Applies to:

Symptoms

You need to take free space from an ASM diskgroup and move it to another host or another diskgroup on the same host. All of the disks are currently used.

Cause

You have allocated LUNs for a DISKGROUP but have not used all the space in that diskgroup and would like to move some of the space to another host or another diskgroup on the same server.

Example:

 
How can you reorganize data to clear out some luns to move them ?

Solution

ASM load balances file activity by uniformly distributing file extents across all disks in a diskgroup.
If you try to downsize a physical LUN and there is Oracle data in that section of the LUN, that will truncate the diskgroup and files inside, causing corruption of an asm diskgroup.
That is why do not downsize a physical LUN.

Dropping an oracle ASM disk and releasing underlying LUNs after rebalance went through is the correct approach to downsize free space in a diskgroup.
Keep in mind to at least leave 10% of free space in each asm disk, otherwise you will get ORA-15041 upon rebalance.

ASM automatically rebalances whenever disks are added or dropped.

For a normal drop operation (without the FORCE option), a disk is not released from a disk group until data is moved off of the disk through rebalancing.
Likewise, a newly added disk cannot support its share of the I/O workload until rebalancing completes.
It is more efficient to add or drop multiple disks at the same time so that they are rebalanced as a single operation. This avoids unnecessary movement of data.
For a drop operation, when rebalance is complete, ASM takes the disk offline momentarily, and then drops it, setting disk header status to FORMER.

You can add or drop disks without shutting down the database. However, a performance impact on I/O activity may result.

To drop disks from a disk group, use the DROP DISK clause of the ALTER DISKGROUP statement.

When a disk is dropped, the disk group is rebalanced by moving all of the file extents from the dropped disk to other disks in the disk group.
A drop disk operation may fail if not enough free space is available on the other disks. If you intend to add some disks and drop others, it is prudent to add disks first to ensure that enough space is available for the drop operation.

The best approach is to perform both the add and drop with the same ALTER DISKGROUP statement. This can reduce total time spent rebalancing.

Below are detailed steps how to drop an oracle ASM disk.

This example drops DATA3 from diskgroup dgroup1.

 
Where 'DATA3' is column 'NAME' from v$asm_disk.

When ASMLIB is used, "/etc/init.d/oracleasm querydisk"  will tell you if you are using correct alias for the drop command.  If it comes as 'Marked as ASM disk' then you are using the right alias name for drop.

The main thing after you issue DROP command, is to make sure rebalance completes.

Rebalance will happen automatically if there is a need. If you add/drop/resize a disk rebalance will take place, there is no need to start rebalance manually.
When you add/drop/resize you can specify rebalance power manually. Power can be 0-11.  0 is stopped, 11 is fastest.
It will put more load on the system, and slow down access to the DG. Power 7 will be faster than 4 and will provide balance between DG vs. rebalance speed.
RBAL background process opens all device files as part of discovery and coordinates the rebalance activity.

The main objective of the rebalance operation is always to provide an even distribution of file extents and space usage across all disks in the diskgroup.
The rebalance is done on a per-file basis to ensure that each file is evenly balanced across all disks. This is critical to ASM's assurance of balanced I/O load.

Is the rebalance still performing the movement of extents?

If yes, you can watch the movement with the following query:

References