RAC failover

Hi
Is there any instructions/documentation available on how to code and configure toplink for RAC failure. I searched the forum and there was a post on this. However the post did not specify what we need to do. It only said that we have solved the problem.
The toplink jar manifest that I am using shows the following information
Specification-Title: TopLink
Specification-Vendor: Oracle
Specification-Version: 10.1.3.3.0
Thanks

I believe that's where you could use an OCI callback or you could also have your application receive and process FAN notifications from ONS.
I haven't implemented them before, but here are some doc links to the details that you might want to review:
http://download.oracle.com/docs/cd/B19306_01/appdev.102/b14250/oci09adv.htm#sthref1501
http://download.oracle.com/docs/cd/B19306_01/appdev.102/b14250/oci09adv.htm#sthref1586
http://download.oracle.com/docs/cd/B19306_01/rac.102/b14197/hafeats.htm#BABEFCIC

Similar Messages

Do I need to change OCI layer to utilize Oracle RAC failover Feature?

Dear Oracle Experts,
From the Oracle documents, the RAC is a great feature providing the scalability and high availablility. I thought the failover features should be transparent to users.
However, from this book:
http://www.oreilly.com/catalog/oressentials2/chapter/ch01.html
"For instance, in high-availability implementations in which multiple systems share disks and implement Real Application Clusters/Oracle Parallel Server, you may want users to reattach to a second server transparently if the first fails. You can write programs that do this using OCI. "
"You can write applications with the Oracle Call Interface (OCI) to provide failover to a second instance transparently to the user. "
I am confused if RAC failover feature have to depend on OCI user application coding?
I have an application developed over Oracle 8.1.6, if I buy Oracle9i to use RAC failover feature, should I modify my application at OCI layer to make it work? Can I switch the oracle instance when the original one fails only by RAC setup instead of changing OCI code?
Thanks.
Wei
P.S. I am running application over Compaq tru64 5.1A.

Hi ,
As answered by Sumit u need not alter ur application for shifting it to RAC.But there are few things which will determine the perfoemance of your application on RAC and depend on the type of application u use .Application partioning,data partitioning,index partitioning ... look for these topics . This may help.
The other thing u mentioned abt. Oracle Libraries,DB name,SID .. . Now when u'll have RAC there will be more than 1 Node I believe. So each Node will have its SID,Oracle Binaries but DB name will be same.
Gud Luck!
Vijay Shanker
Edited by: Vijay Shanker on Feb 14, 2009 6:12 PM

Handling Oracle RAC Failover at the application layer

Hi All,
I am currently researching best practices for handling oracle RAC failover at the application layer since transactional statements (INSERT,UPDATE and DELETE) are not handled transparently. So I have few questions for the community:
1. In case of a node failure would I need to roll back all of transactional statements that are part of the transaction or would I have to re-execute the one that failed only?
2. Does things change with XA 2 phase commit transactions?
Any input and/or architecture suggestions would be much appreciated.
Regards,
Dmitriy Frolov

Hi,
the Oracle RAC stack works very vell on its own, without the need of a third party clusterware. It will be aware of failures on the nodes.
Database will be using ASM or RAW or NFS devices for Shared storage.
Oracle Clusterware can also be configured to monitor your applications and to a failover to other nodes.
However if you need a shared filesystem for your own applications, then you will have to use QFS or similar (which again than requires Sun Cluster).
See: http://www.oracle.com/technology/products/database/clusterware/index.html
Regards
Sebastian

Toplink support for RAC Failover

We need to enable Toplink to support RAC Fast Connection Failover. We need to support
only read query retry functionality and not write query retry. We found one document
(http://www.oracle.com/technology/events/develop2007/presentations/oracle_jdbc_high_availability_load_balancing_best_practices_and_road_map.pdf)
that explains how to configure JDBC connection for RAC failover. However, since we
are using Toplink, we don't explicitly write our own jdbc url and toplink's session/uow
will automatically handle the actual sql query execution.
Is there any way we can configure Toplink to support RAC Fast Connection Failover, such
as possibly in the toplink's sessions.xml or in the oc4j container's connection pools setting?
Just for information, the way we acquire Toplink session is through SessionFactory:
   SessionFactory sessionFactory = new SessionFactory("repository",
                                 "META-INF/repository-sessions.xml");
   Session toplinkSession = sessionFactory.acquireSession();Then we do queries, acquiring unit of work, etc, using that toplinkSession.
Thanks.

