Upgrade :: Migration Without Downtime - Goldengate Approach
Mar 24, 2013
We are facing a project where it is mandatory that the migration (from 9i to 11g) happens without any downtime. We thought about using Goldengate do to this migration. But i would like to listen to somebody who already did such kind of migration (i never used goldengate before). The basic steps to do such migration would be:
1) Install the Goldengate client on both source and target
2) Export only the metadata (structure of the table, for example) from source to target (here is one point of doubt of mine. This export can be only done using exp/imp?)
3) Perform the initial load from source to target (here i have another doubt: It it possible to perform an initial load from a whole database?)
4) Configure manager, export and replicat to perform the migration with the source database open in read-write
With the steps above, would i be able to perform a migration without downtime? What other considerations do you have?
I want to apply divide and conquer approach on the field value, For example I have the field value like this
Value 32,36,12,65,85,9663 43,36,89,65,112,9663 36,63,12,65,95,123 32,36,12,85, 85,9663
Remember that these value are taken from ONLY one field i.e Value. Apply the divide and conquer approach in such a manner that take the value of a single field and each row and divide the value into two half. Continue this process until single value until we reach the single value or single sub problems.
Apply the divide and conquer approach on all the value of field. When we get final value then match the value of two different rows. And check how much field value are similar (match or duplicated). If field values are match more then 80% then delete the duplicated value and keep the original entity.
We are trying eliminate/minimize the downtime for our application. As part of new code deployments sometimes we need to modify DB Structure also. As it is taking time to backup current DB and apply new DDL, the application is down.
Is there a way to eliminate the downtime, if I can leverage Data Guard, Golden Gate or RAC concepts?
Is there any way I can find out what caused the database to crash; either a history of commands executed within the database, I lost my bdump directory before the scheduled backup ran and the only logs available are after I re-created the directory.
SQL> startup ORA-00444: background process "PMON" failed while starting ORA-07446: sdnfy: bad value '' for parameter . SQL> [code]....
I have a question about GG Sequence Replication and Triggers. My main database, which I would like to replicate on another server, is highly dependent on sequences for assigning surrogate keys to every row in every table in the application. I know that I need to add Sequence support to my source database (plus supplemental logging, etc), but I'm curious about the target database.
I do not anticipate allowing Read/Write access to this database - we are migrating from 10.2.0.4 (source) to 11.2.0.3 (target) on a new platform, and I want to keep the 11g database up-to-date with our production data until it is time to begin the actual conversion of our application. My thinking is that if I use the SUPPRESSTRIGGERS dboption in my Replicat session, this should take care of the use of the Sequences for assigning the surrogate key values, and the data should add to the tables normally without any intervention by the sequences/trigger combination. I know I will have to manually "correct" the sequences on my 11.2.0.3 database whenever I want to open this database up for use, but I have a script for this ready to go.
Also, in my source database, I am using Oracle Context indexes for generic name searching - this feature creates a number of DR$ named tables in the main application schema that I am replicating (approximately 50 of them). I am assuming that I should EXCLUDE these tables from the replication, as the context indexing should automatically update them as changes to the underlying data are applied via the replication of the indexed tables.
Using Golden Gate to replicate a database (Encrypted Tablespace, Oracle 11.2.0.1, Windows 2008) to a different database server (No Encrypted Tablespace Oracle 11.1 Linux)
Following error goldengate report ERROR OGG-01771 DBOPTIONS DECRYPTPASSWORD must be used to decrypt TSE data. Use TRANLOGOPTION IGNORETSERECORDS if you do not need to capture any tables that are in an encrypted tablespace.
How use it
GGSCI> ENCRYPT PASSWORD "shared key" Add an entry to the Extract parameter file to decrpt the new shared password
We are getting following error when we are trying to extract data from ASM.
GGS ERROR 500 Oracle GoldenGate Capture for Oracle, ext_1.prm: Getting attributes for ASM file +DATA/testgg/onlinelog/group_1.257.742844671, SQL <BEGIN dbms_diskgroup.getfileattr('+DATA/testgg/onlinelog/group_1.257.742844671', :filetype, :filesize, :lblksize); END;>: (6550) ORA-06550: line 1, column 7: PLS-00201: identifier 'DBMS_DISKGROUP.GETFILEATTR' must be declared ORA-06550: line 1, column 7: PL/SQL: Statement ignoredNot able to establish initial position for begin time 2011-02-16 16:42:05.
I am implementing GG 11g r2 for 12C database. But i am getting below error. My question Why Goldengate needs specific package ... Since this is homogeneous & heterogeneous.
