Data Guard Scenario

Hi,
We have a data guard configuration with 1 primary and 2 standy databases. Two databases are local means physically located next to each other and one is across the WAN. The 2 local databases are candidate of fast start failover and we also have a observer on separate box. For testing purposes, when we pull the network cables from 2 local databases to make third one primary, the whole data guard broker hangs and doesn't allow us to run any command using DGMGRL to manually failover to third database running across WAN. So we have to disable the dg broker process on third database manually and convert it to primary without broker. Then when we put the network cable back for those 2 local servers, the observer thinks that one of that is primary and try to start that database as primary and also try to convert third database as standby by flashing it back to previous scn and it never succeeded. In order to make everything work, we have to manually convert the roles of all the databases and make them in sync and delete the broker configuration and readd it. It's too messy to deal with, I am wondering if there is any clean way to failover back to local databases.
Database Version: 10.2.0.4
Platform: Linux
Thanks
Daljit Singh

Thanks for the reply!
Yes, all the 3 databases are being observed by the observer which is running on 4th box but the third database is not candidate of fast start failover. when we pull the cable from 2 servers, the observer box is still able to ping the third box server. Both of the standby are in sync with primary, there is no problem converting third database to primary manually by shutting down dg broker and doing it using alter database command. But the problem is when we do this then the broker configuration doesn't get updated and when we try to put other 2 databases online, broker think that the local database is primary and then all the problem starts. Before all this test, everything was working, configuration was perfect, and both the standbys were catching up.
Mainly I want to find out if there is any way to failover to third database using broker when we pull network cables from local stores. Right now, as I mentioned, it hang when we pull cables from local servers and it doesn't allow us to run any command from dgmgrl, all commands failed with time out error.
Thanks
Daljit Singh
Edited by: Daljit on Sep 30, 2009 11:17 AM

Similar Messages

Frequent ORA-00308 in Oracle Data Guard Scenario

Hi All,
We have oracle dataguard configured for DR for 9 production systems.
ODG works fine for all the systems, however for SAP SCM system I frequently get ORA-00308 in primary system .
The redo log generation rate in this system is high (Its like 70 logs 600MB each per day)
I get a gap of 2 to 3 logs & have to fix it manually by transfering the logs & register them manually.
Error in alert log of primary
ORA-16055: FAL request rejected
ARCH: FAL archive failed. Archiver continuing
Tue Mar 8 20:36:21 2011
ORACLE Instance SCP - Archival Error. Archiver continuing.
Tue Mar 8 20:36:21 2011
ARC0: Attempting destination LOG_ARCHIVE_DEST_2 network reconnect (3135)
ARC0: Destination LOG_ARCHIVE_DEST_2 network reconnect abandoned
Tue Mar 8 20:36:21 2011
Errors in file /oracle/SCP/saptrace/background/scp_arc0_25158.trc:
ORA-03135: connection lost contact
FAL[server, ARC0]: FAL archive failed, see trace file.
Tue Mar 8 20:36:21 2011
Errors in file /oracle/SCP/saptrace/background/scp_arc0_25158.trc:
ORA-16055: FAL request rejected
ARCH: FAL archive failed. Archiver continuing
Tue Mar 8 20:36:21 2011
ORACLE Instance SCP - Archival Error. Archiver continuing.
Tue Mar 8 20:36:21 2011
Errors in file /oracle/SCP/saptrace/background/scp_arc3_25164.trc:
ORA-00308: cannot open archived log '/oracle/SCP/oraarch/SCParch1_45166_698614348.dbf'
ORA-27037: unable to obtain file status
HPUX-ia64 Error: 2: No such file or directory
Additional information: 3
Tue Mar 8 20:36:21 2011
FAL[server, ARC3]: FAL archive failed, see trace file.
Tue Mar 8 20:36:21 2011
Errors in file /oracle/SCP/saptrace/background/scp_arc3_25164.trc:
ORA-16055: FAL request rejected
ARCH: FAL archive failed. Archiver continuing
Tue Mar 8 20:36:21 2011
ORACLE Instance SCP - Archival Error. Archiver continuing.
Tue Mar 8 20:36:52 2011
Redo Shipping Client Connected as PUBLIC
-- Connected User is Valid
Tue Mar 8 20:36:52 2011
Errors in file /oracle/SCP/saptrace/background/scp_arc2_25162.trc:
ORA-00308: cannot open archived log '/oracle/SCP/oraarch/SCParch1_45166_698614348.dbf'
ORA-27037: unable to obtain file status
HPUX-ia64 Error: 2: No such file or directory
Additional information: 3
Tue Mar 8 20:36:52 2011
FAL[server, ARC2]: FAL archive failed, see trace file.
Tue Mar 8 20:36:52 2011
Errors in file /oracle/SCP/saptrace/background/scp_arc2_25162.trc:
ORA-16055: FAL request rejected
ARCH: FAL archive failed. Archiver continuing
Tue Mar 8 20:36:52 2011
ORACLE Instance SCP - Archival Error. Archiver continuing.
Tue Mar 8 20:36:52 2011
Errors in file /oracle/SCP/saptrace/background/scp_arc3_25164.trc:
ORA-00308: cannot open archived log '/oracle/SCP/oraarch/SCParch1_45167_698614348.dbf'
ORA-27037: unable to obtain file status
HPUX-ia64 Error: 2: No such file or directory
Additional information: 3
v$archived_log query on DR gives the following output
SEQUENCE# ARC APP
     45155 YES YES
     45156 YES YES
     45157 YES YES
     45158 YES YES
     45159 YES YES
     45160 YES YES
     45161 YES YES
     45162 YES YES
     45163 YES YES
     45164 YES YES
     45165 YES YES
SEQUENCE# ARC APP
     45168 YES NO
     45169 YES NO
     45170 YES NO
     45171 YES NO
     45172 YES NO
     45173 YES NO
     45174 YES NO
     45175 YES NO
     45176 YES NO
     45177 YES NO
     45178 YES NO
Now, how to get to the root cause of this issue?
Is it because of the network?
Also, there is an archive policy which runs on primary , it deletes archives every 8 hours
Thanks,
Salman

