Hot Standby Database in 8.05

Does anyone know of a means to recover current online redo logs
from crashed primary database onto standby database ? I looks
like I can only recover with redo archived at the primary.
Has anyone activated a standby without losing contents of
primary's online redo ?. I cannot guarantee the primary will be
available to archive the last set of online redo.
Does anyone know if managed recovery (tranfer of archived redo
via Net8) is available in 8.05?
null

yes that 100032 is present in primary but in zip format ,when i extract and go for manul register its give me courputed archive.
after that i used incremental Rman strategy with the SCN . & i applied all process of rman
From Primary
SQL> Select max(sequence#) from v$archived_log;
MAX(SEQUENCE#)
102074
SQL>
from Physical Standby
SQL> Select max(sequence#) from v$archived_log where applied='YES' ;
MAX(SEQUENCE#)
SQL>
thanks

Similar Messages

  • Apps 11i disaster recovery with hot standby db

    Hello,
    We are trying to formulate a disaster recovery plan for Apps 11i (single node install) utilizing a hot standby database for the db layer and file synchronization
    (rsync, rcp, whatever) of the application stack. We're comfortable with the hot standby db technology, but uncertain that simply keeping the application stack
    sync'd on the standby host is sufficient to have a complete failover solution. Is anyone successfully using this type of configuration and , if so, is there a
    cookbook methodology for promoting the standby system to a primary? Can you simply rename the host on the standby and turn the key on?
    Thanks for any/all suggestions.
    Terry Riddle

    Hi,
    we are planning a similar recovery strategy. Where can I find documentation about your strategy? A consultant has suggested has to install Legato with two products for DBs synchronization.
    Thanks Gigi
    My email: [email protected]

  • DR Server: Standby Database

    Hello All,
    I need to Design a DR Server for my Client.
    I have decided to go for a hot Standby Database.
    I am very new to DB2. Are there any Documents that can help me on designing the procedure for this. If you can let me know the steps/commands in detail that would also be helpful.
    I have read the DB2 Operations guide available in SAP Marketplace which gives a high level Strategy.
    1. On Source System create a Snapshot of the DB. What exactly is a Snapshot?  How to create a Split Image?
    2.On the Target (DR Server) create the db after mounting the split image.
    Can you please help me in understanding how to aplly/execute this.
    Thanks,Bidwan

    Hi,
    Please see the following documentation which should be of help:
    [BACKUP DATABASE command|http://publib.boulder.ibm.com/infocenter/db2luw/v9r5/index.jsp?topic=/com.ibm.db2.luw.admin.cmd.doc/doc/r0001933.html]
    The default behavior for a snapshot backup is a FULL DATABASE OFFLINE 
    backup of all paths that make up the database including all containers,
    local volume directory, database path (DBPATH), and primary log and   
    mirror log paths (INCLUDE LOGS is the default for all snapshot backups
    unless EXCLUDE LOGS is explicitly stated). 
    [Backing up and restoring data is faster with snapshot backups|http://publib.boulder.ibm.com/infocenter/db2luw/v9r5/index.jsp?topic=/com.ibm.db2.luw.wn.doc/doc/c0052793.html]
    [Performing a snapshot backup  |http://publib.boulder.ibm.com/infocenter/db2luw/v9r5/index.jsp?topic=/com.ibm.db2.luw.admin.ha.doc/doc/t0052799.html]
    [Restoring from a snapshot backup image|http://publib.boulder.ibm.com/infocenter/db2luw/v9r5/index.jsp?topic=/com.ibm.db2.luw.admin.ha.doc/doc/t0052798.html]
    [IBM Redbook "High Availability and Disaster Recovery Options for DB2 on Linux, UNIX, and Windows"|http://www.redbooks.ibm.com/redbooks/pdfs/sg247363.pdf]
    Section 14.2 Backup and restore database with snapshot backup  
    o further guides                                                                               
    http://www.ibm.com/developerworks/tivoli/library/t-db2dptsam/    
    http://www.redbooks.ibm.com/redbooks/pdfs/sg246423.pdf           
    Hope this helps,
    Paul

