• "Rebalancing" is the process of moving file extents onto disks
• Purpose of rebalancing is evenly distributing the I/O load of the diskgroup
• In RAC Environment, Rebalancing for a disk group is done within a single ASM instance only and cannot be distributed across multiple cluster node to speed it up
• ASM will automatically rebalance data on disks when disks are added or removed by using
ASM_POWER_LIMIT initialization parameter.
• ASM_POWER_LIMIT parameter default value is 1
• We can change ASM_POWER_LIMIT parameter values by issuing bellow mention command in run time.
ALTER DISKGROUP
• The POWER value can be from 0 to 11 where 0 stops rebalancing and 11 is the fastest.
• The ASM_POWER_LIMIT can be found in the v$asm_operation view.
• The higher the limit, the faster a rebalance operation may complete. Lower values cause rebalancing to take longer, but consume less processing and I/O resources. This leaves these resources available for other applications, such as the database.
• Rebalancing could be done while the databases (ASM clients, in this case) are connected. However, database performance could be impacted but I think the best recommendation would be when the load on databases are low(off load hours, or non peak work times) and there you can use higher intensity setting(asm_power_limit putting higher for example) without degrading performance of databases significantly
Play with Rebalancing:
Check ASM Disk usage, before creating Database:
SQL> select TOTAL_MB, FREE_MB, FAILGROUP,PATH from v$asm_disk;
TOTAL_MB FREE_MB FAILGROUP PATH
--------------------------------------------------------------------------------
2055 2004 A /dev/rdsk/c0d1s0
2055 2004 B /dev/rdsk/c0d1s1
Check ASM Disk Usage, after creating Database:
SQL> select TOTAL_MB, FREE_MB, FAILGROUP,PATH from v$asm_disk;
TOTAL_MB FREE_MB FAILGROUP PATH
--------------------------------------------------------------------------------
2055 1056 A /dev/rdsk/c0d1s0
2055 1056 B /dev/rdsk/c0d1s1
Now we add 2 GB of Disk in failgroup A
SQL> ALTER DISKGROUP DB_DATA01 add failgroup A DISK '/dev/rdsk/c0d1s3';
Check rebalancing Status:
Select * from v$asm_operation
After rebalancing check the Disk Status:
SQL> select TOTAL_MB, FREE_MB, FAILGROUP,PATH from v$asm_disk;
TOTAL_MB FREE_MB FAILGROUP PATH
--------------------------------------------------------------------------------
2055 1367 A /dev/rdsk/c0d1s0
2055 1367 B /dev/rdsk/c0d1s1
2055 1368 A /dev/rdsk/c0d1s3
Conclusion:
Before Adding disk Capacity of each Disk is 2 GB
Total Space: 4 GB (2GB*2 Disk)
Total Used MB: 1998
Each Disk Used MB: 999
Afterb Adding Disk capacity of each disk is 2 GB
Total Space: 6 GB (2GB*3 Disk)
Total Used: 2063
Each Disk Used MB: 688
We had 2x 2 GB disks that where 1998 MB Used and added 1x2 GB disks, the final set should be 3x2 GB disks with 1998 MB of data balance across them, so each would have 688 MB of data. To me that means, at a minimum, we have to move 311 MB of data.