Hi Stefan,
For this system, we have set up Arcn to ship the logs. For all other systems its LGWR async.
Oracle version is :
Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - 64bit
PL/SQL Release 10.2.0.4.0 - Production
CORE    10.2.0.4.0      Production
TNS for HPUX: Version 10.2.0.4.0 - Production
NLSRTL Version 10.2.0.4.0 - Production
Please let me know if you need some more information
Thanks,
Salman

Data guard synchronization after link down b/w primary and physical standby

Hi All,
I have configured data guard on oracle 11gr2 db. Normally switchover between my primary and physical standby happens smoothly and the Apply lag would be zero. Recently We had to test a scenario when the network link between Primary and Physical standby is completely down and Physical standby is isolated completely for more than half an hour.
When we brought up the link every thing worked smoothly but apply lag started increasing from 0 to around 3 hrs. And then it started reducing to 0. Currently Apply lag and transport lag shows 0.
But is this normal behaviour of oracle data guard that when the link between primary and physical standby is completely down, It requires 3-4 hrs for resynchronization ??? Even when during isolation, there were very few transactions happend on primary database ??
Are there any documents available for this scenario??
Thanks

Hi, after the link is up, if there were some transactions and produced archive logs it's normal to take some time for resync. To check if 3-4 hours is normal or not, you can repeat the scenario and this time check
- how many archivelogs does primary produce in this period.
- after the link is up, does archivelog transfer immediately starts from primary to standby? Is primary able to send these archivelogs parallel?
- Is there anything wrong with the apply process?
check primary & standby alert log files, and run this query on standby to check the transport and apply processes:
SELECT PROCESS, STATUS, THREAD#, SEQUENCE#, BLOCK#, BLOCKS FROM V$MANAGED_STANDBY;
regards

Need suggestion on Active data guard or Logical Stand by

Hi All,
Need a suggestion of on below scenario.
We have a production database ( oracle version 11g R2 ) and planning to have a Logical standby or physical standy (Active data guard). Our usage of the standby database is below.
1) Planning to run online reports (100+) 24x7. So might create additional indexes,materialized views etc.
2) daily data feed ( around 300+ data files ) to data warehouse. daily night, jobs will be scheduled to extract data and send to warehouse. Might need additional tables for jobs usage.
Please suggest which one is good.
Regards,
vara.

