Applying recommended patch bundles to a Data Guard environment.

I've learned that there are 3 recommended patch bundles for 10.2.0.4 Data Guard, as follows:
7937113 - 10.2.0.4 Data Guard Logical Recommended Patch Bundle #1
7936993 - 10.2.0.4 Data Guard Physical Recommended Patch Bundle #1
7936793 - 10.2.0.4 Data Guard Broker Recommended Patch Bundle #1
My question is this...should I apply the logical and physical patch bundles to the primary database as well?
Thanks,
Michael Anderson
OCP - Bank of the West

That's what I suspected. Now, in my environment, I could never conceive of a situation where I would switch roles between the primary and the logical standby database, so I would only apply the DG physical patch bundle to the primary database but not the DG Logical patch bundle. Agree?

Similar Messages

Refer to Doc ID 825030 which talks about applying of patch for correction of date format in web adi. Is this patch can be applied on R12.1.3 instance as this patch released date is very old. Please suggest if there is another patch which be applied to cor

Refer to Doc ID 825030 which talks about applying of patch for correction of date format in web adi. Is this patch can be applied on R12.1.3 instance as this patch released date is very old. Please suggest if there is another patch which be applied to correct the date format in web adi. currently the date is coming in number format.

Hello,
This document is for 11i Date Format Changes When Loaded Into Webadi (Doc ID 825030.1)
use this document instead for R12 : Incorrect Date Format in Excel Sheet Generated Via WebADI (Doc ID 828253.1)
<Moderator Edit - deleted MOS Doc content - pl do NOT post such content>
Best Regards,
Eric.

WLS 10.3.x setup in Data Guard environment

Hi all,
We are setting up the following environment:
Oracle Database 11.2.0.2 EE RAC primary and physical standby
WebLogic Server (WLS) 10.3.0
We are using MultiDataSource (MDS) per the documentation which is fine. However, this setup requires the data sources to specify the RAC hostnames (VIPs), instance names, and the service name. Per various documents we are not using the SCAN hostname. Considering this is a phsyical data guard environment, how do we set this up so we can have the best 'seamless' transfer of the WLS envrionment when a switchover or failover occurs from the primary to the standby database?
Thus far we've discussed using instead of the RAC virtual hostname (racprd1), a DNS cname (wls-racnode1) which points to racprd1, but we would modify in the DNS to point to racstby1 when a DG switch or failover occurs.
Any thoughts or recommendations otherwise?
Thanks.

OK...
Alter database rename should be ok As long as both the PRIMARY directory strucures are same or if you have set the parameter db_file_name_convert
While asking a question..
You should post all the necessary information for us to help(is possible)
1.Oracle version
2.Os version
Thread spesific
3.Directory strucutrs of Both Priimary and Standby
4.Init .Ora of both PRIMARY AND STANDBY

Switchover in a data guard environment using Grid Control 10.2.0.3

I've tested switchover in a data guard environment using Data Guard Broker in Grid Control.
However, at times, i receive the message "RemoteOperationException: failed to establish input streaming thread". It looks like it has problems connecting to the remote node using the host credentials supplied. I know the credentials are ok because this worked before. If i test preferred credentials, it's ok too..
The workaround has been to restart the database and this seems to work.
Has anyone experienced this?

