今天是2014-03-19，对oracle TAF技术整理一下学习笔记，记录如下：
####################################################################################
failover_mode 参数 描述
####################################################################################
backup 指定用于创建备份连接的本地服务名，当使用preconnect预创建连接
的时候应该指明这个参数值
method TAF的配置包含如下两种failover切换方式
preconnect：创建到切换实例的预连接，提供快速failover的能力
basic：在发生failover的时候创建连接
retries failover发生后尝试连接的次数，如果指明了delay参数，那么retries默认为5
type Taf的配置包含如下三种failover的类型：
session：如果用户连接丢失，新的会话将自动被创建。这种类型的failover不能
尝试恢复select操作
select：如果用户连接丢失，新创建的会话将继续之前失败之后的select操作
none 这是默认值，不具备failover能力。这个能被明确的指明用于防止failover
的发生。
注意：这些参数只能手动设置，不能在listener.ora文件中SID_LIST_<LISTENER_NAME>条目中设置global_dbname参数，静态配置的全局数据
库名称不能使用TAF功能。另外jdbc thin驱动方式无法使用TAF技术。
实现TAF有两种方式一种为client-side TAF 另一种为server-side TAF ，下面先介绍第一种client-side TAF:
RAC =
(DESCRIPTION =
(FAILOVER=ON)
(LOAD_BALANCE=ON)
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = rac-one)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = rac-two)(PORT = 1521))
)
(CONNECT_DATA =
(SERVICE_NAME = Rac2)
(FAILOVER_MODE=
(TYPE=SELECT)
(METHOD=BASIC))
)
)该方式使用了连接时failover、client-side TAF 和客户端负载均衡，当该客户端尝试连接数据库的时候会在address中随即挑选一个用于连接数据库，假如选择rac-one如果连接失败，那么就会使用rac-two进行连接，如果都失败那么将提出连接错误。当客户端已经连接到数据库的时候，突然rac-one实例关闭，那么该客户端随即创建与rac-two的会话连接这个过程报错select的操作。例外我们可以使用retries和delay参数来指定重新连接次数和延迟重新连接的秒数。如下是重试连接rac-one5次每次120秒。
eg：
RAC =
(DESCRIPTION =
(FAILOVER=ON)
(LOAD_BALANCE=OFF)
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = rac-one-vip)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = rac-two-vip)(PORT = 1521))
)
(CONNECT_DATA =
(SERVICE_NAME = Rac2)
(FAILOVER_MODE=
(TYPE=SELECT)
(METHOD=BASIC)
(RETRIES=5)
(DELAY=120)
)
)
)
另外在failover_mode中的method中有preconnect（预连接），该说明在client-side TAF中分配一个主连接的同时预先分配备用连接。
eg：
RAC1 =
(DESCRIPTION =
(FAILOVER=ON)
(LOAD_BALANCE=OFF)
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = rac-one)(PORT = 1521))
)
(CONNECT_DATA =
(SERVICE_NAME = Rac2)
(FAILOVER_MODE=
(METHOD=PRECONNECT)
(BACKUP=RAC2)
(TYPE=SELECT)
(RETRIES=5)
(DELAY=30)
)
)
)
RAC2 =
(DESCRIPTION =
(FAILOVER=ON)
(LOAD_BALANCE=OFF)
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = rac-two)(PORT = 1521))
)
(CONNECT_DATA =
(SERVICE_NAME = Rac2)
(FAILOVER_MODE=
(METHOD=PRECONNECT)
(BACKUP=RAC1)
(TYPE=SELECT)
(RETRIES=5)
(DELAY=30)
)
)
)
验证client-side TAF：
首先确认/etc/hosts文件如下：
[root@rac-one ~]# more /etc/hosts
127.0.0.1 localhost localhost.localdomain
192.168.2.11 openfiler1
192.168.1.112 rac-two-priv
192.168.1.111 rac-one-priv
192.168.4.111 rac-one rac-one.localdomain
192.168.4.112 rac-two rac-two.localdomain
192.168.4.113 rac-one-vip
192.168.4.114 rac-two-vip
[root@rac-one ~]#
查看客户端tnsname.ora配置：
RAC =
(DESCRIPTION =
(FAILOVER=ON)
(LOAD_BALANCE=ON)
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = rac-two-vip)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = rac-one-vip)(PORT = 1521))
)
(CONNECT_DATA =
(SERVICE_NAME = Rac)
(FAILEOVER_MODE=
(METHOD=BASIC)
(TYPE=SELECT)
(RETRIES=5)
(DELAY=60))
)
)
用户连接数据库，然后查看数据库会话信息，（注意还需要修改windows的hosts名称否则无法识别rac-two-vip或是rac-one-vip）：
rac-one节点：
SQL> r
1 select inst_id,username,failover_type,failover_method,failed_over from gv$session where username in ('SYSTEM','SYS')
2*
INST_ID USERNAME FAILOVER_TYPE FAILOVER_M FAI
---------- ------------------------------ ------------- ---------- ---
2 SYS NONE NONE NO
2 SYS NONE NONE NO
2 SYS NONE NONE NO
2 SYS NONE NONE NO
2 SYS NONE NONE NO
2 SYS NONE NONE NO
1 SYS NONE NONE NO
1 SYS NONE NONE NO
1 SYS NONE NONE NO
1 SYSTEM SELECT BASIC NO
10 rows selected.