We have enhanced our RAC FCF support in 11gR1 leveraging the new Universal Connection Pool to make it more seamless to the application. Which version of TopLink and OC4J are you using?
Doug

11i/10g RAC Failover Test

Hi,
Anybody did 11i/10g RAC Failover test. How to do the test and what steps involved. What is expected when one node down?
my config
web, admin, form - appsnode1
concurrent / db (rac1) - dbnode1
concurrent /db (rac2) - dbnode2
Thanks for your help.
Regards

You can check the below
1. VIP Failover - VIP should failover when node gets rebooted.
2. When one of the node is down you should still be able to logon to applications.
3. Since you have only one node and do not use PCP you CM processes will die.Make sure you start them.
4.Once you bring the node back , test your load balancing is happening fine across both the db nodes.
Let me know if this is helpful/correct and if you need more information.
Regards
Nitin Arora

Optimize rac failover time?

I have 2node RAC and the failover time is taking 4 minutes. Please advice some tips/documents/links that shows, how to optimize the rac failover time?
[email protected]

Hi
Could you provide some more information of what it is you are trying to achieve. I assume you are talking about a the time it takes for clients to start connecting to the available instance on the second node, could you clarify this?
There is SQLnet parameters that can be set, you can also make shadow connections with the preconnect parameter in your fail_over section of your tnsnames.ora on the clients.
Have you set both of your hosts as preferred in the service configuration on the RAC cluster. The impact will be less in a failure as approximately half of your connections will be unaffeced when an instance fails.
Cheers
Peter

Oracle RAC Failover (CTF) support

hi! every developers.
I am running Portal 6.0
In the WAS6.40 JDBC configuration,
Is this JDBC driver THIN?
and Is this connection String effective on failover (CTF) with Oracle RAC?
below is our current configration for reference.
Connection conn =
DriverManager.getConnection("jdbc:oracle:thin:@(DESCRIPTION =
(ADDRESS_LIST = (ADDRESS = (PROTOCOL = TCP) (HOST = *.*.*.11) (PORT = 1521))
(ADDRESS = (PROTOCOL = TCP) (HOST = *.*.*.12) (PORT = 1521))) (CONNECT_DATA =
(SERVICE_NAME = HIDB))) ", "scott", "tiger");
Does anything to need in my configuration WAS6.40?
thanks in advance.

Considering Oracle wrote DIF/DIX support for Linux, yes it does. It requires the use of ASMlib as well.
See Preventing Silent Data Corruption in Oracle Linux for more information.

RAC failover time problem!

I try TAF (transparent application failover) on RAC (9.0.1.3 and 9.2.0.1) and i have the same problem. When i play with "shutdown abort" the failover is fast (about 5-7 sec.). When i turn off the node, the failover is working fine, but it take too mutch time (about 3 minutes). Is there any parameter (tcp/ip or oracle net timeout, or keepalive parameter) that helps?
Thanks: Robert Gasz

Can you confirm that you are able to set up RAC with 9.2.0 on Linux?
Did you use the files downloadable from technet?
I have problems in joining the cluster from the second node (anyone) when
the first is up (anyone).
I hadn't this problem with 9.0.1.x version.
What is the release of Linuz you are using?
Are you using raw devices or lvm or what for your raw partitions?
Thanks in advance.
Bye,
Gianluca

