVE 4026243 10-AUG-17 4050942 10-AUG-17
13 UNASSIGNED 1 0 52428800 512 0 NO UNASSIGNED
21 UNASSIGNED 2 0 52428800 512 0 NO UNASSIGNED
22 2509089778 2 117 52428800 512 2767360 YES ACTIVE 4044272 10-AUG-17 4050945 10-AUG-17
23 UNASSIGNED 2 0 52428800 512 0 NO UNASSIGNED
6 rows selected.
我们发现，使用LGWR进程传输，在备库查询可以看到差异，相比之前ARCH传输RFS中只有ARCH进程而言，又多了LGWR 进程，另外SRLs也被正常使用。
alert日志上也有所区别：
**主库：**
Thu Aug 10 09:36:43 2017
ALTER SYSTEM SET log_archive_dest_2='SERVICE=mynas VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=mynas' SCOPE=BOTH;
Thread 1 advanced to log sequence 150 (LGWR switch)
Current log# 2 seq# 150 mem# 0: +DATA1/jyzhao/onlinelog/group_2.262.919999045
Current log# 2 seq# 150 mem# 1: +FRA1/jyzhao/onlinelog/group_2.258.919999049
Thu Aug 10 09:36:45 2017
Archived Log entry 287 added for thread 1 sequence 149 ID 0x958da9ee dest 1:
******************************************************************
LGWR: Setting 'active' archival for destination LOG_ARCHIVE_DEST_2
******************************************************************
ARC3: Standby redo logfile selected for thread 1 sequence 149 for destination LOG_ARCHIVE_DEST_2
LNS: Standby redo logfile selected for thread 1 sequence 150 for destination LOG_ARCHIVE_DEST_2
Thu Aug 10 09:59:59 2017
Setting Resource Manager plan DEFAULT_MAINTENANCE_PLAN via parameter
Thu Aug 10 10:00:00 2017
Starting background process VKRM
Thu Aug 10 10:00:00 2017
VKRM started with pid=44, OS id=7493
可以看到，相比之前，多了“LGWR: Setting 'active' archival for destination LOG_ARCHIVE_DEST_2”字样。
备库：
Thu Aug 10 09:36:42 2017
Archived Log entry 35 added for thread 1 sequence 148 ID 0x958da9ee dest 1:
Media Recovery Log +FRA/mynas/archivelog/2017_08_10/thread_1_seq_148.301.951644203
Media Recovery Waiting for thread 1 sequence 149
Thu Aug 10 09:36:45 2017
Primary database is in MAXIMUM PERFORMANCE mode
Thu Aug 10 09:36:46 2017
RFS[12]: Assigned to RFS process 28007
RFS[12]: Selected log 11 for thread 1 sequence 149 dbid -1785877518 branch 919999037
RFS[13]: Assigned to RFS process 28005
RFS[13]: Selected log 12 for thread 1 sequence 150 dbid -1785877518 branch 919999037
Thu Aug 10 09:36:47 2017
Archived Log entry 36 added for thread 1 sequence 149 ID 0x958da9ee dest 1:
Media Recovery Log +FRA/mynas/archivelog/2017_08_10/thread_1_seq_149.302.951644207
Media Recovery Waiting for thread 2 sequence 115
Thu Aug 10 09:36:48 2017
Primary database is in MAXIMUM PERFORMANCE mode
RFS[14]: Assigned to RFS process 28009
RFS[14]: Selected log 21 for thread 2 sequence 116 dbid -1785877518 branch 919999037
Thu Aug 10 09:36:49 2017
RFS[15]: Assigned to RFS process 28011
RFS[15]: Selected log 22 for thread 2 sequence 115 dbid -1785877518 branch 919999037
Archived Log entry 37 added for thread 2 sequence 115 ID 0x958da9ee dest 1:
Media Recovery Log +FRA/mynas/archivelog/2017_08_10/thread_2_seq_115.303.951644209
Media Recovery Waiting for thread 1 sequence 150 (in transit)
Recovery of Online Redo Log: Thread 1 Group 12 Seq 150 Reading mem 0
Mem# 0: +FRA/mynas/standbylog/standby_group_12.log
Media Recovery Waiting for thread 2 sequence 116 (in transit)
Recovery of Online Redo Log: Thread 2 Group 21 Seq 116 Reading mem 0
Mem# 0: +FRA/mynas/standbylog/standby_group_21.log
Thu Aug 10 10:02:28 2017
RFS[14]: Selected log 22 for thread 2 sequence 117 dbid -1785877518 branch 919999037
Thu Aug 10 10:02:35 2017
Media Recovery Waiting for thread 2 sequence 117 (in transit)
Thu Aug 10 10:02:35 2017
Archived Log entry 38 added for thread 2 sequence 116 ID 0x958da9ee dest 1:
Recovery of Online Redo Log: Thread 2 Group 22 Seq 117 Reading mem 0
Mem# 0: +FRA/mynas/standbylog/standby_group_22.log
可以看到多了“ (in transit)”字样。
这种LGWR传输，即便是默认的ASYNC，正常延迟也都很低，符合绝大部分场景需要：
SQL> select * from v$dataguard_stats;
NAME VALUE UNIT TIME_COMPUTED DATUM_TIME
-------------------------------- ---------------------------------------------------------------- ------------------------------ ------------------------------ ------------------------------
transport lag +00 00:00:00 day(2) to second(0) interval 08/10/2017 11:43:37 08/10/2017 11:43:35
apply lag +00 00:00:00 day(2) to second(0) interval 08/10/2017 11:43:37 08/10/2017 11:43:35
apply finish time +00 00:00:00.000 day(2) to second(3) interval 08/10/2017 11:43:37
estimated startup time 26 second 08/10/2017 11:43:37
SQL>
可以看到上面的延迟都是0。实际运维经验，一般11g ADG，不出网络等其他问题，这个延迟基本为0。
ARCH和LGWR进程同步DG日志的区别的更多相关文章
- DG日志不应用，GAP,主备切换解决思路与办法
环境ORACLE 10G OS WINDOWS 对于DG故障解决思路,DG日志切换不进行应用,DG出现GAP解决方法,DG主备库切换, 当DG出现故障时,第一时间检测alert日志,服务器OS日志,网 ...
