Logical standby rolling upgrade issue
Hello together,
im experiencing some strange error during the setup of a logical standby environment.
primary system running: 10.1.0.3.0 - 64bit Production
log. standby system running: 10.1.0.5.0 - 64bit Production
compatibility flags set to 10.1.0.3.0 on the standby side.
now everything works fine till it comes to the point where I have to register the logical logfile after building the logminer dictionary on the primary site.
Execution flow on the primary site:
SQL> EXECUTE DBMS_LOGSTDBY.BUILD;
SQL> SELECT NAME FROM V$ARCHIVED_LOG WHERE (SEQUENCE#=(SELECT MAX(SEQUENCE#) FROM V$ARCHIVED_LOG WHERE DICTIONARY_BEGIN = 'YES' AND STANDBY_DEST= 'NO'));
gives me:
/oradata/primary/arch/585226884_1_10.arc
Copied to the standby side and tried to register the logfile:
SQL> alter database register logical logfile '/oradata/standby/arch/585226884_1_10.arc';
alter database register logical logfile '/oradata/standby/arch/585226884_1_10.arc'
ERROR at line 1:
ORA-16225: Missing LogMiner session name for Streams
anyone knows what could cause ?
thanks in advance
Chris
It isn't documented anywhere that I know of but yes you can.
Stop SQL Apply, set your Guaranteed Restore Point (GRP) in the Logical standby, run your replay, get the results, flashback the standby to the GRP, open reset logs and restart the apply.
You would need to figure out at what point you would want to run the capture at the primary and stop the SQL Apply at that point or else you will need to flash the logical standby back to the point you started the capture at the Primary and set your Guaranteed Restore point at that point in the database so that Replay can operate correctly (i.e. on the same set of data).
Bear in mind that you during this process you will have two GRP's. One on the standby and one on the Primary. So you need room for the flashback logs at the standby to hold the flashback data and archive logs generated by the replay as well as room for the archive logs coming in from the Primary. And you will need room for the flashback logs at the Primary because it will be still running production during your replay period and generating flashback data there as well.
Larry
Similar Messages
-
Logical Standby Data Consistency issues
Hi all,
We have been running a logical standby instance for about three weeks now. Both our primary and logical are 11g (11.1.0.7) databases running on Sun Solaris.
We have off-loaded our Discoverer reporting to the logical standby.
About three days ago, we started getting the following error message (initially for three tables, but from this morning on a whole lot more)
ORA-26787: The row with key (<coulmn>) = (<value>) does not exist in table <schema>.<table>
This error implies that we have data consistency issues between our primary and logical standby databases, but we find that hard to believe
because the "data guard" status is set to "standby", implying that schemas' being replicated by data guard are not available for user modification.
any assistance in this regard would be greatly appreciated.
thanks
MelIt is a bug : Bug 10302680 . Apply the corresponding Patch 10302680 to your standby db.
-
Installing Logic Studio 9 upgrade issue
I am installing the Logic Studio 9 upgrade from a copy of Logic Pro 7 which itself is an upgrade from Logic Express 7. I can run it usung the XSKey but when I try to run it without, it asks for a valid serial number (from Logic Pro 7) to verify the upgrade. The problem is I have a serial number for Logic Express but Logic Pro only has an XSKey serial number which isn't a valid number. Whilst I can use it with the XSKey, I would ideally like to use it without as it takes up a much needed usb slot and I hate having to carry it around when I'm out and about with my Macbook Pro as it's small enough to go missing and sticks out of the laptop rather precariously. Can anyone help with a solution?
Just to update, I have to use XSKey every time I run Logic Studio otherwise I am prompted for my Studio serial number and a Logic serial number.
-
Feeling Ripped Off - Logic Express 9 Upgrade Issue
Hi, I have been using Logic Express 8 for about a year. I just bought the upgrade for Logic Express 9. I started installing it, entered the serial number and then I was prompted to enter the serial number for my Logic Express 8.