  • Shadow Database for Hot Standby

    Hi,
    Our company uses embedded Oracle10g Standard and Enterprise Editions. The Oracle database schemas drive real-time applications; such as VoIP, etc. We need to shadow our applications and Oracle database to a failover server.
    We identified several methods to make and maintain a hot standby copy of an embedded Oracle database; including Data Guard, Oracle9i Replication, Streams, Recovery Manager, RAC, etc.
    Can you recommend a light weight, easy to implement, low cost solutions to making and maintaining a shadowed Oracle10g database? The replication is one way, writing from the active database to the standby, without check points, roll backs or any other features.
    Regards,
    Bruce
    LignUp Corporation

    Bruce,
    it looks to me that in order to satisfy the recovery time and affordable data loss requirements you could use either of two products - RAC or DataGuard.
    1. The RAC - comes with Standard edition for free, comes with Enterprise Edition at extra cost (awkward, isn't it?)
    pros -
    will allow you automatically recover in matter of seconds from failure of the node with zero data loss. will allow you automatically distribute the workload among the clustered nodes.
    cons - there is a single point of failure - the shared storage where cluster keeps the data.
    Also you should purchase the additional Oracle database license for the second (third,..) cluster node. Well, not in every configuration. I belive if you use user-based licensing model then clustered configuration will be no different from single node configuration in terms of costs. If you use the CPU based licensing model - you can split your licenses among the clustered nodes - for example if you've purchased 4 CPU licenses you can install 2 node cluster with dual-CPU servers. You have to make sure that dual-CPU server will sustain the load in case of failure of one node. In this case to feel safer you can purchase quad-CPU servers , but allow Oracle to run only on two CPUs on each node - in case of failure just change the affinity mask of the oracle process to let it use the extra 2 CPUs.
    2. DataGuard - comes with EntEd. at no extra cost, is not available with standard edition.
    you can setup Dataguard for zero loss configuration. the switchover time - probably would be closer to minute than to 5 seconds
    pros - doesn't require shared storage thus eliminates the single point of failure. does not require extra licensing costs - the second node is always in mounted state (unless you want to use it for read-only query processing).
    cons - near real-time recovery (compared to real-time of RAC). Requires double storage (to keep the standby database). Requires the extra standby server which is not fully used in production until the failure. Also there are limitations if you are using the direct-path inserts.
    Oracle vison for high availability is to use RAC where you wnat to be protected against failure within a site and use DataGuard to protect yourself against the site failure.
    Other solutions (Replication,Streams) will require you to change your application to incorporate the failure detection and switching logic.
    there are several whitepapers on this topic:
    http://www.oracle.com/technology/deploy/availability/techlisting.html#DR
    Mike

  • How to delete the foreign archivelogs in a Logical Standby database

    How do I remove the foreign archive logs that are being sent to my logical standby database. I have files in the FRA of ASM going back weeks ago. I thought RMAN would delete them.
    I am doing hot backups of the databases to FRA for both databases. Using ASM, FRA, in a Data Guard environment.
    I am not backing up anything to tape yet.
    The ASM FRA foreign_archivelog directory on the logical standby FRA keeps growing and nothing is get deleted when
    I run the following command every day.
    delete expired backup;
    delete noprompt force obsolete;
    Primary database RMAN settings (Not all of them)
    CONFIGURE RETENTION POLICY TO RECOVERY WINDOW OF 9 DAYS;
    CONFIGURE ARCHIVELOG BACKUP COPIES FOR DEVICE TYPE DISK TO 1; # default
    CONFIGURE DB_UNIQUE_NAME 'WMRTPRD' CONNECT IDENTIFIER 'WMRTPRD_CWY';
    CONFIGURE DB_UNIQUE_NAME 'WMRTPRD2' CONNECT IDENTIFIER 'WMRTPRD2_CWY';
    CONFIGURE DB_UNIQUE_NAME 'WMRTPRD3' CONNECT IDENTIFIER 'WMRTPRD3_DG';
    CONFIGURE ARCHIVELOG DELETION POLICY TO APPLIED ON ALL STANDBY;
    Logical standby database RMAN setting (not all of them)
    CONFIGURE RETENTION POLICY TO RECOVERY WINDOW OF 9 DAYS;
    CONFIGURE ARCHIVELOG BACKUP COPIES FOR DEVICE TYPE DISK TO 1; # default
    CONFIGURE ARCHIVELOG DELETION POLICY TO NONE; # default
    How do I cleanup/delete the old ASM foreign_archivelog files?

