5 Replies Latest reply on May 30, 2013 11:05 PM by onedbguru

    rebalance

    Ora_83
    Ora_83 May 29, 2013 8:54 AM
      Hello

      Suppose I`ve added a new disk with rebalance power 0.
      Can the new records inserted to this new disk ? or the new disk is always empty unless I enable rebalance ?
      • 433 Views
      • Tags:
        • 1. Re: rebalance
          swapnil kambli
          swapnil kambli May 29, 2013 9:06 AM (in response to Ora_83)
          from [url http://docs.oracle.com/cd/B19306_01/server.102/b14200/statements_1006.htm] here 

          POWER In the POWER clause, specify a value from 0 to 11, where 0 stops the rebalance operation and 11 permits 
Automatic Storage Management to execute the rebalance as fast as possible. T
he value you specify in the POWER clause defaults to the value of the ASM_POWER_LIMIT initialization parameter.
          • 2. Re: rebalance
            Dude!
            Dude! May 29, 2013 8:52 PM (in response to Ora_83)
            ASM uses disks of a disk group in a round robin fashion. ASM works on a file basis and distributes allocation units among all disks in a disk group when data is created or modified.

            A rebalance is a separate process that attempts to spread data evenly among all disks in a disk group and also manages hot and cold regions for performance. This is normally automatically triggered when adding or removing disks in a disk group. ASM uses an indexing technique to spread extents on available disks and does not need to re-stripe all data, but only the amount of data so that it gets evenly distributed among all disks.

            The affect of setting the rebalance power to 0 (11g) and hence disabling the rebalance process will therefore disable I/O optimization based on device availability in a disk group. In your example, the new disk is still used in a round-robin fashion like any other device in the disk group when creating data, but it won't rebalance existing data and leave the disk group in an unbalanced state.

            ASM also maintains data redundancy (fast mirror resync) among failure groups depending on your settings, but as far as I understand, this is not affected by the ASM_POWER_LIMIT setting.
            • 3. Re: rebalance
              Ora_83
              Ora_83 May 30, 2013 5:29 AM (in response to Dude!)
              Thanks Dude,

              Bottom line is, the new disk will be used by new transactions but existing data will not be rebalanced.

              Is that right?
              • 4. Re: rebalance
                Dude!
                Dude! May 30, 2013 9:52 PM (in response to Ora_83)
                If the "new transaction" demands new disk space to affect the new disk due to ASM writing in a round robin fashion, then yes.
                • 5. Re: rebalance
                  onedbguru
                  onedbguru May 30, 2013 11:05 PM (in response to Ora_83)
                  If you took the time to read the concepts/admin manual you would find that POWER 0 says "don't do anything yet" It will not rebalance the new device until you alter diskgroup rebalance power ( > 1).


                  You can do:

                  add disk power 0
                  add disk power 0
                  delete disk power 0
                  delete disk power 0
                  alter diskgroup rebalance power 4.

                  The result is that the first 4 commands will add and delete the devices, but nothing happens until the alter command is issued.