RAC failover with HP-Serviceguard package over listener

We have two nodes (lcsnew1 and lcsnew3) on which RAC is installed. We are using HP-Serviceguard and HP-SGeRAC in the co-existence model, where RAC derives cluster membership information from SG and SGeRAC.
We are using the dedicated server model, where each node has one listener. We have configured HP-SG packages over RAC listener so as to provide failover capability. The HP-SG packages monitor database listener and on node failure or listener failure, the package moves to the other node and attaches itself with the listener running on the adoptive node. So, the basic setup is
Initially:
lcsnew1 lcsnew3
pkg4 pkg5
On listener / node failure at lcsnew1:
lcsnew1 lcsnew3
Not Available pkg4, pkg5
On listener / node failure at lcsnew3:
lcsnew1 lcsnew3
pkg4, pkg5 Not Available
We want to connect to any of the pkg IP Address, and if it failsover to another node, the connectivity should still be maintained.
I have a small Test programme which uses JDBC connectivity (thin) to connect at pkg4 virtual IP Address. It is able to get a connection and run some queries in a loop on the same connection. Then pkg4 is moved from lcsnew1 to lcsnew3 (at this time the test programme is in sleep mode). On the same old connection, we expect the query to run successfully. But the behaviour we are observing is random. Sometimes, the test programme remains connected to pkg4 IP address even after failover, while sometimes it gives ConnectionError. However, on again connecting to pkg4 IP address (which by now is in adoptive node lcsnew3), it works. So, basically fresh connections always work but the old connections do not failover every time.
I am enclosing all the configuration files for RAC (listener.ora, tnsnames.ora) and package configuration files (config, control). Also, the hosts file is enclosed which gives the details of all names and IP addresses. The test programme is also enclosed.
Pls have a look and suggest what how to move forward.
The listener.ora looks like
==========================================================
# listener.ora.lcsnew1 Network Configuration File: /opt/app/oracle/product/10.2.0/db_1/network/admin/listener.ora.lcsnew1
# Generated by Oracle configuration tools.
LISTENER_LCSNEW1 =
(DESCRIPTION_LIST =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC1))
(ADDRESS = (PROTOCOL = TCP)(HOST = lcsnew1-vip)(PORT = 1521)(IP = FIRST))
(ADDRESS = (PROTOCOL = TCP)(HOST = lcsnew1)(PORT = 1521))
SID_LIST_LISTENER_LCSNEW1 =
(SID_LIST =
(SID_DESC =
(SID_NAME = PLSExtProc)
(ORACLE_HOME = /opt/app/oracle/product/10.2.0/db_1)
(PROGRAM = extproc)
=========================================================
The tnsnames.ora looks like
=========================================================
# tnsnames.ora.lcsnew1 Network Configuration File: /opt/app/oracle/product/10.2.0/db_1/network/admin/tnsnames.ora.lcsnew1
# Generated by Oracle configuration tools.
GMLCDB2 =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = lcsnew3-vip)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = gmlcdb)
(INSTANCE_NAME = gmlcdb2)
GMLCDB1 =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = lcsnew1-vip)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = gmlcdb)
(INSTANCE_NAME = gmlcdb1)
LISTENERS_GMLCDB =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = lcsnew3-vip)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = lcsnew1-vip)(PORT = 1521))
GMLCDB =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = lcsnew3-vip)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = lcsnew1-vip)(PORT = 1521))
(LOAD_BALANCE = yes)
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = gmlcdb)
EXTPROC_CONNECTION_DATA =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC1))
(CONNECT_DATA =
(SID = PLSExtProc)
(PRESENTATION = RO)
========================================================
The package configuration file looks like
========================================================
PACKAGE_NAME     pkg4
PACKAGE_TYPE               FAILOVER
FAILOVER_POLICY               CONFIGURED_NODE
FAILBACK_POLICY               MANUAL
NODE_NAME     lcsnew1
NODE_NAME     lcsnew3
AUTO_RUN               YES
LOCAL_LAN_FAILOVER_ALLOWED     YES
NODE_FAIL_FAST_ENABLED          NO
RUN_SCRIPT               /etc/cmcluster/pkg4/pkg4.cntl
RUN_SCRIPT_TIMEOUT          NO_TIMEOUT
HALT_SCRIPT               /etc/cmcluster/pkg4/pkg4.cntl
HALT_SCRIPT_TIMEOUT          NO_TIMEOUT
The package control file looks like
==========================================================
. ${SGCONFFILE:=/etc/cmcluster.conf}
PATH=$SGSBIN:/usr/bin:/usr/sbin:/etc:/bin
CVM_ACTIVATION_CMD="vxdg -g \$DiskGroup set activation=exclusivewrite"
DEACTIVATION_RETRY_COUNT=0
KILL_PROCESSES_ACCESSING_RAW_DEVICES="NO"
VXVOL="vxvol -g \$DiskGroup startall"
FS_UMOUNT_COUNT=1
FS_MOUNT_RETRY_COUNT=0
CONCURRENT_VGCHANGE_OPERATIONS=1
CONCURRENT_FSCK_OPERATIONS=1
CONCURRENT_MOUNT_AND_UMOUNT_OPERATIONS=1
IP[0]="10.3.10.170"
SUBNET[0]="10.3.10.0"

Why do you want to use the HP-serviceguard even when you are using RAC on 10g Rel 2.
Oracle itself provide facility of TAF and now in 10g the concept of Vip.
all this functionality will be available by using TAF.
Kuljeet

Error on RAC failover process

Dear All,
i am using oracle 10.2.04 RAC configured database and below is my TNS entry.
ORCL =
(DESCRIPTION_LIST =
(LOAD_BALANCE = OFF)
(FAILOVER = ON)
(DESCRIPTION =
(ADDRESS_LIST =
(LOAD_BALANCE = ON)
(FAILOVER = ON)
(ADDRESS = (PROTOCOL = TCP)(HOST = host01.com)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = host02.com)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = host03.com)(PORT = 1521))
(CONNECT_DATA =
(SERVICE_NAME = ORCL)
(FAILOVER_MODE =
(TYPE = SESSION)
(METHOD = BASIC)
(RETRIES = 120)
(DELAY = 5)
Following are the TNS set up for three conection pools to db throught the web logic
jdbc:oracle:thin:@(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=host01.com)(PORT=1521))(CONNECT_DATA=(SERVICE_NAME=ORCL)))
jdbc:oracle:thin:@(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=host02.com)(PORT=1521))(CONNECT_DATA=(SERVICE_NAME=ORCL)))
jdbc:oracle:thin:@(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=host03.com)(PORT=1521))(CONNECT_DATA=(SERVICE_NAME=ORCL)))
today we have faced connectivity issue in one server and failover not worked. due to this we have to restart the manage servers.
can you kindly review and let me know any issues in my configuration? if we create 3 separate connection pools will that ignore the RAC fail over?
Also please tell me the correct way to setup this.
Thanks,

