Another question on Cluster:
DBAsupport.com Forums - Powered by vBulletin
Results 1 to 8 of 8

Thread: Another question on Cluster:

  1. #1
    Join Date
    Jan 2001
    Posts
    642
    I need to do the following IN A CLUSTERED ENVIRONMENT:

    Stop database and start archive logging to the following
    destinations: -

    /u01/oradata/ARC1/ora_data/archive_redo optional reopen 60
    /u02/oradata/ARC2/ora_data/archive_redo


    "Fail Cluster over and check this works on the other Node. "


    What does it mean by the last statement?

    Badrinath

  2. #2
    Join Date
    Feb 2000
    Location
    Washington DC
    Posts
    1,843
    Stop database and start archive logging to the following destinations: -

    /u01/oradata/ARC1/ora_data/archive_redo optional reopen 60
    /u02/oradata/ARC2/ora_data/archive_redo
    Here basically you are archiving to a different location. I guess log_arch_destination_state_n set to defer/enable controls this.

    "Fail Cluster over and check this works on the other Node. "
    What does it mean by the last statement?
    Badrinath
    There are multiple instances running on multiple nodes. Lets say Instance_1 has some problem, so you failover to another instance on another node (say Instance_2) the database activity until you fix the problem with instance_1.

    How ? Sorry! I don't have expertise...even if I have its not something I can explain in this forum.
    Reddy,Sam

  3. #3
    Join Date
    Jan 2001
    Posts
    642
    Thanks Sam,

    I know your unix expertize : Just another ques

    1) Did you say that there will be 2 instances but only one database ? Right

    2) 'YOU FAIL OVER TO ANOTHER INSTANCE' means ?

    When the database is running, Will I be able to see 2 diff set of processes (like dbwr,lgwr,smon) one on both the machines or will they be one. -

    THESE PROCESSES IS WHAT YOU CALL AS INSTANCE - RIGHT?


    Badrinath

  4. #4
    Join Date
    Jan 2001
    Posts
    642
    If I have to install a patch (8.1.7.0.0 to 8.1.7.0.4) on to this, My guess I have to install it only ones. Correct.

    Further one cd (1 of 2) is on Node A and 2 of 2 is on Node B how do I do the Patch installation.

    Thanks in advance/
    Badrinath

  5. #5
    Join Date
    Feb 2000
    Location
    Washington DC
    Posts
    1,843
    Originally posted by badrinathn
    1) Did you say that there will be 2 instances but only one database ? Right
    Yes! you are right! Multiple instances resources contributes to single database. Thats why the database will always available and highly available.
    2) 'YOU FAIL OVER TO ANOTHER INSTANCE' means ?
    When the database is running, Will I be able to see 2 diff set of processes (like dbwr,lgwr,smon) one on both the machines or will they be one. -

    THESE PROCESSES IS WHAT YOU CALL AS INSTANCE - RIGHT?
    Yes! you are right!
    (The database you know)SINGLE instance ---> Single Database ---> Single set of control_files----> Single init.ora ----> Single set of datafiles.
    (The OPS database)MULTIPLE instances ---> Single Database ---> Multiple set of control_files----> Multiple init.ora ----> Single set of datafiles. Its Single machine in clustered environment.

    Hope this Helps. Correct me if I'm wrong ...

    [Edited by sreddy on 07-18-2002 at 05:32 PM]
    Reddy,Sam

  6. #6
    Join Date
    Feb 2000
    Location
    Washington DC
    Posts
    1,843
    If I have to install a patch (8.1.7.0.0 to 8.1.7.0.4) on to this, My guess I have to install it only ones. Correct.

    Further one cd (1 of 2) is on Node A and 2 of 2 is on Node B how do I do the Patch installation.
    You have to install only onceas its SINGLE ORACLE_HOME.

    To be honest with you, You better consult RAC(OPS) docs and Metalink Docs OR the Someone who implemented OPS. I have explored little of OPS couple of years ago at one of my client site when they planned to implement OPS and later they dropped from that idea knowing the cost of maintenance and implemented HACMP high availability solution.
    Reddy,Sam

  7. #7
    Join Date
    Sep 2001
    Location
    NJ, USA
    Posts
    1,287
    -------------------------------------------------------------------
    (The OPS database)MULTIPLE instances ---> Single Database ---> Multiple set of control_files ---->( ?? Multiple init.ora ) ----> Single set of datafiles.
    Its Single machine in clustered environment.
    -------------------------------------------------------------------
    Sam this is correct.

    But had better to create set of init.ora files
    init_inst2.ora -----V
    init_inst1.ora ---> init.ora

    in init_instX.ora u can set instance specific parameters
    in common init.ora u can set common parameters that must be equal for all instances.

    In any way in ops/rac environment ----> Multiple init.ora
    ========================================
    NOTE:
    In rac environment each instance have to has its own subset (no less then 2) of redo log files.
    All instances should know location of all redolog files,
    but only one instance is owner of subset of redolog files.
    But again this is one set on db files.
    ========================================


    [Edited by Shestakov on 07-19-2002 at 12:05 AM]

  8. #8
    Join Date
    Jul 2002
    Posts
    132
    Yes, it is right that each instance will have its own set of redo logs. Besides that each instance can have its own set of rollback segments. The init.ora file will have extra GC* parameters. Might be an addition to what u have asked : when u create a control file using : backup controlfile to trace. U cannot use it as it is if u run it against the database- u have to separate the threads for the redo-logs.

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