    OK, the default is TRUE which is what it is now
    from DBA_LOGSTDBY_PARAMETERS
    LOG_AUTO_DELETE     TRUE          SYSTEM     YES
    I am not talking about deleting the Archive logs files for the Logical database that it is creating, but the Standby archive log files being sent to the Logical Database after they have been applied.
    They are in the alert log as follows under RFS LogMiner: Registered logfile
    RFS[1]: Selected log 4 for thread 1 sequence 159 dbid -86802306 branch 763744382
    Thu Jan 12 15:44:57 2012
    *RFS LogMiner: Registered logfile [+FRA/wmrtprd2/foreign_archivelog/wmrtprd/2012_01_12/thread_1_seq_158.322.772386297] to LogM*
    iner session id [1]
    Thu Jan 12 15:44:58 2012
    LOGMINER: Alternate logfile found. Transition to mining archived logfile for session 1 thread 1 sequence 158, +FRA/wmrtprd2/
    foreign_archivelog/wmrtprd/2012_01_12/thread_1_seq_158.322.772386297
    LOGMINER: End mining logfile for session 1 thread 1 sequence 158, +FRA/wmrtprd2/foreign_archivelog/wmrtprd/2012_01_12/threa
    d_1_seq_158.322.772386297
    LOGMINER: Begin mining logfile for session 1 thread 1 sequence 159, +DG1/wmrtprd2/onlinelog/group_4.284.771760923                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               

  • Impact of force_logging on standby database in 10g?

    Hi there,
    We have a 10.2.0.5 database from which we want to setup a standby database...
    We took a hot backup of the primary and send it to the standby server but at that time we realized that force logging is not enabled at the Primary before we took the hot backup!
    Can you please suggest what are our options at this stage without losing any data and without a need to redo the hot backup again since it is a huge database?
    - If we turn the force_logging on now(on Primary) and then set up the standby using the backup taken prior to force_logging ON - would it cause any issues?
    Any help would be greatly appreciated.
    Thanks.

    No it won't because the hot backup is a copy period. So do it ASAP and the Archive will handle the rest.
    So your hot backup is the Standby copy ( exact copy ). Just make sure forced logging is on and the archive with keep the data in sync.
    If you delay then you are up the creek without a paddle as NOLOGGING actions are bound to occur and then the Standby data will not match the Primary.
    If you have doubts your best option is to roll the standby forward using RMAN Incremental Backups :
    http://www.oracle-ckpt.com/?s=incremental&op.x=31&op.y=11
    Best Regards
    mseberg

