Configuring Non RAC physical standby fro RAC

Hi,
EBS-R12i , Oracle Database:11g
We are in project design phase and following is the hardware setup for Critical banking domain:
Database (2 node RAC)-Solaris 10
Application server server (2 server) -Solaris 10
Web server (2 server for Failover)= RHEL
For DR
Database (1 node)-Solaris 10
need to know
1.What are the best practice to implement this kind of setup and steps to configure DR as Non-RAC physical standby for RAC database server.
2.What are implication if we go for single node except failover node unavailability for DR.
3.What are the steps of Switchover in this case.
4.What's the recommended backup solution for this setup of 8-10 Terabyte database with db growth rate of 50% per year.

1.What are the best practice to implement this kind of setup and steps to configure DR as Non-RAC physical standby for RAC database server.There really aren't any differences in setting up your physical standby in a non-RAC configuration. It is after the physical standby is setup that it is converted to a clustered database, if desired. In your case, you do not desire the conversion.
The best place to start for best practices in setting up your physical standby is the Oracle documentation on the subject.
http://download.oracle.com/docs/cd/E11882_01/server.112/e17022/toc.htm
2.What are implication if we go for single node except failover node unavailability for DR.The major implications are that you will be running with less horsepower after a failover. So depending on your DR resources, you may have to scale back application access. In addition to less resources, you will also have less availability because you only have one node. Should that node then fail, you will not have a node to failover to.
3.What are the steps of Switchover in this case.The steps are identical if the physical standby is a single node or multiple nodes. These steps are outlined in the Oracle documentation.
4.What's the recommended backup solution for this setup of 8-10 Terabyte database with db growth rate of 50% per year.What are your recovery requirements? Without knowing your recovery requirements, a valid backup solution cannot be determined. How much data loss can you afford in the event of a failure? Do you have data that can be read only? In the event of a failure, how long before you are required to be up and running? Are you required to restore and recover at an offsite location?
Cheers,
Brian

Similar Messages

How to Flashback Primary with Physical Standby

We have a dataguard configuration with one physical standby (which is the FSFO). We had an application cause problems in the primary database (basically what would be considered a user error) and we would like to flashback the primary database without having to drop and re-create the physical standby.
Does anyone have a basic set of procedures for doing this?
We are running 10.2.0.3 on windows 2003 server sp2
Thanks!

We have flashback turned on in both the primary and standby. We want to flashback the primary (and subsequently the standby) to some previous point in time.
In a regular non-data guard setup we would just shutdown and startup mount and issue the flashback database to timestamp (sysdate-1). But with dataguard it's a little more complicated because if we shutdown the primary then the observer will initiate a failover which we don't want. And when we startup mount and flashback the primary, how do we flashback the standby? manually or do we re-start the broker and let data guard do it?
I think it's going to require some specific steps and I was hoping someone had gone through this and could offer some suggestions.

DG physical standby 1

Hi
11.2.0.1
We have a DEV database which is also as critical as PROD db since there are lots of critical projects running on it.
This DEV database was configured with dataguard physical standby at the DR server.
Well this is what I read in the documents only, as I am newly taking over as dba. The setup was done outsourced
by 3rd party, and I can not talk to the implentors as they have been long gone.
When I check the DR server only the PROD is running and the DEV was shutdown for the last 12 months.
My question is if the physical standby is shutdown for almost a year, can it still be activated? or will it still be valid?
Will it affect performance of the Primary database? since it might still checking heartbeat to it? or holding up some processes like archivers?
How do you disable physical standby from the primary db?
Thanks
petra k.

