Some doubt on RAC confiuration(IP/listener/TAF)

From: Serol_Luo \(BJ\) <Serol_Luo_at_smics.com>
Date: Mon, 11 Sep 2006 16:13:00 +0800
Message-ID: <B1B44026C565CF4F9753E28AAFB123CD01859B@bjex108.smic-bj.com>

All,
We're planning to implement Oracle RAC 9206 base on VCS3.5 or VCS 4.0(Veritas Cluster System).
But I have some question about the listener/IP configuration/TAF. The situation is:
1. We use 2 instances, with it's own IPMP IP address and own listener. Inst1 listen salmon1-vip
Inst2 listen salmon2-vip  

2. Client use TAF like this:

PROD1 =
(DESCRIPTION=
(ADDRESS = (PROTO = TCP)(HOST = salmon1-vip)(PORT = 1521))
(ADDRESS = (PROTO = TCP)(HOST = salmon2-vip)(PORT = 1521))
(LOAD_BALANCE = yes)
(CONNECT_DATA=
(SERVICE_NAME = PROD1)
(FAILOVER_MODE =
(TYPE = SELECT)
(METHOD = BASIC)

)
)
)

PROD1A =
(DESCRIPTION=
(ADDRESS = (PROTO = TCP)(HOST = salmon1-vip)(PORT = 1521))
(CONNECT_DATA=
(SERVER = DEDICATED)
(SERVICE_NAME = PROD1)
(INSTANCE_NAME = PROD1A)

)
)

PROD1B =
(DESCRIPTION=
(ADDRESS = (PROTO = TCP)(HOST = salmon2-vip)(PORT = 1521))
(CONNECT_DATA=
(SERVER = DEDICATED)
(SERVICE_NAME = PROD1)
(INSTANCE_NAME = PROD1B)

)
)
Refer to :
http://www.oracle.com/technology/global/cn/pub/articles/chan_sing2rac_install.ht ml#taf
But I did not set delay/retry in TAF tnsnames.

---

My problem is:
When server 1 down, or just database resource off lined. The IP "salmon1-vip" also down.
Then the new connection connect to PROD1 will hung for about 3 Min before if realized the IP was down and connect to "salmon2-vip" ¡­¡­

How do I handle this situation?  

Or what¡¯s your current practice configuring IP/listener/TAF in RAC?  

Any input will be very appreciated ¡­¡­  

---

Regards.
Serol Luo, From Peking,China
Or call me RollingPig
http://rollingpig.itpub.net  

--
http://www.freelists.org/webpage/oracle-l

Received on Mon Sep 11 2006 - 03:13:00 CDT