I have since misplaced my LE 8 booklet with the serial number on the back (a bunch of stuff didn't make it to my new apartment in a move). I didn't think this would be a big deal. I thought I would be able to find the serial number of my current version of LE 8 by clicking on the "About Logic Express" tab when I have the program running. However, when you click this all you get is who it is registered to (me) and the version number (8.02).
I called technical support and asked them where I can find the serial number and they said sure, "click on About Logic Express when you have the program open and it will be right there (he also called it the splash page)." It wasn't and isn't. The tech support person said "ah yes, only the previous versions of Logic Express show you the serial number of the version you currently are running." I asked if there was anywhere else that I could find this number and he said "no."
Solutions? He suggested I find my receipt and fax it in and get a new serial number. Impossible, I don't have the receipt. Call the store and see if they can dig it up. Let's get real. It was purchased with Xmas money in cash I got from my parents.
Basically, I'm screwed. I have an opened LE 9 upgrade that I can't return. I have a version of LE 8 that I legally purchased and for some dumb reason there isn't a file or page or something in the program that notes my serial number.
Not fair. I guess I have to tattoo all serial numbers for stuff I purchase from Apple to my arms on the chance that I lose a booklet. Why Apple doesn't include a file or store the serial number of the LE version you are running somewhere on the system is beyond me.
So, I feel ripped off. Apple doesn't care. I thought there would be some way we could work this out (like printing the screenshot of the page that says I am the registered owner) but they would rather try to get another $100 out of a loyal Apple customer and force me to buy LE 9 straight up. I've probably spent about $7,000 on Apple stuff in the past 6 years. I am the definition of a loyal customer.
Not fair.You didn't buy Logic Express, you merely paid for the right to use it. When you install it, you are presented with the license agreement to which you must agree.
At the top of the agreement it reads "IF YOU DO NOT AGREE TO THE TERMS OF THE LICENSE, YOU MAY RETURN THE APPLE SOFTWARE TO THE PLACE WHERE YOU OBTAINED IT FOR A REFUND...."
So, try taking it back to the store and telling them you declined the license agreement and want a refund.
You can find a copy of the Logic Express license agreement here: http://images.apple.com/legal/sla/docs/logicexpress.pdf -
Has anyone performed a successful upgrade from 10.2.0.4 RAC to 11.2.0.2 RAC using a transient logical standby? I found one reference to this white paper "http://www.oracle.com/technetwork/database/features/availability/maa-wp-10gr2-transrollupg-168590.pdf" that describes steps to us transient logical standby to upgrade to 11.1.0.6 but I wasn't sure if it also works for upgrades to 11.2.0.2.
I have a 3 node 10.2.0.4 RAC cluster running on rhel 4.8 with a 3 node dataguard physical standby (which the customer want to keep as a physical standby after the upgrade) that I need to upgrade to 11.2.0.2 with as little downtime as possible. This approach seems like a possible solution but I want to know if it is proven.Hi,
I have a 3 node 10.2.0.4 RAC cluster running on rhel 4.8 with a 3 node dataguard physical standby (which the customer want to keep as a physical standby after the upgrade) that I need to upgrade to 11.2.0.2 with as little downtime as possible. This approach seems like a possible solution but I want to know if it is proven.I suggest you to log a S.R and confirm the same.
thanks,
X A H E E R -
Database rolling upgrade cpu patch with minimal service outage
Oracle 11gR2 Primary Instance
Oracle 11gR2 Active Physical Standby
Requirement of applying CPU and CVE patches with minimal downtime. I believe that I've been able to apply CPU patch to Active Standby after deferring redo transport and applying patch to active physical standby first and then performing a manual switchover to have the standby become the primary and former-primary become the standby and they apply CPU Patch there and then perform another role transition to its original role.
Appreciate your experience with primary and active physical standby rolling upgrades/patches. ThanksHello;
I believe you are over thinking this. All you have to do on the Standby is install the patch. No switchover, no SQL. No nothing.