Thanks for All for replying.
*1. How can I upgrade my Grid control 10.2.0.3 to 10.2.0.4?*
I have upgraded the Grid Control Agent and OMS from 10.2.0.3 to 10.2.0.4
When we upgrade the OMS which also upgrade the Repository database.
Apply the patchset p3731593_10204_Linux-x86-64.zip (Which comes with Grid_Control_10.2.0.4.0_Linux-x86-64.zip)
su - oracle
cd 3731593/Disk1
./runInstaller (Run two times for Agent and oms upgrade)
Agent upgrade_
During upgrade
Specify Home Details: Select the Agent Home to upgrade Grid Control Agent.
Ex: /u01/app/oracle/OracleHomes/agent10g
OMS upgrade+
Specify Home Details: Select the oms home to upgrade oms
Ex: /u01/app/oracle/OracleHomes/oms10g
*2. How can I monitor/Connect my Existing database 10.2.0.4 from Grid Control 10.2.0.3?*
Install Grid Agent on the existing database server 10.2.0.4
Download Linux_x86_64_Grid_Control_agent_download_10_2_0_4_0.zip from OTN http://www.oracle.com/technology/software/products/oem/htdocs/agentsoft.html
su - oracle
cd /u01/software/GridAgent/linux_x64/agent
./runInstaller
Refer:
[To Install an Additional Management Agent Using OUI|http://download.oracle.com/docs/cd/B16240_01/doc/install.102/e10953/installing_em.htm#sthref318]
OracleÂ® Enterprise Manager Grid Control Installation and Basic Configuration
10g Release 4 (10.2.0.4)
Part Number E10953-05
3 Installing Enterprise Manager
Thanks
Mukarram Khan
Edited by: Mukarram Khan on Feb 6, 2009 11:04 PM

ACLs in a Failsafe/data guard environment

Hi,
We have recently upgraded to Oracle Enterprise Edition 11.2.0.3
We had a slight issue with a few batch processes which sent success emails - we hadn't configured ACL permissions (We knew about it, and was on the log, but somehow got missed). All set up fine now in our test environment, but got a couple of queries before I set up in production.
Had a look in the documentation and tried searching
The ACLs I believe are managed as an XML file, on the database server, correct?
If so, how do we manage this on both a clustered, and a data guard environment? Or in the case of a database restore? Is the ACL file automagically regenerated out of some database metadata? Or do we need to recreate in each environment?
Any ideas?
Cheers,
Carl

Hi,
there is two ACL(Access Control List) one at Db level ans one is used at UNIX command for file permission , as you mentioned you have permission issue in batch execution.
so you can check below link
http://www.dartmouth.edu/~rc/help/faq/permissions.html
i am not sure which ACL you required.

How many physical standbys can be created in one data guard environment?

Hi Experts,
How many physical standbys can be created from primary in one data guard environment? Why LOG_ARCHIVE_CONFIG limits up to 9 unique database names? Thanks.

Victor Jin-Oracle wrote:
Hi Experts,
How many physical standbys can be created from primary in one data guard environment? Why LOG_ARCHIVE_CONFIG limits up to 9 unique database names? Thanks.
Post the source of your information abouit the limit of 9 names.
Even the Oracle doc shows 1..30 names in the same provided although that doesn't necessarily indicate a limit.
https://docs.oracle.com/database/121/REFRN/refrn10237.htm

Oracle 11g Goldengate working in a Data Guard environment

Hey Guys, I have the following setup:
Oracle 11.2.0.4.4 running on RHEL 6.6 64-bit
3 node RAC Primary
Single-node Standby with GI (Oracle Restart)
4 DBs on Primary - AA, BB, CC and DD
4 DBs running in Active Standby - AADR, BBDR, CCDR and DDDR
Databases AA and BB replicate schemas into CC, which is a Reporting/Operational Data Store. GoldenGate 11.2.x is used for the replication
XAG used for Goldengate HA
All good here with the Primary DBs and replication humming along quite nicely.
I have questions as part of building the DR environment. How should this work? As part of the real-time apply and active standby, CCDR should keep pace with the replicated primary DB CC. Now from a role transition perspective, I need to be have the GoldenGate software installed and configured on the Standby Server for starting the processes.
Question 1: When the role transition happens, based on the data guard role, I could well have a system trigger to spawn off a shell script that starts off the goldengate processes. Would this be the correct approach?
Question 2: Do I need to worry about setting/starting the extract/replicat from a specific CSN? The checkpoint table will be kept in sync as part of the SQL apply, so would just starting the extract and replicat process on the new Primary (old DR) work?
I am unable to get any pointers on this in Metalink. Would appreciate anyone could point me to a MOS document that talks about using GG and Data Guard together.
Thanks

On the first day, there is no backup to recover. It is only on the third day's run that this command will really be successful.
The white paper does explain this.
Hemant K Chitale

Multiple Location of datafiles in Data Guard Environment

I have Data Guard setup in my oracle 10g database. Uptil now , all the datafiles were at same location, therefore db_file parameter setting in pfile was fine. Now I want to move my few of the datafiles at another location. Say, few of my datafiles would be in F:\ and some of them in G:\. How do I now set db_file parameters in Pfile ?