Hello,
In active dataguad Whig is feature from 11gRx ,
If you choose active dataguard, you have couple of good options, one is you can make a high availability of your production database, which can act as image copy of production, as you are asking in 11g you have more advantage where you can open in read only mode and at the sometime MRP will be active, so you can redirect users to connect standby to perform select operations for reporting purpose. So that you can control much load on production ,
Even uou can perform switchover in case of role change, perform failover if your primary is completely lost. Also you can convert to physical to logical standby databases & you can configure FSFO
You have plenty of options with active dataguard.
Refer http://www.orafaq.com/node/957
consider closing the thread if answered and keep the forum clean.
>
User Profile for user11261773
user11261773
Handle:     user11261773
Status Level:     Newbie
Registered:     Jul 14, 2011
Total Posts:     12
Total Questions:     6 (5 unresolved)
>
Edited by: CKPT on Mar 18, 2012 8:14 PM

Error: ORA-16532: Data Guard broker configuration does not exist

Hi there folks. Hope everyone is having a nice weekend.
Anyways, we have a 10.2.0.4 rac primary and a 10.2.0.4 standby physical standby. We recently did a switchover and the dgbroker files automatically got created in the Oracle_home/dbs location of the primary. Now need to move these files to the common ASM DG. For this, I followd the steps from this doc:
How To Move Dataguard Broker Configuration File On ASM Filesystem (Doc ID 839794.1)
The only exception to this case is that I have to do this on a Primary and not a standby so I am disabling and enabling the Primary(and not standby as mentioned in below steps)
To rename the broker configuration files in STANDBY to FRA/MYSTD/broker1.dat and FRA/MYSTD/broker2.dat, Follow the below steps
1. Disable the standby database from within the broker configuration
DGMGRL> disable database MYSTD;
2. Stop the broker on the standby
SQL> alter system set dg_broker_start = FALSE;
3. Set the dg_broker_config_file1 & 2 parameters on the standby to the appropriate location required.
SQL> alter system set dg_broker_config_file1 = '+FRA/MYSTD/broker1.dat';
SQL> alter system set dg_broker_config_file2 = '+FRA/MYSTD/broker2.dat'
4. Restart the broker on the standby
SQL> alter system set dg_broker_start = TRUE
5. From the primary, enable the standby
DGMGRL> enable database MYSTD;
6. Broker configuration files will be created in the new ASM location.
I did so but when I try to enable the Primary back I get this:
Error: ORA-16532: Data Guard broker configuration does not exist
Configuration details cannot be determined by DGMGRL
Form this link,(Errors setting up DataGuard Broker it would seem that I would need to recreate the configuration....Is that correct ? If yes then how come Metalink is missing this info of recreating the configuration... OR is it that that scenario wouldnt be applicable in my case ?
Thanks for your help.

Yes I can confirm from the gv$spparameter view that the changes are effective for all 3 instances. From the alert log the alter system didnt throw u pany errros. I didnt restart the instances though since I dont have the approvals yet. But I dont think thats required.

DB_FILE_NAME_CONVERT in DATA GUARD

Hi All,
when can i use the below scenario in Data gurad (standby parameter file) ?
DB_FILE_NAME_CONVERT='<DB_UNIQUE_NAME of PRIMARY>','<DB_UNIQUE_NAME of STANDBY>'
LOG_FILE_NAME_CONVERT='<DB_UNIQUE_NAME of PRIMARY>','<DB_UNIQUE_NAME of STANDBY>'
Thanks

Hi,
No worries. In the same doc it states:
Specify the path name and filename location of the primary database datafiles followed by the standby location. This parameter converts the path names of the primary database datafiles to the standby datafile path names. If the standby database is on the same system as the primary database or if the directory structure where the datafiles are located on the standby site is different from the primary site, then this parameter is required. Note that this parameter is used only to convert path names for physical standby databases. Multiple pairs of paths may be specified by this parameter.I've got data guard running on two sites which are physically separate to the primary site. At one site we have the same paths for the data files so we have nothing for the db_file_name_convert parameter. At the other site we have different paths so we specify the full path of the primary, followed by the full path of the standby. Just the way that I have in the example above.
Have you configured your standby yet? If I were you I would keep the file paths the same if possible. It leads to fewer issues in the long run. If that's not possible, for whatever reason, put the full path of the primary followed by the full path of the standby.
Hope that makes sense?
Rob