The SQL will move from the Primary to the Standby.
Check the readme for the patch closely or post the patch number if you want to confirm.
OR
How do you apply a Patchset,PSU or CPU in a Data Guard Physical Standby configuration [ID 278641.1]
( If it helps I had the same thoughts after my first Data Guard setup !! )
If you are trying to avoid downtime entirely the switchover will take about as long as the patch.
If you have RAC this is interesting :
http://javeedkaleem.blogspot.com/2010/05/how-to-apply-oracle-database-patch.html
As is this :
http://docs.oracle.com/cd/B28359_01/server.111/b28282/schedule_outage.htm
Best Regards
mseberg -
Transportable tablespaces with Logical Standby
Does anyone know whether or not transportable tablespaces can be used in a Logical Standby environment? I know they can be used with a Physical Standby, and the documentation covers how to do this, but there's nothing about Logical Standby. This is Oracle 10.2.0.4/SLES 10.
Thanks.Using transportable tablespaces to a logical standby environment. Actually, there is a paragraph in the 10.2 Data Guard Concepts and Administration manual. I will summarize:
After creating the transportable tablespace export file on the source system, copy the datafiles to both the primary and logical standby systems.
On the dataguard logical standby, within SQLplus, issue SQL> alter session disable guard;
Import the tablespaces into the logical standby database
SQL> alter session enable guard;
Import the tablespaces into the primary database.
In 10.1, we imported the tablespaces into an environment with a physical standby and then converted the physical standby to a logical standby.
I tested the procedure for 10.2 and it seemed to work. Has anyone else done this and if so, are you aware of any issues? Like I said, the documentation is one very brief paragraph.
Thanks. -
I'm trying to decide if this rolling upgrade concept is any good in practice. I see the point, the idea is to reduce downtime. I'm looking at:
MAA upgrade document using Data Guard and Physical Standby (Rolling Upgrades):
http://www.oracle.com/technetwork/database/features/availability/maa-wp-11g-upgrades-made-easy-131972.pdf
It's been my experience that most things are more complicated, there are lots of things that can go wrong in practice, and details vary form one release to another. My past experiences make me very nervous about using rolling upgrades. I've just never gone there. Can you advise from you experience? I'm be going from 11.2.0.1 -> 11.2.0.2. expdp/impdp works. Are there MOS docs on the fine details for my versions?Hi,
I currently build my consolidated 11gR2 environment to feel confident about that. My main environment is still a 10.2 one, I tested that but I didn't use it in production since my primary is a 6 nodes RAC on 2 locations and the standby is a single instance on a 3rd location (only used for backup and DSS purposes, I would only activate it in case of a disaster since the server is not powerful enough).
In my tests the scenario described in http://download.oracle.com/docs/cd/B19306_01/server.102/b14239/rollup.htm#SBYDB00830 worked well but there are restrictions and it's not so easy. One difficulty was that some critical application uses the Spatial option etc.
If you have experience with sql apply on this environment, e.g if you have already implemented streams things should look easier for you. I suggest you to review the whole process using a realistic test environment...or perhaps to proceed as usual if you don't have enough time to test this scenario and if you can obtain a downtime.
Best regards
Phil -
Rolling Upgrade of Oracle 10gR2 RAC with Physical Standby (DG)
Hi DBAs,
I want to do a rolling upgrade of Oracle 10.2.0.3 (2 node RAC+ASM) with Physical Standy Dataguard to Oracle 10.2.0.4 also applying CPU (April2009).
Please suggest the best way to perform this upgrade and patching with no downtime.