Hello;
"if the physical standby is shutdown for almost a year, can it still be activated? or will it still be valid?"
No and no. Even a tiny production database would have generated and remove too much archive in a years time for the old standby to have ANY value. Rebuild it.
How do you disable physical standby from the primary db?
You could just do this: ( On the primary side )
alter system set log_archive_dest_state_2=defer;
And you could clean up all the parameters for DG too.
How to remove a Data Guard Configuration from Primary Database (Doc ID 733794.1)
Best Regards
mseberg

One primary and two Physical standby

Hi everyone,
I want configure two separate Physical standby one work in Redo apply and one in log apply mode.
is it posible?

user011232 wrote:
Hi everyone,
I want configure two separate Physical standby one work in Redo apply and one in log apply mode.
is it posible?Yes it can possible very much. You have to configure as below in remote destinations.
Example_:-
Log_archive_dest_2='service= stby1 lgwr async ';
Log_archive_dest_3='service= stby2 arch async ';
What is the plan behind that? just for testing purpose? Even me too not tested but you can try configuring as above. :)
Edited by: CKPT on Feb 14, 2012 5:46 PM

Physical standby database in RAC system

Hi EXperts,
We have a 4 nodes 11.2 cluster system. i will set a data guard as a disaster recovery mothods.
As a physical standby database set, does we need to configure same nodes (4 nodes) in other city as primary database cluster system?
Can we set physical standby database is in 2 nodes cluster system and primary database is in 4 nodes cluster system?
Thanks explaining!
Jin

Can we set physical standby database is in 2 nodes cluster system and primary database is in 4 nodes cluster system?Yes. You can setup 2 node RAC as physical standby and also with different memory parameters. But, the only issue is when DB failover from primary to standby, the applications connect to 2 node RAC in place of 4 node RAC and may be overloaded.

Related to recovery in Physical standby database in RAC

In Physical Standby database all instance involved in recovery processes or any one instance will be their in in recovery mode
If is like this another instance will be for high availabily can any one in forum explain me it will be help full

With your Primary database being RAC, the Physical Standby does not have to be RAC, although, obviously it would be preferred to have scalability as required at a DR site.
Whether your Standby is RAC or non-RAC, the automatic Recovery that is done at the Standby is done by one instance only. It would be applying the ArchiveLogs from all the threads (i.e. instances) of the Primary.
Database Recovery is always done by a single instance.
Hemant K Chitale
http://hemantoracledba.blogspot.com

Cascading physical standby from another physical standby for primary rac

my supplier told me that is possible mount a cascade physical standby from a physical standby for my rac installation.
but according the note "409013.1" seems not:
"Cascading a logical or physical standby database from a physical standby database is supported as long as the primary database is not a RAC database and is not part of a Data Guard Broker configuration - the support status of this configuration is unchanged from previous documentation."
before was possible? do you know why not now?
has anyone mounted? can I allow my provider to mount?
oracle version: 10.2.0.2.0

Is it possible this scenario?YES

RAC & DataGuard (Physical Standby)

Hello all,
I'm trying to get a high level overview of how RAC & DataGuard would behave in the following configuration. I've written down my understanding of how things would work. Please correct me if I'm wrong.
1) 2 node RAC (Primary Database) with a single instance physical standby.
a) Same standby related init.ora parameters would have to configured on both primary rac nodes.
b) The redo apply service at the standby would merge the redo from the 2 threads from the primary and apply it to the standby to keep it in sync.
c) During switch over only one primary RAC instance should be up besides the standby instance.
d) During switch back again only one primary RAC instance should be up besides the standby instance.
e) During failover ofcourse both primary instances would be down which warrants the fail over.
2) 2 node RAC (Primary) with a 2 node physical standby
This where it gets complex and I'm not really sure how a,b,c,d,e would be in this scenario. Appreciate if you could shed some light on this and list what
a,b,c,d,e would look like in this scenario.
I'm assuming that only one instance in the standby RAC should be up when the standby is in RAC configuration for the redo apply to work.
Also, if there is a white paper that details step by step procedure for setting up the above 2 scenarios, please let me know. So far I was able to find only the MAA white paper but that was not very helpful. If you can prescribe a good book for RAC & DataGuard that would be great too..
Thanks for your help