Disable Oracle Data Guard - Urgent

Hello All,
I am in a peculiar situation and any help is appreciated.
I have a Primay Database and one Physical standby database and they are in Sync. Now, my client wants me to disable data guard and run these two databases as independent one. While performing this activity, there should not be any downtime on Primary. There is no data guard broker involved.
PRIMARY SETUP
==========
*.log_archive_config='dg_config=(APR1,DRAPR1)'#used
*.LOG_ARCHIVE_DEST_1='LOCATION=/PR1_u12/oraAPR1/archive VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=APR1'#used
*.LOG_ARCHIVE_DEST_2='SERVICE=DRAPR1 LGWR ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=DRAPR1'#used
*.LOG_ARCHIVE_DEST_STATE_1='ENABLE'
*.LOG_ARCHIVE_DEST_STATE_2='ENABLE'#used
.standby_archive_dest='/PR1_u12/oraAPR1/archive'#used
PHYSICAL STANDBY SETUP
=================
*.fal_client='DRAPR1'#used
*.fal_server='APR1'#used
*.log_archive_config='dg_config=(DRAPR1,APR1)'#used
*.LOG_ARCHIVE_DEST_1='LOCATION=/PR1_u12/oraAPR1/archive VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=DRPR1'#used
*.LOG_ARCHIVE_DEST_2='SERVICE=APR1 LGWR ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=APR1'#used
*.LOG_ARCHIVE_DEST_STATE_1='ENABLE'#used
*.LOG_ARCHIVE_DEST_STATE_2='ENABLE'#used
*.standby_archive_dest='/PR1_u12/oraAPR1/archive'#used
*.standby_file_management='auto'#used
How do i make these two databases as independant database without having downtime in PRIMARY. Please help.

Hi,
++ First steps will be to defer the remote archiving by setting log_archive_dest_state_2 = defer in primary and then unset all dataguard related parameters on primary and standby like fal_server, fal_client, log_archive_dest_2, log_archive_config.
++ After that perform failover steps on standby to activate it as independent database using the steps mentioned in following link:
http://download.oracle.com/docs/cd/B19306_01/server.102/b14239/role_management.htm#i1026491
or
http://download.oracle.com/docs/cd/B19306_01/server.102/b14239/scenarios.htm#i1035282
Regards
Anudeep

Data Guard as a pass through?

Scenario is . . .
Host A is the Primary
Host B is a Standby
Host C is a Standby
Now I know we can set up A->B and A->C
Can we set up A->B->C ?
Essentially using B as a pass through between A and C. Or you can see it as B being in a DMZ.
Would B have to be Active Data Guard or anything special?
I guess what I am really asking is can a Standby be used as the source for another Standby.

See http://download.oracle.com/docs/cd/E11882_01/server.112/e10700/cascade_appx.htm#i638620 for more information on Cascaded Standby Destinations. There are a few restrictions:
Cascading has the following restrictions:
* Logical and snapshot standby databases cannot cascade primary database redo.
* SYNC destinations cannot cascade primary database redo in a Maximum Protection Data Guard configuration.
* Cascading is not supported in Data Guard configurations that contain an Oracle Real Applications Cluster (RAC) primary database.
* Cascading is not supported in Data Guard broker configurations.
Keep an eye on this chapter and Note 409013.1 "Cascaded Standby Databases" when the next patch set for 11.2 comes out :^)
Larry

Data Guard Physical Standby Failover

I need clarification on Physical Standby Failovers. Say I have a Primary db (A) and a physical standby (B), (A) fails and (B) is now the primary, can (A) be setup to become a standby of (B)? I've read 2 different statements in Oracle 10g Data Guard Concepts and Admin
"During failovers involving a physical standby database: In all cases, after a failover, the original primary database can no longer participate in the Data Guard configuration."
vice
"After a fast-start failover occurs, the old primary database will automatically reconfigure itself as a new standby database upon reconnection to the configuration."
Much thanks.