Thanks
-Samar-Unfortunately rolling upgrade is only possible with a logical standby in place:
http://download.oracle.com/docs/cd/B19306_01/server.102/b14239/upgrades.htm#CHDBJAGG
That doesn't help you at this time, but 11g has a way to do it with a physical standby. Even here you need an intermediate logical standby:
http://download.oracle.com/docs/cd/B28359_01/server.111/b28294/rollup.htm#CHDHCBGA
Werner -
Standby db rolling upgrade for EM Grid?
hi,
We are planning on upgrading the repository database to 11gr1 and since we solely use OEM Grid Control to monitor hundreds of database and application services, we want to minimize the Grid control server downtime.
Is standby db rolling upgrade supported for Grid Control? If not, are there any other alternatives we could implement to provide continual database
monitoring via EM Grid while performing the repository upgrade?
Thanks.You won't be able to upgrade the database with zero downtime. At some point in the upgrade process, a script will need to be run which will upgrade the Data Dictionary. That script requires the database to be opened in UPGRADE mode, which means downtime.
If you want near-zero downtime, then employing a transient logical standby is a solution. Even this will have some downtime as things are switched over and then switched back.
Cheers,
Brian -
Oracle 9i database software upgrade issue with standby (Dataguard) database
All,
I have encountered a problem in upgrading the Oracle software on 2 of our databases in-house. The databases in question here are using Dataguard, so one is set to primary and one is set to standby. The version that they are/were starting at is 9.2.0.7.
I was able to install and complete the 9.2.0.8 and 9.2.0.8 OCT CPU patches on the primary database, however when I moved to working on the standby database, I am got an error when attempting to follow step 6 of the post installation tasks:
Enter the following SQL*Plus commands:
SQL> STARTUP MIGRATE
SQL> SPOOL patch.log
SQL> @ORACLE_BASE\ORACLE_HOME\rdbms\admin\catpatch.sql
SQL> SPOOL OFF
(Please note that this is my first experience with working with Dataguard, so my methods in updating might have been a bit off-base.)
If I attempt to startup the standby database in MIGRADE mode, I get the following error (which is expected):
ORA-01666: controlfile is for a standby database
If I try to run the catpatch.sql script, even after mounting the database in standby mode, I receive the additional error:
ORA-01219: database not open: queries allowed on fixed tables/views only
My question here is, how do I get the standby database to allow me to finish the software upgrade process, so that both versions of Oracle match on the primary and standby databases? :)
(I did try some research on Metalink, but only found an article detailing how to upgrade your dataguard database from 9i to 10g, which did not look especially helpful in this case...)
Thank you!
-SabrinaDepending on the kind of Dataguard you currently have, physical or logical, there are two useful metalinks notes:
Upgrading to 10g with a Physical Standby in Place
Doc ID: Note:278521.1
Upgrading to 10g with a Logical Standby in Place
Doc ID: Note:278108.1
~ Madrid
http://hrivera99.blogspot.com/ -
Logical Standby working issues Oracle 9i, Windows
Hi,
Set up Oracle 9i Logical Standby on Windows. (instructions as per Oracle Documentation)
Did not have any issues setting up.
While setting up the Logical Standby, Recovered the Primary Database until Oct 10/09 8:16 pm
Registered the archive log in the logical standby generated hence and the FAL took care of copying/registering the rest of the archivelogs.
Created and inserted some records in Primary database and could see them in Standby.
So far so good.
On Oct11 data was entered into Primary database. Archivelogs were shipped to Standby, I could see them registered in DBA_LOGSTDBY_LOG.
The APPLIED_SCN,NEWEST_SCN were in sync as per DBA_LOGSTDBY_PROGRESS.
Today, we had some issues with data and when we queried the user tables: (no skip settings)
Couldn't see any data in standby past the recovery...
No errors reported in DBA_LOGSTDBY_EVENTS. No errors in Alert log also.
What could be happening?
Thanks,
MadhuriI figured it out...
Today, we had some issues with data and when we queried the user tables: (no skip settings)
Couldn't see any data in standby past the recovery...I was using two tables as random spot check and both did not get updated. So, I was under the impression SQL APPLY did not do anything.
But, it did apply the redo on the rest of the tables.