SQL>
可知目前system用户已经具备failover功能。
注意：在配置client-side TAF的时候尤其注意参数的设置位置，否则无法实现failover。
其实在11G中scan功能也实现了负载均衡的作用，它是从dns解析中的三个地址轮询负载的分配给scan listener进而采去和本地listener进行通信。
另外实现TAF的方式为server-side TAF。说白了，就是通过服务端设置service来实现，先比client-side TAF有很多简便的方式。
eg：
增加服务名rac1和rac2：
oracle@rac-two ~]$ srvctl add service -d Rac -s rac1 -r Rac1 -a Rac2 -P basic -y automatic -e select -m basic -z 5 -w 120
[oracle@rac-two ~]$ srvctl add service -d Rac -s rac2 -r Rac2 -a Rac1 -P basic -y automatic -e select -m basic -z 5 -w 120
查看服务名状态
[oracle@rac-two ~]$ srvctl status service -d RAc
Service rac1 is not running.
Service rac2 is not running.
启动服务资源：
[oracle@rac-two ~]$ srvctl start service -d Rac
[oracle@rac-two ~]$ srvctl status service -d Rac
Service rac1 is running on instance(s) Rac1
Service rac2 is running on instance(s) Rac2
查看配置信息：
[oracle@rac-two ~]$ srvctl config service -d Rac
Service name: rac1
Service is enabled
Server pool: Rac_rac1
Cardinality: 1
Disconnect: false
Service role: PRIMARY
Management policy: AUTOMATIC
DTP transaction: false
AQ HA notifications: false
Failover type: SELECT
Failover method: BASIC
TAF failover retries: 5
TAF failover delay: 120
Connection Load Balancing Goal: LONG
Runtime Load Balancing Goal: NONE
TAF policy specification: BASIC
Edition:
Preferred instances: Rac1
Available instances: Rac2
Service name: rac2
Service is enabled
Server pool: Rac_rac2
Cardinality: 1
Disconnect: false
Service role: PRIMARY
Management policy: AUTOMATIC
DTP transaction: false
AQ HA notifications: false
Failover type: SELECT
Failover method: BASIC
TAF failover retries: 5
TAF failover delay: 120
Connection Load Balancing Goal: LONG
Runtime Load Balancing Goal: NONE
TAF policy specification: BASIC
Edition:
Preferred instances: Rac2
Available instances: Rac1
[oracle@rac-two ~]$
注意这个时候，实例Rac1已经注册了rac1服务，且主要实例为Rac1备用实例为Rac2，实例Rac2注册了rac2服务，且主要实例为Rac2备用实例为Rac1；
本地监听只会注册本地服务名，scan监听将注册所有的监听服务名。
验证：
首先明确客户端配置：
RAC =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = rac-two-cluster-scan.grid.example.com)(PORT = 1521))
)
(CONNECT_DATA =
(SERVICE_NAME = Rac2)
)
)登录数据库查看会话信息如下：
会话一使用system用户登录数据库：
用户没有登录之前状态：
[oracle@rac-one ~]$ sqlplus / as sysdba
SQL*Plus: Release 11.2.0.4.0 Production on Wed Mar 19 20:59:56 2014
Copyright (c) 1982, 2013, Oracle. All rights reserved.