- mysql错误日志及sql日志的区别
my.ini # power by phpStudy 2014 www.phpStudy.net 官网下载最新版 [client] port=3306 [mysql] default-characte ...
- 模拟生产搭建Standby RAC实验环境(11.2.0.4 DG)
模拟生产搭建Standby RAC实验环境(11.2.0.4 DG) 环境:RHEL 6.5 + Oracle 11.2.0.4 GI.DB 1.需求背景介绍 2.准备工作 3.主库配置 4.备库配置 ...
- Oracle 11g DG手工switchover切换标准化流程
Oracle 11g DG手工switchover切换标准化流程 环境:RHEL 6.5 + Oracle GI 11.2.0.4 + Oracle DB 11.2.0.4 Primary RAC(2 ...
- 数据库中的几个概念 - LGWR, ARCH,ASYNC,SYNC,AFFIRM
双机热备(双机容错)就是对于重要的服务,使用两台服务器,互相备份,共同执行同一服务.当一台服务器出现故障时,可以由另一台服务器承担服务任务,从而在不需要人工干预的情况下,自动保证系统能持续提供服务 双 ...
- Oracle DBWR,LGWR,CKPT,ARCH 触发条件 总结
一. DBWR写磁盘数据触发条件 1. 当进程在辅助LRU链表和主LRU链表上扫描以查找可以覆盖的buffer header[空闲缓冲区]时,如果已经扫描的buffer header的数量到达一定的 ...
- 【DG】[三思笔记]一步一步学DataGuard
[DG][三思笔记]一步一步学DataGuard 它有无数个名字,有人叫它dg,有人叫它数据卫士,有人叫它data guard,在oracle的各项特性中它有着举足轻理的地位,它就是(掌声)..... ...
- Oracle DataGuard日志传输
1. 日志传输方式 有两种日志传输方式(ARC和LGWR),第一种是采用ARC进程传输日志,其示意图如下: 注:上图来自<大话Oracle RAC> 其大致过程如下: 1)主库:日志先写入 ...
- oracle 日志学习（转载）
一,重做日志概念 重做日志文件(redo log file)对于Oracle数据库至关重要.它们是数据库的事务日志.通常只用于恢复,不过也可以用于以下工作: q 系统崩溃后的实例恢复 q 通过备份恢复 ...
随机推荐
- Chapter 5. MPEG-4 Visual
本章着重介绍有关MPEG-4 Visual标准的细节. Tool 编码工具集合的子集(比如支持交织等). Object 视频元素(比如一个矩形视频帧,或者一个任意形状的区域,静止的图像). Profi ...
- 4.vbs的循环，switch，判断等语句
1.条件判断语句 If Then Else Sub judge(x) Then MsgBox "the num is 0" Then MsgBox "the num is ...
- iOS- 解决iOS10 App启动时放大铺满App Icon的问题
0.前言 iOS10 App启动时放大铺满App图标 iPad Application shows app icon as launch screen in iOS 10 如图,点击APP后APP图标 ...
- SQL联表查询
数据库中最最常用的语法----select.简单的select语法很直白: select column from table where expression: 从((from)存储数据的地方(tab ...
- Ext & Java 上存图片 Demo
Ext & Java 上存图片 Demo Ext <html> <head> <script id="microloader" type=&q ...
- C#窗体多语言切换（简繁）
多窗体最好继承一个父窗体,在父窗体Load事件中执行此方法 添加引用 using Microsoft.VisualBasic; #region 语言切换 /// <summary> /// ...
- man ssh翻译(ssh命令中文手册)
本文为命令ssh的man文档翻译,翻译了90%的内容,剩余是一些没必要翻译的东西,请见谅. 如此文有所疑惑,希望我的另一篇文章能解惑: SSH(1) BSD Ge ...
- PostgreSQL 使用 PreparedStatement 导致查询慢的分析
实验环境: DB is PostgreSQL version 8.2.15 JDK1.8 测试一 使用JDBC查询一个SQL: public static void test1(String url, ...
- English - Green Peanut Butter
There is a guy. He wants to drink 12 cups of green peanut butter. He needs green peanut butter. So h ...
- OpenStack（企业私有云）万里长征第五步——虚拟机Migrate&Resize
一.前言 上一篇文章讲了OpenStack的部署和简单操作,今天介绍一下如何实现虚拟机的Migrate以及Resize.Migrate操作和Resize操作基本上属于同一种操作,Migrate操作只是 ...