>
1) 2 node RAC (Primary Database) with a single instance physical standby.
a) Same standby related init.ora parameters would have to configured on both primary rac nodes.Usually rac nodes share their spfile on a shared diskvolume or through ASM. So they have identical DG parameter anyway
b) The redo apply service at the standby would merge the redo from the 2 threads from the primary and apply it to the standby to keep it in sync.Correct
c) During switch over only one primary RAC instance should be up besides the standby instance.
Sounds logical to me
Edit: In fact during normal operation both RAC nodes are up, so during switchover they are both active too
It is the configuration that matters. As soon as the standby is becoming primary, it might be that it talks only to one RAC node for its log-apply files
d) During switch back again only one primary RAC instance should be up besides the standby instance.That single was the only one running in step c.
e) During failover ofcourse both primary instances would be down which warrants the fail over.Correct. Keep in mind that during a failover the RAC config would be deconfigured from the DG setup and has to be added back again as standby.
You might have to take a look at: http://www.oracle.com/technology/deploy/availability/pdf/MAA_WP_10g_RACPrimarySingleInstancePhysicalStandby.pdf
>
2) 2 node RAC (Primary) with a 2 node physical standby
This where it gets complex and I'm not really sure how a,b,c,d,e would be in this scenario. Appreciate if you could shed some light on this and list what
a,b,c,d,e would look like in this scenario.
I'm assuming that only one instance in the standby RAC should be up when the standby is in RAC configuration for the redo apply to work.
No. RAC primary to RAC Standby would mean both nodes are up on both side. PrimNode1 send its redo logs to StandbyNode1 and the same for PrimNode2 and StandbyNode2
Have a look here: http://www.oracle.com/technology/deploy/availability/pdf/MAA_WP_10g_RACPrimaryRACPhysicalStandby.pdf
HTH,
FJFranken
My Blog: http://managingoracle.blogspot.com
Edited by: fjfranken on 16-jul-2010 1:13

Can I use data guard to create a RAC standby database for a non RAC primary

Hi,
we need to RAC our production database but the normal methods will mean a long outage. It is possible to create a standby as a single node RAC database and when ready do a graceful failover to the standby database and open it for business. The next step would be to create another RAC node from this on the original server.
servers are already cluster aware, using ASM etc
Oracle 10.2

Yes, you will be able to setup RAC stnadby for a non-RAC Primary. For primary it just needs a available destination for redo shipping it doesn't matter whether it's RAC enabled or not. And ofcourse you are using 10.2 anyway only one node will be running MRP and that is too in standby mount mode.
However since you have are using You may follow below sequence.
1. Setup a new standby as RAC enabled.
2. Perform a switchover.
3. Shutdown the Old primary (which is standby now).
4. Install CRS and RDBMS on the old primary and it's new node.
5. Modify the cluster_database=TRUE and cluster_database_instances=<required number of instances>.
With above modification mount the standby database in standby mode and start MRP.
6. Introduce the database and instances to the OCR using SRVCTL add command.
7. Once you your database is synchronized with Primary do a switchover.
9. Now you can repeat step 3 to 6 on the other site too. <- if you need your secondary site to be RAC enabled too
10. Finally both the sites should be RAC enabled.
Hope this is helpful!!!
Thanks,
Asif Haliyal

How to create 11.2.0.2 physical standby database from 2 node RAC (11.2.0.2)

Hi,
Can any one please help me How to manually create 11.2.0.2 standalone physical standby database from 2 node RAC (11.2.0.2) database which is running in RHEL5 and ASM plugged in.
DB : 11.2.0.2
OS : RHEL5
RMAN duplicate is causing problem with network and we decided to go for manual creation of the same.
Thanks in Advance..

