GoldenGate<一> step by step installation and configuration
最近研究GoldenGate，此文記錄入門第一步--如何安裝。這裡的作業系統為Solaris 10 64bit。
大家可以先從這裡下載各種文件：
1. 根據作業系統和資料庫版本下載相應GG版本。下載地址：
由於我的機器是Solaris 10 64bit的，所以下載的檔案為Oracle GoldenGate v10.4.0.x for Oracle 10g 64bit on Solaris 10 (53 MB)。
到相應目錄解壓。
2. 設定環境變數PATH和LD_LIBRARY_PATH
我在c shell下建立一個檔案.gg，執行source .gg即可。
setenv PATH
setenv LD_LIBRARY_PATH
3. 目錄安裝
在GG的安裝目錄下，執行如下命令：
./ggsci
Oracle GoldenGate Command Interpreter for Oracle
Version 10.4.0.19 Build 002
Solaris, sparc, 64bit (optimized), Oracle 10 on Sep 23 2009 15:19:42
Copyright (C) 1995, 2009, Oracle and/or its affiliates. All rights reserved.
GGSCI (xxx) 1> CREATE SUBDIRS
Creating subdirectories under current directory /xxx
Parameter files /xxx/dirprm: created
Report files /xxx/dirrpt: created
Checkpoint files /xxx/dirchk: created
Process status files /xxx/dirpcs: created
SQL script. files /xxx/dirsql: created
Database definitions files /xxx/dirdef: created
Extract data files /xxx/dirdat: created
Temporary files /xxx/dirtmp: created
Veridata files /xxx/dirver: created
Veridata Lock files /xxx/dirver/lock: created
Veridata Out-Of-Sync files /xxx/dirver/oos: created
Veridata Out-Of-Sync XML files /xxx/dirver/oosxml: created
Veridata Parameter files /xxx/dirver/params: created
Veridata Report files /xxx/dirver/report: created
Veridata Status files /xxx/dirver/status: created
Veridata Trace files /xxx/dirver/trace: created
Stdout files /xxx/dirout: created
4. DDL support安裝
建立一個專為GGS DDL安裝的Oracle user，並賦予相應許可權：
SQL> create user ggs identified by ggs;
User created.
SQL> grant connect,resource to ggs;
Grant succeeded.
SQL> GRANT EXECUTE ON UTL_FILE TO ggs;
Grant succeeded.
SQL> GRANT SELECT ANY DICTIONARY TO ggs;
Grant succeeded.
SQL> GRANT SELECT ANY TABLE TO ggs;
Grant succeeded.
由於GGS_DDL_HIST和GGS_MARKER會持續增長，所以最好單獨建立一個tablespace給他們：
SQL> alter user ggs default tablespace GGSTBS;
User altered.
SQL> alter user ggs quota unlimited on GGSTBS;
User altered.
5. 開啟Oracle supplemental log
SQL> select supplemental_log_data_min from v$database;
SUPPLEME
--------
NO
SQL> alter database add supplemental log data;
Database altered.
6. 啟動Manager
在GGS安裝目錄下：
> ggsci
Oracle GoldenGate Command Interpreter for Oracle
Version 10.4.0.19 Build 002
Solaris, sparc, 64bit (optimized), Oracle 10 on Sep 23 2009 15:19:42
Copyright (C) 1995, 2009, Oracle and/or its affiliates. All rights reserved.
GGSCI (xxx) 1> EDIT PARAMS MGR
PORT 7809
~
"dirprm/mgr.prm" [New File] 1 line, 10 characters written
GGSCI (xxx) 2> start mgr
Manager started.
GGSCI (qadb120) 3> info all
Program Status Group Lag Time Since Chkpt
MANAGER RUNNING
注：以上步驟需要在Source和Target都執行。
7. 在Source新增Extract Group
首先登入Source的Schema，新增EXTRACT程式和放在Target的TRAIL檔案地址
GGSCI (xxx) 1> DBLOGIN USERID , PASSWORD GGS
Successfully logged into database.
GGSCI (xxx) 2> add EXTRACT HAOEXT, tranlog,begin now
EXTRACT added.
GGSCI (xxx) 3> add RMTTRAIL ./dirdat/zh, extract HAOEXT, megabytes 50
RMTTRAIL added.
GGSCI (xxx) 4> info EXTRACT HAOEXT
EXTRACT HAOEXT Initialized 2010-04-21 09:43 Status STOPPED
Checkpoint Lag 00:00:00 (updated 00:00:23 ago)
Log Read Checkpoint Oracle Redo Logs
2010-04-21 09:43:39 Seqno 0, RBA 0
新增EXTRACT的parameter file，然後啟動EXTRACT
GGSCI (xxx) 9> edit param HAOEXT
EXTRACT HAOEXT
SETENV (ORACLE_SID=MOT)
USERID , PASSWORD GGS
RMTHOST xxx.xxx.xxx.xxx, MGRPORT 7809
RMTTRAIL ./dirdat/zh
TABLE HAOZHU_USER.GG1;
~
"dirprm/haoext.prm" [New File] 6 lines, 161 characters written
GGSCI (xxx) 10> start EXTRACT HAOEXT
Sending START request to MANAGER ...