Connected to:
Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
With the Partitioning, Real Application Clusters, Automatic Storage Management, OLAP,
Data Mining and Real Application Testing options
SQL> col username for a20
SQL> set linesize 200
SQL> select inst_id,username,failover_type,failover_method,failed_over from gv$session where username='SYSTEM';
no rows selected
登录之后状态：
SQL> r
1* select inst_id,username,failover_type,failover_method,failed_over from gv$session where username='SYSTEM'
INST_ID USERNAME FAILOVER_TYPE FAILOVER_M FAI
---------- -------------------- ------------- ---------- ---
2 SYSTEM SELECT BASIC NO
2 SYSTEM SELECT BASIC NO
SQL> 这个时候关闭该节点，且在客户端执行select * from dba_objects;语句，
查看节点二 用户会话状态。
SQL> r
1* select inst_id,username,failover_type,failover_method,failed_over from gv$session where username='SYSTEM'
INST_ID USERNAME FAILOVER_TYPE FAILOVER_M FAI
---------- -------------------- ------------- ---------- ---
1 SYSTEM SELECT BASIC YES
SQL>
可以看到用户只在几秒中停顿后继续完成了select操作，且failed_over状态为yes，证明failover已经生效。
查看状态信息如下：
c[grid@rac-two ~]$ crsctl status res -t
--------------------------------------------------------------------------------
NAME TARGET STATE SERVER STATE_DETAILS
--------------------------------------------------------------------------------
Local Resources
--------------------------------------------------------------------------------
ora.DATADG.dg
ONLINE ONLINE rac-two
ora.GIDG.dg
ONLINE ONLINE rac-two
ora.LISTENER.lsnr
ONLINE ONLINE rac-two
ora.asm
ONLINE ONLINE rac-two Started
ora.gsd
OFFLINE OFFLINE rac-two
ora.net1.network
ONLINE ONLINE rac-two
ora.ons
ONLINE ONLINE rac-two
ora.registry.acfs
ONLINE ONLINE rac-two
--------------------------------------------------------------------------------
Cluster Resources
--------------------------------------------------------------------------------
ora.LISTENER_SCAN1.lsnr
1 ONLINE ONLINE rac-two
ora.LISTENER_SCAN2.lsnr
1 ONLINE ONLINE rac-two
ora.LISTENER_SCAN3.lsnr
1 ONLINE ONLINE rac-two
ora.cvu
1 ONLINE ONLINE rac-two
ora.oc4j
1 ONLINE ONLINE rac-two
ora.rac-one.vip
1 ONLINE INTERMEDIATE rac-two FAILED OVER
ora.rac-two.vip
1 ONLINE ONLINE rac-two
ora.rac.db
1 ONLINE ONLINE rac-two Open
2 ONLINE OFFLINE Instance Shutdown
ora.rac.rac1.svc
1 ONLINE ONLINE rac-two
ora.rac.rac2.svc
1 ONLINE ONLINE rac-two
ora.scan1.vip
1 ONLINE ONLINE rac-two
ora.scan2.vip
1 ONLINE ONLINE rac-two
ora.scan3.vip
1 ONLINE ONLINE rac-two
[grid@rac-two ~]$
可知目前ora.ora-one.vip已经failed over，且ora.rac.rac1.svc运行到了rac-two中。
重启节点后查看资源如下：
[grid@rac-one ~]$ crsctl status res -t
--------------------------------------------------------------------------------
NAME TARGET STATE SERVER STATE_DETAILS
--------------------------------------------------------------------------------
Local Resources
--------------------------------------------------------------------------------
ora.DATADG.dg
ONLINE ONLINE rac-one
ONLINE ONLINE rac-two
ora.GIDG.dg
ONLINE ONLINE rac-one
ONLINE ONLINE rac-two
ora.LISTENER.lsnr
ONLINE OFFLINE rac-one
ONLINE ONLINE rac-two
ora.asm
ONLINE ONLINE rac-one Started
ONLINE ONLINE rac-two Started