  • ORA-279 signalled during recovery of standby database

    Hi All,
    I am preparing standby database after taking hot backup and copying those datafiles to standby,
    taken controlfile standby controlfile backup from primary
    mounted the standby database using standby controlfile using
    startup nomount pfile='/u01/stand.ora'
    alter database mount standby database;
    Now I have started applying archives on standby database after copying all archives from primary to standby box the archives generated during backup using..
    recover standby database;
    But I am getting the below warning on the recovery screen...
    ORA-00279: change 51667629050 generated at 07/02/2009 00:59:43 needed for
    thread 1
    ORA-00289: suggestion : /nodal-archive/archive/1_55118_652209172.arc
    ORA-00280: change 51667629050 for thread 1 is in sequence #55118
    ORA-00278: log file '/nodal-archive/archive/1_55117_652209172.arc' no longer
    needed for this recovery
    ORA-00279: change 51667703096 generated at 07/02/2009 01:06:04 needed for
    thread 1
    ORA-00289: suggestion : /nodal-archive/archive/1_55119_652209172.arc
    ORA-00280: change 51667703096 for thread 1 is in sequence #55119
    ORA-00278: log file '/nodal-archive/archive/1_55118_652209172.arc' no longer
    needed for this recovery
    ORA-00279: change 51667767649 generated at 07/02/2009 01:12:28 needed for
    thread 1
    ORA-00289: suggestion : /nodal-archive/archive/1_55120_652209172.arc
    ORA-00280: change 51667767649 for thread 1 is in sequence #55120
    ORA-00278: log file '/nodal-archive/archive/1_55119_652209172.arc' no longer
    needed for this recovery
    ORA-00279: change 51667831821 generated at 07/02/2009 01:19:40 needed for
    thread 1
    ORA-00289: suggestion : /nodal-archive/archive/1_55121_652209172.arc
    ORA-00280: change 51667831821 for thread 1 is in sequence #55121
    ORA-00278: log file '/nodal-archive/archive/1_55120_652209172.arc' no longer
    needed for this recovery
    and I am getting the below warning on the alert log..........
    Sun Jul 5 18:37:36 2009
    ALTER DATABASE RECOVER CONTINUE DEFAULT
    Sun Jul 5 18:37:36 2009
    Media Recovery Log /nodal-archive/archive/1_55256_652209172.arc
    Sun Jul 5 18:40:31 2009
    ORA-279 signalled during: ALTER DATABASE RECOVER CONTINUE DEFAULT ...
    Sun Jul 5 18:40:31 2009
    ALTER DATABASE RECOVER CONTINUE DEFAULT
    Sun Jul 5 18:40:31 2009
    Media Recovery Log /nodal-archive/archive/1_55257_652209172.arc
    Please suggest what to do now.............
    Thanks in Advance,
    Sukanta Paul.

    Hi Sukanta,
    I didn't really understand what the problem is...
    Let me explain, hope it will be clear.
    You created a standby control file, and backed up the database (it doesn't matter in which order). Now, you have a standby database that is ready to apply logs. It will be able to apply log forever, until we will stop it.
    The messages you see say the following:
    ORA-00279: change 51667831821 generated at 07/02/2009 01:19:40 needed for thread 1 {color:blue} - this says that the recovery process is now at scn 51667831821 {color}
    ORA-00289: suggestion : /nodal-archive/archive/1_55121_652209172.arc {color:blue} - the information needed for the recover is in this file (this is only the default name, you can choose other name if the file is called differently) {color}
    ORA-00280: change 51667831821 for thread 1 is in sequence #55121 {color:blue} - same here, the needed archive is sequence 55121 {color}
    ORA-00278: log file '/nodal-archive/archive/1_55120_652209172.arc' no longer needed for this recovery {color:blue} - this message is after Oracle has applied the log, and it means that this log is no longer needed {color}
    Now it will wait for the next log. Again, this process is endless, this is how it is made to allow the standby database to be always synched with the primary.
    Liron Amitzi
    Senior DBA consultant
    [www.dbsnaps.com]
    [www.orbiumsoftware.com]

  • Physical vs Logical standby database

    In the Oracle HA guide, it is mentioned that we must take Logical backup "in addition" to Physical backup (not instead of Physical Backup)
    If I run have two systems i.e a Primary System and a Logical standby database system , is it not sufficient protection ? If the primary fails, can I not recover the primary system using the nightly backup I take from the Logical Standby Database.
    If it is not sufficient protection, what exactly I will lose ?
    Assume I take the nightly complete hot backup backup from the Logical Standby Database instead of the Primary System to save CPU cycles of the primary system.
    In other words, it is said that "physical standby database provides zero data loss without any allowance for data divergence" -- What exactly is meant here ?