Hi;
Can any one please help me How to manually create 11.2.0.2 standalone physical standby database from 2 node RAC (11.2.0.2) database which is running in RHEL5 and ASM plugged in.
DB : 11.2.0.2
OS : RHEL5I had similar issue, what i did
1. Used source oracle_home on standby server
2. Created new asm instance and use same naming
3. I took RMAN full backup on source and move it to target
4. I edit initora file remove RAC setting and restore db(also edited listener file)
Regard
Helios

Rac to non rac standby file name convert option

Hi,
Rac to non rac with ASM both, please find the below details.
Primary Setup.
diskgroup NAME for all files for both primary and standby +DATA
pmon,unique and service name at standby-rac1, rac2
db version -11.2.0.1
platform - HP -UX
$ ps -ef | grep pmon
oracle11 22329 1 0 22:33:55 ? 0:28 ora_pmon_rac1
oragrid 23522 1 0 Jan 2 ? 1:30 asm_pmon_+ASM1
$ ps -ef | grep pmon
oracle11 22329 1 0 22:33:55 ? 0:28 ora_pmon_rac2
oragrid 23522 1 0 Jan 2 ? 1:30 asm_pmon_+ASM2
database and Unique name - RAC, so all files on ASM created under +DATA/RAC
Stanby setup
diskgroup NAME for all files for both primary and standby +DATA
pmon,unique and service name at standby- racdr
$ ps -ef | grep pmon
oracle11 22329 1 0 22:33:55 ? 0:28 ora_pmon_racdr
oragrid 23522 1 0 Jan 2 ? 1:30 asm_pmon_+ASM
database name and unique name is racdr, so all files on ASM created under DATA/racdr when cloned with RMAN from primary backup. (by settling log_file and db_file covert from DATA/RAC to +DATA/racdr )
so now everything ok... but how do we create a new datafiles at primary for the above setting success, ??? and reflect on standby ???.
1.create tablespace tbs size 10g;
2.create tablespace tbs datafile '+/DATA/RAC/test.dbf size 10g; ??
3.create tablespace tbs datafile '+/DATA size 10g;???
4. directories name on asm and specifying on the convert options is case sensitive ???
Thanks in advance...

thats fine john,
log_file and db_file covert from '+DATA/RAC', '+DATA/racdr'
my primary is rac
my standby us racdr
so i mentioned above folders on convert option, so if i created datafile with or without specifying a diskgroup like mentione below, with or without specifying the directory like below
will all create on under '+DATA/racdr' diskgroup ??
create tablespace test DATA size10m; --> not specidying a full path as DATA/RAC' so will it go to DATA/racdr' ?? or creating under DATA ?diskgroup on standby????
Thanks,

Configuring multiple listeners in a non-RAC env.