In all cases, after a failover, the original primary database can no longer participate in the Data Guard configuration.
TRUE ! But this doesn't mean that you cannot make your failed primary a new standby database !.
You can achieve this using FLASHBACK DATABASE option:
From the documentation:
After a failover occurs, the original primary database can no longer participate in the Data Guard configuration until it is repaired and established as a standby database in the new configuration. To do this, you can use the Flashback Database feature to recover the failed primary database to a point in time before the failover occurred, and then convert it into a physical or logical standby database in the new configuration.
Read the following documents:
http://download.oracle.com/docs/cd/B19306_01/server.102/b14239/role_management.htm#sthref995
http://download.oracle.com/docs/cd/B19306_01/server.102/b14239/scenarios.htm#i1049997
In addition to this I'd like to mention that you can achieve the same with doing an incomplete recovery instead of using FLASHBACK DATABASE.
"After a fast-start failover occurs, the old primary database will automatically reconfigure itself as a new standby database upon reconnection to the configuration."
TRUE. If you enable Fast-Start Failover:
http://download.oracle.com/docs/cd/B19306_01/server.102/b14230/sofo.htm#CHDDFFEC
Cheers!

Data Guard vs. Streams

What is the difference between Data Guard and Streams?
Which one should I choose in order that two database systems in different servers are synchronized with each other? There are two situations:
1. One database system is in Tru64 UNIX operating system, the other in Windows 2000 Server OS.
2. Both database systems are in Tru64 UNIX OS.
Thanks,

Sorry, I missed that wrinkle.
You'll need to have machines with the same O/S to have Data Guard work. If your requirements are to have the databases completely synced, you'll need to use Data Guard, so you'll need a new machine.
You can roll your own process using streams without buying a new machine, but it won't be as robust as the Oracle solution and won't handle all the possible failover scenarios, etc. A lot depends on the strictness of your requirements.
Justin
Justin

EMC SRDF or Data Guard

Friends, Is it advisable to have data guard as a replication option (for DR site) for a core banking database sizing close to 4 TB for a financial institution. As of now the customer is using SRDF for replication and its working perfectly fine. Sometimes, we will be failing back to production immediately after doing a DR Failover in case of any issues in in application layer or some other in DR site and is easy to do so with SRDF. What will be the scenario in such circumstances if we use data guard. Your valuable advices please. …

>>In case of failover, do we need to restore the database from DR to production site again to initiate a reverse replication..?
Primary (PD) -- After fail over - Standby (SD) becomes Primary.
After disaster/Failover Primary will be gone. Now you have to build Standby database from current Primary (that was your SD before disaster).
As for as I remember if you doing Switchover and using Flashback database you do not need to build them.
HTH,
Pradeep

Oracle10gR2-Data Guard

I want to configure Oracle10gR2 Data Guard for the following scenario:
•     In a Data Guard environment, we want to configure One Primary Server and Two Standby databases.
•     The Primary server will be using the internal hard drives for storage (operating system, Oracle, production database, logs, etc).
•     The First Standby database server will be connected to the Primary Server on LAN. This server will be using the SAN as its storage. The internal drives will be used for the operating system only.
•     The Second Standby Database would be hosted in a Remote site and connected through Optical Fiber. This database server will also use SAN as its storage.
Any body can help me out if there are any issues regarding SAN in a data guard environment?
Regards

No issues.

OS and DB upgrades to databases under data guard

Hello,
Here is the scenario. I have an OPS db and a physical standby of OPS. Platform is solaris 9 with both a 9i instance and a 10g instance. I want to be able to do a switchover of the ops to the standby and then upgrade the standby. I believe that I can upgrade the OS to solaris 10 with no problems. However, I also want to be able to upgrade the 9i instance to 10gR2. Is it possible to do an upgrade of this degree to a standby under data guard? At the point that the db upgrade has completed data guard will no longer have the same configuration. This scenario is starting to send off alarms to me. Thanks in advance for your responses.

