oracle 10g RAC Failover(二)

原创黄宸宁 2013-04-17

767

oracle 10g RAC Failover(一)

实验准备：
1、创建测试用户

create user scott identified by oracle default tablespace users;
grant create session to scott;
grant resource to scott;
grant select on v_$instance to scott;
grant execute on dbms_lock to scott;
GRANT  SELECT  ANY  DICTIONARY  TO scott;
grant select any table to scott;

2、配置local_listener和remote_listener(为了保证连接都只先连接到其中一个节点上)

ALTER SYSTEM SET LOCAL_LISTENER = '(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.100.112)(PORT = 1521))'  sid = 'orcldb1';
ALTER SYSTEM SET LOCAL_LISTENER = '(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.100.114)(PORT = 1521))'  sid = 'orcldb2'; 

ALTER SYSTEM SET remote_listener = ''  sid = 'orcldb1';
ALTER SYSTEM SET remote_listener = ''  sid = 'orcldb2';

一、Client-side Connection time Failover

1、配置客户端的tnsnames.ora文件

ORCLDB =
  (DESCRIPTION =
    (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.100.112)(PORT = 1521))
    (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.100.114)(PORT = 1521))
    (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = orcldb)
    )
  )

2、通过脚本测试连接

[oracle@hcn10g script]$ cat login.sh
for i in {1..10}
do
echo 'No.'$i
sqlplus -s scott/oracle@orcldb <<EOF
select instance_name from v$instance
/
exit
EOF
done

测试结果：

[oracle@hcn10g script]$ ksh login.sh
No. 1

INSTANCE_NAME
----------------
orcldb1

No. 2

INSTANCE_NAME
----------------
orcldb1

No. 3

INSTANCE_NAME
----------------
orcldb1

No. 4

INSTANCE_NAME
----------------
orcldb1

No. 5

INSTANCE_NAME
----------------
orcldb1

No. 6

INSTANCE_NAME
----------------
orcldb1

No. 7

INSTANCE_NAME
----------------
orcldb1

No. 8

INSTANCE_NAME
----------------
orcldb1

No. 9

INSTANCE_NAME
----------------
orcldb1

No. 10

INSTANCE_NAME
----------------
orcldb1

[oracle@hcn10g script]$

连续10次连接都连接到1号节点

3、进行测试
在运行下面脚本的过程中，对1号节点执行shutdown abort操作，模拟断电

[oracle@hcn10g script]$ 
[oracle@hcn10g script]$ cat loginfail.sh
for i in {1..10}
do
echo 'No.'$i
sqlplus -s scott/oracle@orcldb <<EOF
select instance_name from v$instance
/
set feedback off
exec dbms_lock.sleep(5)
exit
EOF
done

测试结果：

SQL> 
SQL> select instance_name from v$instance;

INSTANCE_NAME
----------------
orcldb1

SQL> shutdown abort
ORACLE instance shut down.
SQL> 

[oracle@hcn10g script]$ 
[oracle@hcn10g script]$ ksh loginfail.sh
No.1

INSTANCE_NAME
----------------
orcldb1

No.2

INSTANCE_NAME
----------------
orcldb1

No.3              ------在这个时刻节点1执行了shutdown abort的操作，下面的连接都连接到了2号节点上

INSTANCE_NAME
----------------
orcldb1

BEGIN dbms_lock.sleep(5); END;

*
ERROR at line 1:
ORA-03113: end-of-file on communication channel


No.4

INSTANCE_NAME
----------------
orcldb2

No.5

INSTANCE_NAME
----------------
orcldb2

No.6

INSTANCE_NAME
----------------
orcldb2

No.7

INSTANCE_NAME
----------------
orcldb2

No.8

INSTANCE_NAME
----------------
orcldb2

No.9

INSTANCE_NAME
----------------
orcldb2

No.10

INSTANCE_NAME
----------------
orcldb2

[oracle@hcn10g script]$

结论：

Client-side Connection time Failover 方式的Failover在实例出席异常后，需要重新连接，不能自动的转到正常的实例上。

oracle

「喜欢这篇文章，您的关注和赞赏是给作者最好的鼓励」

关注作者

oracle 10g RAC Failover(二)

评论