    Thanks Werner for the reply.
    1. If there is a human error (dropping a table by mistake), it will be applied to the standby database irrespective of whether it is physical standby or Logical standby. I am unable to see any advantage with Logical standby database in this case. Am I missing something here ?
    Or are there any other type of user errors where Logical standby provides an advantage ? Can you please name some ?
    2. If I backup from the logical database and if I am willing to "restore" (instead of recover) the primary from this backup in case of a primary failure, is it not the same effort ?
    3. Can't we set up a Logical Standby with Maximum Protection Mode ?
    I am aware physical standby offers a better performance and it has no data type restrictions. I do not care about these two. Also, I am aware of the extra features provided by Logical database.
    Overall, I am trying to see whether physical standby database has any "feature" which I will miss with Logical standby database. I tried my best with Oracle documentation, but I could not get a clear answer.
    Any help is appreciated.

  • Standby database recovery problems....

    HP-UX 11.23
    oracle 10.2.0.2
    I know that this has been discussed quite a bit. I've spent a number of hours on this forum alone today just reading through the search hits. I've tried some of the 'tricks' that have been described in the threads. But I'm still getting the same error....
    My standby database lost connectivity with my primary, and stopped receiving logs. I got no notification, because my scripts were not set up to catch this event...
    The retention policy on my primary is 30 days of archive logs. After that, they are deleted. (I don't agree with it, but not my call)
    The standby got more than 30 days out of sync, so I decided to just refresh the whole standby.
    Since cold backups are not an option, I did an alter system checkpoint to flush all archive logs. Then I issued a complete hot backup.
    Once completed, I went ahead and created a fresh standby control file, as well as a fresh backup control file. In that exact order.
    I transferred all db files, archive logs, and control files over to the standby.
    I ran the commands:
    startup nomount;
    (SUCCESSFUL)
    alter database mount standby database;
    (SUCCESSFUL)
    recover standby database;
    I let it run through all the archive log files until it errored saying it couldn't find "X" file being the next sequence that wasn't in the directory. Which to my understanding, is normal...time to quit recovery.
    When I entered cancel, I get the following error.
    Specify log: {<RET>=suggested | filename | AUTO | CANCEL}
    ORA-00279: change 7313561157126 generated at 12/17/2008 19:01:02 needed for
    thread 1
    ORA-00289: suggestion : /oraarch/swesc/1_13729_640853326.arc
    ORA-00280: change 7313561157126 for thread 1 is in sequence #13729
    ORA-00278: log file '/oraarch/swesc/1_13728_640853326.arc' no longer needed for
    this recovery
    Specify log: {<RET>=suggested | filename | AUTO | CANCEL}
    CANCEL
    ORA-01547: warning: RECOVER succeeded but OPEN RESETLOGS would get error below
    ORA-01152: file 1 was not restored from a sufficiently old backup
    ORA-01110: data file 1: '/ora1/oradata/peregrine/system01.dbf'
    So my understanding of this is that it means that you still need to apply logs.
    So I grabbed the logs that had been auto created during the time that I was rebuilding everything and applied them.
    Same errors.
    I'm at a loss here. How can it still be out of sync? Is there a way to force it to a specific sequence so that I can successfully open it and then shut it down to complete the sequence?
    Thanks.

    Thanks OrionNet. Great info!!
    Here is the response.
    OrionNet wrote:
    select * from v$archive_gap
    NO ROWS SELECTED.
    and all the logs got applied
    select applied , count(*) from v$archived_log;
    -- This should return 1 rows with "YES", number of logs applied.
    I had to change change this as it complained about "not a single-group group function
    I changed it by adding group by applied to the end...
    NO ROWS SELECTED
    Then you should try activating standby database
    SQL> alter database force nomount pfile='....'; -- or without the pfile
    SQL> alter database mount standby database;
    -- If all the archived logs have been applied to your standby then stop recovery
    SQL> alter database recover manated standby cancel;
    -- Note if it fails on activating the database ;shutdown and mount it again and activate (skip stopping recovery step).
    SQL> alter database activate standby database;
    IF successfull, then perform followin
    SQL> alter database open;Regards
    OrionNet;
    I thought that in a primary/standby scenario, thiat this action would automatically shut down the primary.  I can't have that.
    HOLDING OFF ON REST FOR RESPONSE...
    Edited by: WillyB on Dec 17, 2008 9:40 PM