These 2 tables in question were skipped because both of them had Function Based indexes.
They are very huge individual tables .
So, exporting them from Primary database and Importing them into Standby Database. Skipping DML in DataGuard.
That solved the problem.
--Madhuri -
Hi Friends,
I have 4 doubts. Please help me to clear my confusion.
Doubt 1:-
I have one primary and one logical standby database. When i restart the primary database or logical standby database or whenever data is not transffering to standby database, I execute the package "EXECUTE DBMS_LOGSTDBY.BUILD;". Then it starts to transfer the data to standby. My question is executing this package many times will affect the primary database? or my configuration is wrong?
Doubt 2:-
I configured DGMGRL for both database and i can switch over and switch back to logical standby database. It automatically start to apply the data to new logical standby database. But when i switch back to old primary database, It didnt transfer the data automatically. So i execute the package "EXECUTE DBMS_LOGSTDBY.BUILD;" in primary. Then it starts to transfer the data. My question is my switchback is wrong? or it happens like that only?
Note:- I check the verbose of both database before switchback. It show success.
Doubt 3:-
When i check the switchover status in logical standby database, It show "NOT ALLOWED". Even though i can switch over by using dgmgrl. Why logical standby database showing "NOT ALLOWED"?
Doubt 4:-
Is it any possible to view that what commands executing by dgmgrl process in background while it switchovering to logical standby database.?
Because when i try to switchover manually it didnt work. Many times i tried. But in dgmgrl, I simply give "switchover to logdb;", It works.
Please help to me to understand clearly.
Thank you in advance.Wrong forum mate. This forum deals with SQL and PL/SQL languages and related issues - not database administration and configuration issues.
The shared memory error is not unusual though ito SQL and PL/SQL. It is often a result of not using sharable SQL - in other words, SQL statements that are not using bind variables.
This causes additional consumption and fragmentation of shared pool memory. The very worse thing to do in this case would be to increase the size of the shared pool. That is akin to moving the wall a bit further away so that one can run even faster and harder into it.
But I'm merely speculating as you may have another cause for this problem. I suggest that you:
- post this in the [url http://forums.oracle.com/forums/forum.jspa?forumID=61]Database Forum
- research this error on [url http://metalink.oracle.com]Metalink -
Interesting issue with Logical Standby and database triggers
We have a Logical Standby that each month we export (expdp) a schema (CSPAN) that is being maintained by SQL Apply and import (impdp)it to a 'frozen copy' (eg CSPAN201104) using REMAP_SCHEMA.
This works fine although we've noticed that because triggers on the original schema being exported have the original schema (CSPAN) hard-referenced in the definition are imported into and owned by the new 'frozen' schema but are still 'attached' to the original schema's tables.
This is currently causing the issue where the frozen schema trigger is INVALID and causing the SQL Apply to fail. This is the error:
'CSPAN201104.AUD_R_TRG_PEOPLE' is
invalid and failed re-validation
Failed SQL update "CSPAN"."PEOPLE" set "ORG_ID" = 2, "ACTIVE_IND" = 'Y', "CREATE_DT" = TO_DATE('22-JUL-08','DD-MON-RR'),"CREATOR_NM" = 'LC', "FIRST_NM" = 'Test', "LAST_PERSON" ='log'...
Note: this trigger references the CSPAN schema (...AFTER INSERT ON CSPAN.PEOPLE...)
I suspect that triggers on a SQL Apply Maintained schema in a Logical Standby do not need to be valid (since they do not fire) but what if they reference a SQL Apply schema but are 'owned' by a non-SQL Apply schema? This trigger references a SQL Apply table so it should not fire
This is 10gR2 (10.2.0.4) on 64 bit Windows.
Regards
Graeme KingOK, I've finally got around to actually test this and it looks like you are not quite correct Larry in this statement...
'Since this trigger belongs to a new schema that is not controlled by SQL Apply (CSPAN201105) it will fire. But the trigger references a schema that is controlled by SQL Apply (CSPAN) so it will fail because it has to be validated.'
My testing concludes that even though the trigger belongs to a schema CSPAN201105 (not controlled by SQL Apply) and references a schema controlled by SQL Apply - it does not fire. However it DOES need to be valid or it breaks SQL Apply.
My testing was as follows:
Primary DB
Create new EMP table in CSPAN schema on Primary
Create new table TRIGGER_LOG in CSPAN schema on Primary
Create AFTER INSERT/UPDATE trigger on CSPAN.EMP table (that inserts into TRIGGER_LOG table)
**All of the above replicates to Standby**
Standby DB
Create new table TRIGGER_LOG_STNDBY in CSPAN201105 schema on Primary
Create new trigger in CSPAN201105 schema that fires on INSERT/UPDATE on CSPAN.EMP but that inserts into CSPAN201105.TRIGGER_LOG_STNDBY table)
Primary DB
Insert 4 rows into CSPAN.EMP
Update 2 rows in CSPAN.EMP
TRIGGER_LOG table has 6 rows as expected
Standby DB
TRIGGER_LOG table has 6 rows as expected
TRIGGER_LOG_STNDBY table has **0 rows**
Re-create trigger in CSPAN201105 schema that fires on INSERT/UPDATE on CSPAN.EMP but that inserts into CSPAN201105.TRIGGER_LOG_STNDBY table) **but with syntax error**
Primary DB
Update 1 row in CSPAN.EMP
TRIGGER_LOG table has 7 rows as expected
Standby DB
SQL Apply is broken - ORA-04098: trigger 'CSPAN201105.TEST_TRIGGER_TRG' is invalid and failed re-validation -
Logical standby issue after adding a new user at production db server
Hello Experts,
I am facing a issue related to logical standby db server where after adding a new user at production db server logical standby has stopped sql apply. Can anybody give some idea to find out the root cause and give some tips to resolve the issue.
tried below .....
SELECT EVENT_TIME, STATUS, EVENT FROM DBA_LOGSTDBY_EVENTS
ORDER BY EVENT_TIMESTAMP, COMMIT_SCN, CURRENT_SCN;
got error
ORA-16227: DDL skipped due to missing object
ORA-01403: no data found
ORA-00942: table or view does not exist
OS : RHEL 4 (IBM P5)
DB : Oracle 10g (10.2.0.1)
AchyotPlease let me know if can skip the DDL at which it is failing.
ALTER DATABASE STOP LOGICAL STANDBY APPLY;
Use EXECUTE DBMS_LOGSTDBY.SKIP
ALTER DATABASE START LOGICAL STANDBY APPLY;
Thanks,
Renu
Maybe you are looking for
-
This is my first post on the toshiba forums. Recently i have been gettin the dreaded BSOD (blue screen of death) on my toshiba qosmio x775 q7272 with the error code 0x100000b8 'attempted switch from dpc'. So far i have scanned for viruses which was c
-
Oracle Server's harddisk out of space. DBCompress in Oracle 9i
My oracle server's PC is running out a harddisk space because of volume of data. I deleted huge data in order to lessen the occupied space but still out and it doubled the occupied. e.g. DELETE FROM TABLE1 WHERE TBLNO > 100000 I am thinking of log fi
-
Sync iphone 4s to new Macbook Pro
How can I sync my iphone 4s to a new macbook pro without loosing anything The macbook i was syncing with had a hard drive crash. I lost everything
-
Tilde key on Slovenian keyboard?
Heya everyone! I'm from Slovenia and I've been using Illustrator for quite some time now. However I have recently found out that I can't acces the so called "tilde key" because my keyboard doesn't have one. So now after looking for a solution on the
-
I have a lot of controls on the front panel. I want my 3D plot to be updated with any of them changed. Currently only one control's change can result in the update even if I do have done changes on other controls .. that's to say, only one specific c