OK...
Alter database rename should be ok As long as both the PRIMARY directory strucures are same or if you have set the parameter db_file_name_convert
While asking a question..
You should post all the necessary information for us to help(is possible)
1.Oracle version
2.Os version
Thread spesific
3.Directory strucutrs of Both Priimary and Standby
4.Init .Ora of both PRIMARY AND STANDBY

Backuping up a standby database in a Data Guard environment!

Hello everyone,
I have set a Data Guard in Oralce 10.2.0 between 2 databases for testing purposes. Each database is located on a different server. I want to schedule an incremental backup of the standby database to do an incremental level_1 backup everyday. I am following this paper:
[Using Recovery Manager with Oracle Data Guard in Oracle Database 10g|http://www.oracle.com/technetwork/database/features/availability/rman-dataguard-10g-wp-1-129486.pdf]
but it's really unclear, for me, how this command sequence is constructed:
RECOVER COPY OF DATABASE WITH TAG ‘OSS’;
BACKUP DEVICE TYPE DISK INCREMENTAL LEVEL 1 FOR
RECOVER OF COPY WITH TAG ‘OSS’ DATABASE;
BACKUP ARCHIVELOG ALL NOT BACKED UP TO SBT;
BACKUP BACKUPSET ALL;How can I recover a copy of database that wasn't backed up before? When I did that in RMAN, I got the following output:
Starting recover at 03-APR-11
allocated channel: ORA_DISK_1
channel ORA_DISK_1: sid=159 devtype=DISK
no copy of datafile 1 found to recover
no copy of datafile 2 found to recover
no copy of datafile 3 found to recover
no copy of datafile 4 found to recover
no copy of datafile 5 found to recover
Finished recover at 03-APR-11Could anyone please help me understanding this sequence of commands in a simple way to get a better idea of how to backup a standby database?
Thanks in advance...

On the first day, there is no backup to recover. It is only on the third day's run that this command will really be successful.
The white paper does explain this.
Hemant K Chitale

RMAN restore UNTIL SCN in DATA GUARD environment

Hello,
OS=RHL5
DB=11.2.0.3 Primary and 1 Physical STANDBY database.
I have a huge application changes tomorrow and its possible that I have the requirement to ROLLBACK the database 5 hours back because the application.
I am thinking to adopt following method from the official Oracle documentation (untilClause):
STARTUP FORCE MOUNT
RUN
SET UNTIL SCN 1418901; ## Any required SCN number will be used here.
RESTORE DATABASE;
RECOVER DATABASE;
ALTER DATABASE OPEN RESETLOGS;
I will of course take LEVEL0 backup tonight and LEVEL1 backup before the start of work.
My humble question to you is that: If this is the correct method I am adopting? and if yes, then what about Physical standby database? Will there something to restore/recover too? or it's not required or standby database need to create again (which I am afraid of)?
Flash back is not enabled.
Bundle of thanks for the suggestion.
Best Regards

hello John-MK,
My humble question to you is that: If this is the correct method I am adopting?
the method is correct but not the BEST PRACTICE...create a RESTORE POINT instead of using SCN. for example:
SET UNTIL RESTORE POINT before_app_changes;
1. drawbacks of using flashback feature are:
a. you need more disk space
b. there is CPU overhead associated with flashback log operation
c.compulsorily need to use fast recovery area and you can not backup flashback log
d.others..
but my own opinion, i think the advantage out weight the drawbacks, since you are paying for it already
2. if the standby database did not apply redo data past the new resetlogs SCN, then you do not need to re-build. no manual intervention
HTH
Tobi

Active data guard - patching

Hi Gurus,
I have primary db amsprd and active data guard amsprd_stby.
we are planning to apply application patches in amsprd. earlier we used to rely on rman backup for restore in case of any failures.
I would like to make use of standby as primary and keep the business up immediately after application patch failures.(to reduce the downtime - application failures are not having any rollback steps :-( - vendor product).
to achieve this,
before application patch application
do i have to just stop the log shipping?
or do i have to stop just log file apply?
what are the steps involved in case of primary applicaiton patching failures.
Thank you.