  • Downtime for logical or physical standby database

    Hi ,
    Are there any downtime required to refresh the data from the primary database to logical standby or physical standby database. What are the pros and cons of
    using logical and physical standby database approach. which one is advisable to use in production environment.
    Regards,
    Richard

    Hi Richard,
    I'd suggest you read the documentation as a starting point, to answer your questions on here would be quite lengthy,
    http://download.oracle.com/docs/cd/B19306_01/server.102/b14239/toc.htm
    In a nutshell you dont need downtime to maintain a standby, you can create them from a hot backup of your primary and get them recovering whilst it is open, in terms of which type of standby to use (physical or logical) that depends on your requirements for the use of the standby ... as I said read the doco (chapter 2 has a section on the benefits of each type).
    HTH
    Paul

  • Physical Standby Database with slight complication

    Hello Database experts,
    Primary is a multi-node RAC instance , primarily two distinct schemas (for simplicity sake schema A on tablespace A1/A2 and schema B on tablespace B1/B2). There is a proposed physical standby database. Now primary is Enterprise Data warehouse with millions of updates per minute. Is it possible to somehow bypass objects owned by schema A to be propogated on DR instance.
    Any workaround possible?
    Assume that Hot backup of primary database backup has to be done so we can not perform nologging operations on schema A and Logical standby is not a option.
    Kind Regards,
    Sunil

    ***Is it possible to somehow bypass objects owned by schema A to be propogated on DR instance.***
    Yes. but not in physical standby. While bypassing the objects to physical standby the recovery cannot be enabled at the time. So data loss may occur. Try to configure logical standby database. you can get what you expect.
    thanks

  • Manual Standby Database

    Hi,
    I have ORACLE 10.2.0.3.0 64 bit SE1 running on a RHEL4 server. I have a backup server on which I need to create a manual standby database for a (poor man's) high availability environment. We have explored RAC, Dataguard etc and given them up for cost considerations. We understand the potential of up to 15 minutes data loss. I know the general steps necessary to set up the manual standby server. For example, I know I need a cold (or hot) backup of the primary to reset my standby database. Then I need to create the standby control file. I need to write deamon processes to ship the archive logs to the standby server and apply those logs to the database. I am new to ORACLE and would appreciate some specific advice.
    (1) Should I use DBCA to create the database on the standby server first? My impression is that using DBCA will set up the enterprise manager, isqlplus and the relevevant entries in oratab etc on the standby server. If I simply install ORACLE 10g on the standby and copy the backup files from the primary server, I will not have access to the database tools that come with ORACLE 10g.
    (2) Can someone please share the scripts you may have written? or point me to some skeleton scripts?
    (3) What are the failure points for manual standby databases? What do I need to pro-actively monitor? For example, if the logs are not being applied properly to the standby database for a prolonged period, I would like to know about it.
    (4) What are the procedures for recovering from a failed master database?
    (5) Where can I learn more about the subject? I will need specific commands to setup the standby. Are these procedures documented somewhere?
    Thanks!
    - Rajesh

    Try this one Manual Standby Database under Oracle Standard Edition
    Or search the subject on google, many results.
    No, you will not create "standby" using DBCA, you need to use primary's backup to restore on standby server and constantly recover.

  • RMAN backup on the physical standby database.

    Hello,
    11.2.0.3 EE database. Is it a requirement to open the database in read only mode in order to take a hot backup?
    SQL>ALTER DATABASE RECOVER MANAGED STANDBY DATABASE CANCEL;
    SQL> ALTER DATABASE OPEN READ ONLY;
    I've initiated a rman backup without opening the db in read only and the following output is so far:
    RMAN> backup format '/dbstor2/LTS/LTS_rman/%d_set%s_piece%p_%T_%U' check logical database plus archivelog;
    Starting backup at 15-OCT-12
    using target database control file instead of recovery catalog
    allocated channel: ORA_DISK_1
    channel ORA_DISK_1: SID=603 device type=DISK
    specification does not match any archived log in the repository
    backup cancelled because there are no files to backup
    I plan on taking this backup and using it to refresh our Dev and QA dbs. Do you see any issues with taking a backup without opening the db in ready only mode?
    -SteveV

