DBAsupport.com Forums - Powered by vBulletin
Page 1 of 2 12 LastLast
Results 1 to 10 of 11

Thread: difference between dataguard and rac

  1. #1
    Join Date
    Oct 2002
    Posts
    391

    difference between dataguard and rac

    HI

    rac is a cluster of machines (nodes) whereby the oracle instance is being switched around if the node which houses the instance has some problems?

    dataguard is a configuration of 2 machines. one is the production and the other is the standby, exact replica of the production.

    so the difference between the two is that rac is highly available whereas the dataguard ensures data integrity. Since if the one oracle instance in rac has problem, then a restore will be inevitable.

    Please advise.

  2. #2
    Join Date
    Nov 2000
    Location
    greenwich.ct.us
    Posts
    9,092
    Quote Originally Posted by yls177
    HI

    rac is a cluster of machines (nodes) whereby the oracle instance is being switched around if the node which houses the instance has some problems?
    RAC uses a cluster, but there is usually one instance running on each physical node. If one node/instance goes down you can still run in a degraded mode with one node/instance.

    dataguard is a configuration of 2 machines. one is the production and the other is the standby, exact replica of the production.
    more or less, yes.

    so the difference between the two is that rac is highly available whereas the dataguard ensures data integrity. Since if the one oracle instance in rac has problem, then a restore will be inevitable.
    no. If one node goes down, the other node will take over for it. If both nodes go down, then yes, a restore is inevitable. However, your RAC database can be protected by dataguard as well.
    Jeff Hunter

  3. #3
    Join Date
    Nov 2002
    Location
    New Delhi, INDIA
    Posts
    1,796
    Quote Originally Posted by yls177
    HI

    rac is a cluster of machines (nodes) whereby the oracle instance is being switched around if the node which houses the instance has some problems?

    dataguard is a configuration of 2 machines. one is the production and the other is the standby, exact replica of the production.

    so the difference between the two is that rac is highly available whereas the dataguard ensures data integrity. Since if the one oracle instance in rac has problem, then a restore will be inevitable.

    Please advise.
    Data Guard and RAC both are different and serve different purpose.

    RAC is multiple instances accessing a single database with key features availability, scalability and manageability.

    Data Guard software maintains a real time copy of the production database called a standby database.

    You can have 2 or more nodes in a RAC accessing a single database whose copy is maintained by Data Guard. RAC and Data Guard can be combined together to provide Maximum Availability Architecture.

    You can read this excellent white paper on Maximum Availability Architecture (MAA). More can be found here.
    Amar
    "There is a difference between knowing the path and walking the path."

    Amar's Blog  Get Firefox!

  4. #4
    Join Date
    Jul 2005
    Posts
    7
    Data Guard software maintains a real time copy of the production database called a standby database
    In DG (data Guard), does the secondary get started automatically if the primary goes down? Or does a DBA start it up? If the DBA starts it up, what does he have to do? 'startup open'?

    If this is the case, then does a DBA have to be looking at it 24x7? A friend of mine said that is the case; it doesnt look very optimal to me.

    How does the DBA or the DG know when the primary goes down, esp. if it happens in the middle of the night.

    Is it true that once the secondary becomes primary after a failure, we cant use the old primary anymore?

    Rajan.

  5. #5
    Join Date
    Sep 2002
    Location
    England
    Posts
    7,334
    Yes a manual switchover is needed. Automatic would be very bad in my opinion as you dont have the choice then of what you want to do. i.e. you many just be able to bring the primary back up right away.

    And if your database is important enough to have dataguard on it, then it is important enough to have monitoring setup with an on-call dba

    And yes you are right if you do a switchover - you cant switch back, would need to rebuild the primary again from the standby

  6. #6
    Join Date
    Jul 2005
    Posts
    7
    Yes a manual switchover is needed.
    Can you please tell me how a DBA manually switches over.
    Does he open the database and issue recover database and then open reset logs?

  7. #7
    Join Date
    Nov 2002
    Location
    New Delhi, INDIA
    Posts
    1,796
    Quote Originally Posted by RajanR
    Can you please tell me how a DBA manually switches over.
    Does he open the database and issue recover database and then open reset logs?
    Refer to this document to get the instruction on role switch

    If you want to understand how Data Guard works then read the Dataguard Concepts and Administration and Data Guad Broker which automates and centralizes the creation, maintenance, and monitoring of Data Guard configurations.

    HTH
    Amar
    "There is a difference between knowing the path and walking the path."

    Amar's Blog  Get Firefox!

  8. #8
    Join Date
    Jul 2005
    Posts
    7
    Thanks. Does the DBA have to be in the server room? or can he switchover remotely, for example from home by connecting to the server.

  9. #9
    Join Date
    Sep 2002
    Location
    England
    Posts
    7,334
    as long as he has a connection to the database, can do it from anywhere

  10. #10
    Join Date
    Oct 2002
    Posts
    391
    Hi guys...

    Thanks for correcting my thoughts. I will bring to move ahead with this RAC.

    Okay, with my original configuration. I have 2 nodes. From the history of this thread, i understand that there are instances on these 2 nodes which is connected to a single database which is on shared disk, right?

    and the below

    http://www.dbasupport.com/forums/showthread.php?t=48382

    confirms my thought too....


    And now, each instance is a set of processes together.

    1) So the instances in these 2 nodes belong to the same SID ?

    2) The database in the shared disks are actually the oracle binaries and datafiles and other physical files?

    3) At any point of time, there is only one active instance among these 2 nodes?

    4) Finally, if one node has some problems, the other instance in the other node will take over ?

    Thanks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  


Click Here to Expand Forum to Full Width