Fuad Arshad's Blog » dataguard

Physical Dataguard Password change

fuadar — Mon, 14 Sep 2009 12:06:25 +0000

Oracle Performs redo shipping and authentication services between the primary and the standby using Password files and if your security protocols dictate a sys password change you will start having issues with stopped log shipping with errors like .

Error 1017 received logging on to the standby
————————————————————
Check that the primary and standby are using a password file
and remote_login_passwordfile is set to SHARED or EXCLUSIVE,
and that the SYS password is same in the password files.
returning error ORA-16191
————————————————————

The solution to fixing thiscan be

1. Open the physical Standby read only and change the sys password.

2. Use orapwd to create anew password file.

3. ship/sftp the password file from the primary.

11g Brings a new twist to options 1 and 2 .

SEC_CASE_SENSITIVE_LOGON=TRUE by default in Oracle 11g which means the option 1 and 2 will generate different hash when done on different systems

Metalink note #462219.1 provides an explanation and the solution to this problem. The solution includes a new option in the orapwd utility which allows you to ignorecase.

$ORAPWD FILE=filename [ENTRIES=numusers] [FORCE={Y|N}] [IGNORECASE={Y|N}] [NOSYSDBA={Y|N}

The other solution includes changing case sensitivity by changing SEC_CASE_SENSITIVE_LOGON=FALSE . This is a dynamic parameter.

Please evaluate and implement all solutions based on your company security practice.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

User Defined Metrics in Oracle EM Grid Control

admin — Sun, 12 Apr 2009 23:39:00 +0000

Everyone has various ways to manage scripts. I have found UDM’s in Oracle Enterprise manager to be a nice way to catalog and standardize scripts across various platforms.
UDM’s or User Defined Metrics are a great way to track, alert and look at trending information using Oracle Enterprise manager . Oracle provides UDM’s on a server level as well as an instance level. UDM’s can be cataloged and saved into a library for reuse.

I’ve been using SQL UDM’s to monitor and alert on SQL Apply Lag times in minutes as well as restarting the logical dataguard environment .
To see if Sql Apply has stopped i check select count(*) from v$dataguard_stats and based on the count 3 for good and 0 for bad run a corrective action to start the sql apply again. If the corrective action fails . The alert is then escalated to the pager for support. Corrective Actions based on UDM’s if scripted right can help reduce support calls and keep environments in sync.
I’ve used UDM’s to pick up data from DBA_LOGSTDBY_EVENTS and execute alter database datafiles commands to keep my primary and standby in sync.
The UDM creation screen is fairly simple and the wizard walks the user through all the information . If using external scripts i would definitely recommend creating user that manages the scripts and then setting preferred credentials at that user level. This makes it easier to change passwords using the emcli command line interface and be compliant with audit policies.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Applying Datafile resize in logical standby

admin — Mon, 30 Mar 2009 17:39:00 +0000

Well,
with the record_unsupported_operations-true . I was able to capture the commands that oracle skipped. using that info i was able to write a simple piece of code to use that information to get
Alter database datafile commands executed on my logical standby database which in this case has exactly the same structure as my primary.

/* Program Name : Datafile Alter command capture
Author : Fuad Arshad
Purpose : To capture alter database datafile commands from dba_logstdby_events
and execute on logical standby
Prerequisites : dbms_logstdby.apply_set(‘RECORD_UNSUPPORTED_OPERATIONS’,'TRUE’);
dbms_LOGSTDBY.apply_set(‘MAX_EVENTS’,'a value of your choosing based on events ‘);
Licensed : Creative Commons.
*/

set serveroutput on
declare
p_errbuf varchar2(500);
p_retcode number(1);
cursor dba_events is
SELECT to_char(event) event1 FROM dba_logstdby_events WHERE status_code=’16226′and event_timestamp>systimestamp-1
and event like ‘ALTER DATABASE DATAFILE%’;
begin
for
dbevent in dba_events LOOP
DBMS_OUTPUT.PUT_LINE(‘Statement : ‘ || dbevent.event1 );
execute immediate dbevent.event1;
end LOOP;
EXCEPTION
when no_data_found then
null;
when others then
p_errbuf := sqlerrm;
p_retcode := 1;
DBMS_OUTPUT.PUT_LINE(‘Error : ‘ || p_errbuf|| ‘Error Code’ || p_retcode );
end;

This is a very simple way of making sure all the datafiles on the standby look exactly the same as the Primary. I’m sure there are many other ways to do this and if there are please share how u would have taken care of the alter database datafile issue. Comments and improvements as well as ideas are always welcome.

Physical Dataguard Password change Oracle Performs redo shipping and authentication services between the primary...

Related posts brought to you by Yet Another Related Posts Plugin.

DBMS_LOGSTDBY and recording unsupported operations

admin — Tue, 24 Mar 2009 02:40:00 +0000

DBMS_LOGSTDBY is the heart and soul of the logical standby management infrastructure . While trying to figure out how to get my alter database datafile operations . i bumped in a record unsupported operations in $logstdby_stats . There seems to be no mention in the documentation and no reference to this parameter being set anywhere . If you to set it though

>exec dbms_logstdby.apply_set(‘RECORD_UNSUPPORTED_OPERATIONS’,'TRUE’);

This will help capture the unsupported operations as the ORA-16226 operations. I am planning on then reading the clob and re executing the statements onto the logical standy. I still have to figure that part out but stay tuned ….

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Oracle Dataguard Broker and CFC

admin — Thu, 15 Jan 2009 14:56:00 +0000

During a recent implementation of Dataguard on a Active/Passive Cluster. I started receiving weird errors using the installation/configuration phase. The Grid control add standby wizard which is something i like to use for smaller databases started out the the process but would never complete it leaving the primary and the standby in a weird state.
Errors like

DMON: cannot open configuration file “/usr/local/oracle/product/10.2/dbs/dr2test.dat”
ORA-27037: unable to obtain file status

error = ORA-16572
DG 2009-01-15-08:26:48 0 2 0 NSV0: Failed to connect to remote database p
rkmdb1. Error is ORA-12541
DG 2009-01-15-08:26:48 0 2 0 NSV0: Failed to send message to site prkmdb1
. Error code is ORA-12541.
DMON: test unable to contact primary for DRC version check
DMON: status ORA-12541:
~

The problem is that the dataguard broker cannot traverse through a virtual node name which is used in a Active/Passive or CFC configuration,
According to Metalink Note # 413696.1 which is very appropriately named “Data Guard Broker does not support Cold Failover Clusters”
“The Data Guard Broker does not support Cold Failover Clusters (also called Active/Passive clusters) in any version up to and including Oracle Database 10g Release 2. You cannot use the Broker nor Enterprise Manager, you have to create and manage your standby configuration using SQLPlus”

The support for CFC’s come in a limited fashion in Oracle 11gR1 only if Oracle clusterware is used.

This means that a dataguard standby has to be manually created and configured and all instances have to be monitored individually.

dataguard, oracle

Physical Dataguard Password change Oracle Performs redo shipping and authentication services between the primary...

Related posts brought to you by Yet Another Related Posts Plugin.