Hi,
Our database is being made as many as 35 requests per second by the application during peak hours. This is causing the listener to die. Metalink suggested to configure more listeners so that they can balance the load. When I tried to configure 2 listeners on the local server, one of the listeners are not working propoerly.
I am posting the listenerr.ora and their statuses along with the error messages here. Can someone help me understand the problem?
Listener.ora
LISTENER2 =
(DESCRIPTION_LIST =
    (DESCRIPTION =
      (ADDRESS = (PROTOCOL = TCP)(HOST = dbdev.website.com)(PORT = 1522))
SID_LIST_LISTENER2 =
(SID_LIST =
    (SID_DESC =
      (SID_NAME = TESTDB)
      (ORACLE_HOME = /u02/app/oracle)
      (PROGRAM = extproc)
LISTENER1 =
(DESCRIPTION_LIST =
    (DESCRIPTION =
      (ADDRESS = (PROTOCOL = TCP)(HOST = dbdev.website.com)(PORT = 1521))
      (ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC0))
SID_LIST_LISTENER1 =
(SID_LIST =
    (SID_DESC =
      (SID_NAME = PLSExtProc)
      (ORACLE_HOME = /u02/app/oracle)
      (PROGRAM = extproc)
lsnrctl status <listener_name>
oracle@ DBDEV $ lsnrctl status LISTENER1
LSNRCTL for Linux: Version 10.2.0.3.0 - Production on 02-AUG-2009 20:08:03
Copyright (c) 1991, 2006, Oracle. All rights reserved.
Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=dbdev.website.com)(PORT=1521)))
STATUS of the LISTENER
Alias                     LISTENER1
Version                   TNSLSNR for Linux: Version 10.2.0.3.0 - Production
Start Date                29-JUL-2009 22:28:31
Uptime                    3 days 21 hr. 39 min. 32 sec
Trace Level               off
Security                  ON: Local OS Authentication
SNMP                      OFF
Listener Parameter File   /u02/app/oracle/network/admin/listener.ora
Listener Log File         /u02/app/oracle/network/log/arogya1.log
Listening Endpoints Summary...
(DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=dbdev.website.com)(PORT=1521)))
(DESCRIPTION=(ADDRESS=(PROTOCOL=ipc)(KEY=EXTPROC0)))
Services Summary...
Service "PLSExtProc" has 1 instance(s).
Instance "PLSExtProc", status UNKNOWN, has 1 handler(s) for this service...
Service "TESTDB" has 1 instance(s).
Instance "TESTDB", status READY, has 1 handler(s) for this service...
Service "TESTDB_XPT" has 1 instance(s).
Instance "TESTDB", status READY, has 1 handler(s) for this service...
The command completed successfully
oracle@ DBDEV$ lsnrctl status LISTENER2
LSNRCTL for Linux: Version 10.2.0.3.0 - Production on 02-AUG-2009 20:08:07
Copyright (c) 1991, 2006, Oracle. All rights reserved.
Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=dbdev.website.com)(PORT=1522)))
STATUS of the LISTENER
Alias                     LISTENER2
Version                   TNSLSNR for Linux: Version 10.2.0.3.0 - Production
Start Date                31-JUL-2009 02:37:10
Uptime                    2 days 17 hr. 30 min. 56 sec
Trace Level               off
Security                  ON: Local OS Authentication
SNMP                      OFF
Listener Parameter File   /u02/app/oracle/network/admin/listener.ora
Listener Log File         /u02/app/oracle/network/log/arogya2.log
Listening Endpoints Summary...
(DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=dbdev.website.com)(PORT=1522)))
Services Summary...
Service "TESTDB" has 1 instance(s).
Instance "TESTDB", status UNKNOWN, has 1 handler(s) for this service...
The command completed successfully
ps -ef | grep tnslsnr
oracle@ DBDEV $ ps -ef | grep tnslsnr
oracle    7951     1 0 Jul29 ?        00:00:00 /u02/app/oracle/bin/tnslsnr LISTENER1 -inherit
oracle   18080     1 0 Jul31 ?        00:00:00 /u02/app/oracle/bin/tnslsnr LISTENER2 -inherit
Client message
C:\Oracle\product\10.1.0\Client_1\BIN>sqlplus system/manager@asridb1
SQL*Plus: Release 10.1.0.2.0 - Production on Mon Aug 3 09:32:18 2009
Copyright (c) 1982, 2004, Oracle. All rights reserved.
Connected to:
Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - Production
With the Partitioning, OLAP and Data Mining options
SQL>
SQL> Disconnected from Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - Production
With the Partitioning, OLAP and Data Mining options
C:\Oracle\product\10.1.0\Client_1\BIN>sqlplus system/manager@asridb2
SQL*Plus: Release 10.1.0.2.0 - Production on Mon Aug 3 09:32:25 2009
Copyright (c) 1982, 2004, Oracle. All rights reserved.
ERROR:
ORA-28547: connection to server failed, probable Net8 admin error
Enter user-name:Thanks,
Aswin.