EXTRACT HAOEXT starting
GGSCI (xxx) 12> info EXTRACT HAOEXT
EXTRACT HAOEXT Last Started 2010-04-21 10:19 Status RUNNING
Checkpoint Lag 00:00:00 (updated 00:00:02 ago)
Log Read Checkpoint Oracle Redo Logs
2010-04-21 10:20:01 Seqno 59, RBA 157803008
8. 在Target端新增Replicat Group
先新增引數
GGSCI (xxx) 1> edit params mgr
PORT 7809
~
"dirprm/mgr.prm" 2 lines, 27 characters written
再新增REPLICAT Group
GGSCI (xxx) 5> add replicat HAOREP, EXTTRAIL ./dirdat/zh, nodbcheckpoint
REPLICAT added.
GGSCI (xxx) 6> edit param HAOREP
REPLICAT HAOREP
ASSUMETARGETDEFS
SETENV (ORACLE_SID=DC1)
USERID GGS, PASSWORD GGS
map HAOZHU_USER.GG1 , target HAOZHU_USER.GG2;
~
"dirprm/haorep.prm" 5 lines, 128 characters written
GGSCI (xxx) 7> start REPLICAT HAOREP
Sending START request to MANAGER ...
REPLICAT HAOREP starting
GGSCI (qadb121) 8> info all
Program Status Group Lag Time Since Chkpt
MANAGER RUNNING
REPLICAT RUNNING HAOREP 00:00:00 00:00:01
注意：在EDIT PARAMS 的時候，一定要加上最後的逗號，否則會報如下錯，我就在這搞了很久。。
2010-04-21 12:30:45 GGS ERROR 101 Oracle GoldenGate Delivery for Oracle, haorep.prm: Parameter unterminated.
2010-04-21 12:30:45 GGS ERROR 190 Oracle GoldenGate Delivery for Oracle, haorep.prm: PROCESS ABENDING.
9. 測試複製
首先在Source檢測EXTRACT沒有任何資訊：
GGSCI (xxx) 5> stats HAOEXT
Sending STATS request to EXTRACT HAOEXT ...
No active extraction maps.
在Source插入一行但不commit：
SQL> insert into gg1(OBJECT_ID,OBJECT_NAME) values(20122012,'test by hao');
1 row created.
這時，如預期，沒有commit的資訊不會被抓取：
GGSCI (xxx) 6> stats HAOEXT
Sending STATS request to EXTRACT HAOEXT ...
No active extraction maps.
但commit之後：
SQL> commit;
Commit complete.
就有EXTRACT的資料了：
GGSCI (xxx) 7> stats HAOEXT
Sending STATS request to EXTRACT HAOEXT ...
Start of Statistics at 2010-04-21 12:58:23.
Output to ./dirdat/zh:
Extracting from HAOZHU_USER.GG1 to HAOZHU_USER.GG1:
*** Total statistics since 2010-04-21 12:58:07 ***
Total inserts 1.00
Total updates 0.00
Total deletes 0.00
Total discards 0.00
Total operations 1.00
*** Daily statistics since 2010-04-21 12:58:07 ***
Total inserts 1.00
Total updates 0.00
Total deletes 0.00
Total discards 0.00
Total operations 1.00
*** Hourly statistics since 2010-04-21 12:58:07 ***
Total inserts 1.00
Total updates 0.00
Total deletes 0.00
Total discards 0.00
Total operations 1.00
*** Latest statistics since 2010-04-21 12:58:07 ***
Total inserts 1.00
Total updates 0.00
Total deletes 0.00
Total discards 0.00
Total operations 1.00
End of Statistics.
REPLICAT也有相應資料：
GGSCI (xxx) 1> stats haorep
Sending STATS request to REPLICAT HAOREP ...
Start of Statistics at 2010-04-21 13:05:51.
Replicating from HAOZHU_USER.GG1 to HAOZHU_USER.GG2:
*** Total statistics since 2010-04-21 12:58:10 ***
Total inserts 1.00
Total updates 0.00
Total deletes 0.00
Total discards 0.00
Total operations 1.00
*** Daily statistics since 2010-04-21 12:58:10 ***
Total inserts 1.00
Total updates 0.00
Total deletes 0.00
Total discards 0.00
Total operations 1.00
*** Hourly statistics since 2010-04-21 13:00:00 ***
No database operations have been performed.
*** Latest statistics since 2010-04-21 12:58:10 ***
Total inserts 1.00
Total updates 0.00
Total deletes 0.00
Total discards 0.00
Total operations 1.00
End of Statistics.