Hi,
I didn't understand something, you wrote your connection pool strings, but you have 3. How does your application server open these connections? Which string does it use?
JDBC thin client does not use TNS configuration, so if a specific connection from the pool got connected to host2 and this node crashed. How do you expect it to failover if it doesn't have any information other than host2?

Need test documents for RAC failover Scenarios

Hello friends...
By the end of this week i have to produce sum test documents for RAC and Database server including Sun Cluster Failover Scenarios.
Can sumone guide me to a link where i can get enough help.
I have already managed to get enough information.. but i want to see to it that i cover most of the topics.
Thanks, Regards
Monu Koshy

Please check the following links.
http://download-uk.oracle.com/docs/cd/B19306_01/rac.102/b14197/toc.htm
http://download-uk.oracle.com/docs/cd/B19306_01/install.102/b14205/toc.htm
-aijaz

Question on RAC Failover testing

Can someone point me to RAC Fail-over testing strategies?
Thanks,
Subroto

It totally depends on your application and your setup. Using TAF ? Using Load Balancing ? Using OCFS ? Using ASM ?
Kill the sqlnet listener and see what happens.
Kill the database pmon and see what happens.
Kill the ASM dbwr and see what happens.
Kill crsd and see what happens.
Shutdown the VIP interface and see what happens.
Shutdown the interconnect interface and see what happens.
Pull the power plug out and see what happens.

GG with RAC - failover

11.1.1 GG on Oracle 11gR2 - Source is single instance - Target is two node RAC .
We are planning to have HA for the GG process - If one node goes down we have to start processing from the second Node. - We are planning to have a shared file system for GG software - which will have all the metadata required shared across nodes - Can we use the RAC virtual IP for the pump process to ship the trail files - If one node goes down VIP fails over to the other node and we will have the trail files over there - Is it a supported config ? Can you share some metalink notes for configuring GG on RAC Target -
Thanks

Hi,
Please refer to the Oracle White Paper—Oracle GoldenGate high availability with Oracle Clusterware
for more information on configuring OGG with RAC & clusterware.
http://www.oracle.com/technetwork/middleware/goldengate/overview/ha-goldengate-whitepaper-128197.pdf
Thanks & Regards
SK

R12 with RAC - Failover

Hi,
Is it possible to have r12 Single Application tier with two node RAC database tier ?
Suppose if i have a 2 node rac and Node1 goes down, which of the following application services can fail over ( or say connect ) to second node.
1. self service application
2. Forms applicaton ?
3. CM
Can this happend dynamically or do i need to restart the application services ?
Is it manadatory to have multiple applications node if we have RAC at the backend ?

Hi,
You can implement RAC with single application tier node.
It is not mandatory to have multiple application nodes with RAC, but it is recommended (in terms of performance and availability).
I believe all your questions are answered in the following document.
Note: 388577.1 - Using Oracle 10g Release 2 Real Application Clusters and Automatic Storage Management with Oracle E-Business Suite Release 12
https://metalink2.oracle.com/metalink/plsql/ml2_documents.showDocument?p_database_id=NOT&p_id=388577.1
Regards,
Hussein

Oracle RAC installation failover

Hi,
I have an Oracle RAC installation with 2 nodes with the data stored on a shared OCFS partition. I had a client test the connection using jdbc string for RAC failover. I tried shutting down one of the nodes on the RAC installation and the client could not connect to the oracle cluster database for the next 5 to 10mins.
I understand that the client would failover to the next available listener (On the next retry connection) if the node it is currently listening to has failed. Is there any configuration i should make to increase the failover efficiency?
Thanks for any advice.

Hi,
Server side failover is arranged by setting the remote_listener parameter.
Client side failover is set by using T(ransparent) A(pplication) F(ailover) (9i and higher)
or F(ast)C(onnection)F(ailover). Both are documented in the Net administrators manual for the version you didn't care to mention.
As far as I know, both TAF and FCF are not supported by the JDBC thin driver.
Sybrand Bakker
Senior Oracle DBA

RAC failover

Similar Messages

Maybe you are looking for