ice_cold_aswin wrote:
But shared server is not always better than a dedicated server.I did not say it was - only that if you're expecting loads of brand new connection requests to be services fast by the Listener, then shared server will be significantly faster than dedicated server.
Frequency of connections may be handled better by the use of shared server, but how about holding thousands and thousands of connections and serving all their requests by using some preemptive mechanism? Many years ago, I configured an Oracle 8i instance on my desktop PC with shared server and told the Java guys to try their worst. See how many threads they can open against my desktop PC instance versus the big HP server instance. The latter was using dedicated server and they were kind of thinking that this was somehow better than shared server. They could open less than a few hundred connections to the big server. They opened over a 1000 connections to my desktop instance.
Shared Server is and never was a poor and second choice to Dedicated Server as some seems to suggest. It is extremely capable and ideally suited when dealing with a huge number of connections.
In our application, number of connections would reach thousands during peak hours and large number of REALLY BIG queries would be fired by each of them in a small time period. I would say these BIG QUERIES can be termed as OLAP queries, while our main stream application is a OLTP system. It must be apparent to you by now, our database (non-RAC) is supporting an application that gives OLAP and OLTP queries at a very high frequency. A big query ito what? Resource footprint? Source code size of the query? Execution time of the query?
If you have 100's of OLAP type queries (complex and long running queries) hitting the server every few seconds, then your server will die. Unless you are running some seriously expensive hardware.
That aside - it is only a sensible approach to use shared server in an environment that deals with 1000's of OLTP type connections. And if there is an OLAP mix too, then configure those connections to request dedicated server connections.
Why? The shared server scales a lot better than dedicated server. And scalability is what you want in this type of environment.

CREATING A SINGLE INSTANCE PHYSICAL STANDBY FOR A RAC PRIMARY

Hi
Creating a single instance physical standby database for a RAC Primary.
Getting this error.
sql statement: alter database mount standby database
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03002: failure of Duplicate Db command at 01/17/2008 23:05:38
RMAN-03015: error occurred in stored script Memory Script
RMAN-03009: failure of sql command on clone_default channel at 01/17/2008 23:05:38
RMAN-11003: failure during parse/execution of SQL statement: alter database mount standby database
ORA-01103: database name 'PROD' in control file is not 'DPROD'
Any help on this.
Regards
Satish

The problem here is probably with your standby init.ora file.
When you create a standby database, the db_name parameter must NOT change. It has to match the primary database. So in your case, db_name ='PROD' and your db_unique_name='DPROD'...
-peter

How to create a standby redolog from rac to non rac ??

Hi,
How to create a standby redolog from rac to non rac DR setup..???
in rac we can create with specifying thread number for each instances..... but this will be replicating to standby ....so how this will act/create on single DR??
pls help

854393 wrote:
Thanks Shivanandha,
(maximum number of logfiles for each thread + 1) * maximum number of threads
Using this equation reduces the likelihood that the primary instance's log writer (LGWR) process will be blocked because a standby redo log file cannot be allocated on the standby database. For example, if the primary database has 2 log files for each thread and 2 threads, then 6 standby redo log file groups are needed on the standby database.In maximum performance mode you can keep same number of redo logs but having one more redo log member for each thread is recommended and mandtory for other protection modes.
How to create a standby redolog from rac to non rac DR setup..???
in rac we can create with specifying thread number for each instances..... but this will be replicating to standby ....so how this will act/create on single DR??pls help
Yes, even your DR is RAC or non-RAC, you must have both threads, because each instance information exist on own thread, So to perform redo transition from both the instances, you must have both standby redo log threads even though if it is non an RAC DR.
Hope This helps.

Create a physical standby for Rac on Windows and raw devices

Hi buddies:
I need to create a Physical standby database for a 2 node RAC cluster on Windows, that is using RAW devices.
Can someone of you suggest me a note to follow in order to achive this, i can not find to much information about how to do this.
I guess that the best option would be to use RMAN to do this, but as i said before i havent found too much information.
I would really appreciate your help.
Kind regards
MQ

In this case the standby should be created as a single instance on regular files.
Do i have to specify something different in the duplicate rman command to specify that the standby database will be a single instance?
Thanks for your help
MQ

Configuring Non RAC physical standby fro RAC

Similar Messages

Maybe you are looking for