Home > Workload Solutions > Oracle > White Papers > Oracle ASM on ScaleIO Best Practices > ASM diskgroup—adding and removing disks
As databases grow and require more storage, ASM diskgroups can be expanded by adding new disks.
ASM will automatically rebalance allocations across all available disks to maintain a uniform distribution of data throughout the diskgroup.
When adding new disks to a diskgroup, it is advisable to add all new disks in the same command. This allows ASM to rebalance once for all new disks, instead of repeatedly as each new disk is added.
The DBA may assign a rebalance power limit to the operation to limit the impact on a production system of the rebalance operation. The higher the rebalance power limit the more resources will be used to rebalance the disk allocation, and the faster the rebalance operation will complete.
In the following example we add two new disks - /dev/oracleasm/data03 and /dev/ oracleasm/data04 to the existing DATA disk group.
SQL> alter diskgroup DATA add disk '/dev/oracleasm/data03','/dev/oracleasm/data04' rebalance power 1;
The DBA can monitor the progress of the rebalance operation by observing the V$ASM_OPERATION view:
SQL> @operation11
DISKGROUP_NAME OPERATION STATE PWR ACTUAL PCT_DONE EST_MIN ERROR_CODE
--------------- -------------------- -------- ---- ------- -------- ------- ----------
DATA REBAL RUN 1 1 2.8% 2
If disks need to be removed from an ASM disk group, then again all disks should be removed in a single command to minimize the rebalance operations:
SQL> alter diskgroup DATA drop disk DATA_0002,DATA_0003 rebalance power 6;
In the above command, the disk names specified must match those reported in the V$ASM_DISK view.