ora.gsd
OFFLINE OFFLINE rac-one
OFFLINE OFFLINE rac-two
ora.net1.network
ONLINE ONLINE rac-one
ONLINE ONLINE rac-two
ora.ons
ONLINE OFFLINE rac-one STARTING
ONLINE ONLINE rac-two
ora.registry.acfs
ONLINE ONLINE rac-one
ONLINE ONLINE rac-two
--------------------------------------------------------------------------------
Cluster Resources
--------------------------------------------------------------------------------
ora.LISTENER_SCAN1.lsnr
1 ONLINE ONLINE rac-two STOPPING
ora.LISTENER_SCAN2.lsnr
1 ONLINE ONLINE rac-two
ora.LISTENER_SCAN3.lsnr
1 ONLINE ONLINE rac-two
ora.cvu
1 ONLINE ONLINE rac-two
ora.oc4j
1 ONLINE ONLINE rac-two
ora.rac-one.vip
1 ONLINE INTERMEDIATE rac-two FAILED OVER,STOPPING
ora.rac-two.vip
1 ONLINE ONLINE rac-two
ora.rac.db
1 ONLINE ONLINE rac-two Open
2 ONLINE OFFLINE Instance Shutdown
ora.rac.rac1.svc
1 ONLINE ONLINE rac-two
ora.rac.rac2.svc
1 ONLINE ONLINE rac-two
ora.scan1.vip
1 ONLINE ONLINE rac-two
ora.scan2.vip
1 ONLINE ONLINE rac-two
ora.scan3.vip
1 ONLINE ONLINE rac-two
[grid@rac-one ~]$
[grid@rac-one ~]$ crsctl status res -t
--------------------------------------------------------------------------------
NAME TARGET STATE SERVER STATE_DETAILS
--------------------------------------------------------------------------------
Local Resources
--------------------------------------------------------------------------------
ora.DATADG.dg
ONLINE ONLINE rac-one
ONLINE ONLINE rac-two
ora.GIDG.dg
ONLINE ONLINE rac-one
ONLINE ONLINE rac-two
ora.LISTENER.lsnr
ONLINE ONLINE rac-one
ONLINE ONLINE rac-two
ora.asm
ONLINE ONLINE rac-one Started
ONLINE ONLINE rac-two Started
ora.gsd
OFFLINE OFFLINE rac-one
OFFLINE OFFLINE rac-two
ora.net1.network
ONLINE ONLINE rac-one
ONLINE ONLINE rac-two
ora.ons
ONLINE ONLINE rac-one
ONLINE ONLINE rac-two
ora.registry.acfs
ONLINE ONLINE rac-one
ONLINE ONLINE rac-two
--------------------------------------------------------------------------------
Cluster Resources
--------------------------------------------------------------------------------
ora.LISTENER_SCAN1.lsnr
1 ONLINE ONLINE rac-one
ora.LISTENER_SCAN2.lsnr
1 ONLINE ONLINE rac-two
ora.LISTENER_SCAN3.lsnr
1 ONLINE ONLINE rac-two
ora.cvu
1 ONLINE ONLINE rac-two
ora.oc4j
1 ONLINE ONLINE rac-two
ora.rac-one.vip
1 ONLINE ONLINE rac-one
ora.rac-two.vip
1 ONLINE ONLINE rac-two
ora.rac.db
1 ONLINE ONLINE rac-two Open
2 ONLINE ONLINE rac-one Open
ora.rac.rac1.svc
1 ONLINE ONLINE rac-two
ora.rac.rac2.svc
1 ONLINE ONLINE rac-two
ora.scan1.vip
1 ONLINE ONLINE rac-one
ora.scan2.vip
1 ONLINE ONLINE rac-two
ora.scan3.vip
1 ONLINE ONLINE rac-two
[grid@rac-one ~]$
如果当Rac1失效（节点关闭），那么select将再次移动到Rac2上来
SQL> r
1* select inst_id,username,failover_type,failover_method,failed_over from gv$session where username='SYSTEM'
INST_ID USERNAME FAILOVER_TYPE FAILOVER_M FAI
---------- ------------------------------ ------------- ---------- ---
2 SYSTEM SELECT BASIC YES
到了11G R2 使用scan和server-side TAF是最佳选择。
That’s all！！！！！
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++Rhys↖(^ω^)↗Amy+++++++++++++++++++++++++++
扫一扫
2007
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