So the Application Patch may turn out as a "Logical Mistake" from the Database perspective.
One option would be to stop the Apply on the Standby Database. If your Application Patch was no good, you could do an "Immediate Failover" which will open the Standby immediately without applying the received redo first like an ordinary Failover would. You would have to create a new Standby Database on your Ex Primary site then.
Another option would be to turn on Flashback Logging on both sites. If your Application Patch was ok - nothing to do. If not, flashback both sites to before the App Patch. Look here for an example:
http://uhesse.wordpress.com/oracle-database-ha-architecture/2010/08/06/using-flashback-in-a-data-guard-environment/
Second option has the benefit that you do not delay an emergency failover, should your Primary incidentally get destroyed during the App Patch. Also, if you need to rewind the Logical Mistake, you do not need to recreate the Standby completely. If you are on 11gR2, you can even turn on/off Flashback Logging without having to restart:
http://uhesse.wordpress.com/2010/06/25/turning-flashback-database-on-off-with-instance-in-status-open/
Kind regards
Uwe Hesse
http://uhesse.wordpress.com

Data Guard Logical Standby Skip Handler bug

Does anyone know of a documented bug concerning the skip handler not skipping a grant on a view in a schema with a SCHEMA_DDL skip handler defined?
Here is my test case.
On Primary:
Create table scott.junk (x number) tablespace users; (created successfully on logical standby)
On logical standby:
exec dbms_logstdby.skip('SCHEMA_DDL','SCOTT','%'null,true,'\'); (created skip handler on all SCOTT objects)
On Primary:
grant select on scott.junk to public; (successfully skipped due to missing object)
On Primary:
create view scott.v_junk as select * from scott.junk; (successfully skipped on logical standby)
On Primary:
grant select on scott.v_junk to public; (received 'ora-00942 table or view does not exist' on logical standby and SQL Apply stopped).
I'm running version 10.2.0.4 but have not applied the recommended patch bundle for logcal standby #7937113.8.
If anyone knows whether this has been documented or not, please let me know. If not, I will submit an SR to get this documented.
Thanks,
Michael Anderson
OCP - Bank of the West

When I look at the syntax for the package
http://www.morganslibrary.org/reference/dbms_logstdby.html
what I see is:
dbms_logstdby.skip(
stmt        IN VARCHAR2,
schema_name IN VARCHAR2 DEFAULT NULL,
object_name IN VARCHAR2 DEFAULT NULL,
proc_name   IN VARCHAR2 DEFAULT NULL,
use_like    IN BOOLEAN DEFAULT TRUE,
esc         IN CHAR1    DEFAULT NULL);So looking at your code:
exec dbms_logstdby.skip('SCHEMA_DDL','SCOTT','%'null,true,'\'); (created skip handler on all SCOTT objects)
'SCHEMA_DDL' is your statement.
'SCOTT' is your schema name.
'%' indicates all objects
and ... shouldn't there be a comma before the NULL?
In the future don't provide values where the default's will suffice. It makes thing's far simpler.