Hi,
Is it possible to do an upgrade of this degree to a standby under data guard? Yes. Steve Karam does it all the time, and he has these notes:
"Dataguard's physical backups are a GREAT DR situation for a failed upgrade. You will have an exact physical copy of your database ready to switchover to if anything goes wrong with the upgrade. Basically you switch your final archive log and shut down your primary, then begin the standard upgrade.
If something goes wrong, you can perform a failover on your standby server and you're back up with no possibility of inconsistencies since it's a physical backup. In my opinion, the only method better than DataGuard for a backup for upgrade scenario would be to use SAN tools like NetApp's snapMirror or Sun's
Availability Suite in order to back it up; these tools can snapshot a DB in a matter of seconds and are good for this sort of situation: if the upgrade goes wrong, simply restore the snapshot.
If you do it right, you can set up dataguard with only 5 minutes of downtime, then perform your database upgrade with only 5 minutes of downtime. I have done this method for clients in the past and they have been very happy with the results."
Bipul Kumar has a book devoted to using Data Guard, that you may find helpful, and he notes the steps for a Data Guard-based upgrade:
http://www.rampant-books.com/book_2004_2_dataguard.htm
"Upgrading to Oracle10g database software in a Data Guard environment involves the following activities:
If the Data Guard configuration is managed by Data Guard broker, the broker configuration needs to be removed using the existing version of database software. In other words, remove the Oracle9i configuration before upgrading to Oracle10g. Execute the following command after connecting to any site in the configuration using DGMGRL to remove the configuration:
REMOVE CONFIGURATION;
Install Oracle10g database software. The Data Guard feature is not available with the Standard edition, so the Enterprise edition must be installed. Additionally, install Oracle Enterprise Manager (OEM) 10.1 in order to manage the configuration using Data Guard Manager GUI interface, which is tightly integrated with OEM.
The primary and standby databases should be upgraded to Oracle10g. The log management services should be stopped during the primary database upgrade. All of the redo logs generated during the database upgrade should be applied manually on each standby database. For more information on upgrading to Oracle10g, refer to the documentation, “Oracle Database Upgrade Guide 10g Release 1(10.1)”.
Update the initialization parameter files or server parameter files on the primary database and all the standby databases to include the parameter db_unique_name. The database unique name must be identical on all the databases in a configuration.
Once all the databases in the Data Guard configuration are upgraded to Oracle10g, the broker configuration file will have to be created in order to manage it using Data Guard broker. Execute the following command, connecting to the primary database using DGMGRL, to recreate the broker configuration file and to add standby databases"
Also, see the docs, with their instructions:
http://download.oracle.com/docs/cd/B28359_01/server.111/b28294/upgrades.htm
Hope this helps. . .
Donald K. Burleson
Oracle Press author
Author of "Oracle Tuning: The Definitive Reference":
http://www.dba-oracle.com/bp/s_oracle_tuning_book.htm

Cross architecture Data Guard

Hi,
We currently use Oracle 10.2.0.4 on HPUX with data guard and logical
standby. SQL Apply is used
to replicate data to 2 other standby machines/databases. We are
interested adding Redhat linux box into
the mix running 11gR2. Would it be possible to replicate data from
HPUX to Redhat linux?
Thanks for any hint.
Sandeep

Hi,
in such condition,
1) i will consider setting up NTP for all the DBs server
2)Using the same apply time lag for all Standby DB
3) Enable flashback on all the DBs
4) testing all scenario (tedious but worth it)
make sure you can flashback to a specific point in time on all the DBs, if required.
Tobi

Data guard setup for 2 node RAC primary to 2 node RAC standby

Hi All,
I am going to setup data guard for 2 node RAC primary to 2 node RAC standby on Oracle 10.2.0.4. in AIX5L.
Can you please provide the document on the above setup which is having all the steps (details).
Also, the documents on different scenarios like
1) If one node of standby goes down, how the redo logs will be applied. IS there any problem?
2) If both nodes of standby are failed, how to reciver them?
3) If one node of primary fails, is there any issue?
4) If two nodes of primary fails, is there any issue?
Thanks in advance,
Mahi

Have a look at the following location, you may find some similar documents:
http://www.oracle.com/technology/deploy/availability/htdocs/maa.htm
By
http://www.oraxperts.com

Data Guard Scenario

Similar Messages

Maybe you are looking for