    No we are not using Active Dataguard. An issue I am investigating is why can't I perform an archivelog backup from the standby?
    RMAN> backup archivelog all format '/dbstor2/LTS/LTS_rman/arc.rman.bak';
    Starting backup at 15-OCT-12
    using target database control file instead of recovery catalog
    allocated channel: ORA_DISK_1
    channel ORA_DISK_1: SID=603 device type=DISK
    specification does not match any archived log in the repository
    backup cancelled because there are no files to backup
    Finished backup at 15-OCT-12

  • Best way to synchronize DR standby database

    Hi,
    My setup consists of 3 database installations in three data centers on aix 6.1 and oracle 11g release 2 Enterprise editions.
    The first data center will host the primary database, while the second data center will act as a hot standby site, hosting the physical standby database in a data guard configuration. This second data center is just a few hundred meters from the primary site and is connected with dark fiber.
    The 3rd site is a DR site located about 300KM from the first two centers, and with a WAN link of speeds around 2Mbps. This third site should act as a disaster recovery site in case there's a natural calamity that destroys the first two sites.
    We purchased full license for the first two sites for oracle. However, for the 3rd site, we're only allowed to open the database for up to 10 days in a year. My question is how can we always bring this 3rd database into sync with the first two. Can we use full backups to recover the database periodically, like every month? Or is there another option of using archivelogs, considering that the database can only be open up to a maximum of 10 days in a year?
    Both the primary and the DR sites have Tivoli server installed complete with the tape libraries and we're suppose to take backups even on the DR site.
    Regards,
    dula

    I think your potential solutions depend in part on the meaning of 'open' and what kind of database set up you have at the third site. If the third site is a Standby Datbase and open means in use for any purpose other than recovery mode then you should be able to start it and apply the logs so the the database is ready should the need arise.
    If open means the database is started and active then all you can really do is install Oracle at the third site and leave it off. In this case what I would do is use the rman duplicate command to build the initial database, start it, test it, and shut it down. Then you transfer your primary database rman backup to the site every day along with copies of the archived redo logs. In the event of a primary/secondary site failure you just need to run an rman restore of the newest backup set then recovery using the copied archive redo logs and you are up and running.
    There will be the time delay equal to the time to run restore and recovery but you can set up the remote diaster site, test it, and be pretty sure it will work without using up your 10 day limit.
    An alternate license arrangement might be to run RAC at the two local data centers and have a Stand by database at the remote site.
    HTH -- Mark D Powell --

  • 817 simple standby database questions

    i read the standby concept and tried to look for an answer on the web but couldn't get it.
    1. if i will use manage recovery enviroment can i lose transaction when the primary server is down?
    i read that it transfers the archive logs to the standby database but if the primary server "exploded" then i will lose all the transaction that in the online redo log and didn't archived yet.
    is it true ? am i missing something here ?
    2. if i will use a standby database is it important to have a regular hot backup of the primary/standby databases ?
    3. i still don't know if my net8 connection is down and my standby enviroment is manage recovery , what is the efect on the primary server ?
    thanks

    Hi zvika,
    I have make a standby 8i the last week, and I will try to answer your questions :
    1. Yes, you lose the all transaction from the last redolog which is not archived. When you have a standby database, you accept a minimum data lose.
    2. It's not necessary, if the standby is ok, you don't need a cold backup. You can save all archive files since the last cold backup.
    3. In this case, you have a warning (or error message) into the alert.log from primary database. And error message into v$archived_log view at standby line into your primary database. After what, it's depend if the init.ora parameter "log_archive_dest_n" for the standby dest is optional (recommanded) or mandatory.
    If optional, there is nothing to make, Oracle make itself.
    If mandatory (not recommanded), primary database stop all new entries (you need to change log_archive_dest_n to optional) and you need to copy manually all archive files which are created since the network problem and reapply manually it into the standby database.
    Hope this will help,
    Nicolas.

Maybe you are looking for