11.2.0.3 Patch Bundle and opatch installation approach

Hi,
I have installed the below 11.2.0.3 (64 bit) products in my windows 2008 server
1. Oracle Grid Infrastructure for standalone server
2. Oracle database server software and created a standalone database with ASM in this home.
Now i need to apply Update 5 patch bundle and some bug fix patches.
Since there are 2 home in my server what is the right approach for applying the patch bundle and opatch ?
When i went through the patch bundle installation document i could see these 2 options
1. Patch Installation Instructions for Single Instance
2. Patch Installation Instructions for RAC
Since i have 2 homes (OraCrs11g_home1 & OracDb11g_home1 ) in my server i am not sure which patch installation option i should choose and proceed.
Can someone assist me ?
Thanks,
Ashok Kumar.G

no we have to prepare a test POC but we don't have Xterm so had to use silent mode.using VNC server is best option to install or upgrade databases/grid Infrastructure.
VNC server runs on the server and no fear of loosing network connection as in case of connecting thorugh putty.
also you can open GUI in VNC server without need of any xterm or xwindows .
So ask your SA to install VNC on your server and use it.

Patches Uninstall on patch bundle unassign - by design?

Hi,
Recently we began testing ZCM 10.3.3 for deploying Microsoft patches to our Windows XP desktops. We have come across
a situation that whilst after some investigation looks like it might be operating as designed, we think is perhaps
the result of an over sight in how the product works. I thought I would post the scenario here to see if anyone else
had come across this and if so, what your thoughts on it might be. I am tempted to log a SR, but this might be a well
known occurrence?
Essentially what happened was that we created a patch bundle and assigned it to a device group that contained test
machines. The patch deployment went as expected with no issues observed or reported. A week later I removed the
relationship between the patch bundle and the test device group. Immediately (on next device refresh) the machines in
the test device group began to uninstall the patches in the patch bundle that had just been "unassigned". In
addition, interactive "remediate.exe" errors were thrown in any active user sessions on those machines. The errors
seemed to be related to patches that werent technically able to be uninstalled anyway.
Initially we were a little stumped by this as the ZPM patch bundle was not enabled for Uninstall, however looking at
the bundle a little closer we noticed that the actual individual patch bundles being referenced in the install action
DID have Uninstall enabled - if fact it appeared that every patch was enabled for uninstall - even those that werent
technically able to be uninstalled.
When considering this in conjunction with our global ZEnworks Explorer Configuration to Uninstall unassigned bundles
after 0 days, things started to make a little more sense - but I'm not sure this is how it should actually work.
To make a long story short, I personally dont think that the global Zenworks Explorer Configuration to uninstall
unassigned bundles after a certain amount of days should apply to patch bundles, or at least patch bundles should be
able to be excluded from this. We have this setting set to 0 days so that when we unassign licensed applications,
they are immediately removed from the desktops they have been deployed to - this is to ensure we are confident of our
licensing position in terms of how many copies of a application we may be licensed for and how many we actually have
installed in our fleet.
The implication of this situation is that for us, every time we deploy a patch, we need to maintain the assignment between
the patch and the device. On the face of it, this doesnt seem like a viable long term solution for a whole host of
reasons.
So out of this I have a couple of questions:
1. Has anyone else come across this, and if so, a) do you think this is a reasonable way for the product to be
operating, and b) how have you worked with (or around) it.
2. Is it possible to disable uninstall for the individual patch bundles that ZPM creates, or will ZPM simply "fix"
them?
I'd be very interested in peoples thoughts on this.
Cheers,

Hi Shaun,
Thanks for the reply. Yes, the "Ignore chained dependencies" box is checked on the individual patch bundles and "uninstall" is disabled for the parent patch bundle itself. Do you think this sounds like a bug? If thats the case, I'll definitely open an SR for it.
Cheers,

Data Guard Test cases

Hi All we are in process of implementing Data Guard (Physical Standby) setup. Everything has been done. Now we need some good test cases to test our setup. Can you please provide some inputs in testing this.

Simple test would be to create few test tables with some data in primary and see if they are applied in your DR site by Data Guard.
In terms of testing failover & failback or Switchover and Switch back you would need to go through oracle documents or refer the following link.
http://www.oracle.com/technology/deploy/availability/pdf/MAA_WP_10gR2_SwitchoverFailoverBestPractices.pdf
sbs

Applying recommended patch bundles to a Data Guard environment.

Similar Messages

Maybe you are looking for