複製成功：
SQL> select OBJECT_ID,OBJECT_NAME from gg2 where OBJECT_ID=20122012;
OBJECT_ID
----------
OBJECT_NAME
--------------------------------------------------------------------------------
20122012
test by hao
【由此想到的問題：】
對於這種最簡單的配置，是由Source端的EXTRACT程式直接寫Target端的trail檔案，但如果網路出現問題，就會對Source端產生問題。
我實驗在Target端stop replicat and mgr,發現Traget的trail檔案仍被寫入。
> fuser zh000000
zh000000: 11482o
> ptree 11482
18089 zsched
11482 ./server -p 7840 -k -l /xxx/ggserr.log
於是kill -9 11482。
這時我再在Source端產生一點dml，結果發現此時Source端的EXTRACT程式死了。
GGSCI (xxx) 14> info all
Program Status Group Lag Time Since Chkpt
MANAGER RUNNING
EXTRACT ABENDED HAOEXT 00:00:00 00:02:18
然後在ggserr.log中我們發現有146 error：
2010-04-22 03:48:54 GGS WARNING 150 Oracle GoldenGate Capture for Oracle, haoext.prm: TCP/IP error 146 (Connection refused).
如何恢復？
首先在Target start mgr and replicat，然後在Source start extract。一切恢復正常。
但你會發現stats命令重新計數了：
GGSCI (xxx) 20> stats HAOEXT
Sending STATS request to EXTRACT HAOEXT ...
Start of Statistics at 2010-04-22 03:54:01.
Output to ./dirdat/zh:
Extracting from HAOZHU_USER.GG1 to HAOZHU_USER.GG1:
*** Total statistics since 2010-04-22 03:53:47 ***
Total inserts 1.00
Total updates 0.00
Total deletes 0.00
Total discards 0.00
Total operations 1.00
*** Daily statistics since 2010-04-22 03:53:47 ***
Total inserts 1.00
Total updates 0.00
Total deletes 0.00
Total discards 0.00
Total operations 1.00
*** Hourly statistics since 2010-04-22 03:53:47 ***
Total inserts 1.00
Total updates 0.00
Total deletes 0.00
Total discards 0.00
Total operations 1.00
*** Latest statistics since 2010-04-22 03:53:47 ***
Total inserts 1.00
Total updates 0.00
Total deletes 0.00
Total discards 0.00
Total operations 1.00
End of Statistics.
所以我個人比較偏向於再在Source端配置一個data pump程式，用來專門負責網路傳輸。
而由EXTRACT程式抓出來的資訊，先存放在本地的trail檔案裡。
這樣，就和shareplex的capture queue是異曲同工了。
待續。。
來自 “ ITPUB部落格 ” ，連結：http://blog.itpub.net/15415488/viewspace-659749/，如需轉載，請註明出處，否則將追究法律責任。
相關文章
- Step by Step Data Replication Using Oracle GoldenGateOracleGo
- Step-By-Step Installation of 9i RAC on IBM AIXIBMAI
- Step-By-Step Installation of RAC with RAW Datafiles on Windows 2000Windows
- Step-to-Step Installation of 10G RAC on RedHat AS 3.0 – Single Node(二）Redhat
- Step-to-Step Installation of 10G RAC on RedHat AS 3.0 – Single Node（qiangtang）Redhat
- React Step by StepReact
- oracle10g simpe AQ step by step(一)Oracle
- Command 模式 Step by Step模式
- BAPI Step by step GuidanceAPIGUI
- Step by Step TimesTen --- ttIsqlSQL
- Promise的實現(step by step)Promise
- Learn c++ step by step (轉)C++
- 使用RMAN備份集搭建Oracle Dataguard Step by Step（一）Oracle
- Oracle 11gR2 Active DataGuard配置Step By Step（一）Oracle
- Step by Step TimesTen --- DataStore的雙向複製( 一)AST
- Linux Software RAID step by stepLinuxAI
- Git Step by Step (3)：Git物件模型Git物件模型
- Oracle高階複製Step by StepOracle
- 安裝linux(step by step)(轉)Linux
- Learn C++ step by step(2) (轉)C++
- 單步除錯 step into/step out/step over 區別詳解除錯
- 轉載一個step by step change public-ip and vip on RAC
- ABP應用開發（Step by Step）-下篇
- ABP應用開發（Step by Step）-上篇
- TIDB DM資料同步step by stepTiDB
- Git Step by Step (4)：探索.git目錄Git
- ClearCase使用入門--step by step（序） (轉)
- Oracle 12c GI/RAC Step-by-Step安裝指南(一)Oracle
- Step by Step, 為OSRFX2建立一個KMDF驅動程式
- 實時 Linux 抖動分析 Step by stepLinux
- Git Step by Step (6)：Git遠端倉庫Git
- STEP BY STEP INSTALL SSH ON AIX5.3(6.1)AI
- install 11G ASM on RedHat step by stepASMRedhat
- oracle10g simpe AQ step by step(二)Oracle
- linux中配置NFS服務step by stepLinuxNFS
- step by step install netbackup client 6.5 on aixclientAI
- Oracle 10g R2建立ASM例項Step By Step（一）Oracle 10gASM
- STREAMS筆記(1) step by step 建立一個Streams複製環境筆記