/u01/12c_database_software/goldengate/dirtmp. 2013-08-30 05:28:44 INFO OGG-01513 Oracle GoldenGate Capture for Oracle, ext1.prm: Positioning to Sequence 66, RBA 25067536, SCN 0.0. 2013-08-30 05:28:44 ERROR OGG-01028 Oracle GoldenGate Capture for Oracle, ext1.prm: ORA-06550: line 1, column 7: PLS-00201: identifier 'SYS.DBMS_INTERNAL_CLKM' must be declared ORA-06550: line 1, column 7: PL/SQL: Statement ignored. 2013-08-30 05:28:44 ERROR OGG-01668 Oracle GoldenGate Capture for Oracle, ext1.prm: PROCESS ABENDING.
Database details ---------------- SQL> select object_name, object_type from dba_objects where object_name='DBMS_INTERNAL_CLKM' and object_type in ('PACKAGE');
no rows selected
SQL> select object_name, object_type from dba_objects where object_name='DBMS_INTERNAL_CLKM';
Oracle version : 10g, 11g (applicable to both) Consider a schema named 'UNIVERSE' present in the database A running on Linux platform where this schema needs to be moved to another database B running on windows platform or AIX platform with no downtime provided and data needs to be consistent. Is this practically possible?
I'm planning to upgrade a small database (~150GB) from 10.2.0.3 on windows 2003 23bit to 11.2.0.3 RAC on Linux 5.8.The database contains oracle spatial too. A suitable method and link to document to be followed.
I recently performed an upgrade on a new server from oracle 10gr2 to oracle 11gr2 (11.2.0.3).
I take the rman backup from oracle 10g server and restore it on new server where I installed oracle 11gr2.
But on my previous oracle 10gr2 server I enabled the auditing. After doing successful upgrade now when I try to login with any user except sys I receive the following error:
SQL> conn scott/tiger ERROR: ORA-00604: error occurred at recursive SQL level 1 ORA-00904: "OBJ$EDITION": invalid identifier ORA-02002: error while writing to audit trail ORA-00604: error occurred at recursive SQL level 1 ORA-00904: "OBJ$EDITION": invalid identifier
I got the workaround by setting the parameter audit_trail=FALSE (Previous value was DB_EXTENDED) .But I want my auditing to be enabled as per y requirements.
find reference note IDs for DB upgrade from 11.2.0.2 to 11.2.0.3.2, as I am finding only Exadata which I don't want but I want to find for Ebiz database, on OS - Solaris 10 9/10 s10s_u9wos_14a SPARC.
Along with existing RMAN backups we do Exports - of our DB using and OS User and Oracle Wallet.Of the DB's we have upgraded the Data Pump Directory
Select * from dba_directories; (there are other commands to get this info as well).
I captured screens from the DBUA upgrades, but did not see an option to change this information.Is there a way to feed this information to the install moving forward. IE, ./DBUA -silent ?
Also, anyone tracked the percentage of storage increase from 10.2/11.1 to 11.2.
I am going to upgrade database from 11.1.0.7 to 11.2.0.3
1) If compatible is set to 10.2.0 in 11.2.0.3, will it work ? 2) If compatible set to maximum level, will it affect our application ? 3) Whether any code related problem occurred after upgrading like PL/SQL codes ?
We are planning to upgrade our database from Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - 64bi to Oracle Standard Edition 11g . We also have oracle apex installed on Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - 64bi database with oracle apex 3.1
No our plan to upgrade the database and oracle apex to 4.2. Since Oracle Enterprise Edition is licensing is very expensive we though of buying standard edition and upgrade to this version.
can we upgrade the oracle database from enterprise edition to latest standard edition?
I am trying to upgrade the database 32bit from 10.2.0.1 to 10.2.0.4 version in lunux 32 bit . I faced the version incompatibilty error during patchset installation and hence ran the patchset installation with -ignoresysprereqs option.Now during the patchset installation i encountered the below error in in the install logfile.
INFO: Start output from spawned process: INFO: ---------------------------------- INFO: INFO: /u01/app/oracle/product/10.2.0/bin/genclntsh INFO: genclntsh: genclntsh: Could not locate /u01/app/oracle/product/10.2.0/network/admin/shrept.lst [code]....
Need certified os and versions for oracle db 10gr2
After many tries of upgrade'ing oracle I decided to make backup of my database and then remove 11.2.0.1 and install 11.2.0.3 to recover db. Unfortunately I am not sure how should I perform restoration.I backed up data by calling RMAN> backup database include current controlfile;
After that I moved files from fast_recovery_area and cleared my machine from current oracle release.I also did a copy of directories:
