Oracle Data Guard - Physical Standby Databases

From: Bernhard Krautschneider <bernhard.krautschneider_at_acp.at>
Date: 9 Nov 2003 11:47:54 -0800
Message-ID: <dec9f4d1.0311091147.cea7434@posting.google.com>

0) Pre
1) Environment
2) Target
3) Questions
4) Post

0) Pre

Two things before we go to the technical stuff. First of all english is not my native language. I'm sorry for all the mistakes in this post. Second I'm not very experienced with oracle and I'm really new to oracle backup. I got all my information from different oracle whitepapers/manuals/texts. Therefore I understand technical instructions and operations but I have some knowledge gaps with general basic concepts.
sorry about that.

Environment
Right now I use Oracle 9.2 on Windows. I have 2 databases: a primary and a physical standby database. I use LGWR as log-shipping process and the v$archive_dest/status & error Fields say that it works properly.

I'm austrian and therefore my Oracle speaks german language. I try to tranlate the Error's using the error code. Just wanted you to know that there could be some mistakes.

The standby database has been created with rman. I used techniques that are explained within the Oracle9i Recovery Manager User's Guide Chapter 13. The text explains two ways to create the standby database; with or without the use of the DORECOVER rman-clause. Actually I had some troubles at that point. I wasn't able to do the manual recovery-step that should be done after creating the database without the DORECOVER-clause. Although I'm quite sure that I didn't solve the problem with doing the creation with the DORECOVER-clause it seemed to work. The problem was some time-SCN-conflict, that is inaccurately described within the same chapter.

2) Target

My target is to set up an MAXIMIZE AVAILABILITY environment that is prepared for switch -and failovers and to fully understand how it works. Also it is very important to me, to get better knowledge about how to implement physical standby databases.

3) Questions

Recovery Mode Questions
My first question is about the following error I get when I try to put my standby-database in managed Recovery Mode.

SQL> alter database recover managed standby database; alter database recover managed standby database *
FEHLER in Zeile 1:

ORA-00283: Wiederherstellungs-Session wegen Fehlern abgebrochen
ORA-01190: Kontrolldatei oder Datendatei 1 liegt vor letztem RESETLOGS
ORA-01110: Datendatei 1: 'E:\ORACLE\ORADATA\BERLIN\SYSTEM01.DBF'

[Translated Error-Msg - Begin]
ORA-00283 recovery session canceled due to errors Cause: An error during recovery was determined to be fatal enough to end the current recovery session.
Action: More specific messages will accompany this message. Refer to the other messages for the appropriate action.

ORA-01190 controlfile or data file string is from before the last RESETLOGS
Cause: An attempt was made to use a datafile when the log reset information in the file does not match the control file. Either the datafile or the control file is a backup that was made before the most recent ALTER DATABASE OPEN RESETLOGS.
Action: Restore file from a more recent backup.

ORA-01110 data file string: 'string'
Cause: This message reports the file name involved with other messages.
Action: See the associated messages for a description of the problem. [Translated Error-Msg - End]

Is this error resulting from a mistakes that I made during the standby database creation using rman with the DORECOVER option? What exactly is the problem and how do I solve it?
When I do "alter database recover managed standby database disconnect from session;" I get no errors.

b) How can I change the protection mode?

SQL> select protection_mode from v$database; PROTECTION_MODE

MAXIMUM PERFORMANCE The target is to use Maximum availability mode.

c) How exactly are Standby Redo Logs implemented?

I have online redo logs on the primary and the standby database. Actually I can monitor that a log-switch on the primary triggers a log-switch on the standby (via v$archived_log). Now the question is whether the data guard environment uses the online-redo-logs on the standby as standby-redo logs? If not LGWR can not really write information to the standby. I would have to create standby-redo logs by hand (which is perfectly possilbe). But if I do so, how do I tell the LGWR process (on the primary) that there are standby-redo logs on the standby database? This raises the question whether I need online -and standby redo logs on both systems in preperation for a good switch/failover? How can I monitor redo and standby redo logs? How can I be sure that they work properly?

4) Post

I hope I didn't piss anybody off by asking so much questions. I also hope that I did provide enough Information and that I fromulated the questions in an accurate way. Thanks in advance!

faithfully,
Bernhard
. Received on Sun Nov 09 2003 - 13:47:54 CST

Oracle Data Guard - Physical Standby Databases - Questions