Golden Gate配置实例：RHEL 4.7下的Oracle 10g RAC到单实例的单向同步

环境

源端是一个RAC
rac1 rac2: RHEL 4u7 + Oracle 10.2.0.4 + ASM
rac1 外网IP 192.168.47.241 虚拟IP 192.168.47.243
rac2 外网IP 192.168.47.242 虚拟IP 192.168.47.244

目标端是一个单机
ggdb: RHEL 4u7 + ORACLE 10.2.0.4 + ASM
IP: 192.168.47.211

两台主机均已创建数据库，sid分别为oradb 和 dyggdb

配置oradb 到 dyggdb的数据同步

goldengate版本11.1.1.0

时间同步

在RAC中，节点间的时间同步很重要，官方文档里面的解释是，GoldenGate通过时间来做一些关键决策。这里我们通过NTP来配置时间同步
将rac1配置为NTP服务器

rac1:

修改配置文件。配置一个server指向自己

cat>/etc/ntp.conf<restrict default nomodify notrap noquery
restrict 127.0.0.1

server 0.rhel.pool.ntp.org
server 1.rhel.pool.ntp.org
server 2.rhel.pool.ntp.org

server 192.168.47.241
server 127.127.1.0
fudge 127.127.1.0 stratum 10

driftfile /var/lib/ntp/drift
broadcastdelay 0.008
keys /etc/ntp/keys
EOF

启动ntpd服务

service ntpd restart
chkconfig --level 345 ntpd on

rac2:

rac2上的操作与rac1类似，不同的是配置的server指向rac1

cat>/etc/ntp.conf<restrict default nomodify notrap noquery
restrict 127.0.0.1

server 0.rhel.pool.ntp.org
server 1.rhel.pool.ntp.org
server 2.rhel.pool.ntp.org

server 192.168.47.241
fudge 192.168.47.241 stratum 10

driftfile /var/lib/ntp/ntp.drift
broadcastdelay 0.008
keys /etc/ntp/keys
EOF

service ntpd restart
chkconfig --level 345 ntpd on

需要注意的是，配置完ntp后大概需要5-10分钟的时候，才会开始进行同步。可以通过ntpq -p 命令来查看

[root@rac1 ~]# ntpq -p
remote refid st t when poll reach delay offset jitter
==============================================================================
rac1 .INIT. 16 u - 64 0 0.000 0.000 4000.00
*LOCAL(0) LOCAL(0) 10 l 30 64 377 0.000 0.000 0.001

[root@rac2 ~]# ntpq -p
remote refid st t when poll reach delay offset jitter
==============================================================================
rac1 LOCAL(0) 11 u 9 64 1 0.187 0.072 0.001



源端和目标端之间的时间同步则不是那么重要。可以在目标端手动对时间进行一次同步(目标端的ntpd服务没有启动)，并将同步后的系统时间写入bios
[root@ggdb ~]# ntpdate 192.168.47.241
8 Dec 15:40:20 ntpdate[8311]: adjust time server 192.168.47.241 offset 0.003007 sec
[root@ggdb ~]# /sbin/hwclock -w

如果不能成功，可以在ntpdate命令中加上-d参数来排查具体原因

创建系统用户

在源端(包括rac1和rac2)和目标端创建用户，用于管理GoldenGate：

useradd -g oinstall -G dba goldengate
passwd goldengate

配置资源限制：
官方建议尽可能将资源限制放开，其实我们可以根据具体情况进行配置，但不能设得太小

cat >>/etc/security/limits.conf<goldengate soft memlock 3145728
goldengate hard memlock 3145728
goldengate soft nproc 200
goldengate hard nproc 1024
goldengate soft stack -
goldengate hard stack -
goldengate soft fsize -
goldengate hard fsize -
EOF

准备GoldenGate安装环境

源端

RAC环境中，GoldenGate应该安装在一个共享存储中，这样当运行GoldenGate相关进程的节点出现故障时，可以将该目录挂载到另一个节点中继续运行
以下的在rac1上做就可以了

这里有一个分区/dev/sdg1，可以被rac1和rac2访问。我们将其格式化ext3文件系统，并挂载到rac1上(在正常情况下，我们在rac1上启动extract进程)：

[root@rac1 ~]# mkfs.ext3 /dev/sdg1
mke2fs 1.35 (28-Feb-2004)
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
12500992 inodes, 24993115 blocks
1249655 blocks (5.00%) reserved for the super user
First data block=0
Maximum filesystem blocks=4294967296
763 block groups
32768 blocks per group, 32768 fragments per group
16384 inodes per group
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
4096000, 7962624, 11239424, 20480000, 23887872

Writing inode tables: done
Creating journal (8192 blocks): done
Writing superblocks and filesystem accounting information: done

This filesystem will be automatically checked every 24 mounts or
180 days, whichever comes first. Use tune2fs -c or -i to override.
[root@rac1 ~]#

挂载到rac1的/opt/gg下：
[root@rac1 ~]# mkdir /opt/gg
[root@rac1 ~]# mount /dev/sdg1 /opt/gg

然后在rac1中配置开机自动挂载(这一步不要在rac2上做)：
echo "mount /dev/sdg1 /opt/gg" >> /etc/rc.local

目标端

目标端是一个单机，安装到/opt/gg下
mkdir /opt/gg


在源端和目标端上，赋予goldengate用户/opt/gg目录的使用权限
chown -R goldengate:oinstall /opt/gg
chmod -R 775 /opt/gg

解压goldengate安装文件到安装目录

源端(rac1)和目标端都做：
以goldengate用户登录
[goldengate@rac1 goldengateMedia]$ mkdir /opt/gg/goldengate、
[goldengate@rac1 goldengateMedia]$ cp ggs_Linux_x64_ora10g_64bit_v11_1_1_0_0_078.tar /opt/gg/goldengate
[goldengate@rac1 goldengateMedia]$ cd /opt/gg/goldengate
[goldengate@rac1 goldengate]$ tar -xvf ggs_Linux_x64_ora10g_64bit_v11_1_1_0_0_078.tar

配置环境变量

源端(rac1和rac2)和目标端：
修改goldengate用户的环境变量配置文件(ORACLE_SID按实际情况修改)
cat>>/home/goldengate/.bashrc<ORACLE_HOME=/opt/app/oracle/product/10.2.0/db_1
export ORACLE_HOME
ORACLE_SID=oradb1
export ORACLE_SID
GG_HOME=/opt/gg/goldengate
export GG_HOME
PATH=\$ORACLE_HOME/bin:\$PATH
export PATH
GG_HOME=/opt/gg/goldengate
export GG_HOME
LD_LIBRARY_PATH=\$ORACLE_HOME/lib:\$GG_HOME:\$LD_LIBRARY_PATH
export LD_LIBRARY_PATH
EOF

应用刚刚修改的环境变量，然后进入GoldenGate安装目录，执行ldd ggsci，确定需要的库文件都能够找到。如果出现共享库文件无法找到，例如libnnz10.so => not found，检查LD_LIBRARY_PATH环境变量的设置
[goldengate@ggdb goldengate]$ source ~/.bashrc
[goldengate@ggdb goldengate]$ cd $GG_HOME
[goldengate@ggdb goldengate]$ ldd ggsci
libdl.so.2 => /lib64/libdl.so.2 (0x00000037a3900000)
libicui18n.so.38 => /opt/gg/goldengate/libicui18n.so.38 (0x0000002a9558c000)
libicuuc.so.38 => /opt/gg/goldengate/libicuuc.so.38 (0x0000002a958ec000) libicudata.so.38 => /opt/gg/goldengate/libicudata.so.38 (0x0000002a95c25000)
libpthread.so.0 => /lib64/tls/libpthread.so.0 (0x00000037a3d00000)
libxerces-c.so.28 => /opt/gg/goldengate/libxerces-c.so.28 (0x0000002a968fc000)
libnnz10.so => /opt/app/oracle/product/10.2.0/db_1/lib/libnnz10.so (0x0000002a96e13000)
libclntsh.so.10.1 => /opt/app/oracle/product/10.2.0/db_1/lib/libclntsh.so.10.1 (0x0000002a972b4000)
libstdc++.so.6 => /usr/lib64/libstdc++.so.6 (0x00000037a5d00000)
libm.so.6 => /lib64/tls/libm.so.6 (0x00000037a3700000)
libgcc_s.so.1 => /lib64/libgcc_s.so.1 (0x00000037a5b00000)
libc.so.6 => /lib64/tls/libc.so.6 (0x00000037a3400000)
/lib64/ld-linux-x86-64.so.2 (0x00000037a3000000)
libnsl.so.1 => /lib64/libnsl.so.1 (0x00000037a9100000)

数据库方面的准备

在源端数据库中打开归档模式

SQL> archive log list

Database log mode              Archive Mode
Automatic archival             Enabled
Archive destination            +RECOVERY_DG
Oldest online log sequence     120
Next log sequence to archive   121
Current log sequence           121


若处于非归档模式，则改为归档模式：
SQL> shutdown immediate
Database closed.
Database dismounted.
ORACLE instance shut down.
SQL> startup mount;
ORACLE instance started.


SQL> alter database archivelog;

Database altered.

SQL> alter database open;

Database altered.

SQL>

在源端数据库中打开force logging

SQL> select force_logging from v$database;

FOR
---
NO

SQL> alter database force logging;

Database altered.

SQL> select force_logging from v$database;

FOR
---
YES

在源端数据库中打开supplemental log

SQL> select supplemental_log_data_min from v$database;

SUPPLEME
--------
NO

SQL> alter database add supplemental log data;

Database altered.

切换日志，使更改生效
SQL> alter system switch logfile;

System altered.

SQL> select supplemental_log_data_min from v$database;

SUPPLEME
--------
YES

在源端数据库中关闭回收站

官方的说明是，由于一个已知的问题，回收站会对DDL触发器产生影响，因此需要关闭。由此可见，我们只需要在源库中关闭回收站即可。

SQL> show parameter recyclebin

NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
recyclebin string on

SQL> alter system set recyclebin=off;

System altered.

SQL> show parameter recyclebin

NAME TYPE
------------------------------------ --------------------------------
VALUE
------------------------------
recyclebin string
OFF

确保goldengate能够连接到数据库的ASM实例

RAC中所有节点都要配置

在源端TNSNAMES.ORA中配置ASM实例信息
vi $ORACLE_HOME/network/admin/tnsnames.ora

ORADB_ASM =
  (DESCRIPTION =
    (ADDRESS = (PROTOCOL = TCP)(HOST = rac1-vip)(PORT = 1521))
    (ADDRESS = (PROTOCOL = TCP)(HOST = rac2-vip)(PORT = 1521))
    (LOAD_BALANCE = yes)
    (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = +ASM)
    )
  )


在源端LISTENER.ORA中配置ASM实例的相关信息
vi $ORACLE_HOME/network/admin/listener.ora

SID_LIST_LISTENER_RAC1 =
  (SID_LIST =
    (SID_DESC =
      (GLOBAL_DBNAME = oradb)
      (ORACLE_HOME = /opt/app/oracle/product/10.2.0/db_1)
      (SID_NAME = oradb1)
    )
    (SID_DESC =
      (GLOBAL_DBNAME = +ASM)
      (ORACLE_HOME = /opt/app/oracle/product/10.2.0/db_1)
      (SID_NAME = +ASM1)
    )
  )

上面是rac1中的配置，rac2中的SID_LIST_LISTENER_xxx 和SID_NAME要相应修改

重启监听
lsnrctl reload

通过sqlplus sys/xxx@oradb_asm as sysdba来连接asm实例，能连上则说明配置成功

字符集

1. 目标数据库的字符集必须是源数据库字符集的超集
2. 数据库字符集必须为客户端应用程序字符集的超集

SQL> SHOW PARAMETER NLS_LANGUAGE

NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
nls_language string SIMPLIFIED CHINESE

SQL> SHOW PARAMETER NLS_TERRITORY

NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
nls_territory string CHINA

SQL> SELECT name, value$ from SYS.PROPS$ WHERE name = 'NLS_CHARACTERSET';

NAME
------------------------------
VALUE$
--------------------------------------------------------------------------------
NLS_CHARACTERSET
ZHS16GBK

SQL> SHOW PARAMETER NLS_LENGTH_SEMANTICS

NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
nls_length_semantics string BYTE

SQL>

设置终端的字符集：
root用户登录，源(rac1和rac2)和目标端都做
cat >>/etc/bashrc<NLS_LANG="SIMPLIFIED CHINESE_CHINA".ZHS16GBK
export NLS_LANG
EOF

创建goldengate数据库用户

源和目标端都需要

[oracle@gg1 ~]$ sqlplus / as sysdba

SQL*Plus: Release 10.2.0.4.0 - Production on Mon Aug 1 22:31:42 2011

Copyright (c) 1982, 2007, Oracle. All Rights Reserved.


Connected to:
Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options

SQL> create tablespace goldengate;

Tablespace created.

SQL> create user goldengate identified by goldengate default tablespace goldengate;

User created.

SQL> grant connect,resource to goldengate;

Grant succeeded.

SQL> grant execute on utl_file to goldengate;

Grant succeeded.

SQL>


抽取进程使用的数据库用户需要额外的权限，我们将这些权限也授予数据库用户goldengate(在源端数据库中执行)

SQL> exec dbms_streams_auth.grant_admin_privilege('GOLDENGATE');

PL/SQL procedure successfully completed.

SQL> grant insert on system.logmnr_restart_ckpt$ to goldengate;

Grant succeeded.

SQL> grant update on sys.streams$_capture_process to goldengate;

Grant succeeded.

SQL> grant become user to goldengate;

Grant succeeded.

SQL>

为了确保GoldenGate正常运行，特别是在目标端，赋予goldengate用户DBA权限：
SQL> grant dba to goldengate;

UNDO的设置

goldengate使用flashback query从源数据库中读取undo表空间中的数据，以重建基于scn或时间点的读一致性。

建议设置如下：
UNDO_MANAGEMENT=AUTO
UNDO_RETENTION=86400
undo表空间的大小按如下公式估计设置
= * +
❍ is the number of undo blocks.
❍ is the value of the UNDO_RETENTION parameter (in seconds).
❍ is the number of undo blocks for each second.
❍ is the minimal overhead for metadata (transaction tables, etc.).
Use the system view V$UNDOSTAT to estimate and .


该步骤在源端数据库执行即可


SQL> show parameter undo


NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
undo_management                      string      AUTO
undo_retention                       integer     900
undo_tablespace                      string      UNDOTBS1


SQL> alter system set undo_retention=86400;


System altered.


将flashback any table 权限赋给extract用户


SQL> grant flashback any table to goldengate;


